Latest / Logixly Live Unsupervised Learning Podcast / Is AI Making Us Less Skilled? The De-Skilling Debate

Transcript

Is AI Making Us Less Skilled? The De-Skilling Debate · Logixly Live Unsupervised Learning Podcast

• 0:00
  SPEAKER_04: AI is not it's not just taking jobs, man.
• 0:03
  SPEAKER_04: Oh, 100%.
• 0:04
  SPEAKER_04: 100%.
• 0:04
  SPEAKER_00: All right.
• 0:05
  SPEAKER_00: So let's jump into this.
• 0:06
  SPEAKER_00: So welcome everybody to Logic Slee Live, uh, where we try to hit the big stories in AI uh each week.
• 0:12
  SPEAKER_00: You got Katie Bakewell here again and and Matt Burseth.
• 0:15
  SPEAKER_00: And today we're kind of talking about something I I I don't think I realized it until you had brought it up.
• 0:21
  SPEAKER_00: You were showing um kind of the Google trend on the term de-skilling.
• 0:26
  SPEAKER_00: Yeah.
• 0:27
  SPEAKER_00: De-skilling.
• 0:28
  SPEAKER_00: So how I gotta be honest, I'm not sure I really was tracking de-skilling.
• 0:35
  SPEAKER_00: How would you describe de-skilling?
• 0:36
  SPEAKER_00: What is it?
• 0:36
  SPEAKER_04: So de-skilling is losing your abilities, right?
• 0:40
  SPEAKER_04: So you could theoretically de-skill by, I just, you know, we have like the summer slide for kids, right?
• 0:45
  SPEAKER_04: Like they learn to read, they take summer, they go to the water park every day, you know, they're having so much fun, they go back to school, and then all of a sudden they're not quite as good at reading as they used to be.
• 0:54
  SPEAKER_04: So that's sort of the traditional way we've looked at de-skilling.
• 0:57
  SPEAKER_04: You know, at employees have the same thing.
• 0:59
  SPEAKER_04: You might atrophy because you know, you can still be selling.
• 1:01
  SPEAKER_04: You do have it, you haven't done it for a while, so yeah, you you get a little rusty.
• 1:05
  SPEAKER_04: But then the question comes in about de-skilling from AI specifically.
• 1:09
  SPEAKER_04: So it looks like from the Google trends, it's up 60% in the last five years.
• 1:15
  SPEAKER_00: Searches for de-skilling.
• 1:16
  SPEAKER_04: Right.
• 1:17
  SPEAKER_04: And really the spike is just in the last year of people going, uh-huh.
• 1:22
  SPEAKER_04: There's this thing we might have to worry about.
• 1:23
  SPEAKER_04: And then you start, you know, digging into articles.
• 1:26
  SPEAKER_04: And the reason we found this in the first place, um, coworker uh Harsha, so thanks, Harsha, um, sent over an article that's really interesting about colonoscopies and people doing colonoscopies, losing, losing out on a skill.
• 1:39
  SPEAKER_04: And it's like, wait, hold on, I gotta dig into this a little bit.
• 1:43
  SPEAKER_04: And what's interesting, so many of the articles that are on out here on de-skilling, there's a lot about the theoretical problem.
• 1:51
  SPEAKER_04: Is this a problem?
• 1:52
  SPEAKER_04: How do we deal with this problem?
• 1:53
  SPEAKER_04: How are we gonna manage it?
• 1:54
  SPEAKER_01: Right.
• 1:55
  SPEAKER_04: And then there's a lot that are, you know, you see these studies over and over again of like kids that have chat GPT and learn from chat GPT, they don't do as good on a test.
• 2:08
  SPEAKER_04: And I don't really think, you know, it it's kind of de-skilling, I guess, but it's not, right?
• 2:12
  SPEAKER_04: Like that is did you learn in the right way versus like right?
• 2:15
  SPEAKER_04: Yeah, the colonoscopies is the first time I've seen an example where it's like, wow, people actually lost their abilities because of this.
• 2:24
  SPEAKER_00: Right.
• 2:24
  SPEAKER_00: So so the details on that, um, and I've got them right here, but you you feel free to correct me if I summarize wrong.
• 2:29
  SPEAKER_00: So they had this was a study that was done in 2025, and there were roughly 1,500 colonoscopies studied, and the um clinicians were already obviously experts at what they do, and they were um they were did the detections without AI, and then later uh did not were not able to use AI, and they were not able to find as many as they did the first time around.
• 3:02
  SPEAKER_00: Is that right?
• 3:03
  SPEAKER_04: So I think the way the way that I understood it is there's three months before and then three months after, and the thing that they were trying to detect wasn't something the AI detects.
• 3:11
  SPEAKER_03: Okay.
• 3:12
  SPEAKER_04: So the AI is detecting, you know, when you're when you're looking at colonoscopies, I guess there's different things you're looking for.
• 3:17
  SPEAKER_03: Okay.
• 3:18
  SPEAKER_04: And one of the particular types of uh, you know, polyps, I guess, you know, plaques, like different things that are happening in there.
• 3:26
  SPEAKER_04: There's some that are really easy for an AI to identify and others that are more complex.
• 3:30
  SPEAKER_03: Okay.
• 3:30
  SPEAKER_04: And so the AI is is taking these these issues, and so they looked at the same three months before, three months after of the ones.
• 3:38
  SPEAKER_04: Okay, and they detected less.
• 3:40
  SPEAKER_00: So this was the same.
• 3:41
  SPEAKER_00: So it's the same 1500, they they detected them and then they did their jobs with AI assisting them, and then they came back to these original.
• 3:49
  SPEAKER_00: But have to go look.
• 3:50
  SPEAKER_00: And they decide left.
• 3:52
  SPEAKER_00: Okay.
• 3:52
  SPEAKER_04: I think so.
• 3:52
  SPEAKER_00: But I'm interesting because it's like are I feel like this is the feeling.
• 3:55
  SPEAKER_00: I don't know if anybody else is feeling, but um, you know, I look at it like we're doing this, we're doing this um training here at work.
• 4:04
  SPEAKER_00: Okay.
• 4:05
  SPEAKER_00: And I, and you know, we need to read, we need to comprehend, we need to write summaries.
• 4:10
  SPEAKER_00: And I can't tell, but I feel like my retention on content that I am listening to or reading, I just feel like I am not retaining, I get the sense that I am not retaining or comprehending at the same level, not that I was ever all that high, but at the same level as I was previously.
• 4:31
  SPEAKER_04: I 100% agree.
• 4:32
  SPEAKER_04: I think what was the sneaky trick for a college professor?
• 4:36
  SPEAKER_04: It's like, yeah, you can have whatever you can fit on one page for a study guide.
• 4:39
  SPEAKER_01: Right.
• 4:40
  SPEAKER_04: And you mean you're gonna make me learn all of the stuff that I need to put on a one-page thing for a study guide?
• 4:45
  SPEAKER_04: It was like the trick to get you to learn, right?
• 4:47
  SPEAKER_04: Yeah.
• 4:47
  SPEAKER_04: But it's like if AI creates my study guide, yeah.
• 4:50
  SPEAKER_04: You know, I used to I used to make crazy study guides that are in like six-point font.
• 4:55
  SPEAKER_04: Um, and then, you know, people would be like, Can I use your study guide?
• 4:57
  SPEAKER_04: And I'm like, Yeah, you're not gonna get anything out of it though.
• 4:59
  SPEAKER_01: Right.
• 5:00
  SPEAKER_04: Because like there's what all the weird stuff that's going on in here, right?
• 5:03
  SPEAKER_04: Like on a piece of paper.
• 5:05
  SPEAKER_04: Yeah.
• 5:05
  SPEAKER_04: But I the reason that it benefited me isn't that I had the study guide with me, it's that I had you created it.
• 5:10
  SPEAKER_04: Yeah, I've learned.
• 5:11
  SPEAKER_03: Yeah.
• 5:12
  SPEAKER_04: And like we kind of don't do that anymore.
• 5:16
  SPEAKER_03: Right.
• 5:16
  SPEAKER_04: But you know, again, that's that's the concept of learning versus like this is a skill you're already good at.
• 5:22
  SPEAKER_00: Yeah.
• 5:23
  SPEAKER_00: And I I mean, I guess, and uh, you and I were talking about this before, like some of some of the other like you had posed me a pretty good question, which was tilling, yeah.
• 5:32
  SPEAKER_00: Right.
• 5:32
  SPEAKER_00: Or something where it's like like we have developed technology that clearly replaces this thing that we all used to do.
• 5:40
  SPEAKER_00: Tilling was your example.
• 5:41
  SPEAKER_04: Right, yeah, tilling a field, like getting oxen out, hooking them up, like yep.
• 5:45
  SPEAKER_00: Yeah.
• 5:46
  SPEAKER_00: And I think my my response was like, I I that one, like I'm good, like I I think I can in my mind I can easily go back to a mode and say, hey, if we need to relearn or if if there was value in us retilling fields by hand, we could go do that.
• 6:04
  SPEAKER_00: In and I think this is actually relevant to like programming, but but the colonoscopies as well, is I I I genuinely like it's like I do wonder if we are, you know, if we're all kind of atrophine, yeah, and we are really, you know, not just leveraging, but actually like truly dependent upon AI to do these like critical things.
• 6:32
  SPEAKER_00: I I fear that not only is there the risk of us kind of getting to a point of of like like getting to a point where we can't get back to what we could do at humans, whereas I think we could all go out and till fields, maybe not me.
• 6:47
  SPEAKER_03: I can't tell a field.
• 6:48
  SPEAKER_03: But some of us could absolutely not.
• 6:50
  SPEAKER_00: Some of us could, but like I would really worry that could we get back and to do the things we once were able to do.
• 6:58
  SPEAKER_00: And then if like that is truly a requirement to innovate, like like maybe we can do things more efficiently now, but like is in the long term in the 10, 20, 30 year view, like is that actually are we actually missing out on innovations because we're not we're not as expert experts at the things that we were once expert at?
• 7:20
  SPEAKER_00: I don't know.
• 7:20
  SPEAKER_04: So okay, I'm this is my my rebuttal here, yeah full on debate.
• 7:25
  SPEAKER_04: Um usually you and I agree on things, so like it's kind of fun.
• 7:28
  SPEAKER_04: Um I look at this at like it was identifying this one difficult thing, right?
• 7:34
  SPEAKER_04: Like there's stuff that's like okay, you know, you're these are pretty easy to identify, and like that's what AI was taking out.
• 7:40
  SPEAKER_03: Yeah.
• 7:41
  SPEAKER_04: And then, you know, when you're just left with the more difficult, like that's where the skill atrophied.
• 7:45
  SPEAKER_04: But then the question that I would have is when we're studying, right?
• 7:49
  SPEAKER_04: Like when we're teaching this topic, and it sucks for the people that already have that skill, and maybe that skill is gonna atrophy.
• 7:55
  SPEAKER_04: But if I am a person that is teaching people to do colonoscopies, which like I really hope doesn't happen because I sure don't know how to do a colonoscopy, but like, you know, if I'm that person teaching it, I'm gonna spend less time teaching the thing the AI is really good at.
• 8:10
  SPEAKER_04: And then I can spend more time focused on this complex situation that the AI can't pick up.
• 8:15
  SPEAKER_03: Right.
• 8:15
  SPEAKER_04: So can we, is it that we're de-skilling or that we need to adjust what skills we need?
• 8:22
  SPEAKER_04: Like, are the skills changing?
• 8:24
  SPEAKER_04: And you know, you've only got a certain amount of time.
• 8:27
  SPEAKER_04: I think you know, Jack and I looked this up the other day because of you know, we're weirdos and we look stuff up.
• 8:32
  SPEAKER_04: I think it's for like the most complex type of doctors, you need 27,000 hours of hands-on practice.
• 8:40
  SPEAKER_03: Wow.
• 8:40
  SPEAKER_04: Okay.
• 8:41
  SPEAKER_04: Yep.
• 8:41
  SPEAKER_04: And if you're looking for, you know, um I hope they're not spending all 27,000 on a colonoscopy, but maybe they do, right?
• 8:49
  SPEAKER_04: And so, like, let's say I've got all of that time, and 95% of the things that I find are these easy ones.
• 8:56
  SPEAKER_04: 95% of what I'm learning are these easy ones.
• 8:59
  SPEAKER_04: So I'm spending five percent of that time on these complex pieces.
• 9:02
  SPEAKER_04: I want to spend 50% of the time on the complex pieces.
• 9:06
  SPEAKER_04: You know, it's like when I used to tutor people in math, I'll make you do a standard deviation by hand once, I'll never make you do it again.
• 9:12
  SPEAKER_04: You know, you do need to know it.
• 9:13
  SPEAKER_04: So, you know, we teach that skill, but like we could spend more time learning the complex piece.
• 9:19
  SPEAKER_00: So, okay, so the the point there is leverage AI to do the easier stuff, yeah, and then humans focus on the harder stuff, and then that would lead hopefully to more innovation and more um, I guess, you know, combating my concern, more innovation in the long run.
• 9:37
  SPEAKER_04: Right.
• 9:38
  SPEAKER_00: Yeah, yeah.
• 9:38
  SPEAKER_00: Okay, so let me ask you this other one.
• 9:40
  SPEAKER_00: So then how about, and maybe this is a stretch for de-skilling, but for me, I I think it was probably the first thing that popped in my mind when we talked about this topic, which is um, you know, you and I were at a and you know, I'm gonna speak very high level of this scenario, but general terms.
• 9:57
  SPEAKER_00: You and I were at an event one time, and this was a event, it was it was sort of like a hackathon, not really, but there was a lot of presentations and a lot of uh kind of focused work in an area, and it was educational, like it should have been a skill building event.
• 10:18
  SPEAKER_00: Okay.
• 10:18
  SPEAKER_00: And I know what what I saw, and I don't know if it's a de-skilling or if it's like a regression to the mean.
• 10:26
  SPEAKER_00: It's like what I saw was two and a half hours of the same presentations solving the same problems in the same way.
• 10:37
  SPEAKER_00: And it was it was it was a hackathon format, so a lot, you know, a lot of content needed to be created in a short amount of time.
• 10:45
  SPEAKER_00: Right.
• 10:46
  SPEAKER_00: Guess what?
• 10:46
  SPEAKER_00: We have amazing content creation tools today, so that wasn't an issue.
• 10:50
  SPEAKER_00: But I I kind of feel like in the de-skilling, like the people that were in that hackathon, what they did this last year compared to what they did five years ago.
• 11:00
  SPEAKER_00: While the while the if you just took one of the presentations this year, it might be impressive.
• 11:06
  SPEAKER_00: Right.
• 11:07
  SPEAKER_00: If you watched all 10, you would have said, this is mediocrity.
• 11:10
  SPEAKER_00: Like this is literally everybody using generally the same prompts to have AI solve the same problems in the same way to create the same presentations.
• 11:21
  SPEAKER_00: Right.
• 11:21
  SPEAKER_00: And it it to me, I don't know if that's like a different form of de-skilling, but to me, it's like I don't know, that's like all that's like troubling in a way.
• 11:29
  SPEAKER_00: Like I feel like those students in the hackathon, they probably left thinking they did, they built a lot of stuff.
• 11:38
  SPEAKER_04: But it it's no, I'm 100% agree because it's not like that is not an isolated instant, right?
• 11:45
  SPEAKER_04: Like went up um to Georgia Tech because they had a hackathon.
• 11:50
  SPEAKER_03: Yeah.
• 11:50
  SPEAKER_04: And I remember being so incredibly impressed with the presentations and the tools that people had built that were clearly not AI.
• 11:58
  SPEAKER_03: Yeah.
• 11:58
  SPEAKER_04: You know, it's the same as like, I'm not impressed that AI wrote a book.
• 12:01
  SPEAKER_04: I'm impressed that a human wrote a book and like all of that stuff came out of your noodle.
• 12:04
  SPEAKER_04: Like that's incredible.
• 12:06
  SPEAKER_02: Right.
• 12:06
  SPEAKER_04: So, you know, what I'm it I think the the people that differentiated themselves in that hackathon were the ones that it didn't look the same.
• 12:14
  SPEAKER_04: Right.
• 12:15
  SPEAKER_04: Right.
• 12:15
  SPEAKER_04: So it is this thing where, you know, maybe those are skills that were were in that that ugly hiccup period of, you know, colomoscopies have gotten a little bit worse, but it's in service of them getting better.
• 12:29
  SPEAKER_04: Because the actual content of the presentations, like, man, they were great.
• 12:33
  SPEAKER_00: Right.
• 12:33
  SPEAKER_04: For for a yeah, I mean, like they're for where they were at.
• 12:36
  SPEAKER_00: Okay, so let me ask you this thing, because this is de-skilling.
• 12:38
  SPEAKER_00: So the the presentations were great.
• 12:41
  SPEAKER_00: And I'll I'll reference a different presentation you weren't at for another data point, but the presentations were great.
• 12:47
  SPEAKER_00: I'm not confident those students, I I think those were we were talking to Chat GPT, we weren't hearing what's in the students' heads.
• 12:57
  SPEAKER_04: Matt, you know me.
• 12:58
  SPEAKER_04: Like I will complain about it on a probably daily basis of if you cannot explain it, like I don't want to read your paper.
• 13:06
  SPEAKER_04: Like all you're doing is giving me homework.
• 13:08
  SPEAKER_04: Yep.
• 13:08
  SPEAKER_04: I do not want to do homework.
• 13:09
  SPEAKER_04: I didn't do it in school, I'm not doing it now.
• 13:11
  SPEAKER_04: Like, I know, yep, don't give me homework.
• 13:13
  SPEAKER_04: But if you can talk to it, yeah, that's what's gonna be super critical.
• 13:17
  SPEAKER_04: And it's where, you know, the thing that differentiates us is the ability to do those things.
• 13:24
  SPEAKER_04: And so the skills that we need are maybe a little bit different than they were before.
• 13:29
  SPEAKER_04: Yeah.
• 13:29
  SPEAKER_04: Right.
• 13:30
  SPEAKER_04: Like you have to be able to comprehend and you have to be able to talk differently about it, you know, how you need to get better at storytelling, right?
• 13:39
  SPEAKER_04: There's some skills that we really didn't have time to focus on in the past because you had to do all the data analysis and you had to do all the work.
• 13:46
  SPEAKER_02: Right.
• 13:47
  SPEAKER_04: If we go back to the the talk that we had about this data AI report, right?
• 13:50
  SPEAKER_04: Like it was these papers where AI did the same engineering.
• 13:54
  SPEAKER_04: Yeah.
• 13:54
  SPEAKER_04: Like that piece we don't really care about.
• 13:56
  SPEAKER_04: The engineering's fine, right?
• 13:57
  SPEAKER_04: But writing the paper, yeah, man, you gotta you need the human brain to do that.
• 14:01
  SPEAKER_03: Yeah.
• 14:02
  SPEAKER_04: So it's like the when you were, I mean, when you've done research projects for Compsci, right?
• 14:08
  SPEAKER_04: The hard part wasn't writing the paper.
• 14:10
  SPEAKER_04: The hard part was like, I've got to do all this engineering.
• 14:13
  SPEAKER_02: Yep.
• 14:13
  SPEAKER_04: And then when we look at AI, it's the complete opposite.
• 14:16
  SPEAKER_04: Yeah.
• 14:17
  SPEAKER_04: It gives you time to hone that skill even further of the part that was really just kind of an annoyance.
• 14:23
  SPEAKER_04: Why are you making me write a paper?
• 14:24
  SPEAKER_04: Well, now I can get a little bit better, spend a little bit more time, do better storytelling.
• 14:29
  SPEAKER_01: Right.
• 14:29
  SPEAKER_04: But it's sort of I instead of the de-skilling, I think it's a great reskilling, right?
• 14:34
  SPEAKER_04: Like we have to really figure out what are the things that matter, what are the things that differentiate us.
• 14:39
  SPEAKER_00: Yeah.
• 14:39
  SPEAKER_00: No, those are those are great points.
• 14:41
  SPEAKER_00: What what were some of the so I guess the you know, um I I like anecdotes.
• 14:46
  SPEAKER_00: So um my and and maybe it's uh not a great uh parenting technique, but um, but my and I I gotta say, you know, my daughter, my my middle daughter Abby, like she is um she's just like uh goal-oriented.
• 15:05
  SPEAKER_00: So her goal is not to get an A, her goal is to get a hundred percent to know everything on every exam, case in point last Friday.
• 15:13
  SPEAKER_00: So something I would have never done.
• 15:15
  SPEAKER_00: Is she already had an A for the quarter?
• 15:20
  SPEAKER_00: Okay.
• 15:20
  SPEAKER_04: Oh, so you sit back and do nothing, right?
• 15:22
  SPEAKER_00: So Friday we had family in town and everybody else skipped school.
• 15:27
  SPEAKER_00: Abby went to whatever her first period is to take the test to better her A.
• 15:36
  SPEAKER_04: I don't know who does that.
• 15:37
  SPEAKER_04: Okay, that's adorable then.
• 15:38
  SPEAKER_04: That's really sweet.
• 15:39
  SPEAKER_00: It's great.
• 15:39
  SPEAKER_00: But so that but I look at her and I laugh because she is uh very much into using Chat GPT to help her study.
• 15:48
  SPEAKER_00: And but and we share the same account, so I get to um see all of her very innocent chats, which are very funny to read through.
• 15:56
  SPEAKER_00: But she's very much like, uh, I don't understand it.
• 16:01
  SPEAKER_00: Can you explain it to me in a different way?
• 16:03
  SPEAKER_00: Uh here's some examples.
• 16:05
  SPEAKER_00: Can you give me more examples?
• 16:08
  SPEAKER_00: Why is this?
• 16:09
  SPEAKER_00: Like, I I watch her use it, and I and then I see a data point like the AI assisted students not doing as well.
• 16:17
  SPEAKER_00: Yeah, and I know there's lots of ways to use AI good and bad, and I'm not sure Abby's got, you know, I'm sure this is one of many ways to use it.
• 16:24
  SPEAKER_00: Maybe she's not even using it very well, but I but I'm like, I think this is super healthy because if it was my wife, right?
• 16:32
  SPEAKER_00: My wife does way better at this, but like, you know, we don't have the patience that ChatGPT has.
• 16:37
  SPEAKER_00: We don't understand the material, we can't come up with 500 more questions in two minutes.
• 16:42
  SPEAKER_00: Right.
• 16:43
  SPEAKER_00: Right.
• 16:43
  SPEAKER_00: And so I watch her and I'm I'm just like convinced that if you want to use the tools these right ways to get smarter, you can get really smart.
• 16:53
  SPEAKER_00: If you wanted to just like skate by it, you can skate by it.
• 16:56
  SPEAKER_00: Like, like even if we go back to the um to the colonoscopy example, like like I I think that you could use these tools and actually be better at at growing your own knowledge than than if we didn't have these tools available.
• 17:14
  SPEAKER_04: Right.
• 17:14
  SPEAKER_04: I mean, it's it's genuinely like with those tools for your your colonoscopies.
• 17:18
  SPEAKER_04: Man, I could kick back and relax and go, man, this thing's this thing's getting it and it's gonna do everything we need.
• 17:24
  SPEAKER_04: Or I can go, man, let me let me take a closer look at some other things because it's already taken care of that stuff.
• 17:31
  SPEAKER_04: It's let let chat GPT do those things, let the tool do those things.
• 17:35
  SPEAKER_04: Like, I love the way Abby's using it.
• 17:38
  SPEAKER_04: Like that's the goal.
• 17:39
  SPEAKER_04: Yeah, that's amazing.
• 17:40
  SPEAKER_00: Yeah.
• 17:41
  SPEAKER_04: Go Abby.
• 17:42
  SPEAKER_00: Yeah, yeah.
• 17:43
  SPEAKER_00: I mean, I I I won't pull any of her chats out, but but uh, you know, the other thing I I love watching her is like just how um you know jumps right into voice.
• 17:52
  SPEAKER_00: A lot of how she uses it is through voice.
• 17:54
  SPEAKER_00: Yeah, and I'm you know, I'm actually I'm really interested in this.
• 17:57
  SPEAKER_00: Like, I feel like offices today are configured for keyboard work, but the the speed that you can think and talk is so much different.
• 18:09
  SPEAKER_00: But we just, if we're all shouting at our, you know, into our microphones, it's probably not working.
• 18:14
  SPEAKER_00: But I'm also kind of sort of convinced that the future is not the keyboard.
• 18:18
  SPEAKER_00: If the keyboard was like a a stock you could buy, I would be shorting keyboards and I would be buying AirBuds or something, right?
• 18:25
  SPEAKER_04: Do you think the same?
• 18:26
  SPEAKER_04: 100%.
• 18:27
  SPEAKER_04: And I'm gonna use a quote from Jack, who's my partner in crime, sits next to me.
• 18:31
  SPEAKER_04: We talk all day long.
• 18:32
  SPEAKER_03: Yeah.
• 18:32
  SPEAKER_04: And maybe my favorite thing I've ever heard him say was if you as a developer are limited in productivity because of your keyboard throughput, you're kind of a terrible developer.
• 18:42
  SPEAKER_03: Yeah.
• 18:42
  SPEAKER_04: And it's like that's really the hard part about developing isn't typing faster.
• 18:47
  SPEAKER_03: Right.
• 18:48
  SPEAKER_04: Right?
• 18:48
  SPEAKER_04: This that is not the focus.
• 18:50
  SPEAKER_04: And if you can talk through it, yeah, you know, we don't type to a rubber ducky, we talk to a rubber ducky.
• 18:56
  SPEAKER_04: Yeah.
• 18:56
  SPEAKER_00: So, you know Wait, what is that?
• 18:57
  SPEAKER_00: I don't get the ref what is the rubber ducky.
• 19:00
  SPEAKER_04: Really?
• 19:00
  SPEAKER_04: I don't, no, no.
• 19:01
  SPEAKER_04: You've never heard so okay, for software developers, you've got your rubber ducky where you're stuck on a problem and you just need to talk it out.
• 19:07
  SPEAKER_02: I have heard that.
• 19:07
  SPEAKER_04: Yeah, you don't need to actually you just need to talk to somebody about it.
• 19:12
  SPEAKER_04: And like, yeah, you I've never seen somebody type to the rubber ducky, right?
• 19:16
  SPEAKER_04: You you speak because speaking out loud and thinking like that is helpful.
• 19:20
  SPEAKER_04: Yeah.
• 19:20
  SPEAKER_04: So yeah, we're gonna have to switch from I'm I'm short in keyboards with you, Matt.
• 19:24
  SPEAKER_04: I'll leave it there.
• 19:25
  SPEAKER_00: I'm just I'm and I'm just interested.
• 19:27
  SPEAKER_00: Obviously, we we always want to stay on top of things here, but like Fallon was showing me yesterday, you know, the dot that she configured, and she can call the dot.
• 19:35
  SPEAKER_00: And and I'm like, you know what?
• 19:37
  SPEAKER_00: That is it, I know it seems cheeky or cheesy or something, but it's like I do the same thing.
• 19:42
  SPEAKER_00: Like I, you know, if I if I take notes in a meeting or something, I will on the ride home I'll I'll narrate my notes because there's 15 other things that are in my mind that I remember that I didn't write down.
• 19:56
  SPEAKER_00: And I didn't write them down because we're already talking about something else.
• 20:00
  SPEAKER_00: Like the thought came and went, and even if it's on a keyboard, I can only get so much of that before the conversation has moved.
• 20:07
  SPEAKER_04: Right.
• 20:07
  SPEAKER_04: It's genuinely when you're thinking, like, I I say this to people all the time because we go and we do discovery on projects, right?
• 20:14
  SPEAKER_04: And it's like, hey, do you want to give me here's the 10 steps that I do?
• 20:19
  SPEAKER_04: And then I'm gonna say, cool, it's gonna cost this much to automate that.
• 20:22
  SPEAKER_04: Well, it's never gonna be right because the 10 steps are 10 of the 12 steps, but the two other things that you just do in your head because you're like, oh, obviously somebody would do that.
• 20:31
  SPEAKER_03: Yeah.
• 20:31
  SPEAKER_04: And they never get captured.
• 20:32
  SPEAKER_03: Yeah.
• 20:33
  SPEAKER_04: So it's like this thing where like when you speak, when you see, it's so much better.
• 20:38
  SPEAKER_04: Yeah.
• 20:38
  SPEAKER_04: 100% on board.
• 20:40
  SPEAKER_03: Yeah.
• 20:40
  SPEAKER_00: No, it's interesting stuff.
• 20:41
  SPEAKER_00: So, well, I mean, I I like the I feel like de-skilling.
• 20:44
  SPEAKER_00: I mean, I gotta be honest, I hadn't thought about it that deeply.
• 20:47
  SPEAKER_00: And I'm, you know, I know we had a little bit of back and forth.
• 20:50
  SPEAKER_00: I'm still um I think I'm still formulating my thoughts.
• 20:53
  SPEAKER_00: I mean, it's like it feels I gotta be honest, Katie, too.
• 20:56
  SPEAKER_00: I think the other part to me that feels like hollow is the fact that, you know, machine learning leverages data that we humans create for the most part.
• 21:09
  SPEAKER_00: Sure.
• 21:10
  SPEAKER_00: Um but it's like knowing that the models are built on human knowledge at least initially, and then moving to a point where we might not be able to create that original human knowledge.
• 21:25
  SPEAKER_00: I feel like that's the part that I gotta wrap my head around a little bit more.
• 21:29
  SPEAKER_04: For sure.
• 21:30
  SPEAKER_04: I'm right there with you.
• 21:31
  SPEAKER_04: Like it is it is that question of, you know, okay, better storytelling, better presenting for the the hackathon.
• 21:40
  SPEAKER_04: Sure, that's a thing.
• 21:42
  SPEAKER_04: But you know, it wasn't just that the presentations look the same.
• 21:45
  SPEAKER_04: It's that the ideas were the same.
• 21:46
  SPEAKER_04: Right.
• 21:46
  SPEAKER_04: Right.
• 21:47
  SPEAKER_04: And it's yes that hum, you know, to differentiate ourselves, we need that skill of just super critical thinking and super problem solving.
• 21:54
  SPEAKER_03: Yeah.
• 21:54
  SPEAKER_04: And then that storytelling as well.
• 21:56
  SPEAKER_04: So it's like we need to figure out what are the skills humans tomorrow need.
• 22:01
  SPEAKER_04: Right.
• 22:01
  SPEAKER_04: Because honestly, I don't think it's teaching people more AI, right?
• 22:04
  SPEAKER_00: Well, I think that's it, because uh the hackathon, I I feel like I feel like what was missed, the students were speedrunning a really great opportunity to maybe not win the hackathon, but maybe to get a lot of experience and a lot smarter.
• 22:18
  SPEAKER_04: For sure.
• 22:19
  SPEAKER_00: And it's like they sped run right past that to all produce basically the same generic output.
• 22:25
  SPEAKER_00: And it's like, yeah, that output by itself is actually better than any one team could have done last year.
• 22:31
  SPEAKER_03: Right.
• 22:32
  SPEAKER_00: But to get that good output, I just feel like the the students miss the learning in the in the long term, they gotta use AI, but man, they also need some original thinking or ways to like sharpen original thinking.
• 22:44
  SPEAKER_04: Right.
• 22:45
  SPEAKER_04: It's just we've got to figure out what those skills are and we gotta give the kids the skills.
• 22:48
  SPEAKER_04: And I'll I'll stand by it.
• 22:49
  SPEAKER_04: I don't think everybody's going, how do we teach kids AI?
• 22:52
  SPEAKER_04: Kids are gonna figure out AI.
• 22:54
  SPEAKER_04: Oh, like Abby's figured out how to use Chat GPT in a great way.
• 22:57
  SPEAKER_04: Yes.
• 22:57
  SPEAKER_04: We don't need to teach kids to use AI, we do need to teach kids to think, to storytell, like all of those sort of weird soft skills, harder to learn things.
• 23:05
  SPEAKER_04: Yeah.
• 23:06
  SPEAKER_04: That's what we've got to be focusing on.
• 23:07
  SPEAKER_04: Super cool.
• 23:08
  SPEAKER_00: 100%.
• 23:09
  SPEAKER_00: I agree totally.
• 23:10
  SPEAKER_00: Well, thanks everybody for joining us for another edition of Logic Slee Live.
• 23:14
  SPEAKER_00: Anything that uh we missed, feel free to reach out and you know, leave a comment or um reach out to Katie or I or Wendy or Sarah.
• 23:23
  SPEAKER_00: We are happy to get back with you and and um talk one off or bring a topic into the next uh the next show.
• 23:29
  SPEAKER_00: But until next time, keep learning, even when it's unsupervised.