Latest / learning with yacine show / The Science of Learning Math (and Anything Else) with Justin Skycak
Transcript
- 0:00SPEAKER_01: For those that know me, you know I'm a big learning fan.
- 0:03SPEAKER_01: Like, I love learning to a ridiculous degree.
- 0:06I read that one book called The Brain Had Changed Itself way
- 0:10back before I started college.
- 0:12And it was one of the moments that I decided that I really
- 0:15wanted to understand learning.
- 0:16SPEAKER_01: I wanted to learn about learning.
- 0:19It got me into a whole journey of trying to understand how the
- 0:21brain is working and
- 0:23SPEAKER_01: ultimately into neuroscience and AI.
- 0:26The one thing that I realized throughout this journey is that
- 0:28there are better ways
- 0:29SPEAKER_01: to learn than others.
- 0:31There are some fundamental rules that constrain our learning
- 0:34dictated by how our
- 0:35SPEAKER_01: brain is structured.
- 0:37Understanding these rules to learn topics like math,
- 0:39computer science,
- 0:40and physics greatly
- 0:41help in speeding up the learning process now i got the chance
- 0:44to interview the
- 0:45learning legend justin skysack on a whole lot of topics around
- 0:49learning technical
- 0:51stuff like math efficiently what i like about justin learning
- 0:54journey is that he
- 0:55doesn't come from like an academic background and family his
- 0:59way of learning at
- 1:00place at first was very self-directed and throughout his
- 1:04studies he discovered
- 1:05SPEAKER_01: great ways to learn that are actually grounded in the literature.
- 1:08He applied all of these learning techniques throughout his
- 1:11studies while he was
- 1:13SPEAKER_01: teaching and mentoring students while he was doing research,
- 1:16SPEAKER_01: and right now while he's building AI-driven systems for Math Academy,
- 1:20SPEAKER_01: where he's currently the Chief Quant Director of Analytic.
- 1:23SPEAKER_01: Big title.
- 1:24SPEAKER_01: I've put a whole lot of chapters throughout the interview,
- 1:26SPEAKER_01: so feel free to move around the different segments of the discussion.
- 1:29SPEAKER_01: It's around
- 1:31SPEAKER_01: It's around four hours.
- 1:32SPEAKER_01: So please take a day off from school or work,
- 1:34SPEAKER_01: call in sick,
- 1:35and just brew about like two liter of coffee or tea while you
- 1:39watch this thing.
- 1:41We're going to cover his learning backstory,
- 1:43what cognitive science says about the fastest path
- 1:46to mastery why i think learning math is a lot like training for
- 1:49a sport and it's
- 1:50practical advice for adults kids old people people with adhd
- 1:55researcher we've got
- 1:56it covered for everyone in this interview enjoy folks and big
- 2:00thanks to everyone
- 2:00SPEAKER_01: that contributed the question it really helps
- 2:07SPEAKER_01: Justin, first of all, thank you for being here.
- 2:09SPEAKER_01: Really a pleasure.
- 2:11Can you give us like a high level introduction about like
- 2:15yourself and what you're
- 2:17SPEAKER_01: working at the moment?
- 2:17SPEAKER_00: Yeah, sure.
- 2:18SPEAKER_00: So my name is Justin Skysak.
- 2:20SPEAKER_00: I'm the Chief Quant and Director of Analytics at Math Academy.
- 2:24We're an adaptive online math learning platform,
- 2:28elementary school through university.
- 2:30SPEAKER_00: Our goal is to cover all of math and math-adjacent stuff,
- 2:35SPEAKER_00: and we take a full nerd approach to learning.
- 2:38SPEAKER_00: So just trying to optimize at every moment
- 2:43make sure the student is working on the the learning tasks that
- 2:47are the the most
- 2:47efficient use of their time we're all about learning efficiency
- 2:50and so what uh what
- 2:52SPEAKER_00: i'm
- 2:53kind of working on right now and what we're all working on and
- 2:55has consumed my life
- 2:57SPEAKER_00: for the past several months,
- 3:00SPEAKER_00: if not years,
- 3:01is taking us from a workshop to a factory transition where we
- 3:05can really churn out
- 3:06a lot of amazing content at the same level of quality that we
- 3:11have been churning
- 3:12SPEAKER_00: out for years,
- 3:13SPEAKER_00: trying to floor the gas on this thing.
- 3:15So I used to,
- 3:17you know,
- 3:17in the podcast,
- 3:18we'll get into a lot of stories about like our own
- 3:21probably your learning background my learning background how we
- 3:25found uh kind of
- 3:26where we're at right now uh but nowadays it's like uh it's
- 3:30really uh startup early
- 3:32stage company grind mode and that's just like my life has just
- 3:37been totally
- 3:39engulfed by by this tornado in a good way it brings together a
- 3:42lot of stuff but uh
- 3:43SPEAKER_00: nevertheless like uh there's not a whole lot of uh
- 3:47free time for for for learning just just for fun anymore it's
- 3:51like let's go let's
- 3:52go let's go and like uh um it's i'm uh i find it's the most uh
- 3:58rewarding that sort
- 3:59of pressure when it's for something that actually is working
- 4:04like uh like when
- 4:05SPEAKER_01: you're trying it out and it's nothing is giving you feedback
- 4:08SPEAKER_01: It's a bit harder,
- 4:10SPEAKER_01: but once you have this feedback from the environment that like,
- 4:12SPEAKER_01: yes,
- 4:13SPEAKER_01: it's actually is useful and people are loving it.
- 4:15SPEAKER_01: And like there's growth and all that.
- 4:18SPEAKER_01: It can be high pressure, but still super rewarding.
- 4:21SPEAKER_01: Totally.
- 4:21SPEAKER_00: Well, there's that saying, right?
- 4:22SPEAKER_00: Nothing successful.
- 4:23SPEAKER_00: exceeds like success,
- 4:24SPEAKER_00: which I mean,
- 4:25SPEAKER_00: on the surface sounds kind of dumb,
- 4:27SPEAKER_00: kind of like,
- 4:27SPEAKER_00: well,
- 4:27SPEAKER_00: duh,
- 4:28SPEAKER_00: like,
- 4:28SPEAKER_00: but really,
- 4:28SPEAKER_00: it's like,
- 4:29SPEAKER_00: whether it's building a company,
- 4:32SPEAKER_00: working on a project,
- 4:34SPEAKER_00: learning math,
- 4:35SPEAKER_00: learning anything,
- 4:36SPEAKER_00: it's like,
- 4:36SPEAKER_00: the more the,
- 4:38SPEAKER_00: as you experience the small wins,
- 4:39SPEAKER_00: you get a lot of motivation to continue doing it.
- 4:43SPEAKER_00: And that motivation,
- 4:44SPEAKER_00: it's,
- 4:45SPEAKER_00: at least for me,
- 4:45SPEAKER_00: that's been one of the best sources of motivation is just
- 4:49SPEAKER_00: Like knowing that,
- 4:50SPEAKER_00: hey,
- 4:50this is working and it can work even better if I keep doing it
- 4:53and do it more
- 4:53SPEAKER_00: efficiently and just see where this goes.
- 4:56SPEAKER_01: And you see the path clearly also.
- 4:58SPEAKER_01: Like the environment that you're in is rewarding you in some area.
- 5:04SPEAKER_01: It's selling you where you suck like pretty directly.
- 5:08So like you know what to improve and like why you're doing this
- 5:11stuff and why
- 5:12SPEAKER_01: you're learning this thing in order to do that.
- 5:14other element so talking about like learning and efficiency
- 5:19like i saw in your blog
- 5:20that like you self-studied like 3 000 hours of college math in
- 5:25high school right um
- 5:27which is absolutely kind of just just like even if we remove
- 5:32the math just like
- 5:33this amount of hours i think like even the average high
- 5:36schooler might not go
- 5:38through that amount like first of all like i want to start off
- 5:41but but like
- 5:42where did that passion to like actually do that stuff come from
- 5:46like uh like i saw
- 5:47that your background like you're you're you don't have like a
- 5:51like peers around you
- 5:52that were academic or some sort of a researcher or stuff like
- 5:55that what was the
- 5:56main driver here right so this was kind of like a chance event
- 6:01um and it was kind
- 6:02of like you know there's i'd always liked math class uh growing
- 6:07up in middle school
- 6:08SPEAKER_00: and uh like i i
- 6:10SPEAKER_00: It was my favorite class.
- 6:12SPEAKER_00: Math was my favorite class.
- 6:13SPEAKER_00: But am I going to go learn math outside of class?
- 6:15SPEAKER_00: Like that thought just never really occurred to me.
- 6:18SPEAKER_00: I was not that much into math.
- 6:21SPEAKER_00: But at some point in high school,
- 6:24SPEAKER_00: actually,
- 6:26SPEAKER_00: the summer after 10th grade,
- 6:28SPEAKER_00: I just finished up pre-calculus.
- 6:30SPEAKER_00: And it was a summer.
- 6:32SPEAKER_00: And I was actually getting kind of excited.
- 6:33SPEAKER_00: about learning calculus because calculus is like,
- 6:36SPEAKER_00: you know,
- 6:37when you watch like a sci-fi movie and like there's always a
- 6:39scientist on the
- 6:40SPEAKER_00: board,
- 6:41SPEAKER_00: there's gotta be the integral signs,
- 6:43SPEAKER_00: the derivatives,
- 6:44SPEAKER_00: like that's kinda,
- 6:45SPEAKER_00: I mean,
- 6:46SPEAKER_00: just at a very surface level,
- 6:47SPEAKER_00: it's like what's advanced math for somebody who's,
- 6:50going in through high school or like learning pre-calc or
- 6:54algebra,
- 6:54like it's calculus.
- 6:55SPEAKER_00: Like that's the first thing, at least for me, that came to mind.
- 6:58And so knowing that I was just,
- 7:00you know,
- 7:00hey,
- 7:00I'm going to learn this next year.
- 7:01SPEAKER_00: Like I just finished pre-calculus.
- 7:03SPEAKER_00: Calculus is next.
- 7:04SPEAKER_00: We did a little bit of calculus at the end of pre-calc.
- 7:07SPEAKER_00: And I was just like, I was, I was getting ready to go.
- 7:09SPEAKER_00: I was really interested in like science.
- 7:11SPEAKER_00: Like I knew I was going to go into STEM in some form.
- 7:16SPEAKER_00: And it was just like,
- 7:18SPEAKER_00: I was getting a little antsy.
- 7:19SPEAKER_00: So I'm like,
- 7:20SPEAKER_00: why don't I,
- 7:21SPEAKER_00: like,
- 7:21SPEAKER_00: I don't know how this is going to go,
- 7:22SPEAKER_00: but why don't I just try and teach it myself?
- 7:24SPEAKER_00: Like,
- 7:25SPEAKER_00: what's,
- 7:25SPEAKER_00: what's the way,
- 7:26SPEAKER_00: like,
- 7:26SPEAKER_00: I mean,
- 7:27SPEAKER_00: I guess I'll learn it next year at school,
- 7:28SPEAKER_00: but like,
- 7:28SPEAKER_00: if I could just teach it to myself over the summer,
- 7:32SPEAKER_00: that'd be awesome.
- 7:33SPEAKER_00: So, so it was just this idea.
- 7:35SPEAKER_00: It wasn't, it, and it,
- 7:37SPEAKER_00: Nobody really planted that in my head.
- 7:39SPEAKER_00: It was just one of those things where I was just thinking one day.
- 7:43SPEAKER_00: I was like, why don't I just try and do this?
- 7:46But yeah,
- 7:48that ended up turning into this kind of chance event spark that
- 7:54kind of
- 7:55led to the trajectory that I'm in now and my obsession with
- 8:00efficient learning.
- 8:01And so how it kind of ended up is I found this pretty
- 8:07well-sequenced calculus course.
- 8:09SPEAKER_00: It was almost like a proto math academy type thing.
- 8:13It was in some like online course collection,
- 8:18like the national repository of online courses.
- 8:21SPEAKER_00: And it was like this
- 8:24SPEAKER_00: you'd go in and there'd be like a couple,
- 8:26SPEAKER_00: it was broken up into like maybe 50 or 60 or more lessons,
- 8:32SPEAKER_00: the AP calculus course.
- 8:33SPEAKER_00: And you'd go in,
- 8:34SPEAKER_00: you'd read a little bit on some slide of information,
- 8:39SPEAKER_00: and then you'd solve some problems.
- 8:42SPEAKER_00: And they were very concrete problems.
- 8:44SPEAKER_00: You got a lot of reps and stuff.
- 8:45SPEAKER_00: So it was a great course.
- 8:46SPEAKER_00: Unfortunately, it no longer exists.
- 8:49SPEAKER_00: It was like sold to...
- 8:50somebody and revamped and whatever and it's just i mean i i had
- 8:54to yeah yeah so you
- 8:58can't just like uh go learn it from there anymore but um but
- 9:03this ended up working
- 9:04uh really really well actually way better than i thought it was
- 9:07gonna go i thought
- 9:08i was gonna spend like the whole summer just grinding calculus
- 9:11and and maybe
- 9:13you know,
- 9:13maybe one hour a day and come out and have learned like the
- 9:16first third or something.
- 9:18SPEAKER_00: But actually,
- 9:19SPEAKER_00: like,
- 9:20SPEAKER_00: just like we were talking about success breeds success,
- 9:24SPEAKER_00: like having this small band stack up,
- 9:27SPEAKER_00: like got through lesson one,
- 9:28SPEAKER_00: got through lesson two,
- 9:29SPEAKER_00: solving problems.
- 9:30SPEAKER_00: Hey, this is not a big deal.
- 9:31SPEAKER_00: This is actually, I like it.
- 9:33SPEAKER_00: This is fun.
- 9:33SPEAKER_00: And so it was actually so fun that I was,
- 9:39I used to play more video games,
- 9:41and so math took the place of video games.
- 9:46SPEAKER_00: It almost felt like progressing through levels in a video game.
- 9:53SPEAKER_00: And so I just kind of leaned into it,
- 9:55and I just holed up in my room working on calculus problems for
- 9:59like eight hours a
- 10:00SPEAKER_00: day or something,
- 10:01SPEAKER_00: because I was just obsessed.
- 10:02SPEAKER_00: And so my parents are just like,
- 10:04SPEAKER_00: My mom's an art major in school.
- 10:06SPEAKER_00: My dad's a business major, just logistics for a healthcare company.
- 10:11Like,
- 10:11they don't...
- 10:12And nobody in my family really looks technical or mathematical
- 10:16or anything.
- 10:17SPEAKER_00: So I'm just like...
- 10:19SPEAKER_00: It's like I'm talking with other parents and other...
- 10:21SPEAKER_00: Like,
- 10:21SPEAKER_00: you know,
- 10:22SPEAKER_00: like,
- 10:22SPEAKER_00: what's your kid doing over the summer?
- 10:24SPEAKER_00: Like, oh, he's just, like, in his room, like, learning calculus.
- 10:28SPEAKER_00: It's like...
- 10:28SPEAKER_00: Almost like as if I was, like, in the basement playing a video game.
- 10:31SPEAKER_00: Yeah, exactly.
- 10:32SPEAKER_00: And...
- 10:33and and nobody says like like oh yeah my kid's in a basement
- 10:37playing a video game
- 10:38like eight hours a day of the stuff it's great like did they
- 10:41try to tell you like
- 10:42to go out and like do other stuff like what did what they were
- 10:46just really confused
- 10:47honestly it was uh because it's like i mean it started out like
- 10:52like wow my like my
- 10:53kid is learning like taking in his education into his own hands
- 10:57and he's like uh
- 10:59prepping for the upcoming school year that's great and then
- 11:02just like like wait
- 11:03where is he's still he's still in his room he's just like
- 11:05working out these
- 11:06problems all day like it's a kind of weird and yeah i mean they
- 11:09they had they had
- 11:11never learned this kind of stuff uh themselves and and they and
- 11:15nobody had like
- 11:16stem careers or anything but so it was one of those things that
- 11:19was like well it
- 11:21seems like it's a good thing but it also seems like he's like
- 11:23addicted and obsessed
- 11:24SPEAKER_00: to it but it also seems like a good thing so what do you
- 11:28SPEAKER_00: what do you do?
- 11:28SPEAKER_00: So they,
- 11:29SPEAKER_00: they don't really like intervene or anything,
- 11:33SPEAKER_00: which I think ended up being a,
- 11:34SPEAKER_00: being a good thing.
- 11:35SPEAKER_00: I was very, very unbalanced that summer, but it ended up.
- 11:39SPEAKER_01: Can you,
- 11:40SPEAKER_01: can you walk us through like,
- 11:41SPEAKER_01: I guess 3000 hours,
- 11:43SPEAKER_01: like throughout like this whole journey,
- 11:44SPEAKER_01: like,
- 11:47SPEAKER_00: What did it look like day to day?
- 11:49SPEAKER_00: What did you look like day to day?
- 11:51SPEAKER_00: So the 3,000 hours,
- 11:53SPEAKER_00: so after that summer,
- 11:55SPEAKER_00: I spent like maybe 30 days,
- 11:59SPEAKER_00: like eight hours a day,
- 12:00SPEAKER_00: just like grinding this calculus course.
- 12:01SPEAKER_00: And I was just, by the end of this, I was just so sucked in.
- 12:06SPEAKER_00: Like, I want math.
- 12:08SPEAKER_00: Give me more math.
- 12:08SPEAKER_00: What's next?
- 12:09SPEAKER_00: Like, I can't, it's like, yeah, it just felt like I had found this hack
- 12:17or something like nobody told me like you can learn outside of
- 12:21class and I'm like
- 12:22okay let's do this is working out amazingly well I'm having fun
- 12:25this is going to be
- 12:25great for the future I don't really know exactly what it's
- 12:29going to do but it seems
- 12:30like it's going to do good things and so I was like okay so
- 12:33let's keep this going
- 12:33let's learn what's after calculus I finished one month into the
- 12:38summer and I'm like
- 12:39okay I don't know what's after calculus I need to look up
- 12:42online like what course
- 12:43SPEAKER_00: comes after calculus what's
- 12:45SPEAKER_00: math above that.
- 12:46SPEAKER_00: Is there math above that?
- 12:47SPEAKER_00: I'm sure there's math above that, but I don't know what it is.
- 12:50SPEAKER_00: So I stumbled into MIT OpenCourseWare because that has a nice,
- 12:55SPEAKER_00: like,
- 12:55SPEAKER_00: you know,
- 12:55SPEAKER_00: a sequencing of courses.
- 12:57SPEAKER_00: It's like, well, you just, what comes after calculus?
- 13:00SPEAKER_00: Well, just look at the MIT undergrad math major.
- 13:02SPEAKER_00: That kind of tells you what to do.
- 13:04And so that's when I started learning like linear algebra and
- 13:08multivariable calculus.
- 13:11SPEAKER_00: And I would just work through the,
- 13:13SPEAKER_00: I mean,
- 13:13it was much less structured than this original AP calculus
- 13:17course that I had found.
- 13:19SPEAKER_00: because MIT OpenCourseWare,
- 13:20SPEAKER_00: at the time,
- 13:21that was really just a collection of lecture videos and some
- 13:25problem sets.
- 13:26SPEAKER_00: Now,
- 13:26I know they've built out some of their foundational courses a
- 13:29little more since
- 13:30SPEAKER_00: then,
- 13:31SPEAKER_00: augmented it with more learning materials,
- 13:33SPEAKER_00: but at the time,
- 13:34SPEAKER_00: it was very,
- 13:35SPEAKER_00: very rough writing.
- 13:37SPEAKER_00: Like, here's, like, 20 lectures, and here's all the problem sets, go.
- 13:43SPEAKER_00: And so I just did that, and I found...
- 13:47the some pdfs of the of the textbooks and uh yeah i was just
- 13:51doing that so i uh did
- 13:54that the rest of summer into the school year when i got to
- 13:56school uh that that
- 13:58SPEAKER_00: following uh like 11th grade i was just you know um this whole self-study
- 14:04experience was making me realize how much time i was wasting in
- 14:09school just kind of
- 14:11SPEAKER_00: sitting there in class and uh
- 14:14SPEAKER_00: And so I was like, well, I'm not going to go back to that.
- 14:16SPEAKER_00: I'm going to keep leaning into this thing.
- 14:18So I would take the math books with me to like all my classes
- 14:21and just be doing
- 14:22SPEAKER_00: that instead while trying to look like I'm paying attention and stuff,
- 14:28SPEAKER_00: not getting confiscated and stuff like that.
- 14:31SPEAKER_00: But, yes, I kept that going the following year.
- 14:34SPEAKER_00: And, yeah, you know, it was this 3,000 hours thing.
- 14:41It was way more time than I really needed if I were working
- 14:44efficiently on learning this stuff.
- 14:46SPEAKER_00: But I was, I mean, as most teenagers, like, I think I know everything.
- 14:50SPEAKER_00: I know how to learn, like...
- 14:52SPEAKER_00: I did not do my research on how to learn efficiently.
- 14:55SPEAKER_00: I'm just like, okay, I'm going to go do it.
- 14:56And so I ended up hitting my head on basically every sort of
- 15:00ledge there was to hit
- 15:01SPEAKER_00: my head on along the way.
- 15:03SPEAKER_00: Like I kind of just discovering a lot of these principles,
- 15:08SPEAKER_00: like in hindsight,
- 15:10SPEAKER_00: really obvious principles of learning science,
- 15:12SPEAKER_00: but like...
- 15:14you know there was i remember there was like a a day or so that
- 15:17i was that i
- 15:19initially came into it thinking like you know like why am i
- 15:22working all these
- 15:23problems i can just watch all the videos and learn it that way
- 15:26and then so i just
- 15:27like took a day and i just like blew through like like what i
- 15:31thought was like a
- 15:32third of the of the course like watching videos i'm like yeah
- 15:36this is such a hack
- 15:37SPEAKER_00: like i can
- 15:38SPEAKER_00: I don't have to do the problems.
- 15:39SPEAKER_00: I could just do the videos.
- 15:39SPEAKER_00: And then I started like the following day,
- 15:41SPEAKER_00: just try to like work out some problems and realize like,
- 15:44SPEAKER_00: I didn't,
- 15:44SPEAKER_00: I didn't retain anything from it.
- 15:46SPEAKER_00: And so that was just the learning of like, okay, I have to do problems.
- 15:50SPEAKER_00: And so every,
- 15:52SPEAKER_00: basically every single cognitive science principle,
- 15:55SPEAKER_00: um,
- 15:55SPEAKER_00: I,
- 15:56SPEAKER_00: I violated initially during my self study and,
- 15:59SPEAKER_00: and kind of learn from that.
- 16:01SPEAKER_00: And yeah.
- 16:02SPEAKER_00: Yeah.
- 16:02SPEAKER_01: But I mean,
- 16:04SPEAKER_01: like you said,
- 16:05SPEAKER_01: you could compress it to like 70,
- 16:08SPEAKER_01: 750 hours,
- 16:10SPEAKER_01: something like that,
- 16:11SPEAKER_01: along with a good design system.
- 16:13But don't you think like that this other chunk actually was
- 16:18about the meta learning game?
- 16:20SPEAKER_01: And then this chunk, the 700 hours,
- 16:23SPEAKER_01: was about the math skill.
- 16:24SPEAKER_01: Because like, I don't know, because I went to the same thing, right?
- 16:29SPEAKER_01: I was doing like, I was really into learning at the intense level.
- 16:35SPEAKER_01: And I felt like this was the kind of,
- 16:39uh meta-learning that you have to like see for yourself a bit
- 16:44right a bit like um i
- 16:45don't know um in sport or art like people could tell you the
- 16:50stuff right and say
- 16:51SPEAKER_01: like you should do it this way and stuff like that but then
- 16:54you will not understand what they're saying until you actually
- 16:57experience it.
- 16:58SPEAKER_01: And then you're like, yeah, actually, yeah, I need to sleep, right?
- 17:00SPEAKER_01: Or like, I should stay hydrated, otherwise, like nothing works, right?
- 17:04SPEAKER_01: That sort of stuff, right?
- 17:05SPEAKER_01: Like this thing around the learning, which is the core, right?
- 17:09And after that,
- 17:10you can apply this kind of skill that you learn in other stuff,
- 17:15right?
- 17:16SPEAKER_01: Because like,
- 17:16SPEAKER_01: the principles are kind of the same all across,
- 17:19SPEAKER_01: and you're trying to map the principle
- 17:21to like the thing that you have and then the thing that you
- 17:24want to learn.
- 17:26Do you think there's any of this or like you could have learned
- 17:29also the middle
- 17:30SPEAKER_01: learning skill a bit faster?
- 17:31SPEAKER_00: You know, so I mean, on one hand, I think there was a
- 17:38a little bit of value to having tried some inefficient forms of
- 17:43of learning and
- 17:44only to realize like hey this doesn't work and i think um in
- 17:48any discipline that
- 17:49you are like seriously going into that's going to be like your
- 17:52main this thing that
- 17:54you're doing it's worth exploring the space of uh you know what
- 17:59just try things out
- 18:00SPEAKER_00: what works what doesn't but i just think about that
- 18:04spending 3 000 hours to um you know get from from calculus all
- 18:10the way up there i
- 18:11think i i i got through about like real analysis abstract
- 18:14algebra you know junior
- 18:15level um math major courses and also some physics and stuff if
- 18:20i had um you know
- 18:21there was just a lot of time that i'd wasted just kind of being
- 18:24confused doing
- 18:25SPEAKER_00: inefficient techniques that did not
- 18:28SPEAKER_00: contribute to this holistic view of learning.
- 18:32A little bit,
- 18:34a little taste of like,
- 18:35you know,
- 18:35did it wrong,
- 18:36okay,
- 18:37correct to this.
- 18:39SPEAKER_00: There's maybe some small amount,
- 18:40SPEAKER_00: maybe like a couple percent of the time,
- 18:44SPEAKER_00: I would say,
- 18:45SPEAKER_00: was like kind of useful in exploring the space.
- 18:49But there came to a point where I was really like,
- 18:52I just want to go fast and learn this stuff.
- 18:55SPEAKER_00: And it's like,
- 18:57SPEAKER_00: Trudging through mud.
- 18:58SPEAKER_00: And the higher I got,
- 19:00SPEAKER_00: the more things would come back to bite me,
- 19:03like not doing a proper review system and just working on
- 19:08problems that were like
- 19:09SPEAKER_00: way too hard.
- 19:10SPEAKER_00: A lot of this stuff,
- 19:12SPEAKER_00: I mean,
- 19:13SPEAKER_00: some stuff I figured out on the fly,
- 19:14SPEAKER_00: but a lot of it I really did not learn until...
- 19:18years later how to do this really really efficiently and if
- 19:22there's one thing that
- 19:23i wish that i had more during um my my my teenage years and uh
- 19:29just like as a
- 19:30student would have been guidance on just how just tell me just
- 19:34tell me how to do it
- 19:35like i just want to learn math and i'm fighting my way through
- 19:39hitting my head on
- 19:41SPEAKER_00: all these ledges maybe some amount a small amount of that is is um
- 19:46kind of shaping my view of like understanding what works and
- 19:49what doesn't.
- 19:50SPEAKER_00: But most,
- 19:51SPEAKER_00: the vast majority of that time,
- 19:53SPEAKER_00: for me at least,
- 19:54SPEAKER_00: felt like it was just,
- 19:55SPEAKER_00: it had just gone to waste.
- 19:57SPEAKER_00: And yeah,
- 19:58SPEAKER_00: there was some instances of like,
- 20:01SPEAKER_00: I also,
- 20:02SPEAKER_00: like part of my goal in this was like,
- 20:04SPEAKER_00: I want to like just learn
- 20:06the core body of math so I can contribute to like do some cool
- 20:10research projects.
- 20:12SPEAKER_00: And,
- 20:13SPEAKER_00: um,
- 20:13SPEAKER_00: there came a point like around,
- 20:14SPEAKER_00: you know,
- 20:15SPEAKER_00: uh,
- 20:16SPEAKER_00: maybe like 2000 hours in or so,
- 20:18SPEAKER_00: I'm just like,
- 20:19SPEAKER_00: you know,
- 20:19SPEAKER_00: I think I've learned enough,
- 20:21SPEAKER_00: uh,
- 20:21SPEAKER_00: and I'm going to do research.
- 20:22SPEAKER_00: And what was this based on?
- 20:23This was just based on the sheer volume of time that I had put
- 20:27into learning,
- 20:28SPEAKER_00: not based on how far I had gone.
- 20:30SPEAKER_00: And so I ended up doing a research project, which was, um,
- 20:34SPEAKER_00: Just kind of like on my own.
- 20:36It had to do with finding a generalized formula for partial
- 20:40fractions decompositions.
- 20:41SPEAKER_00: And I spent like a whole summer on it.
- 20:43SPEAKER_00: And it was kind of fun.
- 20:44SPEAKER_00: But...
- 20:47What ended up happening is I was kind of groping around in the
- 20:50dark in a direction
- 20:52that I would later find out was just made completely trivial by
- 20:56some results in
- 20:57SPEAKER_00: complex analysis.
- 20:58SPEAKER_00: And complex analysis,
- 20:59SPEAKER_00: I mean,
- 21:00SPEAKER_00: it's a high-level undergrad math thing,
- 21:02SPEAKER_00: but it's not exactly super advanced math.
- 21:04SPEAKER_00: And so it's like, well, if I just...
- 21:07SPEAKER_01: They were trying to rediscover stuff that were already known.
- 21:10SPEAKER_01: Exactly.
- 21:11If you just had like went and then looked at it and you're
- 21:13like,
- 21:13yeah,
- 21:14that's the shape of it.
- 21:15SPEAKER_01: And then you use it like that.
- 21:16SPEAKER_01: And now I understand the setup.
- 21:19SPEAKER_01: Okay.
- 21:19SPEAKER_01: I get the,
- 21:20SPEAKER_01: I get your,
- 21:21SPEAKER_01: your,
- 21:22your thought process here a lot because it seems that you were
- 21:26like almost like
- 21:29SPEAKER_01: discovering the whole cognitive science field on your own.
- 21:35SPEAKER_01: And then, like, try to figure out how to learn on your own.
- 21:39And you,
- 21:39like,
- 21:40mostly rediscovered,
- 21:41like,
- 21:42concepts that were already proven.
- 21:45So,
- 21:45like,
- 21:45did you got into cognitive science at that time or,
- 21:48like,
- 21:48afterward?
- 21:49SPEAKER_01: Like, what happened here?
- 21:52SPEAKER_00: So cognitive science actually was not really on my radar until...
- 21:57SPEAKER_00: Well,
- 21:58SPEAKER_00: I guess after high school,
- 22:00SPEAKER_00: sort of at the beginning of college,
- 22:02SPEAKER_00: I got sort of interested in neuroscience,
- 22:04SPEAKER_00: computational neuroscience,
- 22:06SPEAKER_00: and that led to kind of interest in AI and stuff.
- 22:09SPEAKER_00: And that, I mean, it was kind of like cognitive science adjacent.
- 22:15SPEAKER_00: But it was really,
- 22:16what really kind of flipped the switch for me was once I was
- 22:20responsible for other
- 22:22SPEAKER_00: people's learning.
- 22:23SPEAKER_00: So starting out...
- 22:25SPEAKER_00: You know,
- 22:26I did a lot of tutoring in high school also and got I really
- 22:30liked tutoring and
- 22:31SPEAKER_00: just math education in general.
- 22:32So at one point,
- 22:34I actually during college,
- 22:36I I had a full time data science job.
- 22:39SPEAKER_00: And after graduating college, I was like, you know, screw it.
- 22:42SPEAKER_00: We're not doing data science anymore.
- 22:43SPEAKER_00: We're doing math education and we're going to.
- 22:47find some way to incorporate all of the coding and whatever
- 22:51that I had enjoyed into
- 22:53SPEAKER_00: math education.
- 22:56Once I had a classroom of students who were depending on me to
- 23:01set up some kind of
- 23:02SPEAKER_00: structure that is going to result in them learning this material,
- 23:08SPEAKER_00: that's when I realized a lot of things that were
- 23:13SPEAKER_00: some of the less obvious stuff,
- 23:15SPEAKER_00: like the importance of,
- 23:17SPEAKER_00: you know,
- 23:18SPEAKER_00: spaced review and interleaving and stuff.
- 23:24And so that's about the time that I also met Jason Roberts,
- 23:29founder of Math Academy.
- 23:31SPEAKER_00: He and his wife, Sandy, founded it together.
- 23:32SPEAKER_00: And
- 23:34SPEAKER_00: And so I was teaching in the original school program there.
- 23:39And so he had done a lot of just reading around the subject of
- 23:42cognitive science.
- 23:43SPEAKER_00: It was incorporating those principles.
- 23:45SPEAKER_00: And so I was like, oh, wow, this is really interesting.
- 23:50For once,
- 23:50like,
- 23:51it's like,
- 23:51okay,
- 23:51this is the guidance that I was like,
- 23:53that I didn't have.
- 23:54SPEAKER_00: I was like groping around, discovered maybe like a third of it on my own.
- 23:58And it was like two thirds,
- 24:00at least that was just like,
- 24:01oh my goodness,
- 24:02that makes total sense.
- 24:03SPEAKER_00: I like, okay, let's do that, do that, do that.
- 24:06SPEAKER_00: Yeah.
- 24:09SPEAKER_00: And it was working in the classroom, too.
- 24:12SPEAKER_00: The more that we would leverage these cognitive science principles,
- 24:16SPEAKER_00: the better the students would learn the material.
- 24:21And then at some point after that,
- 24:25actually around maybe like three years ago or so,
- 24:28I had
- 24:30jason was like hey we need to write up about like how the
- 24:33science of math academy
- 24:34works can you just make a little like some information for the
- 24:37web page and i
- 24:39started on that and uh looked up like about 400 pages later we
- 24:44had the math academy
- 24:44way this big book and how we apply the cognitive science
- 24:48learning principles this
- 24:49is cool and that was yeah i think like the when you're teaching
- 24:53the meta learning
- 24:54SPEAKER_01: skills become more obvious to you because like
- 24:58SPEAKER_01: That's the thing you're doing.
- 24:59SPEAKER_01: You're trying to make this other thing learn.
- 25:04SPEAKER_01: It's not really about the skill anymore, the math, biology.
- 25:12This is just something that goes into the scaffold that you're
- 25:15going to go and then
- 25:17SPEAKER_01: make the student learn.
- 25:20SPEAKER_01: Because there's two things in my view.
- 25:23However skilled you are at the thing that you're teaching,
- 25:26if you don't know how to
- 25:28SPEAKER_01: like learning is happening, you will have a hard time teaching.
- 25:32SPEAKER_01: Like I've seen like a research level professor,
- 25:38SPEAKER_01: like being completely useless in the classroom,
- 25:40SPEAKER_01: like actually,
- 25:41SPEAKER_01: actually useless.
- 25:43SPEAKER_01: They will, they will, we're doing more arm than good.
- 25:46SPEAKER_01: I know that because I was tutoring this class and I was like,
- 25:48SPEAKER_01: what the,
- 25:49SPEAKER_01: what is happening,
- 25:50SPEAKER_01: right?
- 25:51And then the second thing is that like the learning is not
- 25:54coming from the teacher
- 25:55SPEAKER_01: to the student.
- 25:56SPEAKER_01: It is emerging from the student.
- 25:58SPEAKER_01: So you have to make this emerge.
- 26:00SPEAKER_01: Right.
- 26:00SPEAKER_01: So like it's not like you're going to hammer them up with like whatever.
- 26:05SPEAKER_01: And then you just get bulldoze them.
- 26:06SPEAKER_01: If they don't want to learn, they want they won't learn.
- 26:08SPEAKER_01: They absolutely will not learn.
- 26:10SPEAKER_01: So there's these two components that comes in,
- 26:12SPEAKER_01: and I think the whole field of science and psychology,
- 26:16SPEAKER_01: neuroscience,
- 26:18they add plenty of time to figure out how to make this thing
- 26:21more efficient.
- 26:24So it's kind of worthless to spend time rediscovering stuff
- 26:29that is already working.
- 26:33SPEAKER_01: I wanted to ask you, sorry, go ahead.
- 26:36SPEAKER_00: Oh,
- 26:36SPEAKER_00: yeah,
- 26:36SPEAKER_00: I was just going to say,
- 26:38SPEAKER_00: yeah,
- 26:38SPEAKER_00: I totally agree with,
- 26:40SPEAKER_00: I've had that same experience with,
- 26:41SPEAKER_00: you know,
- 26:42SPEAKER_00: high-level professors.
- 26:43SPEAKER_00: Like,
- 26:44SPEAKER_00: you kind of,
- 26:45SPEAKER_00: right,
- 26:45SPEAKER_00: you sign up for the class,
- 26:46SPEAKER_00: you're like,
- 26:46SPEAKER_00: wow,
- 26:46SPEAKER_01: this professor's,
- 26:47SPEAKER_01: like,
- 26:48SPEAKER_00: really good.
- 26:48good at research like i'm gonna learn so much and then you you
- 26:52get in there and and
- 26:53then you you get served up these like these slam dunk problems
- 26:59that are like
- 26:59supposed to the professor thinks like oh i'm gonna create these
- 27:02like really just a
- 27:03few really nice like really interesting problems and give them
- 27:07to the students and
- 27:07you get those problems you're like what is this like i don't
- 27:10even know where to
- 27:11SPEAKER_00: begin and you just are flailing around and
- 27:14And the professor is,
- 27:16you know,
- 27:16not really helping you bridge the gap from you to them.
- 27:20SPEAKER_00: And I've seen this, too, in a lot of students I tutored as well.
- 27:23SPEAKER_00: Yeah, it's really unfortunate.
- 27:24SPEAKER_01: And I've seen it a lot in engineering.
- 27:27SPEAKER_01: Like, it was a plague.
- 27:30SPEAKER_01: I was tutoring a lot of engineers.
- 27:33SPEAKER_01: Sorry, student engineers.
- 27:35And,
- 27:35like,
- 27:36for some reason,
- 27:36there was this kind of elitism in engineering where
- 27:42it was okay if the class,
- 27:45some part of the class was not understanding anything.
- 27:48And like they had this kind of weeding out of the people and
- 27:53stuff mentality,
- 27:55SPEAKER_01: which in my view was kind of absolutely nonsensical.
- 27:58Like it is possible that all the class can get 100% in your
- 28:02thing if you're able to
- 28:04SPEAKER_01: teach it properly.
- 28:05SPEAKER_01: Like it is possible.
- 28:07It should be something that is in your mind and that you should
- 28:09be able to kind of
- 28:11like put yourself back into a question whether like what you're
- 28:14teaching is
- 28:15SPEAKER_01: actually good or not.
- 28:16SPEAKER_01: Right.
- 28:18SPEAKER_00: But yeah, we just kind of accepted that, right.
- 28:21The class moves in lockstep and there's a normal distribution
- 28:24of grades at the end.
- 28:26SPEAKER_00: And some of these grades are failing.
- 28:28SPEAKER_00: Some of them are like,
- 28:29just barely understands the material a few grades are like okay
- 28:33they they get it we
- 28:34just kind of accepted this the system but really yeah if you if
- 28:37you have um you
- 28:38know you have every student working on what is the most
- 28:42efficient use of their time
- 28:43um that's the thing like if you look in this sort of system the
- 28:49one that are
- 28:50SPEAKER_01: outlier and they like the the teacher is like look at these get like
- 28:55SPEAKER_01: It could be any teacher.
- 28:56They would have done this because they're studying on their own
- 28:59and they figured that out.
- 29:00SPEAKER_01: Right.
- 29:01SPEAKER_01: And I know that because most of the time it was me.
- 29:03SPEAKER_01: Like I was like going into these classes and I was like,
- 29:07SPEAKER_01: this teacher doesn't know what he's doing.
- 29:09SPEAKER_01: So I cannot just listen.
- 29:11SPEAKER_01: I'm wasting time.
- 29:12SPEAKER_01: So it was like doing the textbook.
- 29:14I was like getting my hands on like any past exam that was
- 29:18legally possible.
- 29:19SPEAKER_01: Any like hint of a problem I could solve.
- 29:23And I was drilling them and then drilling the teacher to show
- 29:26me what he would have
- 29:27SPEAKER_01: done in this case if it was something that you had to work through.
- 29:31And then I was doing okay because I was not just listening to
- 29:34these guys because
- 29:36SPEAKER_01: there was nothing of value here.
- 29:38SPEAKER_01: I was listening to the material,
- 29:39SPEAKER_01: learning the material,
- 29:40SPEAKER_01: and then trying to apply it to how they were grading it,
- 29:43SPEAKER_01: which is totally different than what everybody else was doing.
- 29:46SPEAKER_01: They were waiting for the material to kind of be taught,
- 29:49SPEAKER_01: but it was not taught in any shape or form.
- 29:51Okay,
- 29:52wait,
- 29:52I had a question about your research,
- 29:54because you did some research as a teenager.
- 29:58And I wanted to understand,
- 30:00like,
- 30:01the fundamental math stuff you were doing,
- 30:03right?
- 30:03SPEAKER_01: And the base that you set up,
- 30:05SPEAKER_01: like,
- 30:05SPEAKER_01: did it help,
- 30:07SPEAKER_01: like,
- 30:07SPEAKER_01: push through those research projects,
- 30:09SPEAKER_01: like,
- 30:09SPEAKER_01: especially at the Fermilab and the CERN project you were doing?
- 30:12SPEAKER_00: Yeah, so I'll talk a bit about that.
- 30:14So the Fermilab and CERN projects,
- 30:16these were projects that I did with local universities.
- 30:20SPEAKER_00: Fermilab was with...
- 30:22Indiana University,
- 30:23South Bend,
- 30:25where I grew up,
- 30:26and in the professor Elon Levine's lab.
- 30:30SPEAKER_00: And then CERN project was with a lab at Notre Dame.
- 30:34SPEAKER_00: That's QuarkNet program,
- 30:36which connected like some of the top high school students in
- 30:39the areas with working
- 30:41SPEAKER_00: on projects that were going on with some labs there at Notre Dame.
- 30:46SPEAKER_00: And so the math, it was kind of funny.
- 30:51It helped in a lot of big ways get my foot in the door in terms
- 30:56of these projects
- 30:58SPEAKER_00: because,
- 30:59SPEAKER_00: you know,
- 31:00SPEAKER_00: it's you learn a lot.
- 31:01SPEAKER_00: a bunch of advanced math,
- 31:02SPEAKER_00: you're instantly,
- 31:02SPEAKER_00: you stand out as,
- 31:04SPEAKER_00: in my high school,
- 31:05SPEAKER_00: as like,
- 31:06SPEAKER_00: okay,
- 31:07SPEAKER_00: who are we going to pick?
- 31:08They could pick one student to send to this QuarkNet program at
- 31:12Notre Dame.
- 31:12SPEAKER_00: And they're like, well, who are we going to pick?
- 31:14SPEAKER_00: Well,
- 31:15SPEAKER_00: Justin is,
- 31:16SPEAKER_00: like,
- 31:16SPEAKER_00: just crushing the,
- 31:17SPEAKER_00: like,
- 31:18SPEAKER_00: calculus course as,
- 31:20SPEAKER_00: like,
- 31:20SPEAKER_00: a grade early or whatever.
- 31:22Like,
- 31:23I guess,
- 31:23and he's,
- 31:24like,
- 31:24seems to be,
- 31:24like,
- 31:25learning physics on his own.
- 31:26SPEAKER_00: Like, I don't know.
- 31:26SPEAKER_00: Send him.
- 31:27SPEAKER_00: So that was...
- 31:29SPEAKER_00: Yeah, it was just an obvious decision.
- 31:31SPEAKER_00: Because that's what happens when you get,
- 31:32SPEAKER_00: like,
- 31:32SPEAKER_00: really,
- 31:32SPEAKER_00: like,
- 31:33SPEAKER_00: you know,
- 31:33SPEAKER_00: you find your obsession,
- 31:35SPEAKER_00: you get really ahead.
- 31:36People take notice,
- 31:37and you get a lot of opportunities to pursue that further.
- 31:42SPEAKER_00: And with the other project, so my junior year, I was...
- 31:48SPEAKER_00: working on the Fermilab project.
- 31:50SPEAKER_00: And that was actually just,
- 31:52I was taking this research class in my high school where the
- 31:57point was to reach out
- 31:58to local professors and just try to get your foot in the door
- 32:01working on a research
- 32:02SPEAKER_00: project with them.
- 32:03SPEAKER_00: And so part of the thing that, you know, made me seem like, hey,
- 32:10SPEAKER_00: this kid is,
- 32:11is gonna actually be able to help contribute a little and not
- 32:15just be a total waste
- 32:16SPEAKER_00: of time was the fact that I had,
- 32:18SPEAKER_00: like,
- 32:19SPEAKER_00: I already had my math foundations in place for all this stuff.
- 32:21SPEAKER_00: And I'd taken it upon myself to, to learn.
- 32:24SPEAKER_00: And so when I,
- 32:26SPEAKER_00: when I got to these,
- 32:28SPEAKER_00: um,
- 32:28SPEAKER_00: you know,
- 32:29SPEAKER_00: started working,
- 32:30SPEAKER_00: uh,
- 32:30SPEAKER_00: with these labs,
- 32:31SPEAKER_00: it was,
- 32:31SPEAKER_00: you know,
- 32:32SPEAKER_00: math was just a non issue.
- 32:34I remember for so many,
- 32:35for so many other students in,
- 32:37uh,
- 32:37this,
- 32:37this,
- 32:38um,
- 32:39SPEAKER_00: in my high school research class,
- 32:40SPEAKER_00: like they would go off to their,
- 32:42SPEAKER_00: their labs as well during the school day.
- 32:45SPEAKER_00: And,
- 32:45SPEAKER_00: uh,
- 32:45SPEAKER_00: just for like,
- 32:46SPEAKER_00: uh,
- 32:47SPEAKER_00: at the end of the school day for an hour is the last period of school.
- 32:50SPEAKER_00: So you'd go out and leave early.
- 32:53SPEAKER_00: Um,
- 32:53SPEAKER_00: so a lot of other people would,
- 32:55SPEAKER_00: would be struggling with some of the math involved sometimes,
- 32:58SPEAKER_00: you know,
- 32:58SPEAKER_00: like they,
- 32:58SPEAKER_00: maybe somebody would,
- 32:59SPEAKER_00: would get into a physics lab and,
- 33:02They would not only have to spin up on the context that the lab
- 33:05is operating in,
- 33:06SPEAKER_00: like what research are we doing?
- 33:08SPEAKER_00: What are the materials?
- 33:08But like there's just a big mathematical gap that was placing
- 33:13hard limits on this
- 33:14SPEAKER_00: type of stuff that they could do.
- 33:16SPEAKER_00: Now, I didn't have that gap.
- 33:18SPEAKER_00: And so that was good.
- 33:20SPEAKER_00: Though I did have some gaps in other areas.
- 33:22SPEAKER_00: Like, I mean, most of my projects had to do with experimental physics.
- 33:27SPEAKER_00: So there was like, you know, math was one side of it.
- 33:30SPEAKER_00: But there was also,
- 33:31SPEAKER_00: you know,
- 33:31SPEAKER_00: in the Fermilab project,
- 33:32SPEAKER_00: I was making a,
- 33:33SPEAKER_00: trying to make a material to better transmit sound.
- 33:36and i had never like i was not a very hands-on lab kind of
- 33:41person so i that was a
- 33:43that was definitely a point of like oh man this is challenging
- 33:46but in in in terms
- 33:48of uh you know math just became a a non-problem and um and i i
- 33:53i think um to anyone
- 33:55who who wants to do uh research especially high school research
- 34:00it's like removing
- 34:01SPEAKER_00: skill gaps is such a big um
- 34:05advantage there because you can you know just imagine like say
- 34:08there's some lab
- 34:09SPEAKER_00: it's working in some setting some research setting if you have all your
- 34:12foundational skills in place then you can just like take off
- 34:17running and be a a
- 34:19serious contributing member of lab every every little pillar of
- 34:24foundational skills
- 34:25that you don't have in place is gonna go is gonna slow you down
- 34:29so so I would say
- 34:30in my case like okay the math was a pillar that I just knocked
- 34:33out of the park that
- 34:33SPEAKER_00: was like
- 34:35a non-issue some of the more hands-on experimental physics
- 34:38stuff uh was was
- 34:39definitely a more of a more of a drag factor for me because i'd
- 34:42never um trained up
- 34:44you know that that kind of hands-on but this is super
- 34:47interesting and like did you
- 34:49improve in like uh this experiment experimental physics like a
- 34:53manipulation type of
- 34:54stuff you had to do during that did you try to like learn it
- 34:58and train with it like
- 34:59SPEAKER_01: uh well how did it look like that part
- 35:02SPEAKER_00: You know,
- 35:02SPEAKER_00: so I,
- 35:03SPEAKER_00: I could have leaned into this a lot more so that I,
- 35:06SPEAKER_00: I came in wanting to do theoretical physics.
- 35:09So I was like,
- 35:10I don't want to do experimental physics,
- 35:12but I want to do physics.
- 35:13SPEAKER_00: And like,
- 35:13SPEAKER_00: I mean,
- 35:14SPEAKER_00: you're a high schooler,
- 35:15SPEAKER_00: you're like,
- 35:15SPEAKER_00: Hey,
- 35:15SPEAKER_00: can I get involved in this physics stuff?
- 35:17SPEAKER_00: Like, then they're like, yeah, you'll take whatever.
- 35:20SPEAKER_00: Yeah.
- 35:21SPEAKER_00: They're like,
- 35:21SPEAKER_00: and I'm like,
- 35:22SPEAKER_00: like,
- 35:23I'm not about to like turn down experimental physics
- 35:25opportunity to say like,
- 35:26SPEAKER_00: no,
- 35:27SPEAKER_00: I want to do theoretical like that or nothing.
- 35:29SPEAKER_00: And they're just like, get out of here, kid.
- 35:31SPEAKER_00: What are you talking about?
- 35:34SPEAKER_00: So I was a little like,
- 35:36SPEAKER_00: you know,
- 35:36doing these – I had not quite found my really good fit with
- 35:42these physics projects.
- 35:44SPEAKER_00: It was like a halfway good fit, right?
- 35:45SPEAKER_00: Because I had like that math pillar, like, okay, that was good.
- 35:48SPEAKER_00: But the experimental physics pillar, I was just –
- 35:51SPEAKER_00: You know, I didn't, I wasn't really interested in that kind of stuff.
- 35:56And yeah,
- 35:57I did not take it upon myself to learn this stuff outside a
- 36:01lab.
- 36:02And so it's,
- 36:03it was kind of like,
- 36:04I just did kind of,
- 36:05you know,
- 36:06run of the mill projects there.
- 36:09SPEAKER_00: And this was enough to,
- 36:11SPEAKER_00: I ended up making it a junior year to the Intel International Science and
- 36:15SPEAKER_00: Engineering Theater.
- 36:16SPEAKER_00: And so this project was,
- 36:18SPEAKER_00: it was like,
- 36:19SPEAKER_00: it was good enough to get to that level,
- 36:21SPEAKER_00: but it was not like good enough to like,
- 36:23SPEAKER_00: you know,
- 36:23SPEAKER_00: get place,
- 36:25SPEAKER_00: like get a medal or result in a,
- 36:28SPEAKER_00: like a paper or,
- 36:29SPEAKER_00: you know,
- 36:30SPEAKER_00: stuff like that.
- 36:31SPEAKER_00: Or even like kind of really set a direction for me to like,
- 36:34SPEAKER_00: I want to continue like looking into this more.
- 36:37SPEAKER_01: To be honest,
- 36:37like there's a whole lot of luck involved in like when you are
- 36:43a junior and then
- 36:44SPEAKER_01: you enter in the lab,
- 36:45SPEAKER_01: And then like you get a paper out of it,
- 36:47SPEAKER_01: it's kind of,
- 36:48SPEAKER_01: yeah,
- 36:48SPEAKER_01: you got in there at the right time.
- 36:50SPEAKER_01: I know that because my first paper was that like,
- 36:54SPEAKER_01: I got into a molecular biology lab,
- 36:56SPEAKER_01: no,
- 36:56SPEAKER_01: no experience at all.
- 36:58SPEAKER_01: Zero.
- 36:58SPEAKER_01: Right.
- 36:59I was,
- 36:59I was not even like in,
- 37:00in the,
- 37:01in molecular or like cell stuff yet.
- 37:03SPEAKER_01: That was more in the math physics stuff.
- 37:06When I came into the lab,
- 37:08I just wanted to,
- 37:08to work on like a learning memory.
- 37:11SPEAKER_01: and on the molecular level.
- 37:13SPEAKER_01: And I found a lab that was super interesting, read the paper.
- 37:15SPEAKER_01: And I was like, can I just help?
- 37:18SPEAKER_01: And then they were like, yeah.
- 37:19SPEAKER_01: So I got in.
- 37:20SPEAKER_01: And to be honest,
- 37:21I think I wasted them so much dollar by just fucking up their
- 37:26recipe again and
- 37:27SPEAKER_01: again.
- 37:27SPEAKER_01: It was crazy.
- 37:29SPEAKER_01: There was like so many steps.
- 37:31I had to splice genes,
- 37:33like to change them,
- 37:34then had to go make them sequence and stuff.
- 37:36SPEAKER_01: And then I had to like grow the bacteria and then like,
- 37:39SPEAKER_01: All this, I was messing up like on the daily, right?
- 37:43And then I was like,
- 37:45I can't just mess up their old thing and then like drop out of
- 37:49this.
- 37:50SPEAKER_01: So there was this like lab assistant that was there.
- 37:53SPEAKER_01: She was super old.
- 37:54SPEAKER_01: And I was just like asking her all the questions.
- 37:58SPEAKER_01: And it turned out at the end, I got some good results.
- 38:01SPEAKER_01: And I got into a big paper.
- 38:03SPEAKER_01: My thing was in their gigantic paper that they already was 95% done.
- 38:09SPEAKER_01: So by chance, I got in there.
- 38:11And my first publication was Journal of Biochemistry or
- 38:15something like that.
- 38:17To be 100% honest,
- 38:18I didn't understand like,
- 38:20I would say like 30% of what they were doing.
- 38:24SPEAKER_01: I understand my part and like how it kind of fitted.
- 38:27SPEAKER_01: But the whole story was like, this is that bit beyond me.
- 38:29SPEAKER_01: That's interesting.
- 38:30SPEAKER_01: I actually have a question for you on that.
- 38:33SPEAKER_00: You joined this molecular biology lab,
- 38:36and did you mention you had strong math and physics foundations
- 38:42by that time?
- 38:42So when you were joining the lab,
- 38:47were you leveraging the math and physics in your stuff?
- 38:49SPEAKER_00: Absolutely not.
- 38:51SPEAKER_00: Okay, it was just like very pure biology methodology.
- 38:54SPEAKER_00: Okay, that's interesting.
- 38:56SPEAKER_00: Yeah.
- 38:56To be honest,
- 38:57I could have been like an art major or like just a magician.
- 39:01SPEAKER_01: And then like, it would have made no difference.
- 39:04SPEAKER_01: I had no clue what,
- 39:05SPEAKER_01: like when I got in,
- 39:06SPEAKER_01: I had like,
- 39:07SPEAKER_01: I had some concept about the sales and stuff.
- 39:09SPEAKER_01: I'm like,
- 39:10SPEAKER_01: But I was so many layers of abstraction deep into this thing.
- 39:15Like I was literally like I was running my experiment,
- 39:19make sure I didn't mess it up.
- 39:20And then I had like three hours of incubation or like waiting
- 39:23for the autoclave to
- 39:24SPEAKER_01: finish cleaning my shit.
- 39:25SPEAKER_01: So I had to,
- 39:27SPEAKER_01: I was like blitzing through all their paper,
- 39:30SPEAKER_01: the related paper,
- 39:31SPEAKER_01: like pestering absolutely everybody,
- 39:34SPEAKER_01: including the other lab that was next to us.
- 39:37SPEAKER_01: about the dumbest question you could ever imagine.
- 39:40Literally,
- 39:41they were listening to my question and said,
- 39:43you didn't know that?
- 39:43SPEAKER_01: And I was like, dude, I don't know why they let me into this lab.
- 39:50SPEAKER_01: But by the end of it, I was able to find a foothold.
- 39:54SPEAKER_01: And then I understood a whole lot of things.
- 39:58And I also understood that in molecular biology specifically,
- 40:03it's a bunch of recipes.
- 40:04SPEAKER_01: Right.
- 40:05SPEAKER_01: And like the recipes in certain way.
- 40:06And if you look at the different lab,
- 40:08they do this stuff a bit differently.
- 40:09SPEAKER_01: Right.
- 40:10And all has to do with like the history of the stuff and some
- 40:13mistakes with like
- 40:15SPEAKER_01: living things are just like pure chance.
- 40:17SPEAKER_01: Like your thing is messed up.
- 40:19SPEAKER_01: Right.
- 40:19SPEAKER_01: And like I remember once my thing was messed up three times in a row,
- 40:23SPEAKER_01: but the bacteria culture was trying to like make grow and not die.
- 40:27SPEAKER_01: At some point, like it got really weird.
- 40:29SPEAKER_01: And I was showing this to the supervisor.
- 40:31always in his office like I was just walking in like you need
- 40:34to help me out right
- 40:35you got me in you help me out right and he was like oh just ask
- 40:39this other guy like
- 40:40he's the guy he's a genius about this stuff okay cool and I
- 40:42asked the dude like
- 40:43long dreads like super chill dude and he was looking at my
- 40:46stuff and he was like
- 40:47SPEAKER_01: yeah this is messed up so I was like
- 40:49why like not sure it could be like a hundred different things
- 40:53and i was like so
- 40:54what do i do like try again and that's it right so i was like
- 40:57okay so it's it's a
- 40:59bit different than like the other more like deterministic thing
- 41:03that i was doing
- 41:04before right like we're dealing with living things there's some
- 41:07stuff that if you
- 41:08do it like uh uh like wrong for sure you're gonna contaminate
- 41:12your things but
- 41:13there's other things that like happen because it's like
- 41:15literally like a more
- 41:16SPEAKER_01: complex living thing that is
- 41:18SPEAKER_01: that is trying to grow, right?
- 41:19SPEAKER_01: You're literally,
- 41:20you're shoving a bunch of stuff in bacteria and then trying to
- 41:22make them express a
- 41:23SPEAKER_01: gene and then like create protein,
- 41:25SPEAKER_01: right?
- 41:26SPEAKER_01: Like the bacteria is something living.
- 41:28Like so like sometimes like you don't want in this specific
- 41:31configuration that you
- 41:32SPEAKER_01: didn't account for.
- 41:33SPEAKER_00: Yeah, so I learned a whole bunch on the fly.
- 41:38SPEAKER_00: But that was fun.
- 41:40SPEAKER_00: Well,
- 41:40SPEAKER_00: that's,
- 41:41SPEAKER_00: you know,
- 41:42SPEAKER_00: just you going and,
- 41:43SPEAKER_00: like,
- 41:43SPEAKER_00: pestering these people with all your questions,
- 41:46SPEAKER_00: which I guess initially were a little bit novice questions,
- 41:49SPEAKER_00: but then you built up this form master.
- 41:51SPEAKER_00: That really takes, you know...
- 41:54SPEAKER_00: putting ego aside and like,
- 41:55SPEAKER_00: you know,
- 41:55I'm going to ask the dumb questions because I have to learn
- 41:58this stuff because you
- 42:00SPEAKER_00: don't get results.
- 42:01SPEAKER_00: What's the choice that I have?
- 42:02SPEAKER_01: Like the choice is I'm going to be dumb, right?
- 42:05SPEAKER_01: Literally, like I could instead look dumb, right?
- 42:08SPEAKER_01: And then just tomorrow I'm going to be a bit less, right?
- 42:10Then I'm just going to like climb up with the mud pit that I
- 42:13put myself in.
- 42:15SPEAKER_01: And the interesting thing is that by about the five months mark,
- 42:18SPEAKER_01: because I was during the whole summer,
- 42:21um i was actually pretty good at this like uh the uh the
- 42:24assistant was like oh wow
- 42:25you're running your experiments so well and then like they're
- 42:28so well sequenced i
- 42:28was like i'm just doing whatever the heck you're doing so thank
- 42:32you and then like
- 42:33the supervisor um tacked on to uh like a phd candidate um to
- 42:39take over my work
- 42:40after um i was i was out and she was taking notes and asking me
- 42:44questions i was
- 42:44SPEAKER_01: like wow
- 42:46SPEAKER_01: I can actually,
- 42:46SPEAKER_01: like,
- 42:48provide something now because I've done everything that was
- 42:51messed up for,
- 42:52SPEAKER_01: like,
- 42:53SPEAKER_01: six months at that point.
- 42:54SPEAKER_00: It's very real.
- 42:55SPEAKER_00: Yeah,
- 42:55SPEAKER_00: you know,
- 42:55he has something of value there when they're taking notes on
- 43:00what you've learned.
- 43:01SPEAKER_01: So this is good, man.
- 43:04SPEAKER_01: I think,
- 43:04SPEAKER_01: like,
- 43:04SPEAKER_01: we can maybe switch gear to the stuff that you're working on right now,
- 43:08SPEAKER_01: which is Math Academy.
- 43:09SPEAKER_01: And, like, I've been looking at this stuff.
- 43:12SPEAKER_01: I never actually used it, right?
- 43:13SPEAKER_01: And I'm planning on using it.
- 43:16SPEAKER_01: And I saw that you said that you can teach math four times faster.
- 43:21SPEAKER_01: It's basically like the system that you were using before,
- 43:27SPEAKER_01: but in a more kind of applied and well thought of way.
- 43:31SPEAKER_01: For someone, let's say like me, that never used it, right?
- 43:34SPEAKER_01: And let's say I don't know anything about it.
- 43:36um like what does it look like what the does the learning
- 43:38section look like like
- 43:40how the system works because it's not like a classical you'd
- 43:44watch a video and you
- 43:45SPEAKER_01: do this stuff like what's what's the full looking like here
- 43:48SPEAKER_00: Yeah, so it starts out with the diagnostic test.
- 43:52Figure out what you know,
- 43:53what you don't know in whatever course you're trying to take.
- 43:56It's going to look back for missing prerequisites many years
- 43:59back into math.
- 44:00So if you're placing it into calculus,
- 44:02it's going to test you on a whole bunch of algebra.
- 44:04SPEAKER_00: Like, you don't know completing the square?
- 44:05SPEAKER_00: Well, guess what?
- 44:05SPEAKER_00: You're going to learn completing the square.
- 44:06SPEAKER_00: We're going to add that to your learning plan.
- 44:09SPEAKER_00: makes this custom course for you.
- 44:11SPEAKER_00: Now you got a custom course that you're going to go through.
- 44:14SPEAKER_00: This is kind of like your, your knowledge graph.
- 44:16SPEAKER_00: Uh, and, and so we, we know what, you know,
- 44:20SPEAKER_00: what you're ready to learn and what you don't know.
- 44:22SPEAKER_00: And we're going to serve you stuff that you are ready to learn.
- 44:25SPEAKER_00: And so this is going to be on your knowledge frontier,
- 44:28SPEAKER_00: we call it,
- 44:29SPEAKER_00: your kind of your edge of mastery.
- 44:31SPEAKER_00: And we're going to build up your understanding of these topics.
- 44:35SPEAKER_00: Now, what does a single learning experience on a topic look like?
- 44:40Well,
- 44:40this is a lesson and it's kind of like a sequence of minimum
- 44:45effective doses of
- 44:47SPEAKER_00: Instruction and you actually solving problems.
- 44:50SPEAKER_00: So you think like,
- 44:52SPEAKER_00: you know,
- 44:52this whole situation is very similar to if you were to go with
- 44:57a personal trainer,
- 44:58SPEAKER_00: what are they going to do?
- 44:59Well,
- 44:59they're going to they're going to first just see like,
- 45:01OK,
- 45:01what what are you trying to do?
- 45:03What sport are you prepping for or what are you trying to
- 45:06increase your vertical
- 45:07SPEAKER_00: jump or what are you trying to do?
- 45:09SPEAKER_00: where are your weak points and what can we do to,
- 45:11SPEAKER_00: to,
- 45:12SPEAKER_00: to,
- 45:12SPEAKER_00: to get you,
- 45:13SPEAKER_00: you know,
- 45:13SPEAKER_00: better at those.
- 45:14SPEAKER_00: They're going to have exercises for you and each,
- 45:16SPEAKER_00: each exercise,
- 45:17SPEAKER_00: they're going to like,
- 45:18SPEAKER_00: you know,
- 45:18SPEAKER_00: I'll demonstrate at once.
- 45:20SPEAKER_00: Like, we'll talk like, this is what it is.
- 45:22SPEAKER_00: Here's how you do it.
- 45:22SPEAKER_00: Okay.
- 45:23SPEAKER_00: Now you practice.
- 45:24SPEAKER_00: And so that's what a lesson is like.
- 45:25SPEAKER_00: It's a, starts out with a little, just a brief introduction.
- 45:29SPEAKER_00: And then we initially immediately launch into, into problem solving.
- 45:34SPEAKER_00: You see, okay.
- 45:36SPEAKER_00: you got your introduction.
- 45:37SPEAKER_00: You, this is the type of problem that we're going to focus on solving.
- 45:41SPEAKER_00: Here's an example of like how you, how you solve this type of problem.
- 45:44SPEAKER_00: Okay.
- 45:44SPEAKER_00: Now you do a few problems.
- 45:47SPEAKER_00: If you're just knocking it out of the park,
- 45:48SPEAKER_00: like you,
- 45:48SPEAKER_00: you,
- 45:49SPEAKER_00: you answer the first two questions in a row.
- 45:51SPEAKER_00: It's like, Hey, you got it.
- 45:52SPEAKER_00: Let's move on to the next, the next, uh,
- 45:55more challenging version so um if if you're if you're
- 45:59struggling a little bit
- 46:00you're like oh first question wrong next one right like we're
- 46:03gonna make you solve
- 46:04a few more problems just to just you know you you gotta end on
- 46:08two successes in a
- 46:09row before advancing to make me think of a lesson of a tutor
- 46:13actually like yeah
- 46:14because that's what i was doing i was asking them like yeah
- 46:17like hi nice to meet
- 46:18you like what do you know and then they were blabbering about
- 46:21like everything
- 46:22SPEAKER_01: that's cool let's like do some stuff like and i was looking at like yeah
- 46:25he doesn't know this stuff let's just like go back and then do
- 46:28this thing and oh
- 46:29now you get it okay perfect like we're not gonna waste time
- 46:31because you're
- 46:31literally paying me right now right so we're gonna go and move
- 46:34to the actual stuff
- 46:35that you need to to uh was this inspired by um by that sort of
- 46:40uh so this is this
- 46:42SPEAKER_00: is born out of our hands-on experience teaching and
- 46:47SPEAKER_00: Just like,
- 46:48SPEAKER_00: you know,
- 46:49SPEAKER_00: the tutor,
- 46:50the tutor is really I'm glad you brought that up because the
- 46:52tutor is the that's
- 46:54SPEAKER_00: like our model organism.
- 46:56SPEAKER_00: Basically,
- 46:56we're trying to emulate the the decisions of an expert tutor
- 47:01who knows everything
- 47:02SPEAKER_00: about your experience.
- 47:05SPEAKER_00: your knowledge profile,
- 47:06SPEAKER_00: like what you know,
- 47:07SPEAKER_00: what you don't know,
- 47:08SPEAKER_00: and has unlimited computational bandwidth to like,
- 47:12SPEAKER_00: have you,
- 47:13SPEAKER_00: you know,
- 47:13SPEAKER_00: it can just make up problems.
- 47:14Like it's,
- 47:15it's got a whole inventory of problems that it can match you
- 47:18with.
- 47:18SPEAKER_00: And every single answer that you do,
- 47:21SPEAKER_00: it's going to kind of take that into consideration to,
- 47:23SPEAKER_00: to get you working on the,
- 47:25SPEAKER_00: just what is the most efficient use of your time right at the second?
- 47:29SPEAKER_00: Okay.
- 47:29SPEAKER_00: We're going to, we're going to have you do that.
- 47:31SPEAKER_01: Okay, cool.
- 47:31SPEAKER_01: And you mentioned expert knowledge about the stuff.
- 47:36Behind this whole thing,
- 47:37there's a knowledge graph that is mapping,
- 47:40I think,
- 47:412,000-some topic.
- 47:43But the interesting bit that got my attention here is that they
- 47:46were encoded manually.
- 47:48SPEAKER_01: I think you spent, what, 200-something hours on crafting that stuff?
- 47:55that whole graph so we started we started this stuff i mean
- 47:59it's pre-llms right
- 48:01this is like you know there's just no way to generate any of
- 48:04this information so we
- 48:06started stuff back in uh like 2016 i think ish is is when um
- 48:13you know jason and our
- 48:14uh director of curriculum alex smith uh started working on this
- 48:18thing i i joined
- 48:19SPEAKER_00: around like in
- 48:21SPEAKER_00: in, in, in 2018.
- 48:22And so we,
- 48:23a lot of stuff we're just,
- 48:24um,
- 48:25you know,
- 48:25we're handcrafting it manually.
- 48:27SPEAKER_00: Um,
- 48:28SPEAKER_00: and to this day,
- 48:29SPEAKER_00: everything is,
- 48:30SPEAKER_00: is very,
- 48:31SPEAKER_00: you know,
- 48:31it comes from a human expert just paying very close attention
- 48:35to based on their
- 48:36SPEAKER_00: experiences teaching,
- 48:37SPEAKER_00: um,
- 48:38SPEAKER_00: and,
- 48:38SPEAKER_00: and what they know about,
- 48:40SPEAKER_00: um,
- 48:41SPEAKER_00: you know,
- 48:42SPEAKER_00: uh,
- 48:42SPEAKER_00: max efficiency learning and how to get students through material,
- 48:47SPEAKER_00: um,
- 48:47in in in a very uh in a way that's guaranteed to work uh how do
- 48:51you how do you do
- 48:52it and so we um we build like a lot of tooling over the years
- 48:57to make this process
- 48:58more and more efficient um so we're leveraging like all the
- 49:02tooling that we can but
- 49:03yeah ultimately it comes from from a a human domain expert is
- 49:07is is making sure
- 49:09SPEAKER_00: that uh like all of this is is to their standards and um
- 49:14SPEAKER_00: And this is all mapped out by them.
- 49:19SPEAKER_00: So this is Alex Smith, our director of curriculum.
- 49:23SPEAKER_00: He builds kind of the four words knowledge graph of like,
- 49:26SPEAKER_00: okay,
- 49:27SPEAKER_00: what are the topics that we have?
- 49:29SPEAKER_00: We want to have students learn.
- 49:32SPEAKER_00: What is this course, this calculus course?
- 49:34SPEAKER_00: What is it?
- 49:34SPEAKER_00: What are all the topics in it?
- 49:36SPEAKER_00: We got about 300 something topics in our calculus course.
- 49:39Each one is like an atomic unit of knowledge consisting of
- 49:42roughly like three or
- 49:43four knowledge points where it's like it starts out with the
- 49:47simplest version of
- 49:48SPEAKER_00: the problem.
- 49:49SPEAKER_00: And then...
- 49:51layers on additional complexity as you evidence your ability to
- 49:55solve problems in
- 49:56SPEAKER_00: each successive knowledge point.
- 49:58SPEAKER_00: And so he maps all this out.
- 50:00He connects everything up,
- 50:01the prerequisites,
- 50:02what topics do you need to have learned?
- 50:06before we serve you this new thing and um and i kind of encode
- 50:09this this backwards
- 50:10graph we call the encompassing graph and it and it says what
- 50:14skills encompass what
- 50:16other skills what sub skills are you exercising as components
- 50:19of more advanced
- 50:19skills and this helps with our our space repetition algorithm
- 50:24which is always you
- 50:25SPEAKER_00: know
- 50:26SPEAKER_00: when you do a review on a,
- 50:28SPEAKER_00: on a,
- 50:29SPEAKER_00: on an advanced skill,
- 50:30SPEAKER_00: you're kind of implicitly practicing a lot of skills.
- 50:34SPEAKER_00: Yeah.
- 50:34SPEAKER_00: So we're always trying to,
- 50:36SPEAKER_00: you know,
- 50:36SPEAKER_00: something that we,
- 50:37SPEAKER_00: we,
- 50:37SPEAKER_00: we do is we're always trying to maximize your learning efficiency.
- 50:40So if we can get you practicing your sub skills that are kind
- 50:44of due for some
- 50:44SPEAKER_00: review,
- 50:45SPEAKER_00: we can get you practicing those by learning something new,
- 50:48SPEAKER_00: learn a new lesson that knocks those out.
- 50:51SPEAKER_00: Then, then, then we, um,
- 50:53then the algorithms are so it's like a like a directed acyclic
- 50:57graph right because
- 50:58there's no yeah there cannot be any cycle uh that would not
- 51:01make sense right um and
- 51:02then like if you're working on like this skill that like has
- 51:04the dependency and you
- 51:06like you're super good at this this whole kind of uh like uh uh
- 51:11like a parent
- 51:12SPEAKER_01: lineage is kind of solved so you don't need to go back to that
- 51:15SPEAKER_01: Okay, cool.
- 51:16SPEAKER_01: This is good.
- 51:17SPEAKER_01: Yeah, exactly.
- 51:19SPEAKER_00: So something about just the whole,
- 51:20SPEAKER_00: you know,
- 51:21SPEAKER_00: approaching this as a,
- 51:22SPEAKER_00: so it's like,
- 51:22SPEAKER_00: why not just,
- 51:24SPEAKER_00: you know,
- 51:25SPEAKER_00: generate this,
- 51:27SPEAKER_00: just try to generate all the content with like an LLM or something.
- 51:31SPEAKER_00: Well,
- 51:32SPEAKER_00: so the thing is,
- 51:33SPEAKER_00: a lot of this content,
- 51:35the reason why it works so well is because it's based on all of
- 51:41our learnings
- 51:41SPEAKER_00: throughout the years of doing this manually.
- 51:44SPEAKER_00: And
- 51:45SPEAKER_00: While while alums are like the very I mean,
- 51:48they are really good at traversing this whatever knowledge that
- 51:53they've been
- 51:54SPEAKER_00: trained on.
- 51:54SPEAKER_00: Right.
- 51:54Everything that's been written down on the Internet and and
- 51:59often like applying
- 52:00sequencing of that of that knowledge to kind of go expand
- 52:04beyond the initial
- 52:05SPEAKER_00: knowledge base.
- 52:06SPEAKER_00: There's a lot of stuff that you can.
- 52:09learn by you know just working with students hands-on um and
- 52:14just kind of almost
- 52:15like extracting data from from reality itself not not just a
- 52:20compressed version of
- 52:22like this is reality according to all the text that's been
- 52:24written on the internet
- 52:25SPEAKER_01: there's a lot of stuff that's
- 52:30SPEAKER_01: This is not mapped anywhere.
- 52:33SPEAKER_01: Seeing somebody,
- 52:36an actual human being trying to get toward whatever they're
- 52:40doing in the class and
- 52:42SPEAKER_01: seeing the hundred different ways they can struggle firsthand,
- 52:46this is way different than even the research about teaching or
- 52:51the research about
- 52:52SPEAKER_01: learning.
- 52:53SPEAKER_01: It's way different when you're looking at this kid right here.
- 52:56SPEAKER_01: and how they're struggling and why they're struggling.
- 52:58SPEAKER_01: You can decompose this for sure, right?
- 53:00SPEAKER_01: And there's like emotional component,
- 53:01SPEAKER_01: like resistance to this,
- 53:02SPEAKER_01: like they don't like the material also maybe,
- 53:06SPEAKER_01: right?
- 53:06SPEAKER_01: All of this stuff can go into there.
- 53:10SPEAKER_01: It's really something very complex, right?
- 53:13As you see and you see more of these students,
- 53:15you can kind of start to map it out.
- 53:16But I've never seen anywhere something written about the
- 53:22specific reaction that
- 53:25SPEAKER_01: they can have to the material.
- 53:26Well,
- 53:26actually,
- 53:26just on what you were talking about,
- 53:28like a student's reaction to material.
- 53:29SPEAKER_00: I mean,
- 53:30SPEAKER_00: that's one of the things that,
- 53:32SPEAKER_00: you know,
- 53:33SPEAKER_00: the design of minimum effective doses of instruction and practice.
- 53:38SPEAKER_00: I mean,
- 53:38SPEAKER_00: part of this is just inspired by,
- 53:40SPEAKER_00: you know,
- 53:41SPEAKER_00: cognitive science of like,
- 53:43SPEAKER_00: you know,
- 53:44SPEAKER_00: active practice.
- 53:45Let's maximize the amount of time that you're actively solving
- 53:48problems with guidance.
- 53:49SPEAKER_00: But another part is also just inspired by Jason Roberts, um,
- 53:53founder he he always talks about like how it's not you know
- 53:58when he was sitting in
- 53:58math class um it's not that he hated math it's that he hated
- 54:02being talked at for an
- 54:03hour and so this so many we've realized that so many students
- 54:09um especially
- 54:10students like with um you know adhd or just who are just like
- 54:14really uh can't sit
- 54:16SPEAKER_00: still type of type of students like
- 54:19SPEAKER_00: They don't want to watch a video for an hour.
- 54:21SPEAKER_00: They don't want to watch an hour long lecture video.
- 54:23SPEAKER_00: And if they do,
- 54:24it's often because they just want to like stir off into space
- 54:28and not think about
- 54:28SPEAKER_00: it.
- 54:29It's like what what really keeps the students engaged with the
- 54:33process is.
- 54:34you know actively solving problems that are at the right level
- 54:37of difficulty kind
- 54:38of entering sort of that that that that flow state right at the
- 54:41edge of their
- 54:42ability um but i i want to say in addition to like you know um
- 54:47just designing like
- 54:48the lessons in internally to to try to you know keep keep the
- 54:52experience like um
- 54:53you know maybe a a little challenging but very achievable for
- 54:57for students the the
- 54:59SPEAKER_00: other part is like
- 55:00SPEAKER_00: There's also the whole connectivity of the knowledge graph.
- 55:03So this is stuff that you probably would not think about if
- 55:06you're just like trying
- 55:09to maybe get an LLM to dump out a knowledge graph or just ask
- 55:12somebody else to
- 55:13SPEAKER_00: design a knowledge graph.
- 55:14SPEAKER_00: There's a lot that comes.
- 55:15into play in terms of structuring the knowledge graph to make
- 55:18students successful.
- 55:22We often think about things like quantifying the cognitive
- 55:26congestion of various
- 55:28SPEAKER_00: nodes in the knowledge graph.
- 55:30If you have a topic that has 10 different prerequisites that
- 55:36are all being pulled in
- 55:38SPEAKER_00: that you've never practiced pulling in together before,
- 55:41SPEAKER_00: like that's going to,
- 55:42SPEAKER_00: it's going to inflate the cognitive load a ton.
- 55:45SPEAKER_00: And so there's just a million of these little optimizations that we,
- 55:49SPEAKER_00: that we put into,
- 55:51SPEAKER_00: to make the knowledge graphs like really easy to,
- 55:54SPEAKER_00: to,
- 55:54SPEAKER_00: to learn from.
- 55:55SPEAKER_01: Yeah.
- 55:56and um even when you are teaching right sometimes you don't
- 55:59like you don't remember
- 56:00like there was uh like that much prerequisite and you're
- 56:04teaching it like it's just
- 56:05to be simple you do this like what's that right like uh oh it's
- 56:09taking from this
- 56:11what's this thing right and then you're like okay then we have
- 56:13to go back um the
- 56:16knowledge graph thing is very interesting for me because it is
- 56:19trying to get to
- 56:20like i mean this golden type of learning where you have an
- 56:24expert next to you and
- 56:25SPEAKER_01: it's just like
- 56:26guiding you like with like an infinite amount of patience right
- 56:30and like goodwill
- 56:31they will be there like rain or shine and they will just like
- 56:35help you get through
- 56:36SPEAKER_01: your stuff right and
- 56:40SPEAKER_01: I was thinking that, like, a textbook is also just that, right?
- 56:43SPEAKER_01: A textbook is this,
- 56:45SPEAKER_01: but it's,
- 56:45SPEAKER_01: like,
- 56:46SPEAKER_01: handcrafted with one or two authors,
- 56:49SPEAKER_01: and that's pretty much this,
- 56:50SPEAKER_01: right?
- 56:50SPEAKER_01: And,
- 56:50SPEAKER_01: like,
- 56:50they don't have as much flexibility in how they're presenting
- 56:53the material because
- 56:53SPEAKER_01: they have to make it,
- 56:54SPEAKER_01: like,
- 56:54SPEAKER_01: extremely linear.
- 56:55And you see,
- 56:56like,
- 56:57I've seen this in,
- 56:58like,
- 56:58the more advanced class,
- 56:59right?
- 57:00like at some point the chapter kind of doesn't make any sense
- 57:02anymore because like
- 57:03you like it does branch out and then there's like okay this is
- 57:08like also for these
- 57:09three topic and then by the way this one you require these two
- 57:12and then this other
- 57:13and then like the later chapter is always just like yeah good
- 57:16luck kid right like
- 57:18SPEAKER_01: go ahead and like uh best of luck
- 57:21SPEAKER_01: And you always see,
- 57:22SPEAKER_01: depending on the textbook,
- 57:24some biases from the others and then some like gap that are in
- 57:28there or like some
- 57:29SPEAKER_01: stuff that are more interesting than others that are there,
- 57:32SPEAKER_01: and so applied stuff.
- 57:33SPEAKER_01: But it seems to me like this knowledge
- 57:36graph is like a deconstruction of this stuff and then you're of
- 57:41course injecting
- 57:42like the the knowledge of the the expert because they are
- 57:45acting as a pseudo like a
- 57:47source of truth for the the knowledge of the one-on-one tutor
- 57:50that happens
- 57:51SPEAKER_00: Yeah, exactly.
- 57:52SPEAKER_00: It's kind of like, I think that's a good analogy.
- 57:54SPEAKER_00: It's like if you take a textbook,
- 57:57and let's say that this textbook was constructed not just by
- 58:03one author just doing
- 58:04SPEAKER_00: a brain dump in isolation,
- 58:05SPEAKER_00: but an author,
- 58:08SPEAKER_00: an expert on the subject who's also looking at all the other textbooks,
- 58:12SPEAKER_00: all the other treatments of the subject,
- 58:13SPEAKER_00: trying to just hit all the bases,
- 58:15SPEAKER_00: a comprehensive course.
- 58:18SPEAKER_00: And so they write this textbook, and then
- 58:20SPEAKER_00: Instead of presenting it to the student in this linear sequence,
- 58:23SPEAKER_00: just here's the unit on,
- 58:26SPEAKER_00: I'll just use algebra as an example.
- 58:28SPEAKER_00: Here's the unit on linear equations.
- 58:30SPEAKER_00: Okay, now let's move up to quadratic equations.
- 58:33SPEAKER_00: Okay, now let's do some trigonometry.
- 58:35SPEAKER_00: Instead of presenting...
- 58:37SPEAKER_00: by units,
- 58:37SPEAKER_00: you really chop it up into the atomic bits of knowledge and,
- 58:41SPEAKER_00: and have,
- 58:42SPEAKER_00: that way you can,
- 58:43SPEAKER_00: you can really trace out the student knowledge as a,
- 58:47SPEAKER_00: as a contiguous,
- 58:49SPEAKER_00: as a contiguous area in the knowledge graph,
- 58:52SPEAKER_00: you know,
- 58:52SPEAKER_00: exactly what they're,
- 58:53SPEAKER_00: they're ready to learn.
- 58:53SPEAKER_00: You can deliver the, the next pieces of knowledge.
- 58:56SPEAKER_00: You can kind of adapt the number of problems to the, the student and,
- 59:01SPEAKER_00: And and you're you're also you keep it interesting.
- 59:05SPEAKER_00: Right.
- 59:05SPEAKER_00: You don't have to just do one unit for an entire week or two weeks.
- 59:10You know,
- 59:10at school,
- 59:11I often like,
- 59:12you know,
- 59:12this is this quarter we're doing trigonometry.
- 59:16SPEAKER_00: Nothing but trigonometry for three months.
- 59:19SPEAKER_00: Yeah.
- 59:20yeah i know you get sick of it it's like in addition to just
- 59:23getting sick of it
- 59:24it's like that's not even an efficient way to learn the
- 59:27material it's like you i do
- 59:29you like ideally you want to be spacing it out coming back
- 59:34revisiting it um you
- 59:36don't want to be blocking or like uh having your practice on
- 59:40the consecutive
- 59:41SPEAKER_00: material all the time um interleaves or all the different units yeah
- 59:46But a classroom,
- 59:47I think,
- 59:48is always flawed because it's a one-to-many type of teaching.
- 59:52SPEAKER_01: It's not one-to-one custom-made stuff.
- 59:57SPEAKER_01: I always find it funny.
- 59:59SPEAKER_01: I brought my wife to one of my class.
- 1:00:01it was organic chemistry tree or something like that and it's
- 1:00:06like it's like this
- 1:00:08giant stadium that can hold 700 students and then there's like
- 1:00:12this one tiny
- 1:00:12teacher like over there and yeah it's like an impossible task
- 1:00:17for one to i mean
- 1:00:18that's a good example because like 700 students how are you
- 1:00:21gonna you can't you
- 1:00:22SPEAKER_00: can't just one one teacher cannot deliver
- 1:00:26SPEAKER_00: an optimal earning experience to 700 students simultaneously.
- 1:00:29SPEAKER_01: It's very easy to hijack.
- 1:00:31No,
- 1:00:31because like I was never afraid to ask questions because my
- 1:00:34mentality was like,
- 1:00:36SPEAKER_01: I paid for this stuff.
- 1:00:38SPEAKER_01: I'm going to learn this.
- 1:00:39SPEAKER_01: So,
- 1:00:39SPEAKER_01: like,
- 1:00:40SPEAKER_01: I was raising my hand in the 700 people and asking,
- 1:00:45SPEAKER_01: like,
- 1:00:45SPEAKER_01: the dumbest question,
- 1:00:46SPEAKER_01: like,
- 1:00:47SPEAKER_01: that I didn't get or something like that.
- 1:00:49SPEAKER_01: And,
- 1:00:49SPEAKER_01: like,
- 1:00:50SPEAKER_01: if you think about it,
- 1:00:50SPEAKER_01: there's,
- 1:00:51SPEAKER_01: like,
- 1:00:51SPEAKER_01: 500 people that are like,
- 1:00:52SPEAKER_01: ah,
- 1:00:52SPEAKER_01: shit,
- 1:00:52SPEAKER_01: there we go again,
- 1:00:53SPEAKER_01: right?
- 1:00:53SPEAKER_01: This stuff that, like, you didn't understand this thing.
- 1:00:56SPEAKER_01: But, like, it's very inefficient.
- 1:00:57SPEAKER_01: Like,
- 1:00:58SPEAKER_01: it's absolutely,
- 1:00:59SPEAKER_01: it's just a way to scale to,
- 1:01:01SPEAKER_01: like,
- 1:01:01SPEAKER_01: so that the average can go up or something like that.
- 1:01:04SPEAKER_01: But on an individual level, it's not the best way to learn.
- 1:01:09SPEAKER_01: On the topic of knowledge bases, there's the expert curation here.
- 1:01:15Do you think that generation capability of LLM are useful as an
- 1:01:20interface to the
- 1:01:22SPEAKER_01: knowledge base with the student?
- 1:01:23SPEAKER_01: I'm not talking about generating a knowledge graph.
- 1:01:26SPEAKER_01: Let's say we have a very well-created
- 1:01:28knowledge graph made by expert reviewed and like really well
- 1:01:30crafted in a way that
- 1:01:32like it's very hard for an ai to kind of like figure this out
- 1:01:35do you think that
- 1:01:35like having an interface uh has any use of this or like even
- 1:01:39then it won't be uh
- 1:01:40won't be that useful well i think it's an area that is that is
- 1:01:43uh it's not a radar
- 1:01:44worth paying attention to um at at this moment there's there's
- 1:01:49a lot of challenges
- 1:01:50SPEAKER_00: with that sort of setup it's it's one of those things that's like
- 1:01:53SPEAKER_00: There's a pull.
- 1:01:54SPEAKER_00: You think of efficient learning.
- 1:01:58SPEAKER_00: You think of this personalization.
- 1:02:01You often imagine some tutor having a personalized conversation
- 1:02:05with a student
- 1:02:06around their interests relating the material to their interests
- 1:02:09and stuff like
- 1:02:10SPEAKER_00: that.
- 1:02:12SPEAKER_00: Now, that can...
- 1:02:15potentially work um really well in some cases but you know
- 1:02:19there's there's a lot of
- 1:02:21surface area for things to go off the rails and so just to just
- 1:02:25to name a few a few
- 1:02:26SPEAKER_00: um cases um sometimes students um
- 1:02:31want to kind of get out of scope of a topic you know like
- 1:02:35you're teaching one thing
- 1:02:36and they're like oh what if you do this what if you do that
- 1:02:38blah blah blah and so
- 1:02:40so much of this can be like um can enrich the the learning
- 1:02:43experience but there
- 1:02:45also comes time when it's like you know kid like i know you're
- 1:02:48excited about this
- 1:02:49SPEAKER_00: stuff but we're just talking in the abstract right now you know i i
- 1:02:54SPEAKER_00: I used to,
- 1:02:55SPEAKER_00: back when I was teaching,
- 1:02:56I would have some students sometimes who would ask me like tons
- 1:02:58of questions about
- 1:02:59SPEAKER_00: like,
- 1:02:59SPEAKER_00: we'd be learning like introductory calculus stuff.
- 1:03:01And then they'd start asking me questions about like,
- 1:03:05what's the hardest integral to solve?
- 1:03:07SPEAKER_00: Or just like, what's, are all integrals solvable?
- 1:03:10SPEAKER_00: Like, how do you know, or how do you not know?
- 1:03:12SPEAKER_00: And, you know, pretty quickly you get into a realm where you're no longer
- 1:03:18SPEAKER_00: really like doing skills-based teaching.
- 1:03:21You're just kind of like talking around the subject almost as
- 1:03:25if they were watching
- 1:03:26SPEAKER_00: like a YouTube video,
- 1:03:28SPEAKER_00: like some visualization and like,
- 1:03:30SPEAKER_00: what's the hardest math problem?
- 1:03:31SPEAKER_00: Or like, you know, that kind of stuff.
- 1:03:34SPEAKER_00: And so it can be
- 1:03:36there's some element of like well okay you gotta kind of scope
- 1:03:40it down to know like
- 1:03:42there's there's a balance to be had and and and and you you
- 1:03:46don't want to allow the
- 1:03:48scope of what you're doing to to bleed too much another um part
- 1:03:52is is that um i
- 1:03:54mean so in addition to to make try like making it non-gameable
- 1:03:58and and stuff like
- 1:03:59SPEAKER_00: that there's a challenge in in that um
- 1:04:02SPEAKER_00: Generating content on the fly is always...
- 1:04:05SPEAKER_00: I mean, LLMs are like stochastic generators, right?
- 1:04:09SPEAKER_00: And so you don't...
- 1:04:11SPEAKER_00: I mean, you can...
- 1:04:12SPEAKER_00: They've gotten a lot better at all the hallucinations and stuff.
- 1:04:16SPEAKER_00: There's definitely an order of magnitude improvement in that.
- 1:04:20SPEAKER_00: If you don't have static content underlying, it's a lot harder to...
- 1:04:25SPEAKER_00: Yeah, writing's the gamble.
- 1:04:28SPEAKER_00: And additionally,
- 1:04:31SPEAKER_00: just to name one other thing,
- 1:04:33one thing we do a lot with our knowledge graph is that's an
- 1:04:36advantage of having
- 1:04:37SPEAKER_00: this kind of static content is analytics.
- 1:04:41SPEAKER_00: on,
- 1:04:41SPEAKER_00: you know,
- 1:04:42every student is being served the same exact step-by-step
- 1:04:48lesson that we have been
- 1:04:49SPEAKER_00: optimizing for years to be the most,
- 1:04:53like assuming the students have mastered the prerequisites
- 1:04:55leading into that
- 1:04:56SPEAKER_00: lesson.
- 1:04:56SPEAKER_00: This is the most efficient,
- 1:04:59SPEAKER_00: most robust way of getting a student to learn the material in that topic.
- 1:05:03SPEAKER_00: And so a lot of these optimizations, you know,
- 1:05:06maybe we had three knowledge points and knowledge point number
- 1:05:10two was actually a
- 1:05:11SPEAKER_00: little too aggressive.
- 1:05:13SPEAKER_00: And instead of 98% of the students getting through it on the first try,
- 1:05:20SPEAKER_00: it was only like 80% or 70%.
- 1:05:22SPEAKER_00: And then we break that up into two steps and like,
- 1:05:25SPEAKER_00: just,
- 1:05:26SPEAKER_00: you know,
- 1:05:27SPEAKER_00: we do,
- 1:05:28SPEAKER_00: all this content has gone through so,
- 1:05:30SPEAKER_00: so much refinement,
- 1:05:32SPEAKER_00: analytics refinement.
- 1:05:33SPEAKER_00: And once you kind of,
- 1:05:36get rid of this idea of static content,
- 1:05:39it makes it a lot more difficult to do analytics.
- 1:05:43Not to say that this could never work,
- 1:05:45just to say there's still lots of challenges.
- 1:05:51SPEAKER_01: like AI system for like a production grade and stuff like that.
- 1:05:56The one part I am always looking for when to apply like
- 1:06:02LM-based generation is if I
- 1:06:05SPEAKER_01: can get like this shape of summarization,
- 1:06:08SPEAKER_01: right?
- 1:06:09SPEAKER_01: If you can get like the base ground truth and then in some shape or form,
- 1:06:13SPEAKER_01: right,
- 1:06:14SPEAKER_01: able to kind of go down like this,
- 1:06:16SPEAKER_01: it almost always work, right?
- 1:06:19SPEAKER_01: They almost always work.
- 1:06:20And you have very little guardrail to put because they will
- 1:06:23summarize,
- 1:06:24right?
- 1:06:24And then they're very happy to just summarize and not do
- 1:06:27anything else than do that.
- 1:06:29SPEAKER_01: And even though they can put it in a conversational way,
- 1:06:34SPEAKER_01: like put on different masks,
- 1:06:36SPEAKER_01: they are still just literally summarizing the content.
- 1:06:39So in your case,
- 1:06:40I think having this static graph that no AI is able to touch
- 1:06:45it,
- 1:06:45right?
- 1:06:46and then having this kind of thin layer between the static
- 1:06:53content and the user,
- 1:06:55SPEAKER_01: you can literally leverage one of their best quality,
- 1:06:58SPEAKER_01: which is to be extremely psychophantic.
- 1:07:02they will like rope in the user in the conversation and kind of
- 1:07:08deliver the content
- 1:07:09in like a way that like it's more amenable to this specific
- 1:07:16person because like I
- 1:07:17remember when I was tutoring I could not talk the same way to
- 1:07:20everybody right
- 1:07:22SPEAKER_01: there's some like I had to like
- 1:07:24SPEAKER_01: to do it in a different way.
- 1:07:25SPEAKER_01: I had some,
- 1:07:27SPEAKER_01: I had to take different example,
- 1:07:30SPEAKER_01: or bring them into,
- 1:07:33SPEAKER_01: ask them more questions than others.
- 1:07:36SPEAKER_01: Others just tried, and then we had just this feedback loop.
- 1:07:40SPEAKER_01: Others literally had to be taken by the hand for the first,
- 1:07:43SPEAKER_01: because they were kind of frightened by the problem.
- 1:07:45And then we do it again within a different variation,
- 1:07:46but it's the same stuff,
- 1:07:47right?
- 1:07:47SPEAKER_01: And then we do it again and then they feel comfortable and then like,
- 1:07:49SPEAKER_01: you know,
- 1:07:50SPEAKER_01: we can move on and stuff.
- 1:07:51SPEAKER_01: So it's this part that I was thinking about.
- 1:07:54SPEAKER_01: So food for thought, maybe it's a terrible idea.
- 1:07:58SPEAKER_00: No,
- 1:07:59SPEAKER_00: I mean,
- 1:08:00SPEAKER_00: it's definitely something that's,
- 1:08:02SPEAKER_00: I think will remain on our radar is like,
- 1:08:05SPEAKER_00: you know,
- 1:08:05SPEAKER_00: the idea of,
- 1:08:06SPEAKER_00: I mean,
- 1:08:07SPEAKER_00: it'd be cool to,
- 1:08:09SPEAKER_00: we've talked about the idea of,
- 1:08:13SPEAKER_00: Some people,
- 1:08:13when content is presented in a way like making analogies or
- 1:08:19using word problems in
- 1:08:21SPEAKER_00: a setting that they're just really interested in,
- 1:08:23SPEAKER_00: like,
- 1:08:24SPEAKER_00: oh,
- 1:08:24SPEAKER_00: they're interested in soccer.
- 1:08:26SPEAKER_00: Like, okay, let's make this word problem.
- 1:08:29SPEAKER_00: We're doing all our word problems about soccer.
- 1:08:30SPEAKER_00: Let's just make it fun like that.
- 1:08:33SPEAKER_00: That can definitely...
- 1:08:34SPEAKER_00: elevate the learning experience.
- 1:08:38SPEAKER_00: There's other elements,
- 1:08:41SPEAKER_00: just to say,
- 1:08:42SPEAKER_00: not that you've suggested this,
- 1:08:43SPEAKER_00: but a common confusion,
- 1:08:45SPEAKER_00: though,
- 1:08:46SPEAKER_00: on this point is,
- 1:08:47SPEAKER_00: like,
- 1:08:47SPEAKER_00: some people think that students need,
- 1:08:49SPEAKER_00: like,
- 1:08:50SPEAKER_00: a million different explanations of a topic until one clicks.
- 1:08:56And so that's what I'm used to hearing in terms of suggestions
- 1:09:01for using LLMs in the content.
- 1:09:03But what we found is that if you just make sure students have
- 1:09:08all of their
- 1:09:08prerequisites in place leading into the lesson that you're
- 1:09:12asking them to learn,
- 1:09:14then they're just prepared for it and and oftentimes like what
- 1:09:18you experience as a
- 1:09:19tutor like when you're when you have to explain things a
- 1:09:21million different ways for
- 1:09:22a student to get it is is really yeah there's something missing
- 1:09:26and you're just
- 1:09:26kind of like trying to cover the space to just hit what is this
- 1:09:29missing thing i
- 1:09:30SPEAKER_00: don't know let me like just like rapid fire like to try to
- 1:09:34SPEAKER_00: Hit it,
- 1:09:34SPEAKER_00: which,
- 1:09:35SPEAKER_00: I mean,
- 1:09:36if you don't have a mastery learning system that is guaranteed
- 1:09:39for art because it's
- 1:09:40SPEAKER_00: something in place,
- 1:09:40SPEAKER_00: like,
- 1:09:41SPEAKER_00: that's not a bad heuristic.
- 1:09:42SPEAKER_00: Yeah, true.
- 1:09:43SPEAKER_00: Yeah.
- 1:09:44SPEAKER_00: But we are just on the topic of,
- 1:09:48SPEAKER_00: you know,
- 1:09:48SPEAKER_00: framing things in ways that are,
- 1:09:50SPEAKER_00: like,
- 1:09:50SPEAKER_00: really,
- 1:09:51SPEAKER_00: you know,
- 1:09:51SPEAKER_00: exciting to students.
- 1:09:53One of the things that we actually have coming out in one of
- 1:09:57our upcoming courses
- 1:09:59SPEAKER_00: is projects that are just based in,
- 1:10:03SPEAKER_00: you know,
- 1:10:04just really interesting scenarios,
- 1:10:06at least that we think are very interesting.
- 1:10:08So it was a little,
- 1:10:08you know,
- 1:10:10interesting to people who are interested in astrophysics
- 1:10:14modeling,
- 1:10:16SPEAKER_00: Planning out civilization on Mars,
- 1:10:19SPEAKER_00: translating writings from long lost civilizations,
- 1:10:23SPEAKER_00: stuff like that.
- 1:10:24SPEAKER_00: So we got a lot of cool projects like that.
- 1:10:28SPEAKER_00: But yeah, I mean, the most where this path leads.
- 1:10:32Yeah,
- 1:10:33it's kind of like matching like,
- 1:10:35okay,
- 1:10:35what kind of projects are you really interested in?
- 1:10:37SPEAKER_01: That makes sense.
- 1:10:40SPEAKER_01: A lot of stuff to try out here.
- 1:10:44SPEAKER_01: Before moving on to the rest of the stuff,
- 1:10:45I wanted to dive into your space repetitional algorithm that
- 1:10:51you injected.
- 1:10:52SPEAKER_01: into this system because you're not just going to learn it once,
- 1:10:57SPEAKER_01: see it and check mark,
- 1:10:58SPEAKER_01: it's done.
- 1:10:59You're going to have to practice it again and checking that
- 1:11:01you're moving in the
- 1:11:02SPEAKER_01: right direction.
- 1:11:04SPEAKER_01: So you have this kind of fire-space repetition algorithm.
- 1:11:08SPEAKER_01: which is different than, like, FlashCard.
- 1:11:10I've used FlashCard a lot,
- 1:11:12right,
- 1:11:14in the different classes that I was taking.
- 1:11:17SPEAKER_01: But,
- 1:11:17SPEAKER_01: like,
- 1:11:18SPEAKER_01: in this specific case,
- 1:11:20SPEAKER_01: you're leveraging,
- 1:11:21SPEAKER_01: like,
- 1:11:22SPEAKER_01: the knowledge graph,
- 1:11:23SPEAKER_01: the fact that it's,
- 1:11:24SPEAKER_01: like,
- 1:11:24SPEAKER_01: hierarchical in nature.
- 1:11:25SPEAKER_01: Can you,
- 1:11:25SPEAKER_01: like,
- 1:11:26SPEAKER_01: walk us through,
- 1:11:27SPEAKER_01: like,
- 1:11:27SPEAKER_01: high-level,
- 1:11:27SPEAKER_01: like,
- 1:11:27SPEAKER_01: how this specific algorithm works and how it's different from,
- 1:11:30SPEAKER_01: like,
- 1:11:31SPEAKER_01: just an Enki,
- 1:11:32SPEAKER_01: like,
- 1:11:33SPEAKER_01: FlashCard-based system?
- 1:11:34SPEAKER_00: Yeah.
- 1:11:35SPEAKER_00: Yeah.
- 1:11:35SPEAKER_00: So imagine that you want to,
- 1:11:38SPEAKER_00: you want to learn math,
- 1:11:39SPEAKER_00: you know,
- 1:11:39you got to do some form of space repetition because the spacing
- 1:11:42effect is one of
- 1:11:44SPEAKER_00: the few like free lunches you get in learning.
- 1:11:48SPEAKER_00: If you space out the material and serve it at the right time,
- 1:11:50you can kind of minimize the amount of review that you have to
- 1:11:54do while maintaining
- 1:11:56SPEAKER_00: a certain level of retention.
- 1:11:57SPEAKER_00: So how do you do that?
- 1:11:58SPEAKER_00: Well,
- 1:11:59SPEAKER_00: Option one,
- 1:12:00SPEAKER_00: well,
- 1:12:00SPEAKER_00: just,
- 1:12:00SPEAKER_00: you know,
- 1:12:01SPEAKER_00: this is a thought process that kind of leads to this algorithm.
- 1:12:06SPEAKER_00: You know,
- 1:12:06SPEAKER_00: maybe you say,
- 1:12:08SPEAKER_00: like,
- 1:12:08SPEAKER_00: okay,
- 1:12:08SPEAKER_00: I'm doing all these math problems,
- 1:12:10SPEAKER_00: and I got to make up Anki formulas for them,
- 1:12:14SPEAKER_00: or Anki flashcards.
- 1:12:16And so the first thing I think about is,
- 1:12:18like,
- 1:12:18well,
- 1:12:19what do I put on the flashcard?
- 1:12:21On one hand,
- 1:12:22like,
- 1:12:23yeah,
- 1:12:23there's some formulas and stuff in math,
- 1:12:25but so much of math is,
- 1:12:27like,
- 1:12:27you just memorize the formulas that doesn't mean you know how
- 1:12:30to apply them or how
- 1:12:31to solve problems with them right so it's like well how do you
- 1:12:34do enki for actual
- 1:12:36problem solving well okay the solution to that is just put
- 1:12:40problems on the
- 1:12:40flashcards and just have a system that like instead of serving
- 1:12:44you the same problem
- 1:12:45it just replaces it with a slightly different problem and so
- 1:12:48those are kind of your
- 1:12:50SPEAKER_00: your your repetitions that you're doing so
- 1:12:52So now what you run into pretty quickly in this setting is this
- 1:12:58thing that review
- 1:12:59SPEAKER_00: hell basically is like,
- 1:13:01SPEAKER_00: if you've already get this with flashcards,
- 1:13:03SPEAKER_00: right?
- 1:13:04SPEAKER_00: Just like ones that they can go through in like two seconds,
- 1:13:07SPEAKER_00: three seconds as they pile up,
- 1:13:08SPEAKER_00: you got so many flashcards to do.
- 1:13:11and um even though like you can get through them so quickly now
- 1:13:14you you've just
- 1:13:15made this mathematical problem solving right so now it takes
- 1:13:18you like a minute and
- 1:13:19two minutes to go through the go through each flash card and so
- 1:13:23now you've got this
- 1:13:23massive backlog of review to do so what do you do well well so
- 1:13:27if you start
- 1:13:28thinking about it like okay i got this massive say say you're
- 1:13:31in this situation you
- 1:13:31SPEAKER_00: got this massive backlog right and you gotta practice uh
- 1:13:36SPEAKER_00: Solving a linear equation.
- 1:13:39SPEAKER_00: You've got to practice solving a quadratic equation.
- 1:13:41SPEAKER_00: You've got to practice...
- 1:13:44SPEAKER_00: Wait a second.
- 1:13:46If I solve this quadratic by factoring it into linear pieces
- 1:13:51and then solving each...
- 1:13:53equation arising from the linear piece i've effectively
- 1:13:56practiced solving linear
- 1:13:58equations if i solve a linear equation um 2x plus 3 equals 7 i
- 1:14:06am practicing
- 1:14:07subtracting and dividing or yeah that kind of stuff so you're
- 1:14:11like wait i don't
- 1:14:12have to do all these cards if i do this card if i you just take
- 1:14:16a card at random
- 1:14:17SPEAKER_00: okay i did this card
- 1:14:18And so I can remove these other cards from the pile because
- 1:14:21I've effectively
- 1:14:21SPEAKER_00: reviewed those as sub skills.
- 1:14:23SPEAKER_00: So so that now that that starts to get into,
- 1:14:26SPEAKER_00: you know,
- 1:14:27SPEAKER_00: compressing your your reviews,
- 1:14:29SPEAKER_00: because the next thing that you want to do after that is like,
- 1:14:31SPEAKER_00: wait,
- 1:14:31I don't want to just draw cards at random from the pile that I
- 1:14:35have to do.
- 1:14:36SPEAKER_00: I don't even want to necessarily draw the card that is most overdue.
- 1:14:40SPEAKER_00: I want to draw the card that hits all.
- 1:14:43SPEAKER_00: the most,
- 1:14:43SPEAKER_00: the biggest number of do repetitions,
- 1:14:46SPEAKER_00: uh,
- 1:14:47SPEAKER_00: with the one problem,
- 1:14:48SPEAKER_00: what's the one,
- 1:14:48the golden problem that is going to make the biggest dent in my
- 1:14:52review pile.
- 1:14:53And so that's,
- 1:14:55that's,
- 1:14:55uh,
- 1:14:56where you kind of get into like the knowledge graph and fire.
- 1:15:00SPEAKER_00: It was an acronym for fractional implicit repetition.
- 1:15:03SPEAKER_00: So the, the repetitions, uh, you're getting implicit repetitions on, um,
- 1:15:08SPEAKER_00: and sub skills that you're practicing,
- 1:15:11SPEAKER_00: uh,
- 1:15:11SPEAKER_00: as a,
- 1:15:12SPEAKER_00: as a component of,
- 1:15:13SPEAKER_00: of each,
- 1:15:14SPEAKER_00: uh,
- 1:15:14SPEAKER_00: of each more advanced skill.
- 1:15:16SPEAKER_00: And,
- 1:15:16SPEAKER_00: uh,
- 1:15:17SPEAKER_00: oftentimes these,
- 1:15:18SPEAKER_00: these sub skills are,
- 1:15:19SPEAKER_00: are fractional,
- 1:15:20SPEAKER_00: uh,
- 1:15:20SPEAKER_00: in the sense that like,
- 1:15:21SPEAKER_00: well,
- 1:15:22SPEAKER_00: you solved,
- 1:15:23SPEAKER_00: uh,
- 1:15:24SPEAKER_00: this problem, it didn't encompass 100% of the cases of the subskill.
- 1:15:29Like,
- 1:15:30you know,
- 1:15:30maybe quadratic equation,
- 1:15:32you factor it,
- 1:15:33you solve the linear components.
- 1:15:35SPEAKER_00: Okay, you've handled the like ax plus b,
- 1:15:39SPEAKER_00: B equals C equations,
- 1:15:41SPEAKER_00: or maybe it's like a quadratic equation,
- 1:15:43SPEAKER_00: X squared plus some leading coefficient of one.
- 1:15:46So you've handled the linear equations that are like X plus two
- 1:15:50equals three,
- 1:15:51SPEAKER_00: like stuff like that.
- 1:15:52SPEAKER_00: But what about when there's a coefficient on the X?
- 1:15:54SPEAKER_00: There's some cases that you don't cover.
- 1:15:56SPEAKER_00: So you kind of track the,
- 1:15:57SPEAKER_00: you know,
- 1:15:58SPEAKER_00: what,
- 1:15:58SPEAKER_00: what skills are,
- 1:15:59SPEAKER_00: how,
- 1:15:59SPEAKER_00: how,
- 1:15:59SPEAKER_00: what,
- 1:15:59SPEAKER_00: what fraction of the subskills are being encompassed.
- 1:16:02SPEAKER_00: And then you just,
- 1:16:03SPEAKER_00: as you do each,
- 1:16:04SPEAKER_00: each review problem,
- 1:16:05SPEAKER_00: these,
- 1:16:06SPEAKER_00: these repetitions,
- 1:16:07SPEAKER_00: they,
- 1:16:08SPEAKER_00: they,
- 1:16:08SPEAKER_00: they,
- 1:16:09SPEAKER_00: propagate through the graph and trickle down.
- 1:16:12SPEAKER_00: Almost like you can visualize it almost like strikes of lightning.
- 1:16:15SPEAKER_00: Like you practice this thing.
- 1:16:17SPEAKER_00: It covers a bunch of stuff below.
- 1:16:19It's as if you converted this into like a dynamic programming
- 1:16:25problem,
- 1:16:26right?
- 1:16:27And you're leveraging the structure of the graph to kind of
- 1:16:30like reduce the effort
- 1:16:34SPEAKER_01: complexity on the student.
- 1:16:36SPEAKER_01: Exactly.
- 1:16:37SPEAKER_01: This is pretty nice.
- 1:16:38SPEAKER_00: Yeah.
- 1:16:39SPEAKER_00: Yeah.
- 1:16:39So that's a this is one of the kind of secret ingredients in
- 1:16:44this stew of maximum
- 1:16:45SPEAKER_00: efficiency learning.
- 1:16:46SPEAKER_00: I mean,
- 1:16:47SPEAKER_00: each of these things,
- 1:16:48SPEAKER_00: you know,
- 1:16:49make sure a student has prerequisites in place before you ask
- 1:16:51them to learn the new
- 1:16:53SPEAKER_00: thing.
- 1:16:53Optimize the new thing into a sequence of steps that are just
- 1:16:59like very bite sized
- 1:17:00SPEAKER_00: for the student to learn.
- 1:17:02move the student,
- 1:17:03make the,
- 1:17:04like,
- 1:17:05minimum effective doses of guided instruction and,
- 1:17:11SPEAKER_00: and problem solving.
- 1:17:12SPEAKER_00: Don't make them do 30 problems of the same thing when you know,
- 1:17:16SPEAKER_00: like,
- 1:17:16SPEAKER_00: okay,
- 1:17:16SPEAKER_00: they just nailed the first two,
- 1:17:18SPEAKER_00: ready to move on.
- 1:17:18SPEAKER_00: But at the same time, give them more practice.
- 1:17:21The space repetition,
- 1:17:24periodic quizzes,
- 1:17:25all of this kind of combines together.
- 1:17:27SPEAKER_00: Each of these are the multiplier on the learning efficiency.
- 1:17:30SPEAKER_01: And this knowledge graph topic, like,
- 1:17:33SPEAKER_01: You got one from Matt, right?
- 1:17:35SPEAKER_01: Do you think the same overall structure is transferable to other domain?
- 1:17:40SPEAKER_01: I think there's somebody on Twitter that is asking,
- 1:17:45SPEAKER_01: is there one for graduate-level physics,
- 1:17:48SPEAKER_01: computer science?
- 1:17:50SPEAKER_01: Should it be community-built?
- 1:17:52SPEAKER_01: What's your take on that?
- 1:17:53So I think absolutely there are knowledge graphs for any
- 1:17:58subject that is hierarchical.
- 1:18:02like yeah computer science physics anything math adjacent like
- 1:18:07for sure is like a a
- 1:18:08really clear candidate for a knowledge graph now um this is
- 1:18:12actually on our you
- 1:18:13know we we just released a uh math of mathematical methods for
- 1:18:17for physics course
- 1:18:19um and we are we have a machine learning course that that is uh
- 1:18:24like almost almost
- 1:18:26SPEAKER_00: ready so it's we are um
- 1:18:29we are actively expanding out our knowledge graph beyond math
- 1:18:32proper now how far
- 1:18:33can we expand it uh like how you know into biology or or
- 1:18:37history i mean some just
- 1:18:38get flatter and flatter kind of the the further you move from
- 1:18:42math but um my my gut
- 1:18:44says you can you know take a there's probably uh things that
- 1:18:49that transfer over uh
- 1:18:50SPEAKER_00: to some degree um like you i think it's it's kind of like in um
- 1:18:55SPEAKER_00: You know,
- 1:18:56SPEAKER_00: in math,
- 1:18:58SPEAKER_00: the hierarchy is kind of strict,
- 1:19:00where it's like you have to learn linear equations before you
- 1:19:04learn to solve a
- 1:19:05SPEAKER_00: quadratic equation by factoring.
- 1:19:07SPEAKER_00: Like, there's just no...
- 1:19:09SPEAKER_00: You have to, it's, it's, it's, it's a practice subskill.
- 1:19:11SPEAKER_00: There's no way around it.
- 1:19:12SPEAKER_00: And I think like in other subjects,
- 1:19:14SPEAKER_00: like history and,
- 1:19:16SPEAKER_00: you know,
- 1:19:16SPEAKER_00: there's,
- 1:19:16SPEAKER_00: there's sometimes there's not necessarily one right answer to,
- 1:19:21SPEAKER_00: to like,
- 1:19:21SPEAKER_00: you have to learn this before that,
- 1:19:23SPEAKER_00: but there may be ways that,
- 1:19:25SPEAKER_00: that work out nicely.
- 1:19:26SPEAKER_00: Um,
- 1:19:28SPEAKER_00: I don't know.
- 1:19:28I haven't really thought a whole lot about beyond hierarchical
- 1:19:32subjects too much.
- 1:19:33But I know if subject is hierarchical,
- 1:19:36you can teach it with a knowledge graph.
- 1:19:38SPEAKER_01: I remember when I was pushing my learning stuff,
- 1:19:42SPEAKER_01: I was also looking at like literally other things.
- 1:19:47SPEAKER_01: And I realized that there's some subject that has very different shape,
- 1:19:50SPEAKER_01: like math,
- 1:19:53SPEAKER_01: programming,
- 1:19:54SPEAKER_01: physics,
- 1:19:55SPEAKER_01: they all went into the same bucket,
- 1:19:56SPEAKER_01: right?
- 1:19:57SPEAKER_01: Like of tricks that I could use.
- 1:19:59SPEAKER_01: Biology already there, it's like, it was a bit different.
- 1:20:02Like I remember like most of my stuff was compressing and
- 1:20:06summarization,
- 1:20:07right?
- 1:20:08SPEAKER_01: and then making the stuff as vivid in my mind as possible.
- 1:20:11SPEAKER_01: It was more of a modelization thing, right?
- 1:20:13SPEAKER_01: I was using flashcard a lot for facts, right?
- 1:20:15SPEAKER_01: And then I was trying to kind of bring it down to something I understood.
- 1:20:18SPEAKER_01: It didn't matter if it was not actually
- 1:20:21SPEAKER_01: the stuff, especially biochemistry, because nothing makes sense.
- 1:20:24SPEAKER_01: Like all of the protein have like three letter and then like a bunch of,
- 1:20:29SPEAKER_01: and then some of them are funny.
- 1:20:30SPEAKER_01: They're like called sonic hedgehog, right?
- 1:20:33SPEAKER_01: For some reason, right?
- 1:20:34SPEAKER_01: That was very different.
- 1:20:36SPEAKER_01: And there's some things that like,
- 1:20:38I remember my wife is into archeology and like I was helping
- 1:20:42her into one,
- 1:20:43SPEAKER_01: like studying in one of our classes and was like,
- 1:20:45SPEAKER_01: I'm going to also study this stuff.
- 1:20:46SPEAKER_01: And it was like about Greek columns.
- 1:20:49right and all of my trick was like hey what how do i do that
- 1:20:53how do i learn about
- 1:20:54like all these facts about these columns and like these dates
- 1:20:57and stuff and it was
- 1:20:58it was uh much different like uh it was again mostly about
- 1:21:03summarization and
- 1:21:04bringing like the stuff into my world view um but that that was
- 1:21:08very hard like i
- 1:21:09SPEAKER_01: had to like put more uh i would say thought into um
- 1:21:13bringing that material into my stuff compared to the other
- 1:21:16stuff where I could just
- 1:21:17like go and like do and while I was doing I was able to kind of
- 1:21:22like automatically
- 1:21:23SPEAKER_01: subconsciously ingest the information.
- 1:21:25You had mentioned um yeah just the experience of like learning
- 1:21:29biology like
- 1:21:30archaeology and like the different experience so I I uh last
- 1:21:35fall I started kind of
- 1:21:36SPEAKER_00: dipping my toe into into learning um biology um trying to leverage uh
- 1:21:42SPEAKER_00: These cognitive science principles,
- 1:21:45SPEAKER_00: mastery,
- 1:21:46SPEAKER_00: learning,
- 1:21:46SPEAKER_00: this kind of stuff,
- 1:21:47SPEAKER_00: and just,
- 1:21:47SPEAKER_00: like,
- 1:21:48SPEAKER_00: kind of saying,
- 1:21:49SPEAKER_00: what does that even look like?
- 1:21:50And I don't know that I've figured out the most efficient way
- 1:21:55just yet,
- 1:21:55but I guess I can...
- 1:21:57SPEAKER_00: I can tell you a little bit about what, what I kind of stumbled into.
- 1:22:00SPEAKER_00: And I'd be curious to,
- 1:22:02SPEAKER_00: to know like how that matches or does not match with your,
- 1:22:05SPEAKER_00: your own experience and what's worked for you.
- 1:22:07SPEAKER_00: But so it was kind of so with biology,
- 1:22:12SPEAKER_00: I took a similar approach of just like,
- 1:22:15SPEAKER_00: so my,
- 1:22:16SPEAKER_00: my,
- 1:22:16SPEAKER_00: my wife is doing a PhD in virology.
- 1:22:19SPEAKER_00: And so I, a lot of that kind of comes down to genetics.
- 1:22:24SPEAKER_00: And so I was,
- 1:22:25wanted to be able to talk to her about that kind of work and
- 1:22:31not just have it be
- 1:22:31SPEAKER_00: like a tutoring session where she's like,
- 1:22:33SPEAKER_00: Justin,
- 1:22:34SPEAKER_00: this is how...
- 1:22:35SPEAKER_00: Genetics 101,
- 1:22:36SPEAKER_00: like,
- 1:22:36SPEAKER_00: no,
- 1:22:36SPEAKER_00: I want to like actually be able to,
- 1:22:39SPEAKER_00: you know,
- 1:22:39talk at an intelligent level in the conversation and kind of
- 1:22:45understand her work on
- 1:22:46SPEAKER_00: a higher level.
- 1:22:47SPEAKER_00: And so a lot of this was just like,
- 1:22:51SPEAKER_00: you know,
- 1:22:52SPEAKER_00: it kind of came down to understanding,
- 1:22:55SPEAKER_00: kind of identified this model of like,
- 1:22:57SPEAKER_00: okay,
- 1:22:57SPEAKER_00: I need to understand what
- 1:22:59what exactly is a cell and what's going on inside of it that
- 1:23:03allows it to function
- 1:23:04in the body and and and so a lot of this like you said was just
- 1:23:08compressing like
- 1:23:09there's just a lot of stuff a lot of shit going on inside the
- 1:23:12cell right it's like
- 1:23:14how do you on all these like things that the cell is is
- 1:23:17interacting with other
- 1:23:18cells or manufacturing proteins or whatever it just it like a
- 1:23:23lot of it comes down
- 1:23:23SPEAKER_00: to being
- 1:23:24like being automatic,
- 1:23:26understanding this is about like being automatic in the process
- 1:23:30of it.
- 1:23:30SPEAKER_00: Like what's doing what, who's doing what, where.
- 1:23:33SPEAKER_00: And so I kind of turned this into like a problems style.
- 1:23:39SPEAKER_00: So,
- 1:23:39like just minimum effective dose of this part of the process of
- 1:23:44the cell and then
- 1:23:45like okay ask me questions about this like like what happens if
- 1:23:54this happens in the
- 1:23:55in the ribosome while it's producing a protein what ends up
- 1:23:59happening does the
- 1:23:59protein still get produced or is there a failure or what you
- 1:24:03know things things
- 1:24:04SPEAKER_00: like that a lot a lot of these kind of
- 1:24:06almost like the equivalent of mathematical like problem solving
- 1:24:09questions and so
- 1:24:11that that ended up um working pretty well and and got now it
- 1:24:16worked pretty well
- 1:24:17when i was doing it and then i kind of fell off the wagon have
- 1:24:19not done my space
- 1:24:20repetition so i'm very rusty on it now um but uh at my at my
- 1:24:25peak i was i was
- 1:24:26actually able to like have a lot of like uh conversations with
- 1:24:31with my wife about
- 1:24:31SPEAKER_00: how the
- 1:24:32you know just how you know the various genetic variants of
- 1:24:35whatever are causing
- 1:24:37things to happen viruses infecting the cell in this way
- 1:24:41launching this kind of
- 1:24:41attack oh this screws up that mechanism inside the cell that's
- 1:24:45why it's like that
- 1:24:48that's why this happens um yeah the one thing that i was not
- 1:24:51doing though i'd be
- 1:24:52SPEAKER_00: interested to to hear what you think on this is
- 1:24:56SPEAKER_00: I was learning this from an LLM,
- 1:24:59SPEAKER_00: and I was asking it to,
- 1:25:01SPEAKER_00: like,
- 1:25:01SPEAKER_00: it started initially trying to ask me,
- 1:25:05SPEAKER_00: what are your interests?
- 1:25:06SPEAKER_00: Like, tell me more about this and that.
- 1:25:08And I was just like,
- 1:25:09my interest is in learning this stuff as fast and efficiently
- 1:25:13as possible.
- 1:25:13SPEAKER_00: I don't care.
- 1:25:14SPEAKER_00: Speed up.
- 1:25:16SPEAKER_00: Exactly.
- 1:25:17It was like,
- 1:25:17I was like,
- 1:25:18cause it was trying to like,
- 1:25:19I knew what it was doing.
- 1:25:20SPEAKER_00: Cause this is like,
- 1:25:21you're like kind of what you kind of default to sometimes as a
- 1:25:25tutor is like,
- 1:25:25SPEAKER_00: well,
- 1:25:25SPEAKER_00: what are you interested in?
- 1:25:26SPEAKER_00: Why don't we make this relevant?
- 1:25:28SPEAKER_00: And I'm like, I already have it relevant.
- 1:25:29SPEAKER_00: Like my, I want to be able to talk to my wife about this stuff at
- 1:25:32SPEAKER_00: At a high level, like, let's go.
- 1:25:34SPEAKER_00: Just tell me stuff.
- 1:25:35SPEAKER_00: I want this to just burn it into my brain.
- 1:25:38SPEAKER_00: I'm not going to become a biologist.
- 1:25:40I don't need to know like all the little things of like,
- 1:25:44I don't want to explore the space.
- 1:25:46SPEAKER_00: Just tell me like.
- 1:25:47How it goes,
- 1:25:48what works,
- 1:25:48and I want to capture this automaticity as quickly as possible.
- 1:25:53SPEAKER_00: And so I was periodically having to tell it,
- 1:25:55SPEAKER_00: like,
- 1:25:55SPEAKER_00: don't make,
- 1:25:56SPEAKER_00: like,
- 1:25:56SPEAKER_00: I don't want to hear,
- 1:25:57SPEAKER_00: like,
- 1:25:58SPEAKER_00: don't ask me,
- 1:25:59SPEAKER_00: don't offer analogies unless I specifically ask.
- 1:26:03SPEAKER_00: Yeah,
- 1:26:05SPEAKER_00: I guess I'm curious,
- 1:26:06SPEAKER_00: like,
- 1:26:07when you were kind of getting an understanding of these
- 1:26:09biological processes,
- 1:26:12SPEAKER_00: how would you relate it to your own?
- 1:26:13Or like,
- 1:26:14did you mention something about making it like,
- 1:26:17relevant to your own experience?
- 1:26:19SPEAKER_00: Yeah.
- 1:26:19SPEAKER_00: Like imagery or stuff.
- 1:26:21SPEAKER_00: What did that look like when, when you were doing that?
- 1:26:23SPEAKER_00: Yeah.
- 1:26:24And was that like an efficiency speed up or was it like a
- 1:26:27motivational thing or
- 1:26:28SPEAKER_00: I'm,
- 1:26:28SPEAKER_00: I'm just curious to know more.
- 1:26:29It's more,
- 1:26:29it's more than this,
- 1:26:30like,
- 1:26:31um,
- 1:26:31um,
- 1:26:32like a very,
- 1:26:32very abstract concept.
- 1:26:34SPEAKER_01: are kind of hard for your brain to kind of memorize and grasp,
- 1:26:38SPEAKER_01: especially if they have no like relevance to like the world around you,
- 1:26:42SPEAKER_01: right?
- 1:26:43SPEAKER_01: Like, and this makes stuff not stick at all, right?
- 1:26:46SPEAKER_01: Like, especially like memory-wise, right?
- 1:26:49Like most of your brain is like highly hard word for like a
- 1:26:55space and time type of
- 1:26:56SPEAKER_01: computation,
- 1:26:57SPEAKER_01: right?
- 1:26:57SPEAKER_01: Like if you were to think about it and if you were to close your eyes,
- 1:27:01SPEAKER_01: you would be able to remember everything in your house.
- 1:27:03SPEAKER_00: right?
- 1:27:04Like,
- 1:27:04well,
- 1:27:05where every little spot is and like where,
- 1:27:08like stuff interact.
- 1:27:09SPEAKER_01: And if something was different,
- 1:27:11SPEAKER_01: you,
- 1:27:11one day you like open up the door and then there's like a
- 1:27:15furniture that was not
- 1:27:15SPEAKER_01: there before.
- 1:27:16SPEAKER_01: You will not be just like, meh, and then you move on, right?
- 1:27:19SPEAKER_01: You,
- 1:27:19SPEAKER_01: you already mapped this out and then it's not,
- 1:27:21SPEAKER_01: now this thing is there and it should not be there.
- 1:27:24SPEAKER_01: And then it, like a,
- 1:27:25SPEAKER_01: It goes into this cascading of why is it there?
- 1:27:29SPEAKER_01: What's up?
- 1:27:29SPEAKER_01: And then you start to look around and see stuff.
- 1:27:32SPEAKER_01: So at the base level,
- 1:27:34since this is more of a memorization-heavy type of situation
- 1:27:38where you have to
- 1:27:39SPEAKER_01: remember these facts that are not grounded in the real world,
- 1:27:43SPEAKER_01: I'm trying to bring them into stuff that I can visualize.
- 1:27:47SPEAKER_01: And I spend more time on the visualization aspect.
- 1:27:50SPEAKER_01: But it doesn't have to be very, very...
- 1:27:53SPEAKER_01: concrete, it just needs to make sense for me, right?
- 1:27:56SPEAKER_01: So the ribosome,
- 1:27:57SPEAKER_01: for instance,
- 1:27:58like I was kind of almost looking at it as like a little
- 1:28:01factory and I was drawing
- 1:28:03SPEAKER_01: it like this,
- 1:28:04SPEAKER_01: but with this ribosome shape,
- 1:28:05SPEAKER_01: right?
- 1:28:06And I was looking at the stuff that goes into like the peptide
- 1:28:09as like some sort of coal-based
- 1:28:14SPEAKER_01: chain fuel of some kind, right?
- 1:28:17SPEAKER_01: And I was looking at more into this mechanical sense.
- 1:28:20SPEAKER_01: Then if you were to really look at it,
- 1:28:22SPEAKER_01: it's like,
- 1:28:22SPEAKER_01: is it just a long chain that is folded within a specific configuration?
- 1:28:27SPEAKER_01: If you think about it like this, no way I'm going to make it right now.
- 1:28:30I was kind of reducing the level of abstraction to something I
- 1:28:33could pinpoint in my head.
- 1:28:36SPEAKER_01: And then I was doing this for all of the different kind of category.
- 1:28:41SPEAKER_01: It didn't have to make sense for anybody else than me, right?
- 1:28:43SPEAKER_01: And some stuff, like it was more abstract because I was okay with it.
- 1:28:47SPEAKER_01: And some stuff was way less abstract because like,
- 1:28:49SPEAKER_01: yeah,
- 1:28:49SPEAKER_01: this dude need to,
- 1:28:50like this three-letter name protein with like four numbers
- 1:28:54after it need to be
- 1:28:57SPEAKER_01: You need to be a dude, right?
- 1:28:58SPEAKER_01: And then I was making them like a little characteristic.
- 1:29:01But like one thing that really helped me out when I get into
- 1:29:04this space where you
- 1:29:06have to quickly get like the context and it's not grounded into
- 1:29:10like clear
- 1:29:11SPEAKER_01: boundaries,
- 1:29:12SPEAKER_01: right?
- 1:29:12SPEAKER_01: Like the biology stuff is interesting because there's no clear,
- 1:29:16SPEAKER_01: like we're making these boundaries.
- 1:29:18pseudo boundaries but when you look into it's like yeah
- 1:29:20actually it's like it's
- 1:29:21more a bit wishy-washy like that right uh and when you look at
- 1:29:24the fact it's like
- 1:29:26actually these three experiments says this but these two
- 1:29:28experiments says this
- 1:29:29other thing right i'm trying to ground as much as possible the
- 1:29:32space first
- 1:29:34SPEAKER_01: So like okay,
- 1:29:35what's in the cell go go go and then you look at everything
- 1:29:37that is in the cell,
- 1:29:38SPEAKER_01: right?
- 1:29:38SPEAKER_01: There's nothing else.
- 1:29:39There's not a like a fling flow flow that is in the cell
- 1:29:42There's there's none of that.
- 1:29:44SPEAKER_01: There's like this 75 like bigger things that are in there Fantastic.
- 1:29:49SPEAKER_01: I've like closed the space
- 1:29:51SPEAKER_01: a bit, right?
- 1:29:52And then I can start to sharpen my vision about like what they
- 1:29:56are,
- 1:29:56right?
- 1:29:57SPEAKER_01: And how they interact afterward.
- 1:29:59And at the last layer,
- 1:30:00there's this kind of layer of like,
- 1:30:02we don't really know.
- 1:30:03SPEAKER_01: And like,
- 1:30:05SPEAKER_01: I'm trying to like,
- 1:30:06know a lot about what we don't really know on the stuff that
- 1:30:10I'm actually working
- 1:30:11SPEAKER_01: on.
- 1:30:11The rest,
- 1:30:12I'm going to put in the abstraction of like,
- 1:30:13hey,
- 1:30:14work roughly like this.
- 1:30:15SPEAKER_01: That's the flow usually used.
- 1:30:17SPEAKER_01: And then this is also the flow I
- 1:30:19SPEAKER_01: I use in neuroscientific stuff because it's the same thing.
- 1:30:22SPEAKER_01: You have the cell,
- 1:30:23but then they are in a population and then they're in a region
- 1:30:26and then they're in
- 1:30:26the substrate and then they are getting fed and have to live
- 1:30:30and then they're
- 1:30:31SPEAKER_01: shooting out these vesicles and stuff.
- 1:30:34All of this is way too complicated to take all of the little
- 1:30:37details at the same time.
- 1:30:38So you map the space and then you go into understanding each of
- 1:30:41the components and
- 1:30:42SPEAKER_01: then you understand the interaction and you make it as dumb.
- 1:30:46SPEAKER_01: and as vivid as you can.
- 1:30:49SPEAKER_01: And if you do that, then it's more easy to discuss.
- 1:30:52SPEAKER_01: It can lead you astray.
- 1:30:53SPEAKER_01: If your mental model is something and he has some biases,
- 1:30:57SPEAKER_01: it can literally mess up how you're looking at it.
- 1:31:00SPEAKER_01: And we see this sometimes in biology or neuroscience.
- 1:31:04SPEAKER_01: We're thinking about the model of the brain this way,
- 1:31:07SPEAKER_01: And actually, it's not exactly like this.
- 1:31:10SPEAKER_01: For instance,
- 1:31:11SPEAKER_01: the synapse and then the neurons,
- 1:31:16SPEAKER_01: one of the model is that it's shooting up the stuff,
- 1:31:19SPEAKER_01: the vesicle,
- 1:31:20SPEAKER_01: right?
- 1:31:20SPEAKER_01: And then the vesicle entering a specific postsynaptic terminal.
- 1:31:23But in actuality,
- 1:31:25there are microtubules from one neuron that can stick out to
- 1:31:29the other side.
- 1:31:30SPEAKER_01: So they're kind of connected now.
- 1:31:31SPEAKER_01: So your whole idea of like how this works kind of changed now, right?
- 1:31:35SPEAKER_01: It's not the same.
- 1:31:36SPEAKER_01: So you just have to be careful.
- 1:31:37SPEAKER_00: Information at a lower level of scale.
- 1:31:40SPEAKER_00: Yeah.
- 1:31:40SPEAKER_00: That's kind of broke your abstraction.
- 1:31:43SPEAKER_01: Yeah.
- 1:31:43SPEAKER_01: But then like, does it matter?
- 1:31:45SPEAKER_01: Is another element.
- 1:31:47SPEAKER_01: And then like, can we keep it fuzzy for like 90% of the problem?
- 1:31:49SPEAKER_01: Yes.
- 1:31:50SPEAKER_01: Okay.
- 1:31:50SPEAKER_01: We'll keep it there.
- 1:31:51So there's a lot of this going on versus like the more hard
- 1:31:55type of clear cut,
- 1:31:57like a subject.
- 1:31:59you don't have as much as this right you don't have to kind of
- 1:32:02go into like uh
- 1:32:04these these weird setup where this we're not 100 sure how it
- 1:32:08works but here's our
- 1:32:09model and you're like you have to keep in mind like it works in
- 1:32:1295 of the cases but
- 1:32:14five percent of the case you have to have to go deep deep deep
- 1:32:16in there and figure
- 1:32:17SPEAKER_01: out like uh the the the missing thing or like the
- 1:32:21the thing that are we're not sure so yeah and in biology the
- 1:32:24issue is that like
- 1:32:25it's everywhere like uh neuroscience like it's it's everywhere
- 1:32:29like the if you dive
- 1:32:30deep enough into any of the research and like in any of the
- 1:32:33subject the it's not
- 1:32:34like we know really well about the nucleus and then like
- 1:32:38there's this the the other
- 1:32:39uh shell here and then the upper the upper of golgi is like we
- 1:32:42know 100 like we
- 1:32:44SPEAKER_01: don't know much
- 1:32:45SPEAKER_01: And there's like this big error bar in all of this different direction.
- 1:32:50SPEAKER_01: And it can really chip up your understanding because then you're like,
- 1:32:54SPEAKER_01: what even are we looking at right now?
- 1:32:56SPEAKER_01: Right.
- 1:32:56SPEAKER_01: I remember there was one class that I took.
- 1:32:58SPEAKER_01: It was like lipid 2.
- 1:33:00SPEAKER_01: I was learning about lipids.
- 1:33:02SPEAKER_01: So, and it was like the Golgi apparatus.
- 1:33:06And we had like whole classes during the whole semester just
- 1:33:10about that stuff,
- 1:33:11right?
- 1:33:11SPEAKER_01: And how they were imaging the things.
- 1:33:13SPEAKER_01: And then it was taught by a researcher, which was super cool.
- 1:33:16SPEAKER_01: And they were showing us the research.
- 1:33:18SPEAKER_01: But then I knew like I'm at the edge.
- 1:33:20SPEAKER_01: Like I'm asking questions.
- 1:33:21SPEAKER_01: They're like, I don't know.
- 1:33:23SPEAKER_01: because they literally don't know and they can't like they're running the
- 1:33:26experiment and stuff but like there's there's a lot of stuff
- 1:33:28that we're not
- 1:33:29understanding there and like this helped me a lot get the
- 1:33:32better grasp about these
- 1:33:33material because you're you're really really uh at the edge
- 1:33:37sometimes at the
- 1:33:38edge is like it's it's not like a smooth frontier right it's
- 1:33:41like it's a bunch of
- 1:33:43holes a bit everywhere and it was as soon as you solve one
- 1:33:46there's like six
- 1:33:47different holes that are somewhat bigger that pops up that
- 1:33:51connected to the other
- 1:33:51SPEAKER_01: holes when you discuss to the
- 1:33:53researcher versus like some people that are just like teaching
- 1:33:56the stuff you
- 1:33:57realize how much it affected them and how they're teaching
- 1:34:00because they will be
- 1:34:02much more unsure about stuff and then they will tell you about
- 1:34:04the caveat right so
- 1:34:06you're good you tell me what what's what's going on versus a
- 1:34:09teacher that is
- 1:34:09looking at the curriculum for like these more like a soft
- 1:34:13science right no not soft
- 1:34:14SPEAKER_01: science but like soft certainty right uh science
- 1:34:17they're going to tell you like it's like this it's like that
- 1:34:19it's like that but the
- 1:34:20deeper you you dig into the more you realize that we don't know
- 1:34:24much yeah so i just
- 1:34:26have to yeah keep that into consideration right uh yeah yeah
- 1:34:30definitely yeah it
- 1:34:31SPEAKER_00: makes that like yeah one of the nice things about um you know
- 1:34:36SPEAKER_00: developing math curriculum and,
- 1:34:38SPEAKER_00: you know,
- 1:34:39mostly carries over to computer science and physics is that
- 1:34:44like things don't
- 1:34:44SPEAKER_00: really change.
- 1:34:45SPEAKER_00: Like in math, it's just, you know, it's evergreen.
- 1:34:48SPEAKER_00: Like there are, there's new stuff that you find that builds on top of
- 1:34:53like in in directions i hadn't gone before but on the
- 1:34:56foundations generally like
- 1:34:58it's it's locked in place but um yeah i guess uh right and as
- 1:35:04you kind of go out
- 1:35:05into the fuzzier subjects there's these like just complete
- 1:35:08overhauls of mental
- 1:35:09SPEAKER_00: models of what's happening
- 1:35:10SPEAKER_01: And, like, I remember, like, I was, like, doing consciousness research.
- 1:35:14SPEAKER_01: There was this very famous experiment,
- 1:35:16SPEAKER_01: and then,
- 1:35:16SPEAKER_01: like,
- 1:35:16SPEAKER_01: it would publish,
- 1:35:17SPEAKER_01: and then they were doing all this stuff.
- 1:35:19SPEAKER_01: And then,
- 1:35:19SPEAKER_01: like,
- 1:35:19SPEAKER_01: there were other people that tried to replicate it,
- 1:35:21SPEAKER_01: and they realized,
- 1:35:22SPEAKER_01: like,
- 1:35:22SPEAKER_01: it was not actually,
- 1:35:23SPEAKER_01: like,
- 1:35:24SPEAKER_01: 100%,
- 1:35:25SPEAKER_01: like,
- 1:35:25SPEAKER_01: solid.
- 1:35:25But by the time,
- 1:35:27everybody,
- 1:35:27like,
- 1:35:27kind of built up their mental model on this stuff.
- 1:35:29SPEAKER_01: And now it's, like, it's not there anymore, but it's still there.
- 1:35:33SPEAKER_01: It's a whole new world of confusion, right?
- 1:35:37SPEAKER_01: If we bring it back to Matt...
- 1:35:40SPEAKER_01: I really like some of the analogy that you're making with sport.
- 1:35:44And I think that you're also like,
- 1:35:45you were also into Caliste,
- 1:35:47Caliste,
- 1:35:48I never can say it.
- 1:35:49SPEAKER_01: And you make a lot of parallel within it.
- 1:35:56Do you think there's something specific from math that like
- 1:36:01stuff like sport coach
- 1:36:03can learn from it into learning like the actual skills for
- 1:36:10their students?
- 1:36:11SPEAKER_00: Yeah,
- 1:36:12SPEAKER_00: you know,
- 1:36:12SPEAKER_00: one of my favorite,
- 1:36:15I think one of the most quotable figures in the in the study of
- 1:36:19talent development,
- 1:36:21SPEAKER_00: John Wooden,
- 1:36:22SPEAKER_00: legendary basketball coach,
- 1:36:24SPEAKER_00: he he has so many quotes.
- 1:36:26SPEAKER_00: about sports that apply perfectly to learning in general.
- 1:36:31SPEAKER_00: And so I think just any sort of skills based learning,
- 1:36:36SPEAKER_00: at least mathematics,
- 1:36:39SPEAKER_00: physics,
- 1:36:40SPEAKER_00: coding,
- 1:36:40SPEAKER_00: where you actually,
- 1:36:41SPEAKER_00: you know,
- 1:36:41SPEAKER_00: you're having to
- 1:36:42to perform manipulations on whether it's objects of information
- 1:36:46or physical objects
- 1:36:48SPEAKER_00: or whatever.
- 1:36:48When you're performing these actions and you're having to build
- 1:36:53up to more and more
- 1:36:54SPEAKER_00: and more complex actions,
- 1:36:57SPEAKER_00: This all kind of follows a similar suit where it's like,
- 1:37:00SPEAKER_00: you know,
- 1:37:01SPEAKER_00: all these cognitive science principles of mastery learning,
- 1:37:05SPEAKER_00: like get your prerequisite actions in place before you do.
- 1:37:09Well,
- 1:37:09a lot of these are kind of actually more obvious in sports,
- 1:37:12I think.
- 1:37:12SPEAKER_00: Like, you know, let's before.
- 1:37:17I often like to go to figure skating for sports analogies
- 1:37:20because it's just so.
- 1:37:22dependent on underlying skills so before we get you doing a uh
- 1:37:27skating backwards
- 1:37:27let's have you skating forwards before we get you spinning on
- 1:37:31one leg let's get you
- 1:37:32standing on one leg that's the master of learning let's also um
- 1:37:36you know spaced
- 1:37:37repetition um just because you were able to stand on one leg
- 1:37:41last week does not
- 1:37:42mean that you still um know how to do it we have to continue
- 1:37:46practicing this um
- 1:37:47Things like layering,
- 1:37:49just layering on more advanced skills to lock in the
- 1:37:52fundamental skills.
- 1:37:54It's like,
- 1:37:54well,
- 1:37:55you're good at standing on one leg,
- 1:37:57at balancing on one leg on your skate.
- 1:38:00You've gotten kind of,
- 1:38:02your ability at that has saturated because you've been
- 1:38:04practicing
- 1:38:05in an easy environment but once you kind of like if somebody
- 1:38:08tries to push you a
- 1:38:09little bit or trying to like spin around on one leg this is
- 1:38:13gonna this is gonna
- 1:38:15force you to develop robustness in the underlying skills um
- 1:38:19same thing happens in
- 1:38:20math like in calculus there's a there's a joke that like
- 1:38:23calculus is where students
- 1:38:24SPEAKER_00: actually learn algebra because they are
- 1:38:27uh held account to account for applying a lot of outbreak
- 1:38:31techniques um yeah and uh
- 1:38:35there's interleaving um so doing instead of you know if you're
- 1:38:40going to do a
- 1:38:42skating practice um let's not just practice stopping with your
- 1:38:48right leg the whole
- 1:38:49time let's also practice stopping with your left leg and not
- 1:38:52let's not just do
- 1:38:53stopping let's do a whole bunch of other techniques um there's
- 1:38:58um kind of retrieval
- 1:39:00SPEAKER_00: practice even in sports i think is this like well you think of it
- 1:39:04learning it's like on a on a test or quiz you're retrieving
- 1:39:08something from memory
- 1:39:09with no reference you're not getting a warm start to it by
- 1:39:13solving some easier
- 1:39:14problems leading up to it you're just like right on the dot
- 1:39:17under time constraints
- 1:39:18you gotta pull it out of your brain well yeah in sports similar
- 1:39:21thing you gotta
- 1:39:22SPEAKER_00: develop a a level of uh
- 1:39:26SPEAKER_00: Maybe you can do a spin on your figure skate if you practice a few times
- 1:39:33SPEAKER_00: beforehand,
- 1:39:33SPEAKER_00: and then you can do the spin as you get your balance,
- 1:39:36but really you want to get to the point where you don't need
- 1:39:38that practice leading
- 1:39:39SPEAKER_00: up.
- 1:39:41SPEAKER_00: Stuff like that.
- 1:39:42SPEAKER_00: I think this all carries over really, really well.
- 1:39:45SPEAKER_01: Yeah, true.
- 1:39:45And I think like also like having the skills that you're
- 1:39:48figuring it is important
- 1:39:50because you shouldn't think about them like as you're doing the
- 1:39:52stuff like you're
- 1:39:54always going to use like mental load in order to concentrate
- 1:39:57and do the movement
- 1:39:59SPEAKER_01: that is needed.
- 1:40:00SPEAKER_01: But maybe the opportunity is already gone, right?
- 1:40:04SPEAKER_01: Or like you messed up in this sequence.
- 1:40:06SPEAKER_00: Yeah, exactly.
- 1:40:07SPEAKER_00: If you're playing hockey and basketball, basketball is a better analogy.
- 1:40:11Because if you happen to think about how to dribble or how to
- 1:40:15pass your teammates
- 1:40:17SPEAKER_00: or stuff,
- 1:40:18SPEAKER_00: you're not going to be seeing all your teammates around.
- 1:40:20SPEAKER_00: Yeah.
- 1:40:24SPEAKER_01: In my case, I was doing swimming, right?
- 1:40:29SPEAKER_01: You get that also.
- 1:40:30When you're practicing,
- 1:40:31you have to kind of understand all the movement and there's a
- 1:40:35lot of
- 1:40:36coordination about all the things that are happening at the
- 1:40:39same time but when
- 1:40:40you're in the meet and then you're you're going for it like
- 1:40:42there's no you
- 1:40:44shouldn't be thinking about anything if you're thinking about
- 1:40:47something you're
- 1:40:47SPEAKER_01: gonna you're gonna mess it up uh for sure i think we can like uh move the
- 1:40:51discussion back to it like the sports stuff because it's uh a
- 1:40:55bit less abstract
- 1:40:56than math but there's a lot of similarity right there's a lot
- 1:40:59of similarity about
- 1:40:59like how the learning is is done and like um uh we discussed a
- 1:41:04bit about like the
- 1:41:04SPEAKER_01: um
- 1:41:05the sports coach analogy here where like there is some stuff
- 1:41:10that the learning of
- 1:41:11SPEAKER_01: the math happens that the coach could use also.
- 1:41:15SPEAKER_01: It's kind of a similar type of a phenomenon.
- 1:41:18SPEAKER_01: Um,
- 1:41:19SPEAKER_01: but one thing I've realized,
- 1:41:20SPEAKER_01: like when I was,
- 1:41:21SPEAKER_01: um,
- 1:41:21SPEAKER_01: I was doing like competitive swimming,
- 1:41:23SPEAKER_01: right?
- 1:41:23Also like when I was motivated and I knew what I wanted,
- 1:41:26it was really,
- 1:41:27really easy.
- 1:41:28Like when I was training to understand like,
- 1:41:30yeah,
- 1:41:31like I'm trying to like do the 200 butterfly.
- 1:41:34SPEAKER_01: So like, I don't care about the breaststroke stuff.
- 1:41:36SPEAKER_01: It's not that important.
- 1:41:37SPEAKER_01: I wanted to get focused on this.
- 1:41:39SPEAKER_01: So I need to like, uh, like, uh, really nail these five things.
- 1:41:43SPEAKER_01: Right.
- 1:41:44And I was watching recording and like the people at work
- 1:41:46hitting the time that I needed.
- 1:41:48SPEAKER_01: So like there was all of this that like went into my mind.
- 1:41:52SPEAKER_01: But one thing I realized is that like some of the best swimmer,
- 1:41:56SPEAKER_01: whenever they had to coach,
- 1:41:56SPEAKER_01: they were not that great.
- 1:41:58SPEAKER_01: And then when you were looking at the great coach,
- 1:42:01SPEAKER_01: they were not necessarily like the Olympians and stuff like that.
- 1:42:05SPEAKER_01: They were actually a bit lower in their performance.
- 1:42:09SPEAKER_01: So this got me kind of interested.
- 1:42:11SPEAKER_01: Like there seems to be a discrepancy between like
- 1:42:14SPEAKER_01: doing the movement and being really good at it intuitively,
- 1:42:17SPEAKER_01: right,
- 1:42:18SPEAKER_01: without even thinking,
- 1:42:19SPEAKER_01: and then understanding it enough to be able to explain it,
- 1:42:23SPEAKER_01: right?
- 1:42:24SPEAKER_01: Do you think that,
- 1:42:24SPEAKER_01: like,
- 1:42:24SPEAKER_01: there's,
- 1:42:25SPEAKER_01: like,
- 1:42:25SPEAKER_01: a similar phenomenon in math that is happening underlying or not?
- 1:42:29SPEAKER_01: Absolutely.
- 1:42:30SPEAKER_00: Yeah, I think so.
- 1:42:32SPEAKER_00: So,
- 1:42:32SPEAKER_00: I think we touched on this last time talking about how,
- 1:42:35SPEAKER_00: like,
- 1:42:35SPEAKER_00: you know,
- 1:42:36SPEAKER_00: in college you can...
- 1:42:37get a professor who's like just widely recognized in their
- 1:42:40field and you get really
- 1:42:41excited and you're like oh my god i'm gonna learn so much and
- 1:42:44then they come and
- 1:42:45they're just like proof they're on proof on the chalkboard you
- 1:42:48don't even know what
- 1:42:48they look like because they don't turn to to face it just like
- 1:42:51on the board and
- 1:42:53like they don't really grade the quizzes or exam there are no
- 1:42:56quizzes homework sets
- 1:42:57are graded did you do the homework okay good good enough let me
- 1:43:01get back to my
- 1:43:01SPEAKER_00: research like don't bother me kid
- 1:43:03and like yeah so that can that could totally um yeah i think
- 1:43:08that totally holds and
- 1:43:09and i think a lot of times um the people who i mean we're
- 1:43:14talking kind of about in
- 1:43:15in generalizations here but often uh the the the people who
- 1:43:19rise to the top of the
- 1:43:20discipline um they've often gotten a lot of stuff for free uh
- 1:43:25like a lot of like
- 1:43:25you know cognitive machinery or just life situation or whatever
- 1:43:29there are things
- 1:43:29SPEAKER_00: that they
- 1:43:31SPEAKER_00: they often did not have to figure out a solution for.
- 1:43:35SPEAKER_00: Like,
- 1:43:35SPEAKER_00: you know,
- 1:43:35SPEAKER_00: you can think of like a mathematician who just like,
- 1:43:38SPEAKER_00: you know,
- 1:43:38SPEAKER_00: I never really had challenges like writing my work down.
- 1:43:43SPEAKER_00: Like it always like came naturally.
- 1:43:45SPEAKER_00: What do you mean you struggled to write your work down?
- 1:43:49SPEAKER_00: You just do this.
- 1:43:49SPEAKER_00: Or stuff like,
- 1:43:51SPEAKER_00: you know,
- 1:43:52SPEAKER_00: I never,
- 1:43:53SPEAKER_00: like,
- 1:43:53SPEAKER_00: these concepts,
- 1:43:54SPEAKER_00: like,
- 1:43:54what do you mean you can't just read a textbook and then it
- 1:43:57clicks for you and then
- 1:43:58SPEAKER_00: you do the problem?
- 1:43:59SPEAKER_00: You just read the chapter and do the problems.
- 1:44:01SPEAKER_00: What do you mean the problem's hard?
- 1:44:02SPEAKER_00: Like,
- 1:44:02SPEAKER_00: that's not,
- 1:44:03SPEAKER_00: like,
- 1:44:03SPEAKER_00: you know,
- 1:44:04SPEAKER_00: a lot of these people can,
- 1:44:05SPEAKER_00: like,
- 1:44:05SPEAKER_00: take bigger leaps in generalization of things.
- 1:44:09SPEAKER_00: Like, I got a GPU in their head, basically, right?
- 1:44:14SPEAKER_00: Instead of a CPU.
- 1:44:15SPEAKER_00: They're just running on some different...
- 1:44:17SPEAKER_00: Machinery sometimes.
- 1:44:19Not necessarily different machinery,
- 1:44:21but just more optimized in different ways.
- 1:44:24SPEAKER_00: Individual variation in working memory capacity and retention rates.
- 1:44:30Unfortunately,
- 1:44:30this is a reality of learning that people are kind of like...
- 1:44:34SPEAKER_00: built with different specs like this.
- 1:44:36SPEAKER_00: And yeah,
- 1:44:37SPEAKER_00: so a lot of like learning how to teach effectively comes from,
- 1:44:43SPEAKER_00: I think,
- 1:44:43having to really push the limits of your own specs and make you
- 1:44:47kind of like punch
- 1:44:49SPEAKER_00: above your weight in terms of where you would otherwise be if you...
- 1:44:54SPEAKER_00: Yeah,
- 1:44:54SPEAKER_00: anyway,
- 1:44:55SPEAKER_00: so I would totally agree that this happens all the time.
- 1:44:58I think there's a lot of this,
- 1:44:59but also there's a lot of like,
- 1:45:01they don't realize what they did.
- 1:45:04SPEAKER_01: There was one case.
- 1:45:07SPEAKER_01: The guy was like a menace.
- 1:45:11SPEAKER_01: Super nice dude.
- 1:45:12SPEAKER_01: Very nice.
- 1:45:14SPEAKER_01: I never seen somebody so confused and so confusing.
- 1:45:18SPEAKER_01: He was not able to bring it down to the level
- 1:45:22SPEAKER_01: that a student was able to understand it.
- 1:45:25SPEAKER_01: Right.
- 1:45:25SPEAKER_01: He was always talking into his layer of abstraction.
- 1:45:30And like,
- 1:45:31that's kind of all he did for like 14 years and stuff like
- 1:45:35that.
- 1:45:35SPEAKER_01: Right.
- 1:45:35SPEAKER_01: The guy was smart.
- 1:45:36SPEAKER_01: Yes.
- 1:45:36SPEAKER_01: But like,
- 1:45:37SPEAKER_01: I realized early on,
- 1:45:38SPEAKER_01: like there was no way I could,
- 1:45:40SPEAKER_01: I could like,
- 1:45:42SPEAKER_01: I could like pass this class just with this dude.
- 1:45:44SPEAKER_01: So I did everything else.
- 1:45:46SPEAKER_01: I was doing the exercise.
- 1:45:48SPEAKER_01: I was asking him for exercise and stuff.
- 1:45:50SPEAKER_01: And one thing I realized that like,
- 1:45:52SPEAKER_01: that's also what he did before.
- 1:45:54SPEAKER_01: Like it was not just some, right?
- 1:45:56SPEAKER_01: Like it was just drilling exercise again and again.
- 1:45:59SPEAKER_01: And then he pointed me to other part where the exercise were super good.
- 1:46:04SPEAKER_01: Nowhere in the curriculum, this stuff was there.
- 1:46:07SPEAKER_01: You kind of forgot that like,
- 1:46:09SPEAKER_01: you have to ask the student to drill and they were not drilling it.
- 1:46:13SPEAKER_01: And while I was doing this, I understood nothing, right?
- 1:46:15SPEAKER_01: So I was drilling the exercise, trying to figure my way.
- 1:46:18And then like by drilling them,
- 1:46:19I kind of understood like what he was saying.
- 1:46:22SPEAKER_01: And then I could ask questions, right?
- 1:46:24SPEAKER_01: And then he was like, okay, you understand.
- 1:46:26SPEAKER_01: I was like, my dude, right?
- 1:46:28SPEAKER_01: You failed in like the meta learning aspect of the skill, right?
- 1:46:34SPEAKER_01: You failed in understanding that these poor souls here, right?
- 1:46:38SPEAKER_01: They don't understand that you need to do that stuff.
- 1:46:41SPEAKER_01: So you have to teach that too, right?
- 1:46:42SPEAKER_01: And if you don't, and you don't guide them, what are they going to do?
- 1:46:45SPEAKER_01: You can say, hey, read this page.
- 1:46:46SPEAKER_01: And they're like, okay, read it.
- 1:46:48SPEAKER_01: What's next, right?
- 1:46:49SPEAKER_01: They don't know that skill yet.
- 1:46:51SPEAKER_01: So if you don't also layer it out,
- 1:46:53SPEAKER_01: there's no way they're going to understand what you're saying.
- 1:46:55SPEAKER_01: And then you're going to lose them for real.
- 1:46:57And then as you move through more complicated material and more
- 1:47:02up the hierarchy,
- 1:47:03SPEAKER_01: it just gets worse and worse.
- 1:47:04SPEAKER_01: And then at some point, these skills are completely tuned out.
- 1:47:07SPEAKER_01: And then everything fails, right?
- 1:47:09SPEAKER_01: There's a few outliers that succeed, but...
- 1:47:12um the rest fails so i think yes there are some people that
- 1:47:15like have i don't know
- 1:47:15like a better capacity to adjust material but the poor poor
- 1:47:19teacher in my view like
- 1:47:21most of the time they just fail at the basic of teaching right
- 1:47:24they actually
- 1:47:24absolutely the bar is on the floor they haven't learned that
- 1:47:27skill and um i know
- 1:47:29SPEAKER_01: i'm in this mode
- 1:47:30Whenever I see a teacher start to kind of question the mental
- 1:47:35capacity of a student,
- 1:47:37right?
- 1:47:37SPEAKER_01: Like, hey, like, why you're like this?
- 1:47:41SPEAKER_01: Back in my days, like this, this, this, and that.
- 1:47:43SPEAKER_01: And I'm like, yeah, man, you suck at this.
- 1:47:46SPEAKER_01: This is the truth, man.
- 1:47:48SPEAKER_01: You don't, you're maybe good at the material, but the actual teaching...
- 1:47:52this is not it if year after year this is happening and you see
- 1:47:56sometimes with two
- 1:47:57class right it's the same material and one class is not filling
- 1:48:01as much and the
- 1:48:02other one is like it's a slaughterhouse um yeah and then yeah
- 1:48:06they don't they don't
- 1:48:07have the they don't they haven't learned like yeah they just
- 1:48:10intuitively like did
- 1:48:11the right stuff or like they were taught the right stuff and
- 1:48:14then like yeah this is
- 1:48:14always done but like they're not aware that they're completely
- 1:48:19lacking in this
- 1:48:20SPEAKER_01: whole uh whole domain
- 1:48:22yeah well i actually have two uh kind of funny like concrete
- 1:48:25stories of watching
- 1:48:26this play out one is very brief and it's just um you know i i
- 1:48:30saw differential
- 1:48:31equations class intro differential equations class uh being
- 1:48:36like the the the guy i
- 1:48:37wanted to to to teach the differential equation like
- 1:48:41introduction to it is like
- 1:48:42SPEAKER_00: well the the
- 1:48:44The solution to a differential equation is just the kernel of a
- 1:48:46linear operator.
- 1:48:47SPEAKER_00: That's all it is.
- 1:48:48SPEAKER_00: This is like the students have not done differential equations before.
- 1:48:52They have not like really gone over kernels and linear
- 1:48:56operators aside from maybe a
- 1:48:57SPEAKER_00: little bit in linear algebra.
- 1:48:59SPEAKER_00: So just trying to jump to the highest level of abstraction.
- 1:49:02SPEAKER_00: And the students are all just like, what?
- 1:49:05SPEAKER_00: So I was actually tutoring a bit for this class,
- 1:49:10SPEAKER_00: and the instructor was so out of touch with thinking,
- 1:49:13SPEAKER_00: oh,
- 1:49:13SPEAKER_00: the students are working on their proofs,
- 1:49:16SPEAKER_00: they're learning a lot.
- 1:49:17SPEAKER_00: Meanwhile, everybody's just like, what is this stuff?
- 1:49:20SPEAKER_00: And so I'm trying to...
- 1:49:21fill in all the gaps and like it's just terrible another
- 1:49:25instance is uh you know i
- 1:49:27actually um i worked with my my uh sister-in-law a couple years
- 1:49:32ago and she's
- 1:49:32SPEAKER_00: taking real analysis and like i mean real analysis is the
- 1:49:36the stereotypical like math major wrecker course right because
- 1:49:40that's where it
- 1:49:40transitions from a lot of like you're used to you know linear
- 1:49:44algebra multivariable
- 1:49:45calc differential equations it's kind of concrete and like
- 1:49:48maybe taking a methods
- 1:49:49SPEAKER_00: of proof course kind of dip your toe in
- 1:49:51SPEAKER_00: in the water of like abstract proof writing, but real analysis comes.
- 1:49:55And it's like,
- 1:49:55if you're not like 100% solid on this proof writing,
- 1:49:58you're gonna struggle.
- 1:50:00And so the instructor would basically,
- 1:50:03he would serve up maybe like five homework problems
- 1:50:06SPEAKER_00: every week.
- 1:50:07And those five homework problems,
- 1:50:08we're at the highest level of abstraction.
- 1:50:11So it's like,
- 1:50:12we're not,
- 1:50:12we're going to skip over proving that a set is closed or open
- 1:50:16or stuff.
- 1:50:16SPEAKER_00: We're going to jump right into theorems of like,
- 1:50:18prove this closure property of the n-dimensional sphere under
- 1:50:24these conditions.
- 1:50:25SPEAKER_00: And you're like,
- 1:50:27SPEAKER_00: man,
- 1:50:27can we just like work in like one dimension first with like
- 1:50:30some actual numbers,
- 1:50:31SPEAKER_00: not like just like,
- 1:50:32SPEAKER_00: you know,
- 1:50:32SPEAKER_00: stuff like that.
- 1:50:33SPEAKER_00: And so
- 1:50:34SPEAKER_00: Crazy part, like the homework would not even get like really graded.
- 1:50:39SPEAKER_00: It was kind of like a for completion thing.
- 1:50:41SPEAKER_00: So it's just if you attempt it,
- 1:50:43SPEAKER_00: if it looked like you attempted it,
- 1:50:44SPEAKER_00: then you get like you get points.
- 1:50:46SPEAKER_00: And then there were there was a there were no quizzes.
- 1:50:49SPEAKER_00: There were no assessments, except for a midterm and the final.
- 1:50:53SPEAKER_00: The midterm,
- 1:50:54SPEAKER_00: the midterm was given in the middle of the year and it was handed back,
- 1:50:57SPEAKER_00: I think,
- 1:50:58SPEAKER_00: a week before the final.
- 1:50:59SPEAKER_00: So there's no signal.
- 1:51:00SPEAKER_01: There's no feedback loop at all.
- 1:51:02SPEAKER_01: There's no feedback loop.
- 1:51:03SPEAKER_00: Yeah.
- 1:51:04SPEAKER_00: And so I'm told that during class,
- 1:51:05like the way he would run it is he's like write a theorem up on
- 1:51:08the board,
- 1:51:09SPEAKER_00: write down the proof.
- 1:51:09SPEAKER_00: Like, okay, you guys have any questions?
- 1:51:12SPEAKER_00: So he would mistake the student's silence.
- 1:51:15SPEAKER_00: He would say like, oh, great.
- 1:51:16SPEAKER_00: Okay, this is easy.
- 1:51:17SPEAKER_00: You guys already know this.
- 1:51:18SPEAKER_00: Okay, let's move on.
- 1:51:19SPEAKER_00: Meanwhile, everybody's like, we don't even know what questions to ask.
- 1:51:22SPEAKER_00: Yeah.
- 1:51:23SPEAKER_01: Otherwise, he was like a nice guy, like you have said.
- 1:51:27SPEAKER_00: They're nice.
- 1:51:28SPEAKER_00: Very nice.
- 1:51:28SPEAKER_00: They're nice and talented.
- 1:51:30SPEAKER_00: They just don't understand what's going on.
- 1:51:32SPEAKER_00: In the skill.
- 1:51:33SPEAKER_01: But like the teaching, it's so poor.
- 1:51:36SPEAKER_01: I had a class like this, right?
- 1:51:37SPEAKER_01: And it was my final year of undergrad.
- 1:51:41SPEAKER_01: And I took like a very advanced class.
- 1:51:45uh kind of um time series analysis signal processing course
- 1:51:49it's it was i knew it
- 1:51:51would be hard because it was like one of the worst like class
- 1:51:55and um it was it was
- 1:51:56SPEAKER_01: uh given by a chairman of bio engineering of canada and stuff
- 1:52:02SPEAKER_01: Big dude, right?
- 1:52:03SPEAKER_01: The research on point.
- 1:52:04And like,
- 1:52:05I swear to God,
- 1:52:06I couldn't understand what he was saying,
- 1:52:08right?
- 1:52:09And like there was always,
- 1:52:10at least in this case,
- 1:52:11there was always practice every week.
- 1:52:13SPEAKER_01: So we had the practice.
- 1:52:14SPEAKER_01: The feedback was a bit lagged in time,
- 1:52:16SPEAKER_01: but like I could pester the TA and get the feedback that I needed.
- 1:52:19SPEAKER_01: But in class, I understood nothing, right?
- 1:52:21SPEAKER_01: At some point I was fed up.
- 1:52:22SPEAKER_01: So I put myself first row.
- 1:52:24SPEAKER_01: Straight in like on the next to the chalkboard.
- 1:52:29SPEAKER_01: And then every time he would do this,
- 1:52:31SPEAKER_01: I would just raise my hand and said,
- 1:52:32I understood literally nothing for the last three minutes that
- 1:52:36you started here to
- 1:52:38SPEAKER_01: here.
- 1:52:38SPEAKER_01: Yeah.
- 1:52:39SPEAKER_01: Nothing, nothing.
- 1:52:40SPEAKER_01: Like, what are you saying?
- 1:52:41SPEAKER_01: Right.
- 1:52:42SPEAKER_01: Literally, this is what I was doing.
- 1:52:44And what happened was like the other people in the class were
- 1:52:46like,
- 1:52:46yeah,
- 1:52:46me too,
- 1:52:47man.
- 1:52:47SPEAKER_01: Like, what is this?
- 1:52:48SPEAKER_01: Right.
- 1:52:49And then this is at this point that like close to the end that
- 1:52:53they finally
- 1:52:54SPEAKER_01: realized that like,
- 1:52:55SPEAKER_01: OK,
- 1:52:55SPEAKER_01: this is not working.
- 1:52:56SPEAKER_01: And he was taking us back.
- 1:52:58SPEAKER_01: But I had to force him to take us back.
- 1:53:01It's very hard for a student to do this because like there's
- 1:53:04the peer pressure.
- 1:53:05SPEAKER_01: And then you're like, I don't want to look like an idiot.
- 1:53:07SPEAKER_01: but I don't mind.
- 1:53:08SPEAKER_01: I was there.
- 1:53:09SPEAKER_01: I literally chosen, I could have chosen any other course, right?
- 1:53:12SPEAKER_01: I chosen that course because I wanted to understand the material.
- 1:53:15And now you're presenting this stuff to me that was not in like
- 1:53:17the textbook that we had,
- 1:53:18right?
- 1:53:19SPEAKER_01: And I don't get what you're saying, right?
- 1:53:21SPEAKER_01: Maybe it's a notation problem.
- 1:53:22SPEAKER_01: Maybe I'm just a dumbass, right?
- 1:53:24SPEAKER_01: But I want you to walk it through me, right?
- 1:53:26SPEAKER_01: To bring me to your level.
- 1:53:28SPEAKER_01: And actually the course got better.
- 1:53:30SPEAKER_01: It got much better then, but it was like a ping pong every week.
- 1:53:35SPEAKER_01: It was me like in the front and I'm like,
- 1:53:37SPEAKER_01: I don't understand absolutely anything.
- 1:53:39SPEAKER_01: Can you tell?
- 1:53:39SPEAKER_01: And then I was able to,
- 1:53:41SPEAKER_01: to again,
- 1:53:41SPEAKER_01: like show him like,
- 1:53:42SPEAKER_01: okay,
- 1:53:42SPEAKER_01: I get this because you said that,
- 1:53:44SPEAKER_01: but why is it different than this?
- 1:53:45SPEAKER_01: Oh, I'm using another notation.
- 1:53:47SPEAKER_01: And the whole class is like, oh yeah, of course.
- 1:53:50Like if you use another notation that is in the book,
- 1:53:51like how can I understand this stuff?
- 1:53:53SPEAKER_01: Like, Hey, are you not always engineering?
- 1:53:55SPEAKER_01: And I'm like, I'm a neuroscience dude, right?
- 1:53:58SPEAKER_01: I don't have an engineering background.
- 1:54:00SPEAKER_01: And he's like, oh, okay.
- 1:54:01Like,
- 1:54:01yeah,
- 1:54:01maybe like that will be a good thing to know before we go over
- 1:54:05there.
- 1:54:06SPEAKER_00: Know who you're teaching it to.
- 1:54:08SPEAKER_00: Yeah, absolutely.
- 1:54:09SPEAKER_00: You know,
- 1:54:09SPEAKER_00: another thing that strikes me about the,
- 1:54:11SPEAKER_00: how you're like,
- 1:54:12SPEAKER_00: you know,
- 1:54:12SPEAKER_00: you're sitting in the front row asking a bunch of questions.
- 1:54:14So in addition to kind of like identifying these kind of
- 1:54:18missing prerequisites for,
- 1:54:20SPEAKER_00: you know,
- 1:54:20SPEAKER_00: a lot of these prerequisites shouldn't even be prerequisites.
- 1:54:22SPEAKER_00: It's like a notation change or whatever.
- 1:54:24SPEAKER_00: It's like, why would you do that?
- 1:54:25SPEAKER_00: But
- 1:54:26SPEAKER_00: I'm sure there's other parts that are,
- 1:54:28SPEAKER_00: like,
- 1:54:28SPEAKER_00: legitimate missing prerequisites that are just kind of skimmed over,
- 1:54:31SPEAKER_00: like,
- 1:54:31SPEAKER_00: forcing them to backfill.
- 1:54:32I think another thing that probably made this approach work
- 1:54:37really well is that
- 1:54:38SPEAKER_00: you're kind of,
- 1:54:40SPEAKER_00: you're effectively simulating minimum effective doses of,
- 1:54:43SPEAKER_00: like,
- 1:54:43SPEAKER_00: instruction and active,
- 1:54:45SPEAKER_00: you know,
- 1:54:46SPEAKER_00: you're generating,
- 1:54:47SPEAKER_00: like,
- 1:54:47SPEAKER_00: you're thinking about it,
- 1:54:48SPEAKER_00: kind of answer,
- 1:54:49SPEAKER_00: trying to get your own question answered.
- 1:54:51SPEAKER_00: It's like a reference.
- 1:54:52SPEAKER_00: I took in the class.
- 1:54:53SPEAKER_01: I paid for this.
- 1:54:54SPEAKER_01: Like,
- 1:54:54SPEAKER_01: I'm paying for this.
- 1:54:56SPEAKER_01: I'm here.
- 1:54:57SPEAKER_01: I'm going to get the material I need right now.
- 1:54:59SPEAKER_01: And if I don't need and I'm not
- 1:55:01SPEAKER_01: I'm not socializing.
- 1:55:03SPEAKER_01: I don't care about the dude.
- 1:55:05SPEAKER_01: Great guy.
- 1:55:06SPEAKER_01: Absolutely great guy.
- 1:55:07SPEAKER_01: I don't care about him personally.
- 1:55:09SPEAKER_01: What I care about is getting my learning in.
- 1:55:11And if now it's so easy,
- 1:55:13I might just skip the class and do the textbook and be done
- 1:55:16with it.
- 1:55:17SPEAKER_01: At some point,
- 1:55:21SPEAKER_01: it's a bit sad,
- 1:55:21SPEAKER_01: but with some teacher,
- 1:55:23SPEAKER_01: you have no choice but to literally go and see them
- 1:55:27ask questions and go see them in the office hour and just like
- 1:55:30get the knowledge out of them.
- 1:55:32SPEAKER_01: Like forcefully, almost.
- 1:55:33SPEAKER_00: Well, this is like a broken model of reality.
- 1:55:36SPEAKER_00: You don't teach a student by talking at them for an hour.
- 1:55:40SPEAKER_00: Like that's going to overload.
- 1:55:41SPEAKER_00: And unless the student already knows the material already,
- 1:55:45SPEAKER_00: like they're going to be overloaded with like their working memory,
- 1:55:48SPEAKER_00: their cognitive load.
- 1:55:49SPEAKER_00: You got to be breaking this stuff up.
- 1:55:51Like,
- 1:55:51I mean,
- 1:55:52filling in the prerequisites,
- 1:55:53but also having students like,
- 1:55:55you know,
- 1:55:56SPEAKER_00: like why would you talk at a student for,
- 1:55:58SPEAKER_00: for even like 10 minutes or more before having them like,
- 1:56:01SPEAKER_00: do like just do an exercise.
- 1:56:03SPEAKER_00: And,
- 1:56:03SPEAKER_00: you know,
- 1:56:04SPEAKER_00: some,
- 1:56:04SPEAKER_00: some classes do a kind of like a very,
- 1:56:08like try to move in this direction by doing like clicker
- 1:56:11exercises and stuff.
- 1:56:12SPEAKER_00: But, but that's, uh,
- 1:56:14SPEAKER_00: I mean, it's only a baby step in this direction to what it needs to be.
- 1:56:18SPEAKER_00: But yeah, totally.
- 1:56:20SPEAKER_01: There's something that I read on your blog that,
- 1:56:22SPEAKER_01: like,
- 1:56:22SPEAKER_01: kind of go into the same line,
- 1:56:25SPEAKER_01: which is,
- 1:56:25SPEAKER_01: like,
- 1:56:26automaticity in computation is a prerequisite for conceptual
- 1:56:29understanding,
- 1:56:30SPEAKER_01: right?
- 1:56:30SPEAKER_01: Like, not a substitute for it, but, like, a prerequisite.
- 1:56:33And walk us through your reasoning about why this is so,
- 1:56:36because I think it touched here.
- 1:56:37SPEAKER_01: And, like...
- 1:56:39could you think of ways of like not having this automaticity
- 1:56:43but still having the
- 1:56:44understanding is it even possible in your in your okay so
- 1:56:48here's i i think this
- 1:56:49like i think a lot of people who would um you know clash with
- 1:56:53this or like if
- 1:56:54there's a line drawn in the sand on this there's a battle
- 1:56:56between two sides i think
- 1:56:57SPEAKER_00: the point of contention it really boils down to what what does it mean to
- 1:57:01SPEAKER_00: understand something like there's
- 1:57:03SPEAKER_00: So,
- 1:57:04SPEAKER_00: okay,
- 1:57:04SPEAKER_00: there's a kind of conceptual,
- 1:57:05SPEAKER_00: like,
- 1:57:05SPEAKER_00: you know,
- 1:57:06SPEAKER_00: you watch a video on,
- 1:57:07SPEAKER_00: say,
- 1:57:08SPEAKER_00: neural networks,
- 1:57:09SPEAKER_00: right?
- 1:57:09SPEAKER_00: You don't know really how they,
- 1:57:11SPEAKER_00: you've heard neural networks kind of underpin a lot of like modern AI,
- 1:57:14SPEAKER_00: and you're like,
- 1:57:15SPEAKER_00: you're interested in it.
- 1:57:16SPEAKER_00: And so you watch a video on YouTube.
- 1:57:18And you learn about back propagation and you learn about
- 1:57:23regularization and some of
- 1:57:24SPEAKER_00: these things and train tests,
- 1:57:28SPEAKER_00: whatever.
- 1:57:29SPEAKER_00: And so you get a little familiar with this stuff.
- 1:57:32SPEAKER_00: Do you understand it?
- 1:57:34SPEAKER_00: You definitely don't have automaticity.
- 1:57:35SPEAKER_00: You haven't worked any problems in it.
- 1:57:36SPEAKER_00: You haven't set up anything on your own.
- 1:57:39SPEAKER_00: Do you understand it?
- 1:57:40SPEAKER_00: Well, there's some kind of, I guess, you're familiar with
- 1:57:44SPEAKER_00: with some of the terms.
- 1:57:46SPEAKER_00: I personally,
- 1:57:46SPEAKER_00: I would,
- 1:57:47SPEAKER_00: whenever I talk about understanding,
- 1:57:49SPEAKER_00: I mean,
- 1:57:49like understanding to the level of like conceptual
- 1:57:53understanding to the highest
- 1:57:55SPEAKER_00: degree,
- 1:57:56SPEAKER_00: conceptualizing,
- 1:57:57SPEAKER_00: you know,
- 1:57:58SPEAKER_00: there's a difference between
- 1:57:59how maybe like a a basketball player conceptualizes strategy
- 1:58:04versus somebody who
- 1:58:05just watches basketball games and because the basketball player
- 1:58:08like knows like at
- 1:58:10SPEAKER_00: the fundamental level these strategies are like work or
- 1:58:14SPEAKER_00: don't work because of these movements that I'm doing.
- 1:58:18SPEAKER_00: When they're thinking about these strategies and stuff,
- 1:58:21the parts of their brain that are actually activating movements
- 1:58:24and things.
- 1:58:25SPEAKER_00: It's almost like simulating a simulation
- 1:58:29in their head of all the underlying things versus somebody
- 1:58:32who's just seen this at
- 1:58:33SPEAKER_00: a high level.
- 1:58:34SPEAKER_00: And so I think,
- 1:58:35SPEAKER_00: I mean,
- 1:58:36SPEAKER_00: you can acquire maybe like,
- 1:58:37SPEAKER_00: um,
- 1:58:38SPEAKER_00: some familiarity with just the like,
- 1:58:40SPEAKER_00: okay,
- 1:58:40SPEAKER_00: back prop is something involves hardcore multivariable calculus.
- 1:58:44And you somehow figure out what,
- 1:58:46how to adjust the weights of the network.
- 1:58:48SPEAKER_00: Like,
- 1:58:49SPEAKER_00: I mean,
- 1:58:49SPEAKER_00: if you call that like conceptual understanding,
- 1:58:52SPEAKER_00: but yeah,
- 1:58:52SPEAKER_00: I guess you can,
- 1:58:53SPEAKER_00: you can get that without having automaticity on,
- 1:58:56SPEAKER_00: on the,
- 1:58:56SPEAKER_00: um,
- 1:58:57SPEAKER_00: you know, actually the concrete computations.
- 1:59:00SPEAKER_00: But if you talk about backprop in terms of like,
- 1:59:02SPEAKER_00: you know,
- 1:59:03SPEAKER_00: how do I optimize,
- 1:59:04SPEAKER_00: like,
- 1:59:05SPEAKER_00: say this issue is happening.
- 1:59:07And like the neural network did this,
- 1:59:09this is what the loss curve looks like.
- 1:59:11SPEAKER_00: And this is what layer three reads as like, what's wrong with it?
- 1:59:15SPEAKER_00: Like,
- 1:59:16SPEAKER_00: you know,
- 1:59:16SPEAKER_00: that's,
- 1:59:17I would say that's a level of conceptual understanding that you
- 1:59:21can't get unless
- 1:59:23SPEAKER_00: you have actually worked through the computations.
- 1:59:25SPEAKER_01: I agree here.
- 1:59:26SPEAKER_01: I think there's two things,
- 1:59:30SPEAKER_01: especially on the machining side,
- 1:59:31SPEAKER_01: is that you need to have done the movement,
- 1:59:34SPEAKER_01: right?
- 1:59:35And then,
- 1:59:36in my view,
- 1:59:37you need to have diagnosed enough of the movement around the
- 1:59:42movement.
- 1:59:44in order to be able to to like directly see it and so like oh
- 1:59:48you you look at the
- 1:59:49last curve and you're like this is your problem you should fix
- 1:59:51that and then like
- 1:59:52then you you do it this is like the the point where you're
- 1:59:55you're truly able to
- 1:59:57understand it right you're able to solve problem not even with
- 2:00:01without even kind of
- 2:00:02just looking at the output you're like this is what happened
- 2:00:04for sure yeah i think
- 2:00:05it like in sport also you see that like uh like the i think the
- 2:00:09coach the one that
- 2:00:11are great they are at this level they've already done the
- 2:00:13movement it's not like
- 2:00:14SPEAKER_01: The guy is an NBA and he has no clue and he just watched videos.
- 2:00:18SPEAKER_01: That's not it.
- 2:00:19SPEAKER_01: He already played.
- 2:00:20SPEAKER_01: He knows the movement.
- 2:00:21SPEAKER_01: Right.
- 2:00:22But then he knows also like how the movement comes together to
- 2:00:25get to the final result.
- 2:00:26SPEAKER_01: Right.
- 2:00:27SPEAKER_01: And like, I think like a pro athlete, maybe miss this part.
- 2:00:30Like they miss the,
- 2:00:32how to get to whatever,
- 2:00:33because they're using the coach.
- 2:00:35SPEAKER_01: They don't have to think about this.
- 2:00:36SPEAKER_01: They just need to go and do the stuff.
- 2:00:38SPEAKER_01: Right.
- 2:00:38SPEAKER_01: So,
- 2:00:39they are very good but did not necessarily have like the full
- 2:00:43understanding of like
- 2:00:43yeah you have to go and sleep like this you have to improve
- 2:00:47like this you have your
- 2:00:48diet need to be this stuff right all of this need to like come
- 2:00:51around in order for
- 2:00:52you to get to like the actual actual goal um but that level of
- 2:00:56understanding like
- 2:00:57you like if he was just a guy that like just watched a lot of
- 2:01:00stuff but never kind
- 2:01:01SPEAKER_01: of conceptualize it in its own
- 2:01:03kind of own space and own movement there's very little that
- 2:01:06they can add unless
- 2:01:07like it's like just a surface level like hey maybe maybe try
- 2:01:10this and then you're
- 2:01:11like okay why i don't know and then you move on the
- 2:01:13conversation i've got a
- 2:01:15concrete example for for this again in the machine learning
- 2:01:19setting um so it comes
- 2:01:22SPEAKER_00: from when i was uh teaching this um this really advanced um
- 2:01:27applied math and computer science uh elective sequence within
- 2:01:30math academy's
- 2:01:31original pasadena program and so we got students from the point
- 2:01:35of you know they
- 2:01:35came in like having their core engineering math in place they'd
- 2:01:39learn linear
- 2:01:40algebra multi-variable calc and everything and and we scaffold
- 2:01:43them all the way up
- 2:01:44through like coding exercises building machine learning models
- 2:01:48from scratch no
- 2:01:49SPEAKER_00: libraries it's like re-implementing um
- 2:01:52SPEAKER_00: papers in 90s artificial intelligence.
- 2:01:56And one of the things that I remember just that really stood
- 2:02:00out to me as like just
- 2:02:01validating this kind of like you have to learn the movements in
- 2:02:05order to really
- 2:02:07SPEAKER_00: understand.
- 2:02:07what's going on is when we were talking about um exploding
- 2:02:10gradients and vanishing
- 2:02:12gradients in neural networks and so i i remember having talked
- 2:02:16about uh this
- 2:02:17previously with with you know some people who are who are
- 2:02:20interested in um machine
- 2:02:22learning but had not they didn't have the really the math
- 2:02:25foundations they had not
- 2:02:25SPEAKER_00: gone through the um the computations and it it
- 2:02:30um it was always like a kind of like a fuzzy area where it's
- 2:02:33like wait what do you
- 2:02:34mean why why do they explode or vanish like okay i guess the
- 2:02:38activation function
- 2:02:39the derivative of the activation function if it is that the
- 2:02:43slope is too high or
- 2:02:45it's too low or is it if it's unbounded or like they're just
- 2:02:47kind of like you know
- 2:02:48something about the activation function then it's like it's bad
- 2:02:51and then like bad
- 2:02:53stuff happens all right i understand let's move on let's move
- 2:02:55to the cool stuff but
- 2:02:56SPEAKER_00: with these kids they
- 2:02:58SPEAKER_00: I made them work out some simplified scenarios,
- 2:03:03SPEAKER_00: like small,
- 2:03:04SPEAKER_00: like maybe a two,
- 2:03:06three layer neural net of like two nodes in each layer with
- 2:03:10like some nice numbers.
- 2:03:11SPEAKER_00: So the computation took like maybe like,
- 2:03:13three four minutes or or whatever but they could see very
- 2:03:17clearly the effect of
- 2:03:19different activation functions and so i would like okay what if
- 2:03:23we use let's just
- 2:03:24say we use an exponential function as the activate or just even
- 2:03:28a a linear function
- 2:03:29with with uh you know um or a quadratic or something where the
- 2:03:33the derivative is is
- 2:03:34continually increasing and then they would see in their in
- 2:03:38their computations like
- 2:03:39SPEAKER_00: wait
- 2:03:40When I chain roll it and I take the derivative of this thing
- 2:03:42and then I multiply
- 2:03:43SPEAKER_00: it,
- 2:03:43SPEAKER_00: it just keeps getting bigger and bigger.
- 2:03:45SPEAKER_00: What do you mean?
- 2:03:46Like,
- 2:03:46why are you,
- 2:03:46why did you give me a bad,
- 2:03:48Justin,
- 2:03:48you gave me a bad activation function.
- 2:03:49SPEAKER_00: Why did you do that?
- 2:03:51SPEAKER_00: And then I was like, ah, all right, you see the problem.
- 2:03:53Why don't I give you an activation problem that doesn't have
- 2:03:56that property?
- 2:03:57SPEAKER_00: And then it'll just go, go flat.
- 2:03:58SPEAKER_00: And they're like, okay, great.
- 2:03:59SPEAKER_00: Thanks.
- 2:03:59SPEAKER_00: And then I, and then they, they do that.
- 2:04:01SPEAKER_00: They work with computation.
- 2:04:02SPEAKER_00: They're like, wait, wait a second.
- 2:04:03they just get one round in like and they already anticipate how
- 2:04:06it's going to play
- 2:04:07out they're like wait this thing is already like basically the
- 2:04:10derivative is is
- 2:04:11saturating it's it's gone to to to zero and uh and and and it's
- 2:04:16just all the terms
- 2:04:17are going to go to zero like i don't even have to work out the
- 2:04:19rest i already know
- 2:04:20how it's going to play out because you just chain rule multiply
- 2:04:23chain rule multiply
- 2:04:24it goes to zero goes to zero and then at that point like they
- 2:04:28get it they get the
- 2:04:30SPEAKER_00: exploding and vanishing like in the
- 2:04:32SPEAKER_01: Like some researchers that are super good, they feel it.
- 2:04:37SPEAKER_01: They already kind of feel it.
- 2:04:40And it's very important in research,
- 2:04:43in my view,
- 2:04:44because some of the stuff is not...
- 2:04:47SPEAKER_01: quote unquote true, right?
- 2:04:48SPEAKER_01: Like the ground is there, but like barely, right?
- 2:04:52SPEAKER_01: So some stuff that we,
- 2:04:54SPEAKER_01: it's because of this,
- 2:04:55SPEAKER_01: it's like,
- 2:04:55SPEAKER_01: yeah,
- 2:04:55SPEAKER_01: but there's like four asterisks around it.
- 2:04:58SPEAKER_01: And when you actually like feel the stuff and like understand it,
- 2:05:01SPEAKER_01: you know what you can try and where you can move.
- 2:05:04SPEAKER_01: And then you can get like some really cool result after that.
- 2:05:07Some are super empirical because you use like that kind of like
- 2:05:11empirical based knowledge.
- 2:05:13But the true understanding come here in my view,
- 2:05:15like it's like you do the stuff
- 2:05:17SPEAKER_01: And you're not able to explain it.
- 2:05:19SPEAKER_01: Yeah, okay.
- 2:05:20SPEAKER_01: You can build up the theory and do it in a systematic way in this case.
- 2:05:24SPEAKER_01: And then you're able to explain what's going on.
- 2:05:26SPEAKER_01: But empirically, you can feel which of the direction to go.
- 2:05:31I wanted to move on to a funny question about confusion in
- 2:05:37learning,
- 2:05:38right?
- 2:05:39SPEAKER_01: Do you see any...
- 2:05:40SPEAKER_01: upside of confusion in learning, just any of them.
- 2:05:46SPEAKER_01: I say that because in my own research and professional work,
- 2:05:50whenever I made a very good breakthrough and then it was kind
- 2:05:53of rewarding in some
- 2:05:55SPEAKER_01: safer form,
- 2:05:57SPEAKER_01: usually I had guidance.
- 2:05:58SPEAKER_01: I know where I wanted to go.
- 2:06:00SPEAKER_01: I know, roughly speaking, what was going on.
- 2:06:04SPEAKER_01: But I had to sit in confusion and kind of be just okay.
- 2:06:09with being there for an excellent period of time while I figure
- 2:06:11out my stuff right
- 2:06:13but like that was something I had to do but like the best stuff
- 2:06:16that I produced was
- 2:06:18when I was in this situation what's your what's your take on
- 2:06:20that so I I think I
- 2:06:22think the experience can can vary a lot depending on whether
- 2:06:27you're kind of like
- 2:06:28scaling up on you know known knowledge that's just like yeah
- 2:06:31this is figured out
- 2:06:32SPEAKER_00: like just learn it versus kind of at the edge of the frontier with
- 2:06:36Things are getting really fuzzy and like what's even true and
- 2:06:38stuff like that.
- 2:06:39SPEAKER_00: So I think what the confusion represents like that,
- 2:06:43SPEAKER_00: you know,
- 2:06:43the moment where it clicks in place and you just have that
- 2:06:47insight and you're like,
- 2:06:47SPEAKER_00: oh,
- 2:06:48SPEAKER_00: my goodness,
- 2:06:48SPEAKER_00: this is this is it.
- 2:06:50SPEAKER_00: And then it clicks.
- 2:06:52SPEAKER_00: So I think when you are going through the,
- 2:06:54SPEAKER_00: you know,
- 2:06:55SPEAKER_00: the knowledge graph of just stuff that is known,
- 2:06:58SPEAKER_00: this is foundational material.
- 2:07:01I mean,
- 2:07:02I think what that amounts to is really like you're missing some
- 2:07:05prerequisites.
- 2:07:06SPEAKER_00: prerequisites and you build up all like you,
- 2:07:10maybe there's like this topic sits a bunch or sits atop maybe
- 2:07:15like five different
- 2:07:17SPEAKER_00: prerequisites and,
- 2:07:19SPEAKER_00: and,
- 2:07:19SPEAKER_00: or maybe say like five,
- 2:07:2250,
- 2:07:22100 ancestors,
- 2:07:24ancestor topics,
- 2:07:25prerequisites,
- 2:07:26prerequisites and stuff.
- 2:07:28And there's some there's one learning path in there that you
- 2:07:30haven't really,
- 2:07:31SPEAKER_00: you know,
- 2:07:32SPEAKER_00: it's just missing.
- 2:07:33SPEAKER_00: You haven't filled out.
- 2:07:34SPEAKER_00: So you got most of it filled out.
- 2:07:35SPEAKER_00: But it's there's one pillar that's kind of incomplete.
- 2:07:38SPEAKER_00: And you're just you're banging your head on the stop.
- 2:07:40SPEAKER_00: You know what?
- 2:07:41SPEAKER_00: What?
- 2:07:42SPEAKER_00: what is wrong?
- 2:07:43SPEAKER_00: And then at some point,
- 2:07:44SPEAKER_00: like you kind of,
- 2:07:45whether you kind of stumble into it or somebody directs you to
- 2:07:48it or whatever,
- 2:07:49SPEAKER_00: you kind of realize,
- 2:07:50SPEAKER_00: oh my God,
- 2:07:51SPEAKER_00: wait,
- 2:07:51SPEAKER_00: this,
- 2:07:52SPEAKER_00: that's the,
- 2:07:53SPEAKER_00: oh,
- 2:07:53SPEAKER_00: so this works that way.
- 2:07:54SPEAKER_00: And that's how it works.
- 2:07:56SPEAKER_00: And then it kind of clicks more into place.
- 2:07:59SPEAKER_00: You kind of like fill in that, that missing, that missing pillar.
- 2:08:03SPEAKER_00: And it feels amazing, right?
- 2:08:04SPEAKER_00: It's like the click, it's the aha moment.
- 2:08:07SPEAKER_00: But when you're in that foundational body of knowledge, it's like, well,
- 2:08:11there was a lot of time that you had kind of wasted being
- 2:08:15confused that like,
- 2:08:17SPEAKER_00: if you just had that pillar in place beforehand,
- 2:08:20SPEAKER_00: you wouldn't have had to like sit there for,
- 2:08:22SPEAKER_00: for hours or days,
- 2:08:24SPEAKER_00: just ruminating over this thing.
- 2:08:26SPEAKER_00: And then you would have the pillar.
- 2:08:27SPEAKER_00: It would have,
- 2:08:28SPEAKER_00: you wouldn't have needed,
- 2:08:29SPEAKER_00: you wouldn't have an experience,
- 2:08:31SPEAKER_00: a click and aha moment because it would have just been like,
- 2:08:33SPEAKER_00: wait,
- 2:08:34SPEAKER_00: what's the big deal.
- 2:08:34SPEAKER_00: We just put these prerequisites together.
- 2:08:36SPEAKER_00: Like what was hard, man.
- 2:08:38SPEAKER_00: Like that'd be it.
- 2:08:40SPEAKER_00: And so yeah,
- 2:08:41SPEAKER_00: So I think when you're in the foundational body of knowledge,
- 2:08:48SPEAKER_00: these aha moments are actually something you want to try to avoid.
- 2:08:52SPEAKER_00: I mean,
- 2:08:53SPEAKER_00: if you're in a well-structured curriculum,
- 2:08:54SPEAKER_01: you won't actually have – I know it sounds like a little –
- 2:08:57SPEAKER_00: counterintuitive,
- 2:09:00SPEAKER_00: but it's like,
- 2:09:01SPEAKER_00: I remember somebody,
- 2:09:02SPEAKER_00: we actually had a,
- 2:09:04SPEAKER_00: at Math Academy,
- 2:09:06SPEAKER_00: we were talking to,
- 2:09:07SPEAKER_00: this was years ago,
- 2:09:07SPEAKER_00: there was a,
- 2:09:09SPEAKER_00: somebody we were talking to about like recommendations for growth
- 2:09:12on on social media and stuff and and they had recommended
- 2:09:15something like well you
- 2:09:16gotta give it like a like have a nice name to these like aha
- 2:09:20moments or snap
- 2:09:21moments and like and and like talk about these things and i was
- 2:09:24just thinking like
- 2:09:25wait that's like the opposite we try to smooth it out like so
- 2:09:28you don't have that
- 2:09:29like you're climbing out of the hole and you're like i made it
- 2:09:32i made it like no
- 2:09:33SPEAKER_00: you just make it smooth now
- 2:09:35SPEAKER_00: The thing, eventually you run out of this knowledge that's known, right?
- 2:09:39SPEAKER_00: And you get to the end of the frontier where things are fuzzy.
- 2:09:44SPEAKER_00: And that's kind of,
- 2:09:45SPEAKER_00: I think,
- 2:09:46SPEAKER_00: like you do,
- 2:09:47SPEAKER_00: I think this is more what you're talking about,
- 2:09:48SPEAKER_00: where it's like you're kind of in a setting where...
- 2:09:51SPEAKER_00: This has not been fully mapped out.
- 2:09:53You're going to run into situations where you just have missing
- 2:09:57prerequisites
- 2:09:58because everybody has missing prerequisites because nobody
- 2:10:01knows what are the
- 2:10:02SPEAKER_00: prerequisites.
- 2:10:03SPEAKER_00: That's kind of what we're trying to figure out.
- 2:10:05SPEAKER_00: We're trying to map out the.
- 2:10:07SPEAKER_00: how to extend this body of knowledge further.
- 2:10:09And so part of this mapping it out means like identifying what
- 2:10:13reframes or
- 2:10:14SPEAKER_00: prerequisites are necessary.
- 2:10:16SPEAKER_00: And I think as a researcher,
- 2:10:19that sounds like it would be a very valuable moment where
- 2:10:22you're kind of sitting in
- 2:10:24SPEAKER_00: confusion.
- 2:10:25SPEAKER_00: Why am I confused?
- 2:10:27SPEAKER_00: It's because there's some missing prerequisite.
- 2:10:29SPEAKER_00: Holy crap, if I figure out what this is,
- 2:10:33resolve this everybody around me is confused too if i can
- 2:10:36figure this out that's
- 2:10:37going to be a high value contribution yeah so yeah yeah that's
- 2:10:40that's my take in
- 2:10:41this in this setup um like i i learned to kind of love the the
- 2:10:46company like when i
- 2:10:47SPEAKER_01: see there's like no confusion i'm like
- 2:10:49SPEAKER_01: I actually don't need to be there.
- 2:10:50SPEAKER_01: Like somebody else could take my place here and then like do the stuff.
- 2:10:54SPEAKER_01: We're good to go here.
- 2:10:55SPEAKER_01: But when I see like some area where there's like a lot of confusion,
- 2:10:58SPEAKER_01: I find like there's an opportunity to kind of like do something great,
- 2:11:03SPEAKER_01: right?
- 2:11:03SPEAKER_01: That I kind of push the boundaries.
- 2:11:05SPEAKER_01: So I learned to love these,
- 2:11:06but like what happened was I was like doing the wrong type of
- 2:11:11confusion in school
- 2:11:13SPEAKER_01: by taking like class out of order,
- 2:11:15SPEAKER_01: right?
- 2:11:16SPEAKER_01: Because I was trying to kind of like get this type of,
- 2:11:19SPEAKER_01: I don't know why I did that,
- 2:11:20SPEAKER_01: like this type of skill set,
- 2:11:22SPEAKER_01: a bit of like confusion bending or something like that,
- 2:11:26SPEAKER_01: where I was okay with like not understanding and everybody else actually
- 2:11:30SPEAKER_01: understood.
- 2:11:31And then like trying to like figure my way out of this and then
- 2:11:36like getting the
- 2:11:37SPEAKER_01: prerequisite,
- 2:11:37figuring out like what I was missing and like being able to do
- 2:11:40it with the time
- 2:11:41SPEAKER_01: pressure because I literally put my grade on the line.
- 2:11:44SPEAKER_01: So that was okay.
- 2:11:46SPEAKER_01: And I agree, you shouldn't do that.
- 2:11:49SPEAKER_01: There's better ways of doing it.
- 2:11:51SPEAKER_01: When I was doing research as an undergrad,
- 2:11:53it was much better than when I was doing that because there was
- 2:11:56no point in doing
- 2:11:57SPEAKER_01: this for the outcome.
- 2:11:59SPEAKER_01: It was more for practicing the skill.
- 2:12:01SPEAKER_01: But when I was in a research setting or also in a startup setting,
- 2:12:09SPEAKER_01: Like there's a lot of confusion there too, right?
- 2:12:11There's no one that's going to tell you like,
- 2:12:12oh,
- 2:12:12you want to build this company?
- 2:12:14SPEAKER_01: Well, good news.
- 2:12:14SPEAKER_01: And then here's the steps.
- 2:12:15SPEAKER_01: Like there's none of it.
- 2:12:16Like you can take a look at what others are doing,
- 2:12:19but this is fuzzy because is he faking it?
- 2:12:21SPEAKER_01: Like, can I get access to this financial?
- 2:12:23SPEAKER_01: I'm not sure it's working, right?
- 2:12:25SPEAKER_01: Does he have the same kind of population as me?
- 2:12:27SPEAKER_01: Not really, right?
- 2:12:28SPEAKER_01: Is the stack the same?
- 2:12:29SPEAKER_01: No.
- 2:12:30SPEAKER_01: Okay.
- 2:12:30SPEAKER_01: So we are not sure how to build this.
- 2:12:32SPEAKER_01: So there's a lot of that too.
- 2:12:35SPEAKER_01: And then you have to kind of get there, right?
- 2:12:37SPEAKER_01: While like, like,
- 2:12:39uh being around these confusing signal um throughout but when
- 2:12:43it clicks and when
- 2:12:44you figure out how to make it click kind of click quickly or
- 2:12:46like reduce the
- 2:12:47confusion so that like you you kind of restrict the path of the
- 2:12:51node that you need
- 2:12:52to kind of follow it's uh uh there's a lot of gains uh over
- 2:12:56there so like this i
- 2:12:58think you're right i think like that's that part is is useful
- 2:13:00um but like if if
- 2:13:02there was a known path and you're like beating confusion just
- 2:13:05for the fun of it
- 2:13:06SPEAKER_00: There's better ways to use your time.
- 2:13:10SPEAKER_00: Yeah.
- 2:13:10SPEAKER_00: It's like generally doing hard things is, is, is good, right?
- 2:13:14SPEAKER_00: It's good to push yourself to do hard things,
- 2:13:15SPEAKER_00: but also like when things are,
- 2:13:17SPEAKER_00: they don't need to be hard.
- 2:13:18SPEAKER_00: Like don't, don't make, why, why, why make it harder?
- 2:13:20SPEAKER_00: Like you can, you'll get to the hard things that really matter.
- 2:13:23SPEAKER_00: So spend your time,
- 2:13:24SPEAKER_00: spend your confusion on things that when resolving,
- 2:13:27SPEAKER_00: it's like,
- 2:13:28SPEAKER_01: spend,
- 2:13:29SPEAKER_00: spend your confusion on,
- 2:13:31SPEAKER_00: yeah,
- 2:13:31SPEAKER_00: on things that have high ROI on that confusion.
- 2:13:34Like if you,
- 2:13:34if you're confused,
- 2:13:35you get an aha moment,
- 2:13:36a click moment,
- 2:13:37like,
- 2:13:38How about make it something that actually is a contribution to
- 2:13:41a field instead of
- 2:13:42SPEAKER_00: just like,
- 2:13:44SPEAKER_00: oh,
- 2:13:44SPEAKER_00: this is a result in this math course that I skipped.
- 2:13:47SPEAKER_00: Duh.
- 2:13:48SPEAKER_00: Okay, nobody cares.
- 2:13:49SPEAKER_01: 90 days on it for not that much.
- 2:13:53SPEAKER_01: You hone that skill to be okay with the confusion,
- 2:13:55SPEAKER_01: but other than that,
- 2:13:57SPEAKER_01: it's more of the meta learning process.
- 2:14:00SPEAKER_01: learning happening here.
- 2:14:01SPEAKER_01: If we switch gear to learning math as an adult,
- 2:14:04because everybody has like this nice kind of curve of like
- 2:14:09building stuff.
- 2:14:10SPEAKER_01: So they are maybe already in a position where they need to use the math.
- 2:14:14SPEAKER_01: And then they kind of understand it, but a bit barely.
- 2:14:17SPEAKER_01: So I got a lot of questions about that.
- 2:14:19SPEAKER_01: There's the first one here that IceShady on Twitter asked,
- 2:14:23SPEAKER_01: I'm an adult learning calculus for the first time.
- 2:14:26How do I rebuild the high school foundation while still moving
- 2:14:28forward through new
- 2:14:29SPEAKER_01: material without spending a year going back to basics?
- 2:14:32yeah um so we actually uh we we came up with the solution to to
- 2:14:38this problem at at
- 2:14:40math academy because we were getting like a lot of you know
- 2:14:43adult learners asking
- 2:14:44this question like basically immediately as we kind of emerged
- 2:14:47on the the scene on
- 2:14:48on x twitter um a lot of adult learners are all asking like the
- 2:14:54same question is
- 2:14:55SPEAKER_00: like i what do i do do i i need to kind of start over i'm like
- 2:14:58SPEAKER_00: Like, I'm not even at calculus.
- 2:15:01SPEAKER_00: I took pre-calculus in high school and I forgot, like, all of it.
- 2:15:05SPEAKER_00: Like, should I just start with fractions?
- 2:15:07SPEAKER_00: Like, fourth grade math?
- 2:15:09SPEAKER_00: What do I do?
- 2:15:10SPEAKER_00: And so our answer to that was basically, like,
- 2:15:14Why don't we design a sequence that is the,
- 2:15:19you know,
- 2:15:19we'll start it with fractions.
- 2:15:21SPEAKER_00: So you can go down as far low as you need.
- 2:15:23You can place into the sequence,
- 2:15:24you can place out of material,
- 2:15:25already know.
- 2:15:27SPEAKER_00: But it's going to strip away...
- 2:15:30SPEAKER_00: All the stuff that is not that is there for,
- 2:15:32SPEAKER_00: you know,
- 2:15:33SPEAKER_00: like in schools,
- 2:15:35SPEAKER_00: school common core standards,
- 2:15:36SPEAKER_00: but does not actually really get leveraged in university math.
- 2:15:41SPEAKER_00: So just to name an example in geometry,
- 2:15:44SPEAKER_00: there's this like inscribed circle theorems where you're like,
- 2:15:48SPEAKER_00: OK,
- 2:15:48SPEAKER_00: I got a circle.
- 2:15:49SPEAKER_00: I got an angle in the circle.
- 2:15:51SPEAKER_00: And like, what's the.
- 2:15:52SPEAKER_00: If the vertex lies on the circumference of the circle,
- 2:15:57SPEAKER_00: what's the measurement,
- 2:15:58SPEAKER_00: if you know the arc intercepted by this thing?
- 2:16:01It's one of those things that's like,
- 2:16:03okay,
- 2:16:04maybe it'll come up a little bit if you do like...
- 2:16:07the physics of of um like lenses and stuff but aside from that
- 2:16:12like you're just
- 2:16:12never gonna see this kind of thing not relevant to machine
- 2:16:16learning not relevant to
- 2:16:18most physics not not relevant to like basically and any
- 2:16:24anything else and and so
- 2:16:26there's a ton of these topics that just they're not very
- 2:16:30foundational for for for
- 2:16:32SPEAKER_00: high school math i think
- 2:16:34I think it was maybe like a quarter or a third of the
- 2:16:38traditional school sequence.
- 2:16:40SPEAKER_00: You can you can just kind of like strip away.
- 2:16:42SPEAKER_00: And it's really it's not a big deal.
- 2:16:45And you but you get your foundations,
- 2:16:46the really important stuff that underpins.
- 2:16:49all um like a lot of the the serious math and and physics and
- 2:16:54computer science
- 2:16:55machine like whatever people want to do and so we we made this
- 2:16:58three course
- 2:16:59sequence the call it the mathematical foundation sequence and
- 2:17:02and our director of
- 2:17:03curriculum alex smith was really the the one who put all this
- 2:17:06together and and and
- 2:17:07made it happen um but the sequence is um all together it's this
- 2:17:12three course
- 2:17:13SPEAKER_00: sequence you you
- 2:17:14You take the three courses and you emerge on the other side
- 2:17:17ready for university
- 2:17:20SPEAKER_00: math,
- 2:17:21SPEAKER_00: like linear algebra and multivariable calculus,
- 2:17:23or a composite course like math for machine learning that
- 2:17:26stitches stuff together.
- 2:17:28And to go from this math foundation sequence,
- 2:17:32the very bottom to the very top.
- 2:17:33SPEAKER_00: So assuming that you really forgot how to add fractions,
- 2:17:37SPEAKER_00: like you are starting from almost zero,
- 2:17:41SPEAKER_00: to fill it out all the way takes about
- 2:17:4415,000 minutes now that might seem like a lot but if you
- 2:17:49actually do it like okay
- 2:17:50how long is that actually per per per day it's like well if you
- 2:17:54just do every every
- 2:17:56weekday for an hour for a year you can fill in all your
- 2:18:01foundations um if you're
- 2:18:03starting from the very bottom and now if you if you decide to
- 2:18:05do like i'm gonna do
- 2:18:07SPEAKER_00: two hours per weekday guess what only half a year
- 2:18:10uh three hours per weekday maybe you want to split that up into
- 2:18:14like a morning
- 2:18:14session an evening session or a couple times a day but um if
- 2:18:17you get like really
- 2:18:18really serious like guess what like just a few months um and
- 2:18:23that's assuming that
- 2:18:23you're starting from the very bottom most people don't start
- 2:18:26like just from adding
- 2:18:27fractions they remember some algebra at least like how to solve
- 2:18:30a linear equation
- 2:18:31SPEAKER_00: so guess like that
- 2:18:33so really you can if you're serious about this you can get it
- 2:18:36done in uh in several
- 2:18:38months um yeah but the the path is there and we've had people
- 2:18:42do this and and come
- 2:18:43out the other side you know actually the person who i referred
- 2:18:46to earlier who had
- 2:18:47uh signed up um math academy courses because he wanted to read
- 2:18:51this robotics book
- 2:18:52and do robotics and he just had no like way above him he went
- 2:18:56through that sequence
- 2:18:56SPEAKER_00: and he came out and
- 2:18:57the other end and it was it's just world changing for now this
- 2:19:02is good yeah yeah um
- 2:19:04and like it's true like you don't have to like hop around like
- 2:19:09the like uh the
- 2:19:10extra stuff um which is not directly useful for for your goal
- 2:19:14so you can just trip
- 2:19:15the way to the core of it i really like that there's a two
- 2:19:18other question from um
- 2:19:20SPEAKER_01: nicker cube and zaid on twitter that says
- 2:19:23SPEAKER_01: I'm doing the finishing and learning how to do all this cool stuff,
- 2:19:27SPEAKER_01: but I'm not really sure what it's all used for.
- 2:19:29SPEAKER_01: How do I fill in that motivation gap?
- 2:19:31SPEAKER_01: I think that it ties up to one thing that we discussed earlier.
- 2:19:35SPEAKER_01: Yeah.
- 2:19:36SPEAKER_01: What do you think here?
- 2:19:37SPEAKER_00: Yeah.
- 2:19:37SPEAKER_00: Yeah.
- 2:19:37SPEAKER_00: So it definitely ties into what we were talking about earlier,
- 2:19:41SPEAKER_00: like the end of the pipeline where I mentioned that we were talking.
- 2:19:46Talking about having projects as just really cool projects as
- 2:19:50as kind of this
- 2:19:52SPEAKER_00: anchoring goal for learners for,
- 2:19:54SPEAKER_00: you know,
- 2:19:55SPEAKER_00: you can look and be like,
- 2:19:57SPEAKER_00: like,
- 2:19:57SPEAKER_00: wait,
- 2:19:57SPEAKER_00: you like I just take these math courses.
- 2:20:01And then I get to,
- 2:20:03like,
- 2:20:03training a convolutional neural network to classify digits.
- 2:20:07SPEAKER_00: Like, what?
- 2:20:08SPEAKER_00: Really?
- 2:20:09SPEAKER_00: Or, like, even cooler applications.
- 2:20:12SPEAKER_00: Like, I use a neural net to...
- 2:20:17like approximate some some lengthy physics computation and run
- 2:20:20it really fast or or
- 2:20:22like use a like um clustering of like uh of language symbols
- 2:20:28and then like a
- 2:20:29long-lost civilization whatever like we got a ton of of
- 2:20:32different um projects like
- 2:20:34that coming out in our in our upcoming machine learning course
- 2:20:37um and we want to
- 2:20:38kind of extend this uh to the to the whole curriculum um not
- 2:20:43just machine learning
- 2:20:44SPEAKER_00: but like actually do it in like just the
- 2:20:46SPEAKER_00: you know, your routine calculus course.
- 2:20:48SPEAKER_00: Like,
- 2:20:49SPEAKER_00: why not have calculus with a bunch of really cool projects that involve,
- 2:20:52SPEAKER_00: like,
- 2:20:52SPEAKER_00: you know,
- 2:20:53SPEAKER_00: and be integrated with a little physics and have one of the projects be,
- 2:20:58SPEAKER_00: like,
- 2:20:58SPEAKER_00: figuring out the viability of some situation for...
- 2:21:04SPEAKER_00: setting up a colony on Mars or something,
- 2:21:07or like figuring out like what fuel type of fuel is best for a
- 2:21:12turbo boosted race
- 2:21:13car or figuring out like what ingredients are gonna work or not
- 2:21:18gonna work for like
- 2:21:19SPEAKER_00: a certain recipe given like chemical properties.
- 2:21:23SPEAKER_00: Whatever, stuff like that.
- 2:21:24SPEAKER_00: A goal is that like if we can cover the space with enough really,
- 2:21:29SPEAKER_00: really cool stuff and like allow people to,
- 2:21:32SPEAKER_00: you know,
- 2:21:32SPEAKER_00: anchor on particular projects that they want to do.
- 2:21:35SPEAKER_00: And hopefully like not just the highlights.
- 2:21:38SPEAKER_00: highest level projects,
- 2:21:38SPEAKER_00: but also intermediate projects leading up to there,
- 2:21:42SPEAKER_00: like have some projects in algebra,
- 2:21:44SPEAKER_00: have some projects in calculus,
- 2:21:45SPEAKER_00: have some projects in like linear algebra,
- 2:21:48SPEAKER_00: multivariable calculus,
- 2:21:49SPEAKER_00: math for machine learning,
- 2:21:50SPEAKER_00: all this sort of stuff.
- 2:21:52There's kind of like a,
- 2:21:53a real,
- 2:21:54like an application that,
- 2:21:56that seems really,
- 2:21:57really interesting.
- 2:21:58There's some really cool application that like,
- 2:22:00they don't require that much.
- 2:22:01SPEAKER_01: Like, like even just reproducing a paper, like a result.
- 2:22:05SPEAKER_01: So all the time it's like,
- 2:22:06SPEAKER_01: It's not that much math.
- 2:22:09I was reading a paper recently about trying to do maximum
- 2:22:14likelihood approximation,
- 2:22:17SPEAKER_01: but for reinforcement learning.
- 2:22:18SPEAKER_01: And if you look at the appendix, the proof that they have,
- 2:22:24SPEAKER_01: there's nothing complicated over there.
- 2:22:25SPEAKER_01: There's literally nothing.
- 2:22:26SPEAKER_01: You go step by step as some of the basic stuff to,
- 2:22:29SPEAKER_01: you can,
- 2:22:30SPEAKER_01: you can see,
- 2:22:30SPEAKER_01: and like,
- 2:22:31SPEAKER_01: you just need to understand like the flow of the overall paper,
- 2:22:33but you can reproduce the result directly and you can reproduce
- 2:22:35the derivation and
- 2:22:36SPEAKER_01: explain every step of the way.
- 2:22:38SPEAKER_01: Um, there's just some basic calculus and you're good to go.
- 2:22:41SPEAKER_01: Um,
- 2:22:42SPEAKER_01: there's a lot of the stuff that are very approachable,
- 2:22:44SPEAKER_01: but like people don't know,
- 2:22:45SPEAKER_01: like outside,
- 2:22:46SPEAKER_01: like the boundary of the knowledge.
- 2:22:49SPEAKER_01: So they're like,
- 2:22:50SPEAKER_01: For sure, I'm missing something.
- 2:22:51SPEAKER_01: But when you actually look at the stuff and like, that's it?
- 2:22:54SPEAKER_00: This is the only thing?
- 2:22:57There's one paper that we actually are is going to form the
- 2:23:00basis of one of the
- 2:23:01SPEAKER_00: machine learning projects,
- 2:23:03SPEAKER_00: which I remember,
- 2:23:05I mentioned that I had a high school when I was teaching this
- 2:23:08advanced machine
- 2:23:10SPEAKER_00: learning track within math academies or
- 2:23:12SPEAKER_00: I had the kids reproduce papers.
- 2:23:16SPEAKER_00: And there was one paper back in the 90s, Artificial Intelligence.
- 2:23:21SPEAKER_00: Like you say, it's like it's a cool result.
- 2:23:23SPEAKER_00: And it's surprisingly like, you know, if you it's just tremendous.
- 2:23:28um evolving neural networks to to play games like tic-tac-toe
- 2:23:32or checkers or
- 2:23:33whatever and so like you actually don't even need to do back
- 2:23:36propagation because
- 2:23:37it's an evolutionary procedure and so if you can just forward
- 2:23:40propagate and and i
- 2:23:43mean that's really just the the hard part is not not that much
- 2:23:46math involved um and
- 2:23:48SPEAKER_00: then then you can do it so so yeah i i think um you know having paper uh
- 2:23:55SPEAKER_00: Reproductions is going to be a really cool thing as well.
- 2:23:57SPEAKER_00: And I'm sure there's tons, tons of those papers.
- 2:23:59SPEAKER_01: Yeah.
- 2:24:00SPEAKER_01: There's another question by Mohammed FX that asks,
- 2:24:04how should teaching math for kids or adult balance exploration
- 2:24:07and creativity?
- 2:24:09SPEAKER_01: with what can feel like a repetitive drill.
- 2:24:11SPEAKER_01: He said, like, I remember reading a Metisian Laments.
- 2:24:14SPEAKER_01: I think I also read that.
- 2:24:15SPEAKER_01: And feeling it should be all curiosity and creativity,
- 2:24:18SPEAKER_01: but also,
- 2:24:18SPEAKER_01: no,
- 2:24:19SPEAKER_01: I couldn't have thought deeply without doing the boring problem first.
- 2:24:22SPEAKER_01: I think you answered almost his own question.
- 2:24:25SPEAKER_01: What's your take on this?
- 2:24:26SPEAKER_00: I think it's one of those things where it's...
- 2:24:30SPEAKER_00: You know,
- 2:24:30SPEAKER_00: as you're kind of going,
- 2:24:31SPEAKER_00: we've made this distinction between when you're in the knowledge graph,
- 2:24:36the known knowledge graph versus when you're building on top of
- 2:24:39the knowledge.
- 2:24:39SPEAKER_00: You're like at the frontier of things.
- 2:24:41SPEAKER_00: And so I think I mean,
- 2:24:44SPEAKER_00: just from a pure efficiency standpoint,
- 2:24:46SPEAKER_00: like,
- 2:24:46SPEAKER_00: you know,
- 2:24:47SPEAKER_00: when you're going through the known knowledge graph,
- 2:24:49SPEAKER_00: I mean,
- 2:24:49SPEAKER_00: you.
- 2:24:51SPEAKER_00: any sort of creative thing that you will do it.
- 2:24:54SPEAKER_00: Unfortunately,
- 2:24:55it's not actually that creative because it can be rendered
- 2:24:58trivial by a lot of
- 2:24:59SPEAKER_00: results.
- 2:25:01SPEAKER_00: Like you could study the property,
- 2:25:02SPEAKER_00: the statistical properties of random walks as,
- 2:25:06SPEAKER_00: as a,
- 2:25:06SPEAKER_00: as a,
- 2:25:07SPEAKER_00: you know,
- 2:25:08SPEAKER_00: as a solo creative project or stuff,
- 2:25:10SPEAKER_00: but like,
- 2:25:11SPEAKER_00: guess what?
- 2:25:11SPEAKER_00: That's already figured out.
- 2:25:12SPEAKER_00: Like, so whatever, whatever,
- 2:25:14SPEAKER_00: result you're going to have on that is probably known already.
- 2:25:18SPEAKER_00: Now,
- 2:25:18not to say that there is zero value to this because there's
- 2:25:22motivational value to
- 2:25:23SPEAKER_00: it,
- 2:25:24SPEAKER_00: right?
- 2:25:24SPEAKER_00: Like if the choices are like either I'm going to lock in and
- 2:25:30SPEAKER_00: if you're going to like draw a dichotomy for yourself,
- 2:25:32SPEAKER_00: it's like,
- 2:25:33SPEAKER_00: okay,
- 2:25:33either go monk mode for like all my studies until I reach the
- 2:25:36frontier and only
- 2:25:36SPEAKER_00: then will I allow myself like any sort of creative exploration.
- 2:25:41SPEAKER_00: Like that's like just mechanically,
- 2:25:43SPEAKER_00: that's going to be the most efficient way to get the frontier.
- 2:25:45SPEAKER_00: But if you're going to, if that's going to,
- 2:25:47SPEAKER_00: force you off the rails where you're just like, man, I quit.
- 2:25:50SPEAKER_00: I'm done.
- 2:25:51Like,
- 2:25:51well,
- 2:25:52then like,
- 2:25:52okay,
- 2:25:53like just do,
- 2:25:54do what you need to do to stay on the rails.
- 2:25:57SPEAKER_00: And if that involves like just going down some,
- 2:25:59SPEAKER_00: some rabbit hole or whatever,
- 2:26:01SPEAKER_00: just to like have some kind of like fun with it,
- 2:26:03SPEAKER_00: have like experience,
- 2:26:04SPEAKER_00: like some,
- 2:26:05SPEAKER_00: some tastes of the rewards that you want to like get once you,
- 2:26:10SPEAKER_00: once you get to the frontier of it,
- 2:26:12SPEAKER_00: then like,
- 2:26:12SPEAKER_00: okay,
- 2:26:13SPEAKER_00: do that,
- 2:26:13SPEAKER_00: do,
- 2:26:14SPEAKER_00: do what you need to do to,
- 2:26:15SPEAKER_00: to,
- 2:26:16SPEAKER_00: to make it fun.
- 2:26:17um and everyone's going to be different in in this capacity and
- 2:26:21like how much of
- 2:26:22that they they feel like they want to need or what type of
- 2:26:25rabbit holes they want
- 2:26:26to go down or whatever um but i i think the real the the the
- 2:26:30point where this like
- 2:26:32you know the exploration and everything becomes very valuable
- 2:26:36like a high roi sort
- 2:26:38SPEAKER_00: of thing is really
- 2:26:40SPEAKER_00: Once you reach the frontier,
- 2:26:41SPEAKER_01: that's totally like,
- 2:26:43SPEAKER_01: um,
- 2:26:43SPEAKER_01: there's a bunch of actually undergrad that I've met,
- 2:26:47SPEAKER_01: uh,
- 2:26:47SPEAKER_01: found really cool machine learning results at the frontier.
- 2:26:51Um,
- 2:26:52and like the,
- 2:26:52the,
- 2:26:53because like a machine learning also is very empirical in
- 2:26:56nature,
- 2:26:56right?
- 2:26:56SPEAKER_01: Like,
- 2:26:57SPEAKER_01: uh,
- 2:26:57SPEAKER_01: yes,
- 2:26:57SPEAKER_01: there's all a theory,
- 2:26:59SPEAKER_01: but if you look at the frontier enough,
- 2:27:01there's a lot of people like messing around and just like
- 2:27:03trying stuff up.
- 2:27:04SPEAKER_01: Right.
- 2:27:04SPEAKER_01: And you have this feedback loop about like,
- 2:27:07SPEAKER_01: Did it work?
- 2:27:07SPEAKER_01: Yes.
- 2:27:08SPEAKER_01: Okay.
- 2:27:09SPEAKER_01: This is good.
- 2:27:09SPEAKER_01: Positive thing.
- 2:27:10SPEAKER_01: So what I've seen a lot of them,
- 2:27:13like the very bright one do is that like they go into the
- 2:27:17frontier and they look at
- 2:27:19SPEAKER_01: the frontier and they understand jack shit,
- 2:27:21SPEAKER_01: literally nothing.
- 2:27:22SPEAKER_01: Right.
- 2:27:23And then they drill and then like intensely and then they do
- 2:27:25stuff like at the
- 2:27:27SPEAKER_01: knowledge gap that they have.
- 2:27:28SPEAKER_01: And then they kind of build this up.
- 2:27:29SPEAKER_01: Right.
- 2:27:30SPEAKER_01: But time to time they go back to this and then they tinker with it.
- 2:27:33SPEAKER_01: Right.
- 2:27:33SPEAKER_01: So it's not like tinkering with a known thing.
- 2:27:36SPEAKER_01: For this, they completely eliminated from their kind of regime.
- 2:27:39SPEAKER_01: They go...
- 2:27:41and tinker in the frontier and try to see stuff like and then
- 2:27:44but then they learned
- 2:27:45a bunch of things they have and then they're able to kind of
- 2:27:48reapply it and get
- 2:27:49ideas and then the blocks a lot of creativity yeah exactly like
- 2:27:54some of them like
- 2:27:54they're doing some wild stuff like they they try things that
- 2:27:58like they have no
- 2:27:59biases so like they're like yeah why not this why why not this
- 2:28:02other thing and then
- 2:28:02SPEAKER_01: they try a bunch of stuff and then they find
- 2:28:04some cool surface way to simplify the stuff and then they get
- 2:28:09good results that
- 2:28:09SPEAKER_01: they can then like iterate with.
- 2:28:11SPEAKER_01: But when they go back into their kind of learning,
- 2:28:14SPEAKER_01: they're anchoring onto this goal and they have this kind of desire to
- 2:28:19up their skills so that they can kind of go back to this nice
- 2:28:22project that is at
- 2:28:23the frontier and um and do stuff with it i like the best
- 2:28:27student that i've i've
- 2:28:28seen we're doing a mix of that like uh they they were doing
- 2:28:32both the frontier stuff
- 2:28:33to for their own motivation and their curiosity um and then
- 2:28:37they were never bored
- 2:28:38into the basic because they're like yeah i know i need that
- 2:28:40because i don't
- 2:28:41understand this this part absolutely no clue what's going on so
- 2:28:44a better like skill
- 2:28:45SPEAKER_01: up fast
- 2:28:46so i can get that and do that fun thing so this approach of
- 2:28:51like you know doing a
- 2:28:53like go to some kind of interesting project at the frontier do
- 2:28:57what you can and
- 2:28:58then like backfill to to try to unlock more creativity uh or
- 2:29:03more just um ideas or
- 2:29:05SPEAKER_00: whatever to to to do with this thing or to understand it better um i i i
- 2:29:10SPEAKER_00: I totally see how this can work for people who have,
- 2:29:16SPEAKER_00: you know,
- 2:29:16some amount of like a large chunk of foundational knowledge
- 2:29:19already in place.
- 2:29:21SPEAKER_00: And it's I just it I think it kind of depends on somebody.
- 2:29:25somebody's the distance between what they're trying to do and
- 2:29:28what they're doing
- 2:29:29because like you can there's there's other this may seem like a
- 2:29:32caricature but this
- 2:29:33this this happens a lot where there's some somebody who doesn't
- 2:29:37even know how to
- 2:29:38solve a linear equation right and and they want to do like wow
- 2:29:42this ai stuff is
- 2:29:43cool i want how do i get involved how do i do something and
- 2:29:45it's like well i mean
- 2:29:46SPEAKER_00: you can i guess you can like
- 2:29:48download and run like a transformer model or something um but
- 2:29:53if if the gap is that
- 2:29:54big like when it comes into like like backfilling prerequisites
- 2:29:58you're gonna be
- 2:29:59like well okay i need to learn like back propagation what is
- 2:30:02that that's squiggly
- 2:30:04symbol oh is that calculus you're kind of like so it's like if
- 2:30:09if the gap to fill
- 2:30:10SPEAKER_00: in is not very
- 2:30:11SPEAKER_00: very big, then you can be efficient in terms of backfill.
- 2:30:16SPEAKER_00: Like, I just need this, that, that, learn, learn, learn.
- 2:30:19SPEAKER_00: Okay, new unlock.
- 2:30:20SPEAKER_00: But the gap is, like, gigantic, where it's like you're essentially...
- 2:30:25missing like all of algebra all of calculus and and that then i
- 2:30:30i think it gets to
- 2:30:31a point where it can be inefficient to try to like trace your
- 2:30:35weight all the way
- 2:30:37back down to like fractions or algebra the guy doesn't know how
- 2:30:40to count yeah
- 2:30:41SPEAKER_01: exactly
- 2:30:44And it's like,
- 2:30:46well,
- 2:30:46I guess you're going to need to learn this stuff anyway.
- 2:30:49SPEAKER_00: It's not like you can just take like,
- 2:30:51SPEAKER_00: oh,
- 2:30:51SPEAKER_00: there's like,
- 2:30:52there's three secret topics in all of like 300 algebra topics
- 2:30:56that you need to
- 2:30:56SPEAKER_00: learn to unlock machine learning.
- 2:30:58SPEAKER_00: It's like, no, you got to learn basically all of algebra.
- 2:31:02SPEAKER_00: It's kind of foundational.
- 2:31:04SPEAKER_00: And so there's,
- 2:31:04SPEAKER_00: I think for these foundational,
- 2:31:06SPEAKER_00: but like when a body is foundational like that,
- 2:31:08SPEAKER_00: I think it's,
- 2:31:09SPEAKER_00: you know,
- 2:31:10SPEAKER_00: if you try to just like go on backfill,
- 2:31:12you're going to be ultimately just kind of trying to construct
- 2:31:14your own algebra
- 2:31:15SPEAKER_00: graph.
- 2:31:16SPEAKER_00: And, and this is a known thing.
- 2:31:18SPEAKER_00: This is, this is,
- 2:31:20you just like take the algebra take an algebra course uh you
- 2:31:23can get through it
- 2:31:24like a lot a lot faster but but i agree as you like reach the
- 2:31:27like closer to you
- 2:31:29know like okay you're kind of within striking distance of this
- 2:31:32thing and you can
- 2:31:32kind of see the full the full path between the project that
- 2:31:36you're doing and the
- 2:31:37math that you need that the you know the backfilling strategy
- 2:31:40can i think it
- 2:31:41becomes less inefficient yeah there's there's a lot of this for
- 2:31:45like most people
- 2:31:46but i think also like the the folks that i i see even young
- 2:31:51that are um very
- 2:31:53comfortable doing that stuff they understand really well like
- 2:31:56the mental learning
- 2:31:57skill like they kind of um already uh they know that like ah i
- 2:32:03can't just like
- 2:32:03SPEAKER_01: watch lectures right i actually have to drill problems
- 2:32:06SPEAKER_01: And I'm so far down that like,
- 2:32:08SPEAKER_01: yeah,
- 2:32:08SPEAKER_01: I can look at this,
- 2:32:09SPEAKER_01: try a bunch of stuff,
- 2:32:10SPEAKER_01: but like,
- 2:32:11SPEAKER_01: I can't spend like six hours a day here.
- 2:32:13SPEAKER_01: It will be useless.
- 2:32:14SPEAKER_01: I need to quickly go and like build these skills.
- 2:32:17So then they go in and they see like the most like efficient
- 2:32:19pattern and they go
- 2:32:20SPEAKER_01: through it as fast as humanly possible.
- 2:32:23SPEAKER_01: And it just kind of blitz through it.
- 2:32:25So this is something that I've seen is like,
- 2:32:28even if the gap is big in these people,
- 2:32:30it kind of doesn't matter because like they will be able to
- 2:32:33like regulate what
- 2:32:34they're doing right like the goal is over there i know i'm
- 2:32:37missing a ton of stuff
- 2:32:38but this is what i want to do and i tinker a bit with it i'm
- 2:32:41like wow this is so
- 2:32:42cool but then when they go back and then they actually try to
- 2:32:46like for real do it
- 2:32:47and not just like have this aspirational goal that is not
- 2:32:50attainable they work on
- 2:32:51SPEAKER_01: the right stuff because they know how to get like
- 2:32:54SPEAKER_01: the learning they need in order to get there.
- 2:32:57SPEAKER_01: People that are kind of flail in this,
- 2:33:00SPEAKER_01: I see also a lot of people like this because they want to do the stuff,
- 2:33:03SPEAKER_01: but like they have no concept of really how to get the prayer quiz
- 2:33:09knowledge in order to do this thing they go into a weird path
- 2:33:13and they kind of get
- 2:33:14lost within themselves and even if you try to nudge them back I
- 2:33:18don't know they
- 2:33:18have some learning psychosis or something they're completely
- 2:33:23spread out and then
- 2:33:24SPEAKER_01: they have the most monstrous kind of setup imaginable
- 2:33:28SPEAKER_01: Um, and most of it is just wasted time.
- 2:33:30SPEAKER_01: Like it's just wasted.
- 2:33:32SPEAKER_01: Uh, it's hard to bring them back after that.
- 2:33:34SPEAKER_01: So there's the, yeah, I agree.
- 2:33:35SPEAKER_01: If there's like this, this kind of discrepancy, um, yeah, definitely.
- 2:33:39SPEAKER_00: Yeah.
- 2:33:39SPEAKER_00: That's, I mean, that's a, I think that's, that's right.
- 2:33:41SPEAKER_00: There's also the variable of like how,
- 2:33:43SPEAKER_00: how much does this person understand how to learn effectively?
- 2:33:46SPEAKER_00: Yeah.
- 2:33:47SPEAKER_00: That plays the same role as like a gap in knowledge.
- 2:33:51SPEAKER_00: Yeah.
- 2:33:51SPEAKER_00: Yeah.
- 2:33:53SPEAKER_01: Last question for the adult learning stuff.
- 2:33:56SPEAKER_01: There's two people actually in Clifton that ask,
- 2:34:00SPEAKER_01: is it ever too late to start learning seriously?
- 2:34:02I know it's not,
- 2:34:03but give me a compelling answer with example for reassurance.
- 2:34:07SPEAKER_01: Let's do a bit of a psychology here with the guy.
- 2:34:12SPEAKER_00: Yeah.
- 2:34:12SPEAKER_00: Yeah.
- 2:34:13SPEAKER_00: I,
- 2:34:13SPEAKER_00: I have,
- 2:34:14SPEAKER_00: I actually have a,
- 2:34:15SPEAKER_00: a very concrete example,
- 2:34:16SPEAKER_00: but before I say the example,
- 2:34:18SPEAKER_00: I will say like,
- 2:34:19SPEAKER_00: there have been a number of,
- 2:34:21SPEAKER_00: um,
- 2:34:21SPEAKER_00: adult students who I've,
- 2:34:23SPEAKER_00: I've seen go from like,
- 2:34:24SPEAKER_00: even in their thirties,
- 2:34:26SPEAKER_00: like just go from,
- 2:34:27SPEAKER_00: you know,
- 2:34:27SPEAKER_00: I don't really,
- 2:34:28SPEAKER_00: I want to get involved in some kind of like cutting edge thing,
- 2:34:31SPEAKER_00: or maybe not even a cutting edge.
- 2:34:33Like I just want to become a software engineer and,
- 2:34:35and,
- 2:34:35and they,
- 2:34:36they start this,
- 2:34:38um,
- 2:34:38SPEAKER_00: in their 30s or even their 40s.
- 2:34:41SPEAKER_00: I'm sure there's people who do it even later in life.
- 2:34:44SPEAKER_00: If you're willing to put in
- 2:34:46SPEAKER_00: The work it's I mean, it's the same amount of work.
- 2:34:49The amount of work does not change whether you start this at
- 2:34:52the age of 15 and like
- 2:34:55SPEAKER_00: emerge where you want to be at the age of 20 or 25 or whatever.
- 2:34:59SPEAKER_00: Like this,
- 2:35:00it's the same amount of work if you started at 25 or if you
- 2:35:03started at 35 or or
- 2:35:05SPEAKER_00: whatever.
- 2:35:06And some of the things that makes it more more challenging for
- 2:35:10why I think.
- 2:35:12Sometimes people think like,
- 2:35:14oh,
- 2:35:14it's too late for me to do this kind of thing.
- 2:35:17SPEAKER_00: And I think that tends to be more of just like,
- 2:35:21SPEAKER_00: I'm not willing to put in the work given that I have less free time now.
- 2:35:27You go through life,
- 2:35:29you got like a family,
- 2:35:30you got kids,
- 2:35:32you got more responsibility.
- 2:35:33SPEAKER_00: It's just harder to make time decisions.
- 2:35:36to do things aside from,
- 2:35:37especially if you have a job too,
- 2:35:39like a separate job.
- 2:35:41SPEAKER_00: It's like, where's the time in the day to learn?
- 2:35:44When you were a teenager,
- 2:35:44you could just like,
- 2:35:45you just come home from school.
- 2:35:46SPEAKER_00: Like, what are you doing?
- 2:35:47SPEAKER_00: Like play video games or like go to the basketball court with friends?
- 2:35:52SPEAKER_00: Like, you know, you got all the time in the world to do this stuff.
- 2:35:56but the amount of work does not change even if your your time
- 2:35:59to do it does but the
- 2:36:00time that you devote to doing the work is is within your
- 2:36:04control you can choose to
- 2:36:06be more efficient in in in various areas of your life to to
- 2:36:10open up time and it's
- 2:36:11not going to be easy and there's always going to be like lots
- 2:36:14of constraints and
- 2:36:16and stuff and probably not going to open up as much time as as
- 2:36:19you did when when
- 2:36:20SPEAKER_00: you were younger but if you but the the the
- 2:36:23SPEAKER_00: The takeaway is if you put in the work, you will get to the result.
- 2:36:27SPEAKER_00: If you find some way to put in that work.
- 2:36:29SPEAKER_00: And so I've seen this happen.
- 2:36:31SPEAKER_00: In particular,
- 2:36:33SPEAKER_00: back when I was working in data science,
- 2:36:36I started out as an intern for like six months before
- 2:36:41converting to just full-time
- 2:36:43SPEAKER_00: intern.
- 2:36:44SPEAKER_00: And so there was another guy who was actually interning with me.
- 2:36:50SPEAKER_00: And he was, I think he was actually in his 30s.
- 2:36:55SPEAKER_00: Yeah, somewhere in his 30s or maybe 30 years old on the dime side.
- 2:36:58SPEAKER_00: So he had a,
- 2:37:00SPEAKER_00: his story was kind of like that he,
- 2:37:03SPEAKER_00: you know,
- 2:37:04SPEAKER_00: in high school,
- 2:37:04SPEAKER_00: he didn't really,
- 2:37:05he wasn't really paying attention to what he wanted to really
- 2:37:08get out of like
- 2:37:10SPEAKER_00: career-wise.
- 2:37:11SPEAKER_00: He was more concerned about like, I just want to have like kids
- 2:37:14have a family um and like that was his main goal and that's a
- 2:37:17fine main goal and so
- 2:37:19he did that he accomplished that had kids had a family and he
- 2:37:22was working um as a
- 2:37:24as as um call receiver at a at a emergency center like the if
- 2:37:30you 911 who answers
- 2:37:31SPEAKER_00: the phone him um
- 2:37:33and and so um he wanted to get into like just you know a a kind
- 2:37:38of career that
- 2:37:40would that would offer more um money and better just uh like
- 2:37:44like balance and stuff
- 2:37:45and just so he could like give more to his family and and stuff
- 2:37:49like that so so he
- 2:37:50SPEAKER_00: was like you know um
- 2:37:52SPEAKER_00: Like, I want to go into software engineering now.
- 2:37:54SPEAKER_00: I want to get into, like, data science.
- 2:37:57SPEAKER_00: And so he's...
- 2:37:58So this was,
- 2:38:00you know,
- 2:38:00he was,
- 2:38:00like,
- 2:38:01in his late 20s when he even,
- 2:38:03like,
- 2:38:03this...
- 2:38:04SPEAKER_00: He started even thinking about this.
- 2:38:06SPEAKER_00: And he didn't know, like, any...
- 2:38:08SPEAKER_00: math, no coding background, like none, none of this stuff.
- 2:38:12SPEAKER_00: And so he just signed up at his community college and in his late 20s,
- 2:38:18SPEAKER_00: like one of the oldest students there.
- 2:38:21SPEAKER_00: And he's like,
- 2:38:23whatever man like i like this is this i gotta learn this stuff
- 2:38:27so i'm gonna learn
- 2:38:27this stuff like so and he's like and at the same time he's like
- 2:38:33he's got his job
- 2:38:35that he's like holding down as well he's got his family and
- 2:38:38stuff so he's just this
- 2:38:39is not an easy like sort of like he he's pressed for time um
- 2:38:45and and but he's
- 2:38:47SPEAKER_00: making it happen um
- 2:38:49SPEAKER_00: And and so he goes there,
- 2:38:51he gets his computer science degree from community college and
- 2:38:55then gets an
- 2:38:56SPEAKER_00: internship at this company.
- 2:38:58SPEAKER_00: And and so he and I interned in the same around the same time.
- 2:39:03SPEAKER_00: And he also.
- 2:39:04converted to full-time as well and he he took it very seriously
- 2:39:08and and and really
- 2:39:10got good at that at the the kinds of um software engineering
- 2:39:13they were doing like
- 2:39:14and data flows and stuff got the nickname like the lord of the
- 2:39:18data flow over there
- 2:39:19because he was like it's just he was so like into it yeah and
- 2:39:22he's like because i
- 2:39:24remember talking to him and he's like man i really need this to
- 2:39:27work like there's
- 2:39:28SPEAKER_00: no but it's like he was at a point where it's like there was no
- 2:39:32what's the alternative plan he wasn't thinking about like
- 2:39:34alternative plans he's
- 2:39:36like no this is what i'm gonna make happen there's there's no
- 2:39:39hedging of like like
- 2:39:41oh if if this doesn't work out i guess i'll do this or whatever
- 2:39:46i'm sure he would
- 2:39:47there would be plan b's but he's not thinking about he's
- 2:39:50focusing on the thing in
- 2:39:51front of him making it happen and um yeah and and you know one
- 2:39:55thing that i um that
- 2:39:57SPEAKER_00: i that i think is true is like
- 2:39:58SPEAKER_00: Initially,
- 2:40:00SPEAKER_00: this journey,
- 2:40:01SPEAKER_00: especially if you're embarking on it later in life,
- 2:40:03the hardest part is kind of the upskilling phase when you're
- 2:40:09trying to prepare
- 2:40:11yourself to get paid to do something that's more aligned with
- 2:40:16what you want to do.
- 2:40:17SPEAKER_00: Because at that moment, you're not good enough to get paid for it yet.
- 2:40:22SPEAKER_00: So probably you have to be doing a different thing to get paid.
- 2:40:26SPEAKER_00: That's going to...
- 2:40:28eat up a lot of your time so you have to spend time doing this
- 2:40:30other thing and
- 2:40:31spend time preparing for this new thing it's like where you're
- 2:40:34just pressed for
- 2:40:35time but once once you can make that transition to like
- 2:40:38actually earn a living from
- 2:40:40the thing that's in the direction that you want to do it's like
- 2:40:43a just total
- 2:40:45phase change because now you have like maybe eight hours of
- 2:40:49your life every weekday
- 2:40:50that you're actually doing the thing that's closer to what you
- 2:40:53want to do improving
- 2:40:54in that direction and it's it's like you get a rocket thruster
- 2:40:57yeah and if you can
- 2:40:58just make it to that point then yeah the the start is difficult
- 2:41:02right like uh yeah
- 2:41:04you don't have enough time and like if you're going into the
- 2:41:06wrong stuff and you're
- 2:41:07like learning it wrong because you're gonna eat up your time
- 2:41:10and then like the
- 2:41:11SPEAKER_01: motivation will start to
- 2:41:13uh to falter a bit um and like also like for adults generally
- 2:41:17like there's a bit
- 2:41:18less neuroplasticity um uh in compared to like kids right like
- 2:41:23adolescent there's
- 2:41:24SPEAKER_01: just generally less neuroplasticity uh but it's not uh it's kind of
- 2:41:29counterintuitive because it's not a set thing like you can get
- 2:41:33more neuroplasticity
- 2:41:35SPEAKER_01: right like you can make your brain more plastic
- 2:41:37SPEAKER_01: If you learn like new stuff, right?
- 2:41:39SPEAKER_01: And if you exercise also,
- 2:41:40SPEAKER_01: but like if you learn new stuff,
- 2:41:42SPEAKER_01: your brain becomes a bit more plastic.
- 2:41:44You have neurogenesis that is like start kicking and you're
- 2:41:47able to kind of learn better.
- 2:41:49SPEAKER_01: So,
- 2:41:49but this first like hill that you're climbing and trying to
- 2:41:52learn the new stuff,
- 2:41:54SPEAKER_01: but it's hard and then you don't have time.
- 2:41:56SPEAKER_01: If you like push through it and then you're able to do it well,
- 2:42:00every day afterward gets a bit better because like you learn
- 2:42:03new stuff now you have
- 2:42:04to learn more new stuff but your brain is better at learning
- 2:42:07new stuff now and then
- 2:42:08you kind of snowball into that and if the motivation is in the
- 2:42:11right direction you
- 2:42:12actually see that you're moving then it gets better and better
- 2:42:15but the first few
- 2:42:16SPEAKER_01: steps are the most critical here yeah it's like it's kind of like you let
- 2:42:21SPEAKER_00: life ossify around you.
- 2:42:23SPEAKER_00: You let the concrete harden.
- 2:42:24SPEAKER_00: You got to break the concrete.
- 2:42:26SPEAKER_00: Yes.
- 2:42:26SPEAKER_00: But once the concrete is broken, it's a lot easier.
- 2:42:28SPEAKER_00: This reminds me of just one more concrete example.
- 2:42:31SPEAKER_00: There was a student that we had, well, that we had and started last year.
- 2:42:37SPEAKER_00: And he was talking about how,
- 2:42:39SPEAKER_00: you know,
- 2:42:40he posted some video talking about like the experience of like
- 2:42:44just getting a habit
- 2:42:45SPEAKER_00: in place of doing math every day.
- 2:42:47SPEAKER_00: And he was talking about how like when he started,
- 2:42:49SPEAKER_00: He just was not really into it.
- 2:42:52SPEAKER_00: Like math is like, I think he wants to go into software engineering or,
- 2:42:57SPEAKER_00: data science, something in that direction, in the tech direction.
- 2:43:03SPEAKER_00: Math is something like, okay, I got to learn this material.
- 2:43:07SPEAKER_00: He was doing a degree in computer science.
- 2:43:12There was a bunch of math courses that he was just like,
- 2:43:14man,
- 2:43:15I can't do it.
- 2:43:16SPEAKER_00: I've got to fill in my foundations.
- 2:43:19SPEAKER_00: He started on this and initially he was saying, man, every day I wake up,
- 2:43:24SPEAKER_00: And my brain just tells me like,
- 2:43:27SPEAKER_01: don't do it, man.
- 2:43:28SPEAKER_00: Don't do it.
- 2:43:29SPEAKER_00: Don't, don't start working on the problems.
- 2:43:31SPEAKER_00: This is not going to be fun.
- 2:43:32SPEAKER_00: This is, you should run away from it.
- 2:43:33SPEAKER_00: And so it's like every day.
- 2:43:36SPEAKER_00: And he just like, there was the beginning.
- 2:43:38SPEAKER_00: There's just a lot of like, you know, just like, shut up brain.
- 2:43:42SPEAKER_00: I'm doing this.
- 2:43:43And we're going to do it like in the morning before you're
- 2:43:45really like awake enough.
- 2:43:46SPEAKER_00: I'm going to,
- 2:43:47SPEAKER_00: I'm going to beat you to like,
- 2:43:49SPEAKER_00: before you start yelling at me that I don't like this,
- 2:43:52SPEAKER_00: we're just already going to be doing this.
- 2:43:54SPEAKER_00: And,
- 2:43:55SPEAKER_00: And so he said that, like, it's continued for weeks.
- 2:43:59It was like,
- 2:44:00I can't remember,
- 2:44:00he said like 30 or 60 days or something like that.
- 2:44:04SPEAKER_00: And he was like,
- 2:44:04SPEAKER_00: he was like,
- 2:44:05I just accepted that this feeling is probably just never going
- 2:44:09to go away,
- 2:44:09SPEAKER_00: but I'm going to get through it anyway.
- 2:44:11SPEAKER_00: And then...
- 2:44:12Like a week after he,
- 2:44:13he just accepted this,
- 2:44:15like,
- 2:44:16I'm just,
- 2:44:16this is what we're doing.
- 2:44:18SPEAKER_00: He said like his,
- 2:44:18SPEAKER_00: his,
- 2:44:19SPEAKER_00: he started like just not having all those,
- 2:44:22SPEAKER_00: those bad thoughts anymore about that trying to dissuade him from him.
- 2:44:26It's like he'd wake up and he'd be waiting for his brain to be
- 2:44:29like,
- 2:44:29don't do it,
- 2:44:29man.
- 2:44:30SPEAKER_00: Don't do it.
- 2:44:30SPEAKER_00: Don't you dare start us on math problems.
- 2:44:32SPEAKER_00: Yeah.
- 2:44:33SPEAKER_00: And it's like the brain, it's just given in.
- 2:44:35SPEAKER_00: It was like, just like, okay, yeah, I guess this is what we're doing.
- 2:44:38SPEAKER_00: So like, fine.
- 2:44:39SPEAKER_00: Like,
- 2:44:39SPEAKER_00: it's like,
- 2:44:40SPEAKER_00: if you get yourself,
- 2:44:41you just go through the habit enough to get your brain to just
- 2:44:45accept that this is
- 2:44:48SPEAKER_00: happening.
- 2:44:49This is the path of least resistance is really like,
- 2:44:52you're going to be doing the stuff anyway.
- 2:44:54SPEAKER_00: So why would I dig in my heels?
- 2:44:55SPEAKER_00: Yeah.
- 2:44:55SPEAKER_01: But this is like pure energy conservation from the brain, right?
- 2:44:59SPEAKER_01: Because it is using a lot of energy, right?
- 2:45:02SPEAKER_01: Compared to the rest of the organs.
- 2:45:03SPEAKER_01: So like anything you can not do, it will fucking do it, right?
- 2:45:07SPEAKER_01: So if he sees that you're doing something novel,
- 2:45:10SPEAKER_01: especially later in life,
- 2:45:11SPEAKER_01: it's like,
- 2:45:11SPEAKER_01: dude,
- 2:45:14SPEAKER_01: I have to do all this effort.
- 2:45:17SPEAKER_01: And at some point,
- 2:45:17SPEAKER_01: it's just able to reconfigure itself to kind of lose a bit less energy,
- 2:45:22SPEAKER_01: right?
- 2:45:22SPEAKER_01: And he has to spend a bit more.
- 2:45:24SPEAKER_01: But now it's better at, like, doing that stuff.
- 2:45:27But if you keep doing this,
- 2:45:28actually,
- 2:45:28like,
- 2:45:28your brain is just more flexible in general.
- 2:45:31SPEAKER_01: Because you're not,
- 2:45:32SPEAKER_01: like,
- 2:45:32SPEAKER_01: most of the time,
- 2:45:32SPEAKER_01: you're not in a resource-scarce environment where,
- 2:45:35SPEAKER_01: like,
- 2:45:35SPEAKER_01: you won't have enough to eat,
- 2:45:37SPEAKER_01: right?
- 2:45:38SPEAKER_01: So, like, this kind of remnant of our evolution is kind of not that.
- 2:45:42SPEAKER_01: You have to kind of fight it a bit.
- 2:45:44SPEAKER_01: And then your brain is a bit better.
- 2:45:46SPEAKER_01: I think, so, if you're an adult here, like, you can do it.
- 2:45:51SPEAKER_01: It's never too late.
- 2:45:53I want to switch gears into maybe it's a bit too niche of a
- 2:45:58topic because I really
- 2:45:59SPEAKER_01: like learning.
- 2:45:59SPEAKER_01: Go for it.
- 2:46:01SPEAKER_00: I'm happy to talk about anything.
- 2:46:02SPEAKER_01: I really like it.
- 2:46:06I've seen you and the rest of the Metacame folks talk about
- 2:46:10mastery learning a lot.
- 2:46:11SPEAKER_01: And I was reading your 3000 hour, like the self-directed intense stuff.
- 2:46:18SPEAKER_01: And the first thought that came into my mind was ultra learning.
- 2:46:23SPEAKER_01: It's like by this dude, Scott Young, right?
- 2:46:26SPEAKER_01: And yeah, like he branded the whole thing.
- 2:46:28But like at the core of it is that it's kind of an intense way
- 2:46:32of learning to do like a specific
- 2:46:35SPEAKER_01: specific skill.
- 2:46:36SPEAKER_01: And he went through the same thing as you,
- 2:46:38like he did the MIT courseware and he was trying to blitz it
- 2:46:41all in the year and
- 2:46:42SPEAKER_01: all of this thing.
- 2:46:43And then there was some stuff that like he was able to kind of
- 2:46:46like do all the
- 2:46:48SPEAKER_01: exercise,
- 2:46:49SPEAKER_01: all the exam and all this stuff.
- 2:46:50SPEAKER_01: And when he was actually doing the rigorous type of grading,
- 2:46:55SPEAKER_01: he got a bit lower result than he had before when he was self grading.
- 2:46:59SPEAKER_01: but it was still pretty good.
- 2:47:00SPEAKER_01: So I think there's a lot of connection.
- 2:47:03Well,
- 2:47:03first of all,
- 2:47:04are you familiar with like this whole ultra learning stuff or
- 2:47:06not at all?
- 2:47:07SPEAKER_00: I'm a little bit familiar with it.
- 2:47:09SPEAKER_00: I have the book on my reading list, Ultra Learning book.
- 2:47:13SPEAKER_00: I've read through a few blogs on his journey through this.
- 2:47:20SPEAKER_00: This was a while ago, so I'm a little rusty on all the details of it.
- 2:47:25SPEAKER_00: This comes up very frequently when I talk to people.
- 2:47:29SPEAKER_00: They're like, oh, this is very similar to the Ultra Learning thing.
- 2:47:32SPEAKER_00: I'm like, yeah, that's right.
- 2:47:34SPEAKER_00: I need to read this thing again.
- 2:47:36SPEAKER_00: I need to...
- 2:47:37SPEAKER_01: It's funny because it started the same way, exactly the same way as you.
- 2:47:41SPEAKER_01: But you went the other way at the end.
- 2:47:47You went through mastery learning,
- 2:47:48which is a bit different than what this thing is about.
- 2:47:52SPEAKER_01: which is like you have structured,
- 2:47:54SPEAKER_01: like efficient,
- 2:47:56SPEAKER_01: like no real autonomy about the sequence,
- 2:47:59SPEAKER_01: about where you're going.
- 2:48:01SPEAKER_00: And then you build this as fast as possible.
- 2:48:03SPEAKER_01: But like when you were doing the 2000 hours,
- 2:48:07it was kind of a bit the opposite where like you had this goal
- 2:48:09and then you were
- 2:48:10SPEAKER_01: kind of doing like the course in a more exploratory fashion.
- 2:48:15um how do you do these two actually so when i was when i was uh
- 2:48:20doing my own
- 2:48:21version of the the ultra learning marathon a 3 000 hour study
- 2:48:25marathon so i was um
- 2:48:27i was actually not really going in my own interest areas this
- 2:48:32was i think that's a
- 2:48:33SPEAKER_00: little
- 2:48:33SPEAKER_00: this is a difference between me and my program,
- 2:48:38SPEAKER_00: the ultra learning program,
- 2:48:39SPEAKER_00: how we kind of started out.
- 2:48:40SPEAKER_00: So my, my, and I also did not have a very concrete goal at the end of it.
- 2:48:44SPEAKER_00: So my,
- 2:48:45SPEAKER_00: my goal was actually not to go through the entire,
- 2:48:49SPEAKER_00: or it was not like a specific project that I wanted to do.
- 2:48:54SPEAKER_00: It was not,
- 2:48:55SPEAKER_00: it was just like,
- 2:48:55SPEAKER_00: I,
- 2:48:56SPEAKER_00: I,
- 2:48:57SPEAKER_00: this started with me realizing that,
- 2:48:59SPEAKER_00: like just that earth shattering,
- 2:49:01realization that the pace of school is comically slow and if i
- 2:49:05just take this into
- 2:49:06my own hands then i can learn way faster that paired with like
- 2:49:10just the idea of
- 2:49:11like okay all this um like so initially it was like it started
- 2:49:15i was like okay i'm
- 2:49:16gonna learn calculus that was like the the initial goal um once
- 2:49:21once i got through
- 2:49:21SPEAKER_00: that then it was like i'm just gonna like all this math stuff is
- 2:49:25it seems like it's it's going to be really useful for for me in
- 2:49:28the future and i'm
- 2:49:29enjoying it like this is a lot of fun like i just and improving
- 2:49:34at the stuff it's
- 2:49:35so i i didn't have a a concrete area but i was just like okay i
- 2:49:39just want to learn
- 2:49:40like this this subject is as well as i can as quickly as i can
- 2:49:44and so it's not like
- 2:49:45SPEAKER_00: i was like sitting there like being like okay
- 2:49:48SPEAKER_00: I learned calculus.
- 2:49:49SPEAKER_00: I learned some linear algebra.
- 2:49:51SPEAKER_00: Why don't we go into differential equations?
- 2:49:53SPEAKER_00: How about that differential equation?
- 2:49:54SPEAKER_00: How about that one?
- 2:49:55SPEAKER_00: Oh, that's cool.
- 2:49:56SPEAKER_00: It's like,
- 2:49:57SPEAKER_00: what about a cross-pollination between differential equations and some,
- 2:50:01SPEAKER_00: like,
- 2:50:01SPEAKER_00: in a multiple equation?
- 2:50:02multivariate okay the partial difference like i wasn't like um
- 2:50:06like going on
- 2:50:06wikipedia and going down the rabbit hole on this stuff and
- 2:50:09backfilling it was a
- 2:50:10very structured kind of like i'm gonna like i i just just tell
- 2:50:15me what to do just
- 2:50:16find a resource that just tells me what to do in the most
- 2:50:18efficient way possible
- 2:50:19SPEAKER_00: and i want to do it and so initially like you know
- 2:50:22this i mean i was coming into it kind of like i i i wanted to
- 2:50:26be efficient but i
- 2:50:27wasn't always going about that the right way i was hitting my
- 2:50:30head on like i my my
- 2:50:32end goal was like i wanted to learn this stuff really really
- 2:50:36like i want to get as
- 2:50:36far as i can you want you you you were going for the mastery of
- 2:50:40the skill right
- 2:50:41yeah yeah for the sake of the mastery of the skill uh which i
- 2:50:45think i think you're
- 2:50:46right that this is the core difference because like um in the
- 2:50:48ultra learning kind
- 2:50:49SPEAKER_01: of paradigm it's kind of the
- 2:50:52uh um the opposite in some sense where you always have to have
- 2:50:55a goal if you have
- 2:50:56no goal like yeah it's not gonna work like you're you're
- 2:50:59there's nowhere to anchor
- 2:51:00you're just gonna go endlessly around right so there's a goal
- 2:51:03and then there's like
- 2:51:04this kind of prerequisite chaining kind of movement where you
- 2:51:08try to do this stuff
- 2:51:08it doesn't work you go back like what's the prerequisite for
- 2:51:11this and then this
- 2:51:12SPEAKER_01: doesn't work you go back and then then you go back and you go back so
- 2:51:15SPEAKER_01: So you're right.
- 2:51:17With the discussion that we had before,
- 2:51:20there is some issue with the top-down approach.
- 2:51:24If the goal is not well-defined and if the gap is way too big,
- 2:51:27like if you've never seen
- 2:51:30SPEAKER_01: like addition and subtraction,
- 2:51:33SPEAKER_01: and you're trying to do the top-down learning,
- 2:51:35SPEAKER_01: at the end of the day,
- 2:51:37it's going to be a waste of time because there's just so much
- 2:51:39prerequisite that you
- 2:51:40SPEAKER_01: have to go through.
- 2:51:42And I think that's where there's the disconnect between what
- 2:51:48you were doing and
- 2:51:50SPEAKER_01: what something like Ultra Learning and Scott Young was doing.
- 2:51:53They had a specific goal in mind and they were trying to blaze
- 2:51:56through it as fast as possible.
- 2:51:58SPEAKER_00: You know,
- 2:51:59SPEAKER_00: one thing I want to say is that,
- 2:52:01SPEAKER_00: like,
- 2:52:01sometimes I'm never if anyone has a particular goal that they
- 2:52:07want to achieve,
- 2:52:07SPEAKER_00: like in machine learning,
- 2:52:08SPEAKER_00: I want to reimplement this paper or whatever.
- 2:52:10SPEAKER_00: Like, I never want to say, like, don't even try it.
- 2:52:13SPEAKER_00: Like.
- 2:52:14SPEAKER_00: go fill in your foundation.
- 2:52:17SPEAKER_00: Go, try it.
- 2:52:18SPEAKER_00: Give it a whirl.
- 2:52:19SPEAKER_00: If you are close enough within striking distance that you can do it,
- 2:52:23SPEAKER_00: then who am I to say that you shouldn't?
- 2:52:25I'm just saying if somebody is like,
- 2:52:27man,
- 2:52:28I tried it,
- 2:52:28but I can't do this thing,
- 2:52:31it seems I...
- 2:52:32what what the hell do i do like am i just too dumb to do it
- 2:52:36then that's what i'm
- 2:52:36like okay no stop stop what you're doing get your foundations
- 2:52:41in place this is
- 2:52:41going to make all the difference in in the world yeah and i
- 2:52:45just like get the
- 2:52:46foundation like really slowly yes as efficiently as possible so
- 2:52:50like i i don't even
- 2:52:51think like there's um so much disconnect between like a mastery
- 2:52:55learning and ultra
- 2:52:56learning like in in mastery you're trying to go to for the
- 2:52:59mastery of the skill so
- 2:53:00SPEAKER_01: you have to make sure that you have the
- 2:53:01right foundation at the right place and you're honing that
- 2:53:03stuff but like maybe
- 2:53:05you're already mastering enough to do the thing so you can just
- 2:53:07do it and then like
- 2:53:08you continue and then build up the mastery um and like uh this
- 2:53:12whole question set
- 2:53:13like uh got sparked because of that biology stuff that you were
- 2:53:17doing right um it
- 2:53:19really was looking like uh some ultra learning blitz that you
- 2:53:23were doing and then
- 2:53:24SPEAKER_01: you were doing the prerequisite chaining
- 2:53:26SPEAKER_01: And like every time the LM was like,
- 2:53:27SPEAKER_01: Hey,
- 2:53:28SPEAKER_01: what do you want to like,
- 2:53:29SPEAKER_01: you're like,
- 2:53:29SPEAKER_01: dude,
- 2:53:30SPEAKER_01: I'm focused on this and we're going to that goal.
- 2:53:33SPEAKER_01: In my view,
- 2:53:33SPEAKER_01: this was more on the,
- 2:53:35SPEAKER_01: on the,
- 2:53:36SPEAKER_01: on the edge of a,
- 2:53:37of ultra learning kind of sprint project than the mastery type
- 2:53:41of a skillset
- 2:53:43SPEAKER_01: building.
- 2:53:43SPEAKER_00: Yeah.
- 2:53:43SPEAKER_00: That's interesting.
- 2:53:44SPEAKER_00: So I,
- 2:53:45SPEAKER_00: when I was doing the biology stuff,
- 2:53:47I had kind of viewed it as like the math Academy approach to
- 2:53:50just what do you do if
- 2:53:52SPEAKER_00: there's like literally no infrastructure.
- 2:53:54SPEAKER_00: And so I,
- 2:53:55SPEAKER_00: I kind of viewed it as like,
- 2:53:56SPEAKER_00: OK,
- 2:53:56SPEAKER_00: I just want to spin up on biology as fast as possible,
- 2:54:00SPEAKER_00: as efficiently as just master the subject.
- 2:54:03SPEAKER_00: But but yeah, now that you bring it up, it's true.
- 2:54:05SPEAKER_00: I kind of scoped it down in particular to,
- 2:54:07SPEAKER_00: you know,
- 2:54:08SPEAKER_00: genetics,
- 2:54:09SPEAKER_00: because that was that that was the area.
- 2:54:11SPEAKER_00: And and and so I think overall, like they're they're.
- 2:54:15there is a reasonable amount of scoping down that you can often
- 2:54:18do through top-down
- 2:54:20SPEAKER_00: planning whenever you have a goal.
- 2:54:22SPEAKER_00: And so that's something that we've actually...
- 2:54:25SPEAKER_00: We've created courses on Math Academy specifically for that purpose,
- 2:54:28SPEAKER_00: like Math for...
- 2:54:29machine learning math for mathematical methods for physics or
- 2:54:33it's like you know if
- 2:54:35somebody wants to do machine learning you know you're gonna
- 2:54:38have to learn linear
- 2:54:39algebra multivariable calculus probability and statistics but
- 2:54:43you don't need to
- 2:54:43learn all of each of those courses like there's a lot of
- 2:54:47multivariable calculus
- 2:54:48like green's theorem divergence theorem stokes theorem like at
- 2:54:51the that come like
- 2:54:53SPEAKER_00: towards the end that's like you know that's not the type of
- 2:54:55SPEAKER_00: multivariable calculus you need.
- 2:54:57Like it really just need up to like the chain rule and like
- 2:55:00hyperplanes and stuff
- 2:55:02SPEAKER_00: like that,
- 2:55:03SPEAKER_00: which is really,
- 2:55:04SPEAKER_00: I mean,
- 2:55:04SPEAKER_00: it's a large chunk of multivariable calculus,
- 2:55:06SPEAKER_00: but it's not the whole thing.
- 2:55:07SPEAKER_00: And so you can, there's a lot of reasonable scoping down
- 2:55:10SPEAKER_00: that you can do with that.
- 2:55:12SPEAKER_00: That's kind of like a,
- 2:55:13SPEAKER_00: almost like a high,
- 2:55:15I guess it sounds almost like a midway between like the entire
- 2:55:19subject mastery
- 2:55:21SPEAKER_00: learning versus the only prerequisite chaining ultra learning to like,
- 2:55:27SPEAKER_00: just kind of like a scope it down,
- 2:55:30SPEAKER_00: top down and then fill it up,
- 2:55:32SPEAKER_00: bottom up sort of thing.
- 2:55:33SPEAKER_01: And like,
- 2:55:34SPEAKER_01: to be very honest,
- 2:55:35SPEAKER_01: like the ultra learning,
- 2:55:36SPEAKER_01: like type of learning is very efficient in order to get to your stuff.
- 2:55:40but like your foundation is still brittle right like uh there's
- 2:55:43no escaping you're
- 2:55:44not you've not become master in math right or like master of
- 2:55:49like language because
- 2:55:49now you can speak it um you just are able to do it right um
- 2:55:54which is very very
- 2:55:54SPEAKER_01: different
- 2:55:55than like a full-blown mastery it's more efficient to get the
- 2:56:00goal and it's faster
- 2:56:01it's more rewarding in my view also if you're going at it at
- 2:56:04the right way and then
- 2:56:05you understand like you actually have to learn stuff like
- 2:56:07you're not going to swing
- 2:56:08it and then like it'd be done um yeah so um cool perfect so i
- 2:56:13got my information
- 2:56:14here the last topic i wanted to bring in is um kind of more on
- 2:56:18the like uh
- 2:56:20SPEAKER_01: at the French, right?
- 2:56:22SPEAKER_01: At the not typical type of stuff, right?
- 2:56:26SPEAKER_01: I have a question specifically for like the neurodivergent people,
- 2:56:30SPEAKER_01: because there's a lot of,
- 2:56:32SPEAKER_01: I'm not neurodivergent,
- 2:56:33SPEAKER_01: like actually,
- 2:56:35I don't have any problem like focusing for a very,
- 2:56:38very extended period of time.
- 2:56:40And it was always a trouble when I was like speaking to people
- 2:56:44with ADHD and that
- 2:56:45they were with me and stuff because what I was doing was not
- 2:56:49fitting at all like
- 2:56:51SPEAKER_01: what they could do.
- 2:56:52SPEAKER_01: And not just in like, hey, just try better.
- 2:56:54SPEAKER_01: It was like almost physically impossible
- 2:56:57to do the same amount of like I could stay for like six hours
- 2:57:01like not moving and
- 2:57:02just like churning like a whatever it take in term of like math
- 2:57:07problem right it
- 2:57:07didn't matter was reading and stuff the only my only issue is I
- 2:57:11had to eat and at
- 2:57:12some point I was getting tired so I knew like my brain was all
- 2:57:15fucked up but I need
- 2:57:16to sleep for 10 hours that was the stuff that I could do right
- 2:57:18I could do that for
- 2:57:19a week no issue right and I was uh showing that to like my
- 2:57:24brother and like my my
- 2:57:25SPEAKER_01: The other people in my extended family that were more on the ADHD side,
- 2:57:31it was kind of impossible for them to even conceptualize like
- 2:57:33this is possible.
- 2:57:35How would these type of people that have more of an ADHD type
- 2:57:43of neural profile
- 2:57:44SPEAKER_01: would fit in your model?
- 2:57:46SPEAKER_00: Yeah.
- 2:57:46SPEAKER_00: So,
- 2:57:47SPEAKER_00: you know,
- 2:57:47SPEAKER_00: the interesting thing is a lot of these people with ADHD think that the,
- 2:57:53SPEAKER_00: you know,
- 2:57:54SPEAKER_00: just a lot of the education system is not cut out for them.
- 2:57:58SPEAKER_00: And that they have like a,
- 2:58:00SPEAKER_00: you know,
- 2:58:00SPEAKER_00: that they're just not geared for optimal learning and stuff.
- 2:58:04SPEAKER_00: And so only one of those things is true.
- 2:58:06The education system as it stands is not friendly to people
- 2:58:10with disabilities.
- 2:58:11with attention challenges like people who can't sit still like
- 2:58:15actually i mean it's
- 2:58:16it's really not friendly to to anyone but especially people
- 2:58:19with attention
- 2:58:20challenges and and it's like nobody likes being talked at for
- 2:58:24an hour on a subject
- 2:58:26but a lot of like yeah i'm kind of similar to you in that like
- 2:58:30i can just like you
- 2:58:30SPEAKER_00: know
- 2:58:31I can grind on something for like six hours before I even
- 2:58:34realize that any time has passed.
- 2:58:39SPEAKER_00: But it's like, yeah.
- 2:58:41SPEAKER_00: But if you take somebody who has like just,
- 2:58:45SPEAKER_00: you know,
- 2:58:46SPEAKER_00: is not like you and me in this opposite way,
- 2:58:51like just the impact of being talked at for an hour or like it
- 2:58:55just kind of limits
- 2:58:56SPEAKER_00: the amount of,
- 2:58:57SPEAKER_00: you know,
- 2:58:58How much can you communicate to somebody before they start
- 2:59:01doing something?
- 2:59:02SPEAKER_00: And, you know, initially that might seem like some kind of disadvantage.
- 2:59:06SPEAKER_00: Like,
- 2:59:06SPEAKER_00: well,
- 2:59:07I just can't sit still for like for the hour it takes to go
- 2:59:11through a lecture or
- 2:59:11SPEAKER_00: whatever.
- 2:59:12SPEAKER_00: Like, you know, but really like if you look into like the
- 2:59:16SPEAKER_00: you start going down a rabbit hole in the science of learning.
- 2:59:18SPEAKER_00: It's like,
- 2:59:19SPEAKER_00: well,
- 2:59:19that's not even a good thing to do in the first place to be
- 2:59:21talked at for an hour.
- 2:59:22SPEAKER_00: That's like,
- 2:59:23SPEAKER_00: it's just people have a,
- 2:59:25SPEAKER_00: some other people have a higher tolerance for it,
- 2:59:27SPEAKER_00: but it's not good for anybody.
- 2:59:28SPEAKER_00: What's,
- 2:59:28SPEAKER_00: what's good is these minimum effective doses of just like,
- 2:59:32SPEAKER_00: okay,
- 2:59:32SPEAKER_00: just tell me a little bit of what to do and then let me start doing it.
- 2:59:35SPEAKER_00: And then like, you know, just active, active, active.
- 2:59:38The more you stay just like in a sort of passive state or just
- 2:59:42like not doing anything.
- 2:59:44SPEAKER_00: It's like, you're not,
- 2:59:44SPEAKER_00: You know,
- 2:59:45if you go to tennis lessons and the coach just talks that for
- 2:59:48you an hour,
- 2:59:48SPEAKER_00: like,
- 2:59:49SPEAKER_00: I don't care if you have ADHD or not.
- 2:59:50SPEAKER_00: Like, you're not getting better at tennis.
- 2:59:52SPEAKER_00: Like...
- 2:59:53SPEAKER_00: If you have ADHD, then you might just be expressing more discomfort.
- 2:59:57SPEAKER_00: Like, come on, let's go.
- 2:59:59SPEAKER_00: Let's hit some balls.
- 3:00:00SPEAKER_00: Let's do something, right?
- 3:00:01SPEAKER_00: But either way, the learning's not happening.
- 3:00:03So I think a lot of times the neurodivergence stuff just makes
- 3:00:08you more sensitive
- 3:00:09SPEAKER_00: to things that are like,
- 3:00:11SPEAKER_00: there's an issue here.
- 3:00:14SPEAKER_00: You feel the issue more earlier on.
- 3:00:18SPEAKER_00: But so,
- 3:00:19SPEAKER_00: you know,
- 3:00:20SPEAKER_00: I mentioned I am very similar to you in that I can,
- 3:00:25SPEAKER_00: you know,
- 3:00:26SPEAKER_00: the passage of time is just it's almost it's hard to keep on track of.
- 3:00:32SPEAKER_00: I can just like.
- 3:00:33SPEAKER_00: sink into somebody.
- 3:00:35SPEAKER_00: Our founder of Math Academy,
- 3:00:37SPEAKER_00: Jason Roberts,
- 3:00:38SPEAKER_00: he is actually on the exact opposite end of the spectrum.
- 3:00:42SPEAKER_00: We're in the ADHD spectrum.
- 3:00:44SPEAKER_00: So a lot of Math Academy is built in terms of
- 3:00:48of what would jason have want what did jason hate about math
- 3:00:52education growing up
- 3:00:53and and like you know and that was one of the things like being
- 3:00:56talked at for a
- 3:00:57long time before actually doing problems like just hated it
- 3:01:01because like you know
- 3:01:02he could he's got these stories about like well he would just
- 3:01:04be sitting in class
- 3:01:05he could never sit still the teacher's like jason pay attention
- 3:01:08like what are you
- 3:01:08doing and it's like well but the problem isn't jason not paying
- 3:01:11attention the
- 3:01:12problem is that like he's being talked at for too long without
- 3:01:15doing anything and
- 3:01:16SPEAKER_01: yeah
- 3:01:17SPEAKER_00: Yeah.
- 3:01:17And so Jason actually has three kids who are all across the
- 3:01:23different spectrum.
- 3:01:24One was just like very highly computationally mathematically
- 3:01:31gifted and trained.
- 3:01:32SPEAKER_00: just like blows through all this stuff.
- 3:01:34SPEAKER_00: And he has another kid with dyscalculia,
- 3:01:38and then another kid who's kind of just more of your typical
- 3:01:42honors math student
- 3:01:43SPEAKER_00: level.
- 3:01:44SPEAKER_00: And so the kind of a
- 3:01:47approach that we take is is kind of you know making sure that
- 3:01:50that this efficient
- 3:01:52learning cycle is working for for all the the students all
- 3:01:56these kids you know he
- 3:01:57and i are he and i are so different um too and we just agree on
- 3:02:01like all the same
- 3:02:02things like it's just whether you come at it from like okay kid
- 3:02:05with adhd can't
- 3:02:07SPEAKER_00: It doesn't want to sit still for 15 minutes for an explanation,
- 3:02:11or you come at it from just a calculate the learning efficiency
- 3:02:14of this setting
- 3:02:15SPEAKER_00: versus that setting.
- 3:02:15SPEAKER_00: You get to the same answer, I think.
- 3:02:17SPEAKER_01: I think the optimal learning flow, I think it's kind of the same in all.
- 3:02:23SPEAKER_01: I think the issue with most ADHD folks is that at the ground level,
- 3:02:31SPEAKER_01: they have this dopaminergic dysregulation.
- 3:02:34But there's also all the other emotional components that are
- 3:02:37layered on top.
- 3:02:37SPEAKER_01: I guess this is a very...
- 3:02:39SPEAKER_01: It's not just like, hey, I have a problem.
- 3:02:41SPEAKER_01: Your brain is not the same as neurotypical...
- 3:02:45SPEAKER_01: Like, it's not thinking the same way, right?
- 3:02:48SPEAKER_01: And it's fine to have, like, all of these different profiles.
- 3:02:51SPEAKER_01: But what I see often is that,
- 3:02:52SPEAKER_01: like,
- 3:02:52SPEAKER_01: they have this shame of,
- 3:02:55SPEAKER_01: like,
- 3:02:55SPEAKER_01: not being good or,
- 3:02:56SPEAKER_01: like,
- 3:02:56SPEAKER_01: not being able to sit there and then,
- 3:02:57SPEAKER_01: like,
- 3:02:58SPEAKER_01: listen to it.
- 3:02:59So then they develop all of these kind of,
- 3:03:02like,
- 3:03:02bad emotions with the thing.
- 3:03:05SPEAKER_01: And, like, it's not about the thing anymore, right?
- 3:03:07SPEAKER_01: It's about how the thing make them feel.
- 3:03:09SPEAKER_01: And this is an issue.
- 3:03:11Because also,
- 3:03:12like,
- 3:03:12if you look at these people,
- 3:03:15it's not they can't focus,
- 3:03:17right?
- 3:03:17SPEAKER_01: The focus is not, like, as evenly spread as a neurotypical.
- 3:03:21SPEAKER_01: Like, in my case, I could be, like, learning about anything.
- 3:03:25SPEAKER_01: And it's going to be fine.
- 3:03:26SPEAKER_01: Like, we're going to, like, learn about ballet.
- 3:03:29SPEAKER_01: I'm going to listen.
- 3:03:30SPEAKER_01: And we're going to try to figure it out.
- 3:03:32SPEAKER_01: We're going to learn about, like, rocks.
- 3:03:34SPEAKER_01: We're going to listen.
- 3:03:35SPEAKER_01: I'm going to figure this out, right?
- 3:03:37SPEAKER_01: But like in most of these cases,
- 3:03:40SPEAKER_01: if it doesn't fit into like the cone of interest that they have,
- 3:03:45SPEAKER_01: it's going to be very difficult.
- 3:03:46SPEAKER_01: But when it fits in there or you make it fit for some reason, right?
- 3:03:50SPEAKER_01: Making it easier to ingest into that little cone.
- 3:03:53SPEAKER_01: Yeah.
- 3:03:55SPEAKER_01: Like some of them could sit still for like eight hours, no issue, right?
- 3:03:58SPEAKER_01: Like I've seen people that like their thing was coding, right?
- 3:04:02SPEAKER_01: Like this is the stuff.
- 3:04:04SPEAKER_01: they really liked front-end web development, right?
- 3:04:06SPEAKER_01: For some reason.
- 3:04:07SPEAKER_01: This is like something that they were highly worried in.
- 3:04:09SPEAKER_01: You talk to them about operating system or other topic,
- 3:04:12SPEAKER_01: like they kind of zone out,
- 3:04:14SPEAKER_01: like,
- 3:04:14SPEAKER_01: yeah,
- 3:04:14SPEAKER_01: but like,
- 3:04:14SPEAKER_01: it's not visual.
- 3:04:15SPEAKER_01: I want to click on the stuff and like see the things.
- 3:04:18SPEAKER_01: They could sit down
- 3:04:19SPEAKER_01: Like there's no issue about it.
- 3:04:21SPEAKER_01: It's just like not as evenly spread out.
- 3:04:23SPEAKER_01: So if you run into like these,
- 3:04:25these like talking session and they can't see and how to kind
- 3:04:29of ingest into their
- 3:04:30SPEAKER_01: stuff that doesn't fit in there.
- 3:04:31SPEAKER_01: And then they pile up all these emotions.
- 3:04:34um then it's you have to deconstruct that before doing anything
- 3:04:37else like i had
- 3:04:38students like this like they like it was a like a psychiatry
- 3:04:41session for like the
- 3:04:42first three four sessions like no it's not you it's like this
- 3:04:46and like you see and
- 3:04:48SPEAKER_01: then like after some
- 3:04:49time like then we pick up pace and then we were able to do it
- 3:04:51in a one-on-one
- 3:04:52session but if they didn't have that or like they didn't have
- 3:04:54like someone with
- 3:04:55patience um they just they just will feel dumb and that's the
- 3:04:59end of it absolutely
- 3:05:00yeah it's kind of funny the feeling of it's like the feeling of
- 3:05:04dumb is often the
- 3:05:05real impediment to the learning not not the not the being dumb
- 3:05:09like usually the
- 3:05:10issue is not that you're dumb it's just that you'd like just
- 3:05:14some other settings
- 3:05:16SPEAKER_00: that are exactly not
- 3:05:17SPEAKER_00: you know,
- 3:05:18SPEAKER_00: causing the information to flow,
- 3:05:19SPEAKER_00: whether that's a prerequisite gap or just a,
- 3:05:22SPEAKER_00: like,
- 3:05:23SPEAKER_00: you know,
- 3:05:23SPEAKER_00: just not wanting to participate in the,
- 3:05:26SPEAKER_00: in the process or,
- 3:05:28SPEAKER_00: or,
- 3:05:28SPEAKER_00: or whether it's,
- 3:05:29SPEAKER_00: you're just feeling like this is not you,
- 3:05:31SPEAKER_00: you don't identify with this stuff.
- 3:05:34SPEAKER_00: This is not something you're good at.
- 3:05:36SPEAKER_00: It's all these kinds of blockades going to be put up and it's just,
- 3:05:40SPEAKER_00: this is separate from the actual,
- 3:05:41SPEAKER_00: like,
- 3:05:42SPEAKER_00: yeah.
- 3:05:42how good you are at that i think that's actually one of my
- 3:05:45superpowers like um i'm
- 3:05:47okay being like looking like a dumbass because i know that like
- 3:05:51that's just normal
- 3:05:52like it's you're you're new to the field right you're we have
- 3:05:54to learn all this
- 3:05:55stuff yeah just just be dumb fast right and then you're gonna
- 3:05:58be good and it's not
- 3:05:58gonna be okay like the guy would laugh at you because your
- 3:06:01question is like
- 3:06:03SPEAKER_01: absolutely basic okay now i know it and like
- 3:06:06SPEAKER_01: I don't care about this guy, I'm here for this stuff, right?
- 3:06:08SPEAKER_01: So let's learn it.
- 3:06:09And if we switch gears to how the new technologies are entering
- 3:06:17the picture now,
- 3:06:18SPEAKER_01: do you think that if,
- 3:06:20SPEAKER_01: let's say,
- 3:06:20there's a gifted amateur mathematician and he's doing math
- 3:06:24academy and stuff,
- 3:06:25and then he has all these emerging tooling and can use AI to
- 3:06:30probe a bit on the
- 3:06:31SPEAKER_01: frontier,
- 3:06:32Do you think we can see like some more like amateurish
- 3:06:36mathematician be able to
- 3:06:38contribute to the frontier or still like we will need to have
- 3:06:42like this kind of
- 3:06:43SPEAKER_01: formal mastery learning process?
- 3:06:46SPEAKER_00: Yeah, that's a really good question.
- 3:06:49SPEAKER_00: All right.
- 3:06:49SPEAKER_00: So I'm not a research mathematician.
- 3:06:51So there is a I don't want to like overstep in my kind of
- 3:06:55answer to to this making
- 3:06:56SPEAKER_00: claims that I don't actually know what I'm talking about.
- 3:06:59SPEAKER_00: So I think I think this is
- 3:07:01SPEAKER_00: this question would be best posed to somebody who has a track record of
- 3:07:06SPEAKER_00: mathematical publications.
- 3:07:09Let me generalize the scope here a little bit about just
- 3:07:13amateurs contributing to
- 3:07:15SPEAKER_00: the edge of fields in general.
- 3:07:17SPEAKER_00: One thing,
- 3:07:18SPEAKER_00: at least in my experience,
- 3:07:20SPEAKER_00: and that I hear from a lot of other people confirmed,
- 3:07:22SPEAKER_00: is that this tooling that exists nowadays,
- 3:07:26SPEAKER_00: like LLM stuff,
- 3:07:29SPEAKER_00: Cloud,
- 3:07:30SPEAKER_00: OpenAI,
- 3:07:31SPEAKER_00: whatever,
- 3:07:32SPEAKER_00: all these kinds of tools are,
- 3:07:34SPEAKER_00: they're kind of multipliers on what you're able to do initially.
- 3:07:39SPEAKER_00: And so,
- 3:07:40SPEAKER_00: now,
- 3:07:40SPEAKER_00: like,
- 3:07:41SPEAKER_00: if you're not able to do a whole bunch,
- 3:07:44SPEAKER_00: you don't have,
- 3:07:44SPEAKER_00: like,
- 3:07:45SPEAKER_00: technical chops,
- 3:07:46SPEAKER_00: and
- 3:07:47SPEAKER_00: And then you,
- 3:07:47SPEAKER_00: you,
- 3:07:48SPEAKER_00: you,
- 3:07:48SPEAKER_00: you ask like some stuff,
- 3:07:50SPEAKER_00: ask questions to the LMS,
- 3:07:51SPEAKER_00: have it like code up stuff for you that you couldn't otherwise do.
- 3:07:54SPEAKER_00: This can feel like sort of an,
- 3:07:57SPEAKER_00: like an equalizing,
- 3:07:58SPEAKER_00: like,
- 3:07:58SPEAKER_00: Oh,
- 3:07:58SPEAKER_00: Hey,
- 3:07:58SPEAKER_00: like I'm at the same caliber as like everybody.
- 3:08:01SPEAKER_00: Like, Hey, I don't have to,
- 3:08:03learn how to how to code very well because i can just vibe code
- 3:08:05everything and like
- 3:08:07i don't have to learn pedagogical like like cognitive science
- 3:08:11and stuff i can just
- 3:08:12tell like tell tell tell the llm to to generate like a a course
- 3:08:18for me or like
- 3:08:19stuff like that it's like well okay that it is i guess leveling
- 3:08:24you up but now like
- 3:08:25SPEAKER_00: think about how much it's leveling up people who actually know like these
- 3:08:30SPEAKER_00: foundations of what's going on
- 3:08:32SPEAKER_00: The thing is,
- 3:08:34SPEAKER_00: when you have a high level of mastery in something,
- 3:08:38SPEAKER_00: at least in my experience,
- 3:08:39SPEAKER_00: these tools can scale you.
- 3:08:43SPEAKER_00: You are the zero to one,
- 3:08:45SPEAKER_00: and then these tools put the zeros behind the one,
- 3:08:48SPEAKER_00: go from 10 to 100 to 1,000.
- 3:08:50SPEAKER_00: And if you don't really know what you're doing,
- 3:08:53SPEAKER_00: you know,
- 3:08:53SPEAKER_00: you can't,
- 3:08:54SPEAKER_00: it's like you start being led by the LLM.
- 3:08:57SPEAKER_00: Exactly.
- 3:08:58SPEAKER_01: It will lead you in a weird direction.
- 3:09:00Like it will lead you like cheerfully into like some psychosis
- 3:09:06area where you're
- 3:09:06SPEAKER_01: like,
- 3:09:07SPEAKER_01: what the fuck is happening here?
- 3:09:09SPEAKER_01: I receive a lot of these DMs like,
- 3:09:10SPEAKER_01: hey,
- 3:09:10SPEAKER_01: look,
- 3:09:10SPEAKER_01: I made this in like 26 pages of like,
- 3:09:14SPEAKER_01: what the hell are you even saying?
- 3:09:17SPEAKER_01: Is it good?
- 3:09:17SPEAKER_01: I see that too.
- 3:09:19SPEAKER_01: Did you try it?
- 3:09:20SPEAKER_01: Did you run this stuff?
- 3:09:21SPEAKER_01: And you're like, no, but am I into something?
- 3:09:24SPEAKER_01: And then you see on the opposite side,
- 3:09:28SPEAKER_01: when you do have the skills,
- 3:09:29SPEAKER_01: if you look at somebody like Terence Tao,
- 3:09:32he's actively using these tools and getting a lot out of them
- 3:09:35because he's really
- 3:09:36SPEAKER_01: good at this stuff.
- 3:09:37SPEAKER_01: He's one of the best, right?
- 3:09:39And he's able to leverage them to help him scaffold and move a
- 3:09:44bit faster.
- 3:09:45um but still like it's his intellect and it's like his mastery
- 3:09:49of the material that
- 3:09:50is able to like figure out when they're going wrong or like
- 3:09:53like the the question
- 3:09:55is not well posed or like uh this the there is something
- 3:09:58missing and what he's
- 3:10:00saying or like there's a gap or something i didn't know that i
- 3:10:03need to bring into
- 3:10:04the picture and okay yes true i remember that stuff is for this
- 3:10:08and then you can
- 3:10:08you can kind of pile this stuff together but um i agree like on
- 3:10:11their own like uh
- 3:10:13SPEAKER_01: uh it's very difficult for uh
- 3:10:15SPEAKER_01: for a beginner to not get like absolutely swayed out, right?
- 3:10:20Like they will not know where to go and like what's truly
- 3:10:23novel,
- 3:10:23what is at the edge.
- 3:10:25SPEAKER_00: Yeah, yeah, absolutely.
- 3:10:27SPEAKER_00: Yeah,
- 3:10:27SPEAKER_00: you know,
- 3:10:27a lot of the efficiency gains that I've seen personally using
- 3:10:30these tools have
- 3:10:32SPEAKER_00: been,
- 3:10:32SPEAKER_00: it's like,
- 3:10:33SPEAKER_00: I mean,
- 3:10:33SPEAKER_00: I'm not...
- 3:10:34It's more about delegating work that I already know how to do
- 3:10:39that I can automate
- 3:10:41SPEAKER_00: more and just free me up to do additional stuff.
- 3:10:46But it's not a way to get me to offload work that I don't know
- 3:10:52how to do because it
- 3:10:52SPEAKER_00: has the same failure modes.
- 3:10:54SPEAKER_00: I've done this before.
- 3:10:55I've made this mistake before of trying to...
- 3:11:00You have a junior employee,
- 3:11:01human employee,
- 3:11:03SPEAKER_00: And you got a project that you're like, oh, this sounds pretty cool.
- 3:11:06SPEAKER_00: You haven't really scoped it out.
- 3:11:07SPEAKER_00: You haven't gotten in the weeds on it or anything.
- 3:11:09SPEAKER_00: You're just like, oh yeah, this sounds like a good thing for you to do.
- 3:11:12And then you look up a day later,
- 3:11:15a week later,
- 3:11:17and you're just like,
- 3:11:17wait,
- 3:11:19hold on.
- 3:11:19SPEAKER_00: What the hell is happening?
- 3:11:20SPEAKER_00: Like this is all, oh my God.
- 3:11:21SPEAKER_00: And then you have, anyway.
- 3:11:24SPEAKER_01: You didn't prompt it well, but it's true.
- 3:11:26SPEAKER_01: They go into the weird direction.
- 3:11:28SPEAKER_01: And if they're not asking for feedback,
- 3:11:30um they're gonna go into like some weird stuff and then you're
- 3:11:33gonna unblock them
- 3:11:34into like two or three sentences and they're like oh and then
- 3:11:37you're like yeah man
- 3:11:38should like and if you can't give the feedback if you don't
- 3:11:40know enough yeah yeah
- 3:11:42if you don't know the feedback you're like yeah just go and do
- 3:11:45it and they're gonna
- 3:11:45try more and they're just gonna burn out like yeah yeah yeah i
- 3:11:49also use these tools
- 3:11:52um like when i know it's it's good but also when i don't know i
- 3:11:55it's useful to get
- 3:11:56SPEAKER_01: like a
- 3:11:57SPEAKER_01: lay of the land, like about 80% correct, right?
- 3:12:01SPEAKER_01: And like, yes, they have their own tone and their own assertiveness.
- 3:12:05If I strip that all away,
- 3:12:06I'm like,
- 3:12:06okay,
- 3:12:07there's like these stuff,
- 3:12:07perfect.
- 3:12:08And then you go and you drill into the right things and you
- 3:12:10learn them properly.
- 3:12:11But then it quickly gives you like a good approximation of like
- 3:12:15what this stuff is
- 3:12:16SPEAKER_01: actually looking like.
- 3:12:17SPEAKER_00: Yeah, yeah.
- 3:12:18SPEAKER_00: It gives you like a little, okay, here's our options here.
- 3:12:21SPEAKER_00: What do you think?
- 3:12:22And then you can kind of like evaluate all those and totally,
- 3:12:24totally agree.
- 3:12:25SPEAKER_00: Yeah,
- 3:12:25SPEAKER_00: you know,
- 3:12:26SPEAKER_00: one of the big unlocks I'd say that I've experienced recently,
- 3:12:29SPEAKER_00: at least with the,
- 3:12:30SPEAKER_00: you know,
- 3:12:31SPEAKER_00: agentic coding stuff was kind of just scoping these things.
- 3:12:36SPEAKER_00: scoping stuff down enough that I can get it into like,
- 3:12:39SPEAKER_00: that it fits into my own working memory of what we're doing here.
- 3:12:44SPEAKER_00: That was one of the challenges that I experienced before is like,
- 3:12:47SPEAKER_00: I'd have this whole plan for like,
- 3:12:51SPEAKER_00: okay,
- 3:12:52SPEAKER_00: here's the feature that we're going to build.
- 3:12:53SPEAKER_00: Here's the thing that we're going to do.
- 3:12:54SPEAKER_00: You ask it to do too much and then
- 3:12:57It's going to go off in directions that you don't anticipate
- 3:13:00because you have not
- 3:13:01like verified everything that's going to be doing because you
- 3:13:03can't keep it in your
- 3:13:04SPEAKER_00: working memory.
- 3:13:05SPEAKER_00: But if you like split this up into like if you are able to like sign off,
- 3:13:09SPEAKER_00: give expert feedback and everything.
- 3:13:10SPEAKER_00: OK, yeah, this all looks good.
- 3:13:12SPEAKER_00: I've seriously looked at every little thing and it looks good.
- 3:13:15SPEAKER_00: You could keep the reins tight like that.
- 3:13:17SPEAKER_00: Then, yeah, it's amazing.
- 3:13:20SPEAKER_01: And it helps,
- 3:13:21SPEAKER_01: like,
- 3:13:21SPEAKER_01: I spend a lot of time,
- 3:13:22SPEAKER_01: like,
- 3:13:23SPEAKER_01: working with a thing to get the right blocks,
- 3:13:27SPEAKER_01: like,
- 3:13:27SPEAKER_01: mapped out,
- 3:13:28SPEAKER_01: right?
- 3:13:29SPEAKER_01: And I usually,
- 3:13:30SPEAKER_01: like,
- 3:13:30SPEAKER_01: make it the right report and stuff,
- 3:13:32SPEAKER_01: and then I look at the report,
- 3:13:33SPEAKER_01: I trim down,
- 3:13:33SPEAKER_01: I remove,
- 3:13:34SPEAKER_01: like,
- 3:13:34SPEAKER_01: bullshit it says,
- 3:13:35SPEAKER_01: and,
- 3:13:35SPEAKER_01: like,
- 3:13:36SPEAKER_01: simplify.
- 3:13:36SPEAKER_01: They want it to go into a big direction,
- 3:13:38SPEAKER_01: and,
- 3:13:38SPEAKER_01: like,
- 3:13:38SPEAKER_01: if you do that,
- 3:13:39SPEAKER_01: like,
- 3:13:39SPEAKER_01: it's going to cost us,
- 3:13:40SPEAKER_01: like...
- 3:13:40thousands of dollars from nothing you just do this thing and
- 3:13:42it's gonna be good and
- 3:13:43then iterate like this and i iterate with it to kind of revise
- 3:13:47the plan and start
- 3:13:48to chunk it out but then i take a fresh session i'm like we're
- 3:13:51only gonna do this
- 3:13:52SPEAKER_01: stuff
- 3:13:52right this is the only thing we're doing and then it does it
- 3:13:55and you review it you
- 3:13:57change a bit of thing and you're like done right then you take
- 3:14:00the other things but
- 3:14:01like all of the architecting and like the the connection to
- 3:14:04like what you're
- 3:14:05actually trying to do is literally you done like uh you done
- 3:14:09all the duct tape to
- 3:14:10put to pull everything together that's much more efficient than
- 3:14:13uh than just trying
- 3:14:14to to let something go when you have absolutely no clue you're
- 3:14:18gonna get something
- 3:14:19SPEAKER_01: out but like
- 3:14:20SPEAKER_01: It's very brittle in structure.
- 3:14:24SPEAKER_01: Perfect, man.
- 3:14:24I had another question,
- 3:14:26but I think we covered this in six different ways already.
- 3:14:31SPEAKER_01: I wanted to ask you to close,
- 3:14:34SPEAKER_01: what are you most excited about at the moment in terms of learning,
- 3:14:38SPEAKER_01: math academy,
- 3:14:39SPEAKER_01: math in general?
- 3:14:40SPEAKER_01: What's the thing that excites you the most in this area?
- 3:14:43SPEAKER_00: Yeah.
- 3:14:44SPEAKER_00: You know,
- 3:14:44there's one thing that I think about as soon as I wake up late
- 3:14:49into the night all
- 3:14:50SPEAKER_00: the time.
- 3:14:50SPEAKER_00: It's what I'm working on.
- 3:14:51SPEAKER_00: And I think I mentioned at the very beginning when he asked,
- 3:14:55SPEAKER_00: like,
- 3:14:55SPEAKER_00: what am I working on right now?
- 3:14:56SPEAKER_00: It's taking Math Academy from the
- 3:15:00workshop to a factory transition it's a huge team lift with me
- 3:15:04and jason and uh
- 3:15:06alex our director of curriculum and like just figuring out how
- 3:15:11to you know take the
- 3:15:13it's like it interfaces with learning with like um all the ai
- 3:15:18tooling everything
- 3:15:19how to how to deliver this this this learning experience and in
- 3:15:23particular this
- 3:15:24SPEAKER_00: this this content uh all the courses at a much
- 3:15:29SPEAKER_00: Higher level of scale,
- 3:15:30SPEAKER_00: how to turn out more courses like a factory style,
- 3:15:33SPEAKER_00: but not like,
- 3:15:34SPEAKER_00: you know,
- 3:15:35SPEAKER_00: not just like,
- 3:15:36SPEAKER_00: hey,
- 3:15:36SPEAKER_00: I'll build a course for me and then it's slop.
- 3:15:40SPEAKER_00: Like, how do you take the expert, take the human expertise and apply?
- 3:15:44these tools as multipliers to just speed up that level of
- 3:15:49expertise without
- 3:15:50dropping standards and we've made so much progress on this in
- 3:15:54the uh in the past
- 3:15:56few months and i'm really excited about this but this is really
- 3:15:59our kind of like
- 3:16:00you know the tesla elon sleeping on the factory floor kind of
- 3:16:04stuff or just like
- 3:16:05SPEAKER_00: the production hell of like oh my god this is
- 3:16:08SPEAKER_00: we got ourselves into this game and,
- 3:16:10SPEAKER_00: and now,
- 3:16:10SPEAKER_00: and now we have to like scale up and really deliver and just,
- 3:16:14SPEAKER_00: it's going to suck for the next,
- 3:16:16SPEAKER_00: however long it takes to make this factory work.
- 3:16:20SPEAKER_00: That's kind of where we're, we're kind of doing the equivalent.
- 3:16:23SPEAKER_00: That's, that's where we are in our journey right now.
- 3:16:25SPEAKER_00: So that's just, that's, that's what I'm doing.
- 3:16:28SPEAKER_00: Like every waking hour, basically.
- 3:16:31SPEAKER_01: Cool, man.
- 3:16:32SPEAKER_01: Best of luck.
- 3:16:33Honestly,
- 3:16:35I think there's nothing more satisfying than to build something
- 3:16:39truly at the edge.
- 3:16:41SPEAKER_01: It's not done before,
- 3:16:42SPEAKER_01: and then you're trying your best,
- 3:16:43and then you're using all of your skill set in order to get
- 3:16:46there and do the thing.
- 3:16:48um that's absolutely i always felt like this was the most
- 3:16:51exciting and this is also
- 3:16:53where i learned the most like i learned most about like how to
- 3:16:56do it faster do it
- 3:16:57better like uh uh and then just asked for the right help at the
- 3:17:01right time also um
- 3:17:02SPEAKER_01: yeah well fantastic thank you assistant
- 3:17:06SPEAKER_01: Thank you for having me.
- 3:17:07SPEAKER_00: It's my honor to be on.
- 3:17:10SPEAKER_00: This is great.
- 3:17:11SPEAKER_00: This is a lot of fun talking to you.
- 3:17:12And,
- 3:17:13you know,
- 3:17:13like I said earlier,
- 3:17:14it's only I came in with reflections today.
- 3:17:17And that only happens when you have a really good conversation
- 3:17:19with somebody.
- 3:17:20SPEAKER_00: It's so a testament to you as an interviewer here.
- 3:17:23SPEAKER_00: This is to us.
- 3:17:24Thank you for putting so much effort into this,
- 3:17:27coming up with great questions.
- 3:17:28SPEAKER_00: And, well, I also...
- 3:17:31SPEAKER_00: including so many people's questions that they sent to you.
- 3:17:35I never imagined that we would be addressing this many people's
- 3:17:38questions.
- 3:17:39SPEAKER_00: And it's awesome that we were able to.
- 3:17:40SPEAKER_01: I think we've got all the possible questions.
- 3:17:42SPEAKER_01: This will be well-packed.
- 3:17:43SPEAKER_01: We could reuse this stuff for the next 10 years.
- 3:17:47SPEAKER_01: We're going to be covered here.
- 3:17:52SPEAKER_01: And that's it for today, guys.
- 3:17:53SPEAKER_01: I hope that you enjoyed the video.
- 3:17:55SPEAKER_01: And don't forget to like and subscribe if you enjoyed this stuff.
- 3:17:59SPEAKER_01: And I wish you all a great week.