Latest / Core Memory / The History And Future Of Brain Implants - EP 31 Sumner Norman

Transcript

The History And Future Of Brain Implants - EP 31 Sumner Norman · Core Memory

0:00
SPEAKER_01: It's kind of cool, right?
0:01
SPEAKER_01: I got it.
0:01
SPEAKER_01: I can just do it myself.
0:02
SPEAKER_01: Usually there's like four people that do all this stuff.
0:05
SPEAKER_01: Let's pretend that there is, though.
0:06
SPEAKER_03: Just be like, hey, John, Bob, you guys ready?
0:09
SPEAKER_03: Guys.
0:11
SPEAKER_03: We're rolling.
0:11
SPEAKER_03: Be quiet.
0:13
SPEAKER_03: Be quiet.
0:13
SPEAKER_03: Keep the cat down.
0:15
Be quiet.
0:17
SPEAKER_01: If you're ready, do you want to chill or are you good?
0:22
SPEAKER_03: I don't know what I would do in the chill time.
0:24
SPEAKER_01: We are just talking, so it's not overly complicated otherwise.
0:32
SPEAKER_01: Okay, well then I will introduce you.
0:35
SPEAKER_02: Okay.
0:52
SPEAKER_01: Sumner Norman, you're a co-founder and CEO of Forrest Neuro.
0:58
SPEAKER_01: More than that,
0:59
SPEAKER_01: you are,
1:00
SPEAKER_01: you're someone I met,
1:01
SPEAKER_01: I feel like,
1:02
SPEAKER_01: I don't know,
1:03
SPEAKER_01: three,
1:03
SPEAKER_01: four years ago now?
1:07
It's either six months or six years,
1:08
but either way,
1:09
the time dilation is real.
1:11
SPEAKER_01: One of those.
1:14
SPEAKER_01: We got introduced,
1:14
SPEAKER_01: I suspect,
1:15
SPEAKER_01: through Adam Marblestone and James Fickle or some brain network,
1:23
SPEAKER_01: and then you're someone I just ended up discussing.
1:26
You've been immensely helpful in so many things I do in the
1:30
BCI,
1:30
brain-computer interface space.
1:32
SPEAKER_03: That's all any of us hope to be is helpful.
1:35
SPEAKER_03: Um, I think I just like nerding out with you.
1:37
SPEAKER_03: We did it, we did some good nerding out.
1:38
SPEAKER_01: We've done good nerding out,
1:39
SPEAKER_01: but you're,
1:40
SPEAKER_01: you're,
1:40
SPEAKER_01: I always feel like you're,
1:42
SPEAKER_01: um,
1:43
SPEAKER_01: there's so many things going on in this brain implant,
1:45
SPEAKER_01: brain computer interface world.
1:46
SPEAKER_01: I always feel like you have a balanced take on things.
1:49
SPEAKER_01: And so it's been helpful to me to calibrate what is going on.
1:53
SPEAKER_01: And then, so we'll get into Forrest Neuro in a second.
1:56
SPEAKER_01: And, um, but I always like to, uh,
1:59
SPEAKER_01: let people get to know folks a little bit, you know?
2:03
SPEAKER_01: And so I was gonna dive in a little biographical journey for a second.
2:11
One thing,
2:11
when I was...
2:12
I hadn't,
2:13
like,
2:13
properly researched you in a while.
2:15
SPEAKER_01: I'm sure I did the first time I was coming down to L.A.
2:19
SPEAKER_01: to interview you.
2:19
SPEAKER_01: But then, yeah, there's something...
2:22
SPEAKER_01: I don't know,
2:22
SPEAKER_01: I just had not filed this away properly,
2:24
SPEAKER_01: but you're like a mechanical engineer by training,
2:28
SPEAKER_01: as opposed to neuroscience,
2:29
SPEAKER_01: and then you seem to know,
2:30
SPEAKER_01: like,
2:30
SPEAKER_01: everything about neuroscience as well.
2:34
SPEAKER_03: Yeah, I don't know everything about neuroscience.
2:37
And luckily for us,
2:38
actually,
2:39
the neuroscientists also don't know everything about
2:41
neuroscience.
2:42
SPEAKER_03: That's how I get away with it.
2:44
SPEAKER_03: But yeah,
2:44
I was actually trained as a mechanical and aerospace engineer
2:47
and some background
2:48
SPEAKER_03: in physics as well.
2:49
SPEAKER_01: And you got all the way up to a PhD in mechanical engineering.
2:53
SPEAKER_01: Yep, yep, yep.
2:54
SPEAKER_03: So yeah,
2:55
SPEAKER_03: undergrad in physics and mechanical specifically,
2:57
and then my master's and PhD were in mechanical and aerospace
3:00
by division,
3:01
SPEAKER_03: like the actual department of the school.
3:03
SPEAKER_03: But my focus is in control theory and brain-computer interfaces.
3:08
SPEAKER_03: And where that really marries is the control of exoskeleton robotics with
3:13
SPEAKER_03: brain-computer interfaces for therapeutic use.
3:16
SPEAKER_01: When you were either high school years or the beginning of college years,
3:22
SPEAKER_01: did you know you were gonna brain stuff,
3:24
SPEAKER_01: or you thought you would be making rockets then?
3:26
SPEAKER_03: I thought I'd be making rockets.
3:28
SPEAKER_03: Um, yeah, I think when... So, just like, I'll back way up for you.
3:32
SPEAKER_03: Yeah, yeah, let's do it.
3:33
I grew up in a very,
3:34
very small town in Utah,
3:35
something like 5,000 people.
3:36
It was like Mormons,
3:37
cows,
3:38
and then this like one nerdy brown kid that didn't fit in.
3:42
SPEAKER_03: And so I was the first person in my family to go to college.
3:46
SPEAKER_03: I was trying to find my bearings, and one thing I knew I loved was math.
3:51
SPEAKER_03: So I was just a huge math nerd.
3:53
I knew I wanted to be a mechanical engineer because we went on
3:56
a family trip to the
3:56
SPEAKER_03: Boeing factory when I was 11 years old.
3:59
SPEAKER_03: My dad was a pilot.
4:00
SPEAKER_03: And so,
4:02
SPEAKER_03: um,
4:02
SPEAKER_03: I met a mechanical engineer,
4:04
SPEAKER_03: and I remember this,
4:04
SPEAKER_03: and I was like,
4:05
SPEAKER_03: that guy's,
4:05
SPEAKER_03: like,
4:06
SPEAKER_03: got powers,
4:07
SPEAKER_03: because there's something on his computer screen,
4:08
SPEAKER_03: and on his desk is the thing he designed.
4:10
SPEAKER_03: And these days, it was, like, green lines on black screen, you know?
4:13
So I had my mom write it down,
4:15
mechanical engineer,
4:15
and I was like,
4:16
that's what I'm gonna do.
4:17
SPEAKER_01: Was that a professional pilot or a hobbyist pilot?
4:21
SPEAKER_03: Actually, it went in both directions.
4:23
SPEAKER_03: So he's professional,
4:24
SPEAKER_03: but like really,
4:25
SPEAKER_03: really small aircraft because in Utah,
4:28
like farm boys used to fly airplanes for delivery and crops and
4:33
stuff like that.
4:34
SPEAKER_03: And then when he got older and that became like less common,
4:38
SPEAKER_03: then he flew private just like himself for a while.
4:41
But then eventually even that became kind of costed out for
4:44
many people,
4:44
including us.
4:45
SPEAKER_01: So he's from Utah and grew up in the same sort of setting.
4:49
SPEAKER_03: Yeah.
4:49
SPEAKER_03: He grew up in Spanish Fork, Utah, which is a really small town.
4:53
SPEAKER_01: Okay.
4:53
SPEAKER_01: He's just flying out in the country.
4:57
SPEAKER_01: Yeah, out in the country.
4:59
SPEAKER_01: And did your mom work or stay at home?
5:02
SPEAKER_03: My mom was a bank teller.
5:05
SPEAKER_03: My dad was a small businessman and my mom was a bank teller.
5:10
SPEAKER_03: And the one thing I definitely got was math.
5:13
SPEAKER_03: And I still love math.
5:15
SPEAKER_01: It sounds like your dad was,
5:17
SPEAKER_01: if he's taking you to the Boeing factory and stuff,
5:19
SPEAKER_01: I mean,
5:19
SPEAKER_01: I know he's in the aviation,
5:21
SPEAKER_01: but it must have been inclined that way a little bit.
5:24
SPEAKER_03: Yeah, so I wanted to go that direction.
5:26
SPEAKER_03: I was like, I want to design stuff that goes fast.
5:28
I'm a huge car nerd,
5:31
love old cars,
5:32
and I thought that was the direction I wanted to go.
5:35
SPEAKER_03: It didn't take very long for that illusion to be broken.
5:39
SPEAKER_03: Because as I actually got into the math of it in undergrad,
5:42
SPEAKER_03: it was very much like,
5:43
SPEAKER_03: well,
5:44
SPEAKER_03: this guy came up with this equation,
5:45
SPEAKER_03: and this woman came up with that equation,
5:47
SPEAKER_03: and now,
5:48
SPEAKER_03: but we'll teach you all that theory,
5:49
SPEAKER_03: but in the end,
5:49
SPEAKER_03: you're just going to go to the lookup table in the back of the textbook,
5:52
SPEAKER_03: because that's been solved.
5:53
SPEAKER_03: You don't need to redo all that work.
5:55
SPEAKER_03: Like,
5:55
SPEAKER_03: NASA did that in the 70s,
5:56
SPEAKER_03: so just go look up there,
5:57
SPEAKER_03: airfoil designs,
5:57
SPEAKER_03: they're categorized very well.
6:00
And I was just like,
6:00
man,
6:01
where...
6:01
All the creativity has been,
6:03
like,
6:03
sucked out of this.
6:04
SPEAKER_03: Like,
6:04
SPEAKER_03: there's,
6:04
SPEAKER_03: of course,
6:05
SPEAKER_03: now,
6:05
SPEAKER_03: like,
6:05
SPEAKER_03: a resurgence,
6:06
SPEAKER_03: but that didn't feel like it was there for me.
6:08
SPEAKER_03: So I was... And I don't think I did this in a logical way.
6:12
SPEAKER_03: Just intuitively started pushing into,
6:14
SPEAKER_03: like,
6:15
SPEAKER_03: well,
6:15
SPEAKER_03: where are the corners of this space that there is a lot of creativity?
6:19
SPEAKER_03: And robotics was the big one.
6:21
SPEAKER_03: And control theory was kind of the area of that that just like,
6:24
SPEAKER_03: for me,
6:24
SPEAKER_03: my math brain just like hooked onto that immediately.
6:26
SPEAKER_03: Like, how do you control these machines?
6:28
SPEAKER_03: And why is it so hard to get a robot that is,
6:31
SPEAKER_03: you know,
6:32
SPEAKER_03: more powerful in many ways than a human and stronger and greater,
6:35
SPEAKER_03: whatever,
6:35
SPEAKER_03: but...
6:36
we just can't get these things to act naturalistically or
6:39
interact with humans in
6:40
SPEAKER_03: an interesting way.
6:41
SPEAKER_03: That really grabbed my brain.
6:43
SPEAKER_03: So I joined a haptics lab, actually.
6:45
SPEAKER_03: Like in undergrad?
6:46
SPEAKER_01: In undergrad, yeah.
6:47
SPEAKER_01: Okay, and you did undergrad in, was it LA or University of Utah?
6:51
SPEAKER_01: University of Utah, yeah.
6:52
But actually,
6:52
my first research experience,
6:54
I didn't know even how PhDs worked.
6:57
SPEAKER_03: to be honest with you, or how labs work.
6:59
SPEAKER_03: But credit to them,
7:01
there was a paid research experience for undergraduate thing at
7:05
Brigham Young
7:05
SPEAKER_03: University in Frogo.
7:07
I had to go through the Mormon screening process as a
7:10
non-Mormon,
7:10
which is a whole thing.
7:12
um but that was the first lab i joined uh and that was a
7:15
theoretical physics lab so
7:16
i was doing x-ray crystallography and that was really where i
7:19
learned a program and
7:20
then i was like that's just too theoretical and then the like
7:23
mechanical stuff was
7:24
too solved and then i found my middle ground and that was a
7:26
robotics haptics lab at
7:27
SPEAKER_03: the university of utah okay
7:29
SPEAKER_01: Not to,
7:29
SPEAKER_01: like,
7:29
SPEAKER_01: totally sidetrack us,
7:30
SPEAKER_01: but what are they screening for in the non-Mormon screening program?
7:36
SPEAKER_03: Um, they just, like, wanna make sure that you're not a menace.
7:40
SPEAKER_03: You know,
7:40
SPEAKER_03: you can't just be,
7:41
SPEAKER_03: like,
7:42
SPEAKER_03: doing drugs and fornicating with all the students that they've,
7:45
SPEAKER_03: like,
7:45
SPEAKER_03: very carefully curated to be,
7:47
SPEAKER_03: like,
7:47
SPEAKER_03: model Mormons.
7:49
SPEAKER_01: Okay, and were your parents Mormon?
7:51
SPEAKER_01: Uh, no, huh?
7:52
SPEAKER_01: No?
7:52
SPEAKER_01: Oh, so you just, you guys were just out there.
7:54
SPEAKER_03: Yeah, we were just out there doing our thing, just trying to get along.
7:57
SPEAKER_03: Okay.
7:59
SPEAKER_01: All right, what do we do at Core Memory?
8:00
SPEAKER_01: We cover innovative,
8:02
SPEAKER_01: fast-moving,
8:03
SPEAKER_01: forward-thinking companies,
8:05
SPEAKER_01: which is why Core Memory is sponsored by Brex,
8:09
because Brex is the intelligent finance platform for many of
8:14
these companies,
8:14
SPEAKER_01: 30,000 companies from startups to the world's largest corporations.
8:20
SPEAKER_01: Rely on Brex's technology for their finances.
8:22
SPEAKER_01: They've got smart corporate cards,
8:25
high yield business banking and expense automation tools that
8:30
are fantastic.
8:31
I hate doing my expenses and Brex's AIs and software run right
8:36
through those
8:36
SPEAKER_01: expenses,
8:37
figure out where we're spending money and take care of so much
8:40
stuff for you so you
8:41
SPEAKER_01: don't have to waste your time on it yourself.
8:45
Go to brex.com slash core memory to learn more and just,
8:49
you know,
8:50
get with the program.
8:50
SPEAKER_01: Let's get going.
8:51
Let's get out of this archaic finance software and move toward
8:56
the future.
8:56
SPEAKER_01: Core memory and Brex.
8:59
SPEAKER_01: As you're going through this journey,
9:01
when you start gravitating towards brains and neuroscience and
9:06
adding that in,
9:08
SPEAKER_03: Yeah,
9:08
SPEAKER_03: that really kind of came towards the end of undergrad,
9:11
SPEAKER_03: where I was working again on these kind of haptic robotic interfaces.
9:14
SPEAKER_03: So I was thinking about teleoperated surgical robotics.
9:18
SPEAKER_03: And we had come up with these...
9:19
SPEAKER_03: The lab there had come up with these ways of stretching your skin around.
9:23
It turns out that your fingertips,
9:26
especially your index finger and thumb,
9:27
are so sensitive.
9:29
So one of the papers I wrote was about how you could actually
9:31
discriminate 16
9:32
SPEAKER_03: different types of tactile stretch cues,
9:35
SPEAKER_03: which is like pulling your finger around.
9:36
SPEAKER_03: This is actually really cool for surgical robotics,
9:39
SPEAKER_03: where you don't want to have force feedback,
9:40
SPEAKER_03: because you can imagine a doctor is working on a patient,
9:44
SPEAKER_03: carefully cutting with the scalpel,
9:46
SPEAKER_03: and then they,
9:46
SPEAKER_03: like,
9:47
SPEAKER_03: slip off the robot a little bit,
9:48
SPEAKER_03: and the force feedback kicks,
9:49
SPEAKER_03: and now the scalpel's flying.
9:50
SPEAKER_03: So, you can't have force feedback.
9:52
SPEAKER_03: So,
9:52
SPEAKER_03: we were developing this way of actually,
9:53
SPEAKER_03: like,
9:55
SPEAKER_03: giving you tactile sensation,
9:56
SPEAKER_03: but with forces that are minuscule.
9:59
Wait,
9:59
on the stretching the skin part,
10:00
you would be feeling that level of detail?
10:05
SPEAKER_03: Mm-hmm.
10:06
SPEAKER_01: Okay.
10:07
SPEAKER_03: And so,
10:07
SPEAKER_03: yeah,
10:08
SPEAKER_03: imagine like,
10:08
you know those little red trackpad things that are in the
10:11
middle of the IBM
10:12
SPEAKER_02: laptops?
10:12
SPEAKER_02: Yeah, yeah, yeah.
10:13
SPEAKER_03: It's like that in reverse.
10:14
So instead of you like pushing the red thing around,
10:16
it's pushing you around.
10:17
But these tiny little forces can tell you a great deal about
10:20
what you're doing with it.
10:21
Oh,
10:21
I see what you mean,
10:22
but you could feel it in your fingertips for the surgeon.
10:25
Okay,
10:25
so there's so much sensitivity there that it gives you this
10:28
finer control.
10:30
SPEAKER_01: Okay,
10:31
SPEAKER_01: so you're starting at this from this robotics lens and digging in,
10:35
SPEAKER_01: and then you have to start at some point.
10:39
SPEAKER_01: learning?
10:40
SPEAKER_01: Neuroscience is part of this?
10:43
What happened was just totally serendipity that I built this
10:46
experimental device.
10:47
It was a psychophysics experiment to understand how people used
10:51
the sense of touch
10:52
SPEAKER_03: in robotic environments.
10:55
SPEAKER_03: Someone mentioned me from another lab,
10:56
SPEAKER_03: like,
10:57
SPEAKER_03: hey,
10:58
SPEAKER_03: that device in and of itself could actually be really useful for stroke
11:01
SPEAKER_03: rehabilitation.
11:03
SPEAKER_03: because it's providing really good feedback that's real-time,
11:07
SPEAKER_03: it requires proprioception and tactile sensation,
11:10
SPEAKER_03: and you can teach people to move again in high-repetition movements.
11:14
SPEAKER_03: It's a low-cost device.
11:15
Maybe you should write a paper for this special IEEE
11:18
transactions journal on rehab.
11:22
SPEAKER_03: And so I said, okay.
11:23
SPEAKER_03: I just thought it was a great path to publication.
11:24
SPEAKER_03: And then I got my mind blown as I'm writing this thing,
11:27
SPEAKER_03: because now I'm reading all these publications about stroke rehab,
11:30
SPEAKER_03: and I'm going,
11:31
SPEAKER_03: hey,
11:31
SPEAKER_03: there's a control system of a robotic actuator of some type.
11:35
SPEAKER_03: There's a control system of the human being.
11:37
SPEAKER_03: And these interfacing in and of itself is really interesting.
11:40
SPEAKER_03: But then when you damage one, but it has plasticity, right?
11:44
SPEAKER_03: The brain can change and learn and adapt.
11:47
And now you can use the robot to assist and teach and protect
11:51
that other control system.
11:53
It just created this more intimate intersection of human and
11:58
machine interaction
11:59
SPEAKER_03: that's really what drives me deep down.
12:02
SPEAKER_03: So that was where I really hooked on to it.
12:05
SPEAKER_03: And so as I considered my options and thought about grad school,
12:09
SPEAKER_03: I knew I wanted to do something where it was like a...
12:12
SPEAKER_03: the most intimate form of human-machine communication,
12:14
SPEAKER_03: and that's how I found my way to brain-computer interfaces eventually.
12:17
SPEAKER_01: And then what do you have to,
12:19
SPEAKER_01: like,
12:19
SPEAKER_01: I would feel,
12:19
SPEAKER_01: if I was doing this,
12:21
SPEAKER_01: like,
12:21
SPEAKER_01: I was so far behind on the neuroscience stuff,
12:24
SPEAKER_01: and,
12:24
SPEAKER_01: I mean,
12:25
SPEAKER_01: to your point earlier,
12:26
SPEAKER_01: I've interviewed enough neuroscientists to know,
12:29
SPEAKER_01: like,
12:30
SPEAKER_01: there is this degree to which we still don't know that much,
12:33
SPEAKER_01: but you would have to go through the whole...
12:36
SPEAKER_01: corpus of key texts, I assume.
12:41
SPEAKER_01: I mean, did you feel like you had to catch up and then start?
12:45
SPEAKER_03: Oh, yeah, for sure.
12:46
SPEAKER_03: I felt very behind in the beginning.
12:49
SPEAKER_03: But I think there were a couple things that helped.
12:51
One is I just had an excellent PhD advisor that was patient and
12:55
thoughtful and
12:56
SPEAKER_03: pointed me to all the right resources,
12:59
SPEAKER_03: and was like,
12:59
SPEAKER_03: if you read these,
13:00
SPEAKER_03: you'll start to understand.
13:03
And I do think that a PhD really,
13:06
what you actually learn in a PhD is how to teach yourself.
13:09
SPEAKER_03: And so he was very good at teaching people to teach themselves,
13:12
SPEAKER_03: hands-on without being overbearing by any stretch of the imagination.
13:16
SPEAKER_03: And it was just a lot of reading papers.
13:21
SPEAKER_03: There's no secret to it.
13:21
SPEAKER_03: It's just like lots and lots and lots of exposure.
13:23
SPEAKER_03: A few classes, but honestly, not that many.
13:26
SPEAKER_03: A lot of it was really just exposure to the literature itself.
13:29
SPEAKER_01: Alright,
13:29
SPEAKER_01: this may be an unfair question,
13:30
SPEAKER_01: but if I,
13:31
as a journalist or somebody who just wants to start getting
13:36
into this stuff,
13:37
SPEAKER_01: I mean,
13:37
SPEAKER_01: are there two or three key texts that get you the farthest?
13:41
SPEAKER_03: Well,
13:41
SPEAKER_03: I mean,
13:42
SPEAKER_03: there's the Bible,
13:43
SPEAKER_03: Eric Kandel's Principles of Neuroscience,
13:45
SPEAKER_03: which I admittedly have not read,
13:47
SPEAKER_03: but I've read parts of it.
13:50
And I think that like that's gonna tell you about neurobiology
13:53
all the way from
13:53
SPEAKER_03: kind of the atomic level up to like what systems neuroscience is.
13:59
SPEAKER_03: I don't have like a magic bullet here.
14:03
SPEAKER_03: I'll think about it.
14:03
SPEAKER_03: Can we, like, put it in the pod notes?
14:05
SPEAKER_03: If I've got, like, a Neuroscience for Dummies?
14:07
SPEAKER_03: Maybe that's it.
14:08
SPEAKER_03: Is it Neuroscience for Dummies?
14:09
SPEAKER_01: We will do this.
14:10
SPEAKER_01: I'm sure that probably exists.
14:11
SPEAKER_01: It probably does, right?
14:13
SPEAKER_01: No, I just, you know, I think you answered the question even.
14:16
SPEAKER_01: Just,
14:16
SPEAKER_01: there's always one,
14:17
it feels like there's always one seminal text that seems to get
14:21
people pretty far,
14:22
SPEAKER_01: so...
14:23
SPEAKER_03: Yeah, I think I got lucky.
14:24
SPEAKER_03: I had a great mentor.
14:26
SPEAKER_03: And then,
14:26
SPEAKER_03: you know,
14:26
SPEAKER_03: when I joined Caltech later on,
14:29
SPEAKER_03: the group that I joined was kind of led by a legendary neuroscientist.
14:33
SPEAKER_03: I was actually just scared of that.
14:36
SPEAKER_03: He was like, oh, don't worry, actually.
14:38
And I asked during my job talk,
14:39
like,
14:39
hey,
14:40
is this something I should be like brushing up on or
14:42
SPEAKER_03: He's like,
14:43
SPEAKER_03: don't worry,
14:44
I generally find that hiring engineers and teaching them
14:46
neuroscience works really
14:47
SPEAKER_03: well.
14:48
SPEAKER_03: He said,
14:48
SPEAKER_03: I have a very hard time hiring neuroscientists and trying to teach them
14:51
SPEAKER_03: engineering,
14:52
SPEAKER_03: and you need both for neurotech.
14:54
SPEAKER_03: And so that was my entry into that space.
14:56
SPEAKER_01: Was that Mikhail Shapiro?
14:58
SPEAKER_03: That was Richard Anderson that said that.
15:00
There's many,
15:01
many,
15:02
many different approaches to bringing computer interfaces at
15:05
the moment.
15:05
I would kind of categorize,
15:07
I think,
15:08
Neuralink as maybe the most aggressive they're doing.
15:12
SPEAKER_01: They have an implant with tons of electrodes on it.
15:15
SPEAKER_01: They're doing a full-on surgery, going into the brain.
15:18
There's other people that want to sit right above,
15:22
like,
15:22
sort of the first layers of the brain.
15:24
There's even people that I just interviewed this week who are,
15:28
you kind of dig a little,
15:30
SPEAKER_01: groove in the skull and you're just sitting there.
15:33
SPEAKER_01: And then there's people going in blood vessels like Synchron.
15:36
And so there's all these different approaches,
15:38
some more invasive than others.
15:40
And then a lot of them are doing this electrodes into the motor
15:46
cortex and things like that.
15:47
SPEAKER_01: When we get to Forrest, you know, you guys are doing
15:51
SPEAKER_01: um, this different approach with ultrasound, with an implant.
15:56
SPEAKER_01: But before we got there,
15:57
SPEAKER_01: I just wanted to give people some context on,
15:59
SPEAKER_01: like,
15:59
SPEAKER_01: where we're heading.
16:00
SPEAKER_01: And you're kind of looking to treat,
16:03
SPEAKER_01: um,
16:04
to understand and treat some different conditions than a
16:07
company like Neuralink is
16:09
SPEAKER_01: with movement through the motor cortex.
16:12
SPEAKER_01: But one of the things that I've always enjoyed talking with you about is,
16:15
SPEAKER_01: like,
16:15
SPEAKER_01: the history of how we got to this point.
16:19
SPEAKER_01: Um, it can feel...
16:21
just because there's been so much startup capital and
16:23
everything that's gone in
16:24
SPEAKER_01: this field,
16:25
SPEAKER_01: all of a sudden,
16:25
SPEAKER_01: it sort of feels new.
16:27
SPEAKER_01: A lot of the work that's being done is piggybacking off this Utah array,
16:31
SPEAKER_01: which is kind of like 20,
16:33
SPEAKER_01: 25 years in the making.
16:34
But we go,
16:35
it actually,
16:36
you know,
16:36
people messing with the brain goes back even farther.
16:39
Yeah,
16:40
I mean,
16:40
what were,
16:41
like,
16:41
when did people start poking at brains with electrodes?
16:46
SPEAKER_03: Well, I mean, yeah, with electrodes, I mean, it's well over 100 years.
16:50
SPEAKER_03: I know the EEG just had its 100-year birthday.
16:53
SPEAKER_03: So electroencephalography,
16:54
SPEAKER_03: or non-invasive recording of electrical activity,
16:57
SPEAKER_03: which people often talk about as brainwaves,
16:59
SPEAKER_03: right?
17:01
SPEAKER_03: It was Hans Berger in 1924.
17:03
SPEAKER_03: So we're just over 100 years there.
17:05
SPEAKER_03: And I think that measuring extracellular electrical activity potentially
17:12
SPEAKER_03: Oh,
17:13
SPEAKER_03: jeez,
17:13
SPEAKER_03: that was the large squid axon,
17:15
SPEAKER_03: or giant squid axon,
17:16
SPEAKER_03: and I forget the exact year,
17:17
SPEAKER_03: but it wasn't too long after that,
17:19
SPEAKER_03: 40s,
17:19
SPEAKER_03: 50s,
17:19
SPEAKER_03: somewhere in there.
17:19
SPEAKER_03: So, yeah.
17:21
SPEAKER_01: Like, what happens?
17:22
SPEAKER_01: People just realize the brain,
17:25
SPEAKER_01: like the body,
17:26
SPEAKER_01: has electricity running through it,
17:28
SPEAKER_01: and they start to get interested in...
17:30
SPEAKER_03: That goes all the way back to Luigi Galvani.
17:32
SPEAKER_03: So that's like the 1700s.
17:34
It was when I think that he stimulated some dead frog legs and
17:38
they twitch,
17:39
SPEAKER_03: which now sounds obvious.
17:40
But at the time,
17:41
we didn't understand how movement was generated in the body.
17:43
SPEAKER_03: And so now you have, okay, these muscles are twitching.
17:46
SPEAKER_03: What's going on here?
17:48
SPEAKER_03: What is it about electricity and musculature that's connected?
17:51
SPEAKER_03: And so that's where we start to understand the peripheral nervous system.
17:54
SPEAKER_03: And then over the past hundred years,
17:55
we start exploring electrical activity in the nervous system in
17:58
a real way.
17:59
SPEAKER_03: Or sorry, in the central nervous system and in the brain.
18:01
And so that's kind of where our foray into brains starts,
18:06
really,
18:06
neurotech proper.
18:07
SPEAKER_01: And in the 1920s, they're doing what?
18:09
SPEAKER_01: They're doing...
18:11
SPEAKER_03: Yeah,
18:11
SPEAKER_03: so there,
18:12
SPEAKER_03: they're just measuring the activity and saying,
18:14
SPEAKER_03: like,
18:14
SPEAKER_03: you know,
18:15
SPEAKER_03: look,
18:15
SPEAKER_03: this person's asleep,
18:16
SPEAKER_03: this person's awake.
18:18
It's very crude,
18:18
and honestly,
18:19
EEG hasn't really progressed that far beyond that these days.
18:23
SPEAKER_01: They're just putting some sensor on the outside of the skull,
18:25
SPEAKER_01: and they're just seeing that there's stuff going on.
18:28
SPEAKER_03: Yep, there's a voltage.
18:30
SPEAKER_03: It's called a montage.
18:31
SPEAKER_03: You have these electrodes that you place over the skull.
18:33
SPEAKER_03: You can put them in different arrangements.
18:36
SPEAKER_03: And then a voltage has to be across some distance.
18:41
SPEAKER_03: One electrode is going to be a reference.
18:42
SPEAKER_03: Maybe you use it over here in the mastoid or something.
18:44
And then you have another one on top of the head that you're
18:46
measuring a signal.
18:47
SPEAKER_03: And then you might have 16, 32, 64 channels.
18:49
SPEAKER_03: And then you see these voltage shifts.
18:54
SPEAKER_03: So,
18:54
SPEAKER_03: for example,
18:55
SPEAKER_03: when someone falls asleep,
18:56
SPEAKER_03: you see these alpha waves,
18:57
SPEAKER_03: which are roughly 5 Hz and lower.
19:00
You see these relatively slow-moving waves,
19:03
and they increase in magnitude.
19:05
SPEAKER_03: And then when someone's awake, those kind of go away.
19:07
SPEAKER_03: And then there are some other effects, like when someone's...
19:10
SPEAKER_03: preparing to move, you see this beta.
19:13
So this is like this 12 to 15 hertz signal actually disappears
19:19
right before someone
19:20
SPEAKER_03: moves,
19:20
SPEAKER_03: and it's a movement preparation signal.
19:22
We actually don't know that much about it,
19:25
but we think there's a kind of constant...
19:28
SPEAKER_03: inhibition of movement that's happening between your cortex,
19:31
SPEAKER_03: the kind of surface of the brain,
19:32
SPEAKER_03: and the thalamus.
19:33
SPEAKER_03: And this corticothalamic loop is just constantly going,
19:36
SPEAKER_03: inhibiting you from movement,
19:37
SPEAKER_03: which is probably an evolutionary thing that it's better to do that.
19:40
SPEAKER_03: And then it actually gets interrupted and then you move.
19:44
SPEAKER_03: So you get these kind of correlates,
19:45
SPEAKER_03: but they're really kind of crude and not that specific,
19:49
SPEAKER_03: but you can get something out of EEG.
19:52
SPEAKER_03: And then,
19:53
SPEAKER_03: as time goes on,
19:54
SPEAKER_03: extracellular electrophysiology,
19:56
SPEAKER_03: which is kind of what you're describing,
19:57
SPEAKER_03: you put an electrode down into the tissue itself,
20:00
SPEAKER_03: get really close to the neuron.
20:02
SPEAKER_03: Now,
20:02
instead of just getting the summed voltage of millions or
20:05
billions of neurons,
20:06
you're actually getting the activity of one or maybe a couple
20:09
nearby neurons.
20:11
And that tells you something really interesting because instead
20:13
of these waves,
20:14
SPEAKER_03: what you see or hear are neuron spikes.
20:17
SPEAKER_03: And so you hear like neurons firing or spiking or depolarization.
20:21
SPEAKER_03: That's what we're talking about.
20:22
SPEAKER_03: So you see as a neuron becomes active, it does a little pop.
20:25
SPEAKER_03: It has this big voltage jump and then recovery.
20:28
SPEAKER_03: And this happens over just a couple of milliseconds.
20:32
And what's interesting is that some neurons are active for
20:35
certain types of
20:36
SPEAKER_03: activity and then otherwise completely silent.
20:39
And then sometimes they're active for many types of activities,
20:42
but not others.
20:43
SPEAKER_03: And so there's this kind of what's called mixed selectivity.
20:46
SPEAKER_03: They actually pick up certain aspects of different behavior.
20:48
SPEAKER_03: So to make that concrete,
20:50
SPEAKER_03: let's imagine that I told you to move your finger or thumb,
20:54
SPEAKER_03: let's do thumb,
20:55
SPEAKER_03: to the right.
20:56
SPEAKER_03: When you move your thumb to the right,
20:57
there's a set of neurons in your motor cortex that are
21:00
connected through all these
21:01
fibers into your thumb that are doing that and that are
21:04
preparing that movement and
21:05
SPEAKER_03: executing that movement.
21:07
And so if I'm lucky and I put my electrode in the right spot in
21:10
the motor cortex,
21:12
SPEAKER_03: I can actually find the neuron that pops every time you...
21:15
SPEAKER_03: or depolarizes or fires every time you move your thumb to the right.
21:19
SPEAKER_03: And now I have a little, like, thumb-right neuron.
21:21
And then if I find another one that's a thumb-left neuron and
21:24
up and down and so on,
21:26
Now if I have a couple hundred electrodes recording from a
21:29
bunch of different
21:30
SPEAKER_03: neurons,
21:30
SPEAKER_03: I can start to mix those together and say,
21:32
SPEAKER_03: not only can I know when he's going to move to the right,
21:34
SPEAKER_03: I see that the right one was firing and the up one was firing,
21:37
SPEAKER_03: he's probably going at a 45-degree angle.
21:40
SPEAKER_03: You can use simple machine learning methods,
21:41
SPEAKER_03: or decoders,
21:42
SPEAKER_03: we call them,
21:43
SPEAKER_03: to decode that thumb movement.
21:45
This is precisely what groups like BlackRock and Neuralink and
21:49
the BrainGate
21:50
SPEAKER_03: studies that go all the way back into the early 2000s have done.
21:53
SPEAKER_03: is record from the motor cortex of human beings who are paralyzed,
21:56
SPEAKER_03: unable to move,
21:58
and then reconstruct the intention or the willed intent of
22:01
movement that they just
22:03
can't do because their spinal cord is not allowing that signal
22:06
to pass through.
22:06
SPEAKER_01: So the person is still thinking,
22:08
SPEAKER_01: I want to move my thumb to the right,
22:10
SPEAKER_01: and we can pick up that intention.
22:14
SPEAKER_01: Yep, exactly.
22:15
SPEAKER_03: So we're just correlating intended behavior.
22:17
SPEAKER_03: So we have to train these decoders.
22:18
SPEAKER_03: So we tell the person, hey, try to move to the right.
22:20
SPEAKER_03: They can't.
22:22
SPEAKER_03: We assume that they do the best they can.
22:24
And we record the neural activity and we look for neurons that
22:26
are firing.
22:26
And then after a while,
22:27
our decoder learns what the person is trying to do.
22:30
SPEAKER_03: And then, yeah, exactly.
22:31
SPEAKER_03: We can recreate through an external device.
22:33
SPEAKER_03: It could be a cursor on a screen.
22:34
SPEAKER_03: It could be a robot.
22:35
SPEAKER_03: It doesn't really matter.
22:36
SPEAKER_03: It's easy to turn it into whatever once you have the control signal.
22:40
SPEAKER_01: When we first started going into the brain,
22:44
SPEAKER_01: I assume this was like in animals first,
22:46
SPEAKER_01: or were people just diving right into humans?
22:49
SPEAKER_03: Animals.
22:50
SPEAKER_03: But the translation path wasn't that long, though.
22:54
SPEAKER_03: So we see Ebb Fetz at the University of Washington.
22:57
SPEAKER_03: I think he was a postdoc at the time, I think.
23:00
SPEAKER_03: But I think that paper was 1969.
23:02
SPEAKER_03: He taught a monkey to control the firing rate of a single neuron.
23:07
SPEAKER_03: This actually taught us a couple things.
23:09
SPEAKER_03: So this was before we could do multicellular recordings.
23:11
SPEAKER_03: We're just listening to one neuron.
23:13
And then you have a little meter that's counting how many
23:15
neuron spikes it sees over time.
23:18
SPEAKER_03: And he teaches the monkey if he can tip this needle up in the right way,
23:22
SPEAKER_03: he'll get a juice reward.
23:23
SPEAKER_03: And so this teaches us a couple things.
23:25
SPEAKER_03: One, that we can actually detect these signals.
23:28
SPEAKER_03: And two,
23:29
SPEAKER_03: that they're not only controllable,
23:31
SPEAKER_03: but in this case,
23:31
SPEAKER_03: it's sort of abstracted away from any kind of specific behavior.
23:35
The monkey just learns to control this through like,
23:37
hey,
23:38
I want to make that needle move.
23:39
SPEAKER_03: And over time, the monkey connects that.
23:42
SPEAKER_03: So 1969 is when we see the first kind of,
23:43
SPEAKER_03: like,
23:44
SPEAKER_03: BCI and the motor neuroprosthetic way we think of BCI.
23:48
SPEAKER_03: And it's like a wire that's in the brain?
23:51
SPEAKER_01: Okay.
23:52
SPEAKER_01: I'm just picturing electronics at that stage,
23:54
SPEAKER_01: but I guess it's,
23:55
SPEAKER_01: it's,
23:55
SPEAKER_01: you could still make thin wires with a little electrode at the end.
23:59
SPEAKER_03: Yeah, yeah, thin wires have been around a long time.
24:01
SPEAKER_03: I mean,
24:02
I think where we actually see the biggest jumps in groups like
24:05
Neuralink and others
24:06
SPEAKER_03: today is in the commercializability of these things.
24:11
SPEAKER_03: So they need to last 10 years in the human body.
24:14
SPEAKER_03: That's an incredibly hard problem.
24:17
SPEAKER_03: Like,
24:17
making a wire that you stick in the brain for a day and making
24:20
a wire that you
24:20
stick in the brain for 10 years and you know it's gonna work
24:23
and not shock somebody
24:24
SPEAKER_03: and not break down or be toxic,
24:27
SPEAKER_03: that's an infinitely harder problem.
24:30
And so,
24:30
like,
24:31
you mentioned earlier that,
24:32
like,
24:32
what Neuralink is doing is aggressive.
24:34
And what I would actually say is that it is very technically
24:37
aggressive and it's
24:38
SPEAKER_03: medically aggressive,
24:39
SPEAKER_03: but it's scientifically actually very conservative.
24:42
SPEAKER_03: Because the science goes back 50-plus years in some ways,
24:46
SPEAKER_03: but it's really the technical aspect of making these things work.
24:49
SPEAKER_01: Well, and people ding them for that sometimes.
24:51
They're like,
24:52
oh,
24:52
you're just replicating stuff that's gone on for years and
24:55
years and years.
24:56
SPEAKER_01: And yeah,
24:57
SPEAKER_01: I always come down,
24:58
SPEAKER_01: I think,
24:58
SPEAKER_01: where you are,
24:59
SPEAKER_01: where it's...
25:01
SPEAKER_01: I don't want to put words in your mouth.
25:03
SPEAKER_01: It's a blend, right?
25:04
SPEAKER_01: Because yeah,
25:05
SPEAKER_01: okay,
25:06
SPEAKER_01: you've proven some science out,
25:07
SPEAKER_01: but the machines you did that on were enormous,
25:11
SPEAKER_01: and the person had to be in a hospital,
25:14
SPEAKER_01: and there had to be like eight people managing this process.
25:16
SPEAKER_01: And then Neuralink, like you said, I mean, it's got this,
25:19
tiny coin that's essentially a consumer electronics device,
25:23
and I've been with the patients.
25:24
SPEAKER_01: They're at their house using this thing on their own,
25:26
SPEAKER_01: and,
25:27
SPEAKER_01: you know,
25:27
SPEAKER_01: the...
25:29
SPEAKER_01: Like,
25:29
SPEAKER_01: sometimes I get it,
25:30
SPEAKER_01: because Neuralink has a big press conference,
25:32
SPEAKER_01: and they're making a big song and dance about what they've done,
25:34
SPEAKER_01: and so people in the field are like,
25:36
SPEAKER_01: oh,
25:36
SPEAKER_01: we did that a long time ago.
25:38
SPEAKER_01: But, you know, then I...
25:40
SPEAKER_01: So I understand their point of view,
25:41
but then I also feel like they're so into their field that
25:45
they're discounting all
25:46
the work that goes into making this thing work in these
25:50
particular circumstances.
25:52
I think wanting groups like Neuralink to push the frontier of
25:56
science is really fair.
25:58
SPEAKER_03: I'm with all of those people that are like, yeah, let's do more.
26:00
SPEAKER_03: Let's see where the edge of this really is.
26:02
SPEAKER_03: And I'm 100% certain that Neuralink feels the same way.
26:06
SPEAKER_03: I think it's just a completely wrong argument.
26:09
SPEAKER_03: It's like saying,
26:10
SPEAKER_03: like,
26:10
SPEAKER_03: you know,
26:10
SPEAKER_03: a 747 flies overhead or F18,
26:13
SPEAKER_03: and people go,
26:14
SPEAKER_03: wow,
26:15
SPEAKER_03: holy shit,
26:15
SPEAKER_03: did you see that?
26:16
SPEAKER_03: And someone next to them goes, the Wright brothers did that.
26:19
SPEAKER_03: They came up with that concept years and years ago.
26:22
SPEAKER_03: It's like, okay, true.
26:23
SPEAKER_03: Technically,
26:24
SPEAKER_03: the concept's been around a long time,
26:25
SPEAKER_03: but turning that into something,
26:27
SPEAKER_03: like you said,
26:27
SPEAKER_03: where...
26:28
SPEAKER_03: A person can go home,
26:29
SPEAKER_03: they don't need a team of scientists hiding behind the camera,
26:32
SPEAKER_03: which is what the academic studies really are.
26:35
SPEAKER_03: And racks full of computers and equipment that cost millions of dollars.
26:39
Like,
26:39
to shrink that down into something that fits inside the skull
26:42
is an insanely hard problem.
26:44
SPEAKER_01: Yeah.
26:44
SPEAKER_01: Okay,
26:45
SPEAKER_01: if we...
26:45
Am I fast-forwarding too far if we go from some of the
26:49
experiments you were talking
26:50
SPEAKER_01: about to the Utah Array?
26:53
SPEAKER_03: Not really.
26:53
SPEAKER_03: Yeah,
26:54
SPEAKER_03: so if we go from like Ed Fetz,
26:56
SPEAKER_03: like 1969,
26:57
SPEAKER_03: and then you start to see...
27:00
SPEAKER_03: So you see tetrodes and multi-electrode arrays coming on the scene.
27:04
SPEAKER_03: I think it was 1992,
27:08
SPEAKER_03: seven,
27:08
SPEAKER_03: somewhere in there that the Utah Ray,
27:09
SPEAKER_03: I forget exactly,
27:10
SPEAKER_03: Richard Norman at the University of Utah.
27:12
SPEAKER_03: No affiliation, by the way.
27:14
SPEAKER_03: Two N's in his name.
27:16
SPEAKER_03: creates this little,
27:17
SPEAKER_03: for anyone that hasn't seen it,
27:18
SPEAKER_03: it looks like a tiny little bed of spikes,
27:20
SPEAKER_03: but it is,
27:20
SPEAKER_03: I mean,
27:21
SPEAKER_03: really tiny.
27:21
SPEAKER_03: It's four by four millimeters like that.
27:25
SPEAKER_03: And it has a 10 by 10 array of spikes that stick out of it.
27:29
SPEAKER_03: And then you basically stick that into the brain,
27:32
SPEAKER_03: and it's just right in kind of the surface of the cortex.
27:35
SPEAKER_03: And now you have 100 channels of multi-unit electrophysiology,
27:38
SPEAKER_03: and so recording single units,
27:40
SPEAKER_03: single neurons.
27:41
SPEAKER_03: and you can start to mix together these signals.
27:43
SPEAKER_03: So if you put it in motor cortex,
27:45
SPEAKER_03: for example,
27:46
SPEAKER_03: you can start to decode these kind of fine motor controls.
27:49
SPEAKER_03: They didn't go straight into humans, they went into monkeys first.
27:53
SPEAKER_03: So we see groups like Richard Anderson, my former mentor at Caltech,
27:58
SPEAKER_03: And Andy Schwartz and John Donahue,
28:01
SPEAKER_03: these kind of like early big names,
28:03
SPEAKER_03: Miguel Nicolaylis,
28:05
SPEAKER_03: all around the same time in the 90s and early 2000s,
28:09
start to do these multi-unit recordings in monkeys and start to
28:13
show monkeys
28:14
SPEAKER_03: controlling robotic arms,
28:15
feeding themselves with a robotic arm or controlling a cursor
28:18
on a screen,
28:19
SPEAKER_03: and not too dissimilar from what we eventually see Neuralink doing.
28:24
SPEAKER_03: And in fact, then the BrainGate group pops up.
28:27
SPEAKER_03: And this is John Donahue and a few others.
28:30
And this is the effort to translate this from monkeys into
28:34
humans for the first time.
28:36
SPEAKER_01: Okay.
28:36
SPEAKER_01: And John Donahue, is that Brown?
28:37
SPEAKER_01: Mm-hmm.
28:38
SPEAKER_01: Yeah.
28:39
SPEAKER_01: Okay.
28:40
SPEAKER_01: And that's the beginning of human trials.
28:43
Yeah,
28:43
I think that's where I think people would go,
28:45
oh,
28:45
I recognize that as a BCI.
28:47
SPEAKER_03: Because it's like,
28:47
SPEAKER_03: there's a human doing the thing that Neuralink is doing today,
28:50
SPEAKER_03: not in nearly the same level of performance.
28:53
SPEAKER_03: Like it's, you know, a lot jankier, a lot more expensive.
28:57
It's a long ways away from commercialization,
28:59
but that's where it's like,
29:00
I recognize that.
29:01
SPEAKER_01: All right,
29:02
SPEAKER_01: I will bastardize this,
29:03
SPEAKER_01: and you can help fix it,
29:04
SPEAKER_01: but that small,
29:05
SPEAKER_01: when you talk about the Utah Ray,
29:06
SPEAKER_01: I mean,
29:08
SPEAKER_01: can we roughly say,
29:09
SPEAKER_01: like,
29:09
SPEAKER_01: fingernail-sized,
29:11
SPEAKER_01: maybe?
29:12
SPEAKER_01: Smaller than a fingernail, yeah.
29:13
SPEAKER_01: With spikes on the end,
29:15
SPEAKER_01: and it gets kind of,
29:16
SPEAKER_01: like,
29:16
SPEAKER_01: pneumatically hammered into the brain with some taps,
29:21
SPEAKER_01: right?
29:23
SPEAKER_03: Yeah.
29:24
SPEAKER_03: Although,
29:26
SPEAKER_03: I won't say where or how I heard this,
29:29
SPEAKER_03: but in many cases,
29:30
SPEAKER_03: the pneumatic applicator just doesn't work that well,
29:32
SPEAKER_03: and the surgeon goes.
29:33
SPEAKER_01: Just gives it a shot of a final, a final poke.
29:37
SPEAKER_01: Okay, okay.
29:38
SPEAKER_01: And then,
29:38
SPEAKER_01: but then to read the signals out,
29:40
SPEAKER_01: I mean,
29:40
SPEAKER_01: that thing is small,
29:41
SPEAKER_01: but then you've got a hole in the brain.
29:43
And then on top of that,
29:44
there's like,
29:45
I always compare it to like a fist.
29:47
I mean,
29:47
maybe it's a little,
29:48
it depends on the size of your fist,
29:49
I guess.
29:50
SPEAKER_01: But I mean,
29:51
SPEAKER_01: there's like this hardware element,
29:53
SPEAKER_01: I guess they call it a pedestal,
29:55
SPEAKER_01: that's sitting on top of your head,
29:57
SPEAKER_01: and that is quite clunky.
29:59
SPEAKER_01: And then that has traditionally had cables that stretch into a computer,
30:04
and so you've got a person who's maybe paralyzed or has ALSing
30:07
quite dire.
30:09
SPEAKER_01: circumstances,
30:10
SPEAKER_01: sitting in a hospital room with this thing coming out of their head,
30:13
SPEAKER_01: wired into like a rack of computers.
30:16
SPEAKER_01: I mean, and that was probably sort of 20, 15 years of that.
30:22
SPEAKER_03: Yeah, at least.
30:24
SPEAKER_03: Yeah, at least, yeah.
30:25
SPEAKER_03: So I think the first paper was in 2006,
30:28
SPEAKER_03: but I think the patient was implanted before that,
30:29
SPEAKER_03: 2003,
30:30
SPEAKER_03: 4,
30:31
SPEAKER_03: 5,
30:30
SPEAKER_03: somewhere in there.
30:32
SPEAKER_03: So yeah, we're, you know, over 20 years of, we're still doing this.
30:36
SPEAKER_03: There are still many patients implanted with Utah Arrays.
30:38
SPEAKER_01: Right,
30:39
SPEAKER_01: and then,
30:39
SPEAKER_01: like you said,
30:40
SPEAKER_01: for those 20 years,
30:40
we see these things where people who are paralyzed are
30:44
controlling robotic arms.
30:45
SPEAKER_01: They're sending out...
30:48
SPEAKER_01: words of what they would like to see happen, right?
30:52
SPEAKER_01: And translating some of this.
30:53
And then when we have,
30:56
if I remember right,
30:57
Neuralink comes around around,
30:58
like,
30:58
2016.
30:58
SPEAKER_01: In the,
31:02
SPEAKER_01: at least from where I was sitting,
31:03
SPEAKER_01: the push forward is,
31:04
SPEAKER_01: okay,
31:04
SPEAKER_01: we've had this thing in this research setting for 20 years.
31:08
SPEAKER_01: It's been incredibly useful.
31:10
People haven't really advanced the underlying microelectronics
31:15
behind this.
31:17
um there wasn't really a business use case to do this so there
31:21
wasn't the same kind
31:22
of money pouring in and then all of a sudden brian johnson gets
31:26
interested in
31:26
brains elon gets interested in brains then everybody's
31:30
interested in brains and and
31:32
so money floods in and now we've seen these different attempts
31:35
to um either
31:38
miniaturize tweak um different takes on on what amounts to utah
31:44
right and different
31:45
SPEAKER_01: flavors
31:47
SPEAKER_03: I think that's sort of true.
31:50
SPEAKER_03: I would say that,
31:50
SPEAKER_03: like,
31:52
SPEAKER_03: during that entire time,
31:54
SPEAKER_03: there was a ton of effort going into making better,
31:57
SPEAKER_03: thinner electrodes.
31:58
SPEAKER_03: And so,
31:59
SPEAKER_03: you know,
31:59
some of the tech that went into Neuralink came out of Lawrence
32:01
Livermore National
32:02
SPEAKER_03: Labs,
32:02
SPEAKER_03: the electrodes,
32:03
SPEAKER_03: and Vanessa Tolosa's group there.
32:06
Flip Sabes and Tim Hansen are working on a sewing machine to
32:09
actually get these
32:09
SPEAKER_03: things into the brain in a way.
32:11
SPEAKER_03: So they're literally like sewing these thin wires in.
32:14
SPEAKER_03: People are working on longevity to make them last a long time.
32:17
SPEAKER_03: So all that's happening.
32:18
SPEAKER_01: So people are moving away from this kind of rigid, these rigid, um,
32:23
rigid nodes,
32:24
I guess,
32:25
and the Utah Ray itself being kind of like this- Like a chip.
32:29
SPEAKER_01: Yeah, they're like a fixed to this chip.
32:31
SPEAKER_01: So the people, they were trying to move away from that.
32:33
SPEAKER_03: And I think like right back to the very beginning,
32:36
SPEAKER_03: even in the late 90s,
32:37
SPEAKER_03: early 2000s,
32:38
SPEAKER_03: people knew that that was the way forward.
32:40
SPEAKER_03: It's a hard engineering problem.
32:41
SPEAKER_03: They were working on it.
32:42
But the fact is like academia is set up to solve three to $5
32:47
million problems.
32:48
SPEAKER_03: There's like chunked in that way.
32:50
SPEAKER_03: And at some point,
32:52
SPEAKER_03: these just become very expensive engineering problems,
32:55
SPEAKER_03: and they're not really as interesting as fundamental research.
32:59
And even if they are,
33:01
there's only so much national funding to support it along the
33:04
way.
33:05
And translating that into human use is a very expensive,
33:09
difficult,
33:09
long proposition.
33:11
SPEAKER_03: So once the Utah array makes it into humans, that really actually...
33:15
SPEAKER_03: A lot of that early success came from DARPA funding.
33:18
SPEAKER_03: And so you see this big injection that was kind of one big catalyst to,
33:21
SPEAKER_03: like,
33:21
SPEAKER_03: okay,
33:22
SPEAKER_03: now we've got electrodes in humans and,
33:24
SPEAKER_03: like,
33:24
SPEAKER_03: we've moved the field forward.
33:26
SPEAKER_03: And then you see that kind of,
33:27
SPEAKER_03: like,
33:27
SPEAKER_03: fade a bit,
33:28
SPEAKER_03: and the NIH is driving things forward,
33:30
but there's really not a lump sum sitting around anywhere that
33:33
people can say,
33:34
I'm going to go take those really fine threads that are much
33:37
more biocompatible,
33:38
SPEAKER_03: and I'm going to take that delivery method,
33:40
SPEAKER_03: I'm going to take that processing and that integrated circuit.
33:42
SPEAKER_03: That didn't exist.
33:44
SPEAKER_03: There's no way to do that.
33:46
Neuralink came along and said,
33:47
hey,
33:47
look,
33:47
all the pieces of the puzzle exist.
33:50
The science is far enough along that we really kind of know
33:53
it's going to work.
33:54
SPEAKER_03: That's what I meant by it's not aggressive on the science front.
33:56
SPEAKER_03: They knew it would work from day one.
33:59
SPEAKER_03: With a large enough investment,
34:01
we could actually put these together and get a killer BCI out
34:05
there into humans.
34:08
SPEAKER_01: Okay,
34:08
SPEAKER_01: since day one,
34:09
the Core Memory podcast has been supported by the fine people
34:13
at E1 Ventures.
34:14
SPEAKER_01: They are a young and ambitious VC firm in Silicon Valley,
34:18
SPEAKER_01: investing in young and ambitious companies and people.
34:22
SPEAKER_01: Thank you so much to E1 Ventures for all your support.
34:25
SPEAKER_01: Actually, it's not that dissimilar from SpaceX in a lot of ways.
34:28
SPEAKER_01: It's like, we knew how rockets worked.
34:30
SPEAKER_01: We've been doing them for a long time.
34:33
There'd been this consumer electronics revolution that the
34:36
aerospace industry
34:37
SPEAKER_01: hadn't really...
34:37
SPEAKER_01: Control systems,
34:38
SPEAKER_01: man.
34:39
SPEAKER_01: Yeah, you know, they put... Inverted pendulum.
34:41
SPEAKER_01: And they just hadn't... There was a...
34:45
It's like stuff got a little bit fixed in time where you had
34:48
space-grade radios.
34:50
It's like,
34:50
we can't trust regular radios,
34:53
and software had obviously modernized.
34:55
SPEAKER_01: I mean,
34:55
SPEAKER_01: I've always thought of SpaceX as,
34:59
SPEAKER_01: you know,
34:59
SPEAKER_01: the physics of rockets are just well understood,
35:02
SPEAKER_01: and there's limitations,
35:03
but they assembled all these modern parts and started with this
35:07
clean slate and
35:08
SPEAKER_01: kind of integrated everything in a way the other folks hadn't,
35:13
SPEAKER_01: and modernized it that way.
35:14
SPEAKER_03: Yeah, I think that's right.
35:17
SPEAKER_03: I think Neuralink takes some of those same pages out of the playbook,
35:20
SPEAKER_03: which is like,
35:21
SPEAKER_03: let's take the pieces,
35:22
SPEAKER_03: the engineering problem that we know,
35:24
SPEAKER_03: if we put it together,
35:24
SPEAKER_03: we can solve it with enough people,
35:26
SPEAKER_03: talent,
35:27
SPEAKER_03: time,
35:27
SPEAKER_03: and capital.
35:29
SPEAKER_03: We don't need to revolutionize the neuroscience,
35:31
SPEAKER_03: but maybe once we have the tool,
35:33
SPEAKER_03: we can start to push the forefront of neuroscience.
35:36
SPEAKER_03: And I would argue in some ways they have.
35:38
I mean,
35:38
the fastest bit rate decoding that we have with a motor BCI
35:41
today came from that group.
35:44
And,
35:44
you know,
35:45
it's built on science that came from speech decoding groups.
35:48
SPEAKER_03: So,
35:48
SPEAKER_03: like,
35:49
SPEAKER_03: Sergei Staviskyi's group at UC Davis is really pushing on this and,
35:52
SPEAKER_03: like,
35:52
SPEAKER_03: making incredible progress as an academic group.
35:56
And before that,
35:57
Krishna Shanoi's group at Stanford,
35:58
just down the road,
35:59
was doing similar work.
36:01
SPEAKER_03: So,
36:02
SPEAKER_03: yes,
36:02
SPEAKER_03: it's building on science that's coming out of the lab,
36:05
SPEAKER_03: but it's translating it faster and faster,
36:06
SPEAKER_03: and it's driving it further and further to the edge.
36:09
SPEAKER_03: Okay.
36:10
SPEAKER_01: All right.
36:10
I want you to explain what Forrest does because I have a ton of
36:12
BCI questions about
36:15
SPEAKER_01: all these various players.
36:16
SPEAKER_01: And so I want to place you guys in where all this fits.
36:22
SPEAKER_01: I can do my journalist stab at this for a second.
36:25
SPEAKER_03: Yeah, yeah, let's do the journalist step.
36:27
SPEAKER_01: Okay, okay.
36:27
SPEAKER_01: I mean,
36:28
SPEAKER_01: it is an implant in the vein of,
36:32
SPEAKER_01: there's a surgery required,
36:34
SPEAKER_01: and you're putting,
36:35
SPEAKER_01: it's like not horribly dissimilar in shape and size a bit from Neuralink.
36:41
SPEAKER_01: I mean, the circular,
36:43
SPEAKER_01: device going into the skull,
36:45
and then you have a chip on there that is able to pulse
36:53
ultrasound through the
36:55
SPEAKER_01: brain.
36:55
SPEAKER_01: And so,
36:56
SPEAKER_01: whereas like a neural lig implant sitting in the motor cortex,
36:59
SPEAKER_01: and it's just a couple millimeters in there,
37:01
SPEAKER_01: and it's just able to see that really specific group of neurons,
37:06
SPEAKER_01: With the ultrasound,
37:06
SPEAKER_01: you can see,
37:07
I think it's like 80% to 90% of coverage across the brain and
37:13
the activity.
37:13
SPEAKER_01: We'll get there.
37:15
SPEAKER_03: We're not quite there yet, but we're closer to like 20%, 25% now.
37:18
SPEAKER_01: 20%, 25%.
37:18
SPEAKER_01: But with a single device.
37:19
Okay,
37:19
but the hope is that you're not just seeing this tiny clump of
37:22
neurons you're seeing.
37:23
SPEAKER_01: That's right.
37:24
SPEAKER_01: Broader-based brain activity,
37:26
SPEAKER_01: and then instead of focusing in the near term on things like movement,
37:32
SPEAKER_01: and having the motor cortex generate those kinds of outputs,
37:37
SPEAKER_01: you're looking more at,
37:38
SPEAKER_01: like,
37:39
SPEAKER_01: how does our brain operate if I'm depressed?
37:42
SPEAKER_01: What does depression look like?
37:44
SPEAKER_01: And hopefully...
37:47
SPEAKER_01: Again, a lot of these studies end up being, okay, somebody's depressed.
37:51
SPEAKER_01: We have them in this very clinical machine in the hospital.
37:55
SPEAKER_01: I'm gonna study how your mental state,
37:58
while we have all these people watching you in this very
38:00
clinical setting.
38:01
Instead,
38:02
you could have an implant that you're wearing for a month,
38:04
for two months.
38:06
SPEAKER_01: You could be at home,
38:07
SPEAKER_01: and now,
38:08
SPEAKER_01: hopefully,
38:08
we're getting a more accurate baseline of what your brain looks
38:12
like when you're...
38:14
SPEAKER_01: Am I doing okay?
38:15
SPEAKER_03: Yes, and you can use it to then stimulate.
38:18
So you can actually try to correct these brain states into more
38:21
normative,
38:22
healthy ones.
38:23
SPEAKER_03: And then you have a closed-loop monitor, stimulate.
38:27
SPEAKER_03: Did that have the effect I want?
38:28
SPEAKER_03: Yes or no?
38:28
SPEAKER_03: Back to the control systems, right?
38:30
SPEAKER_03: Control system theory here.
38:31
SPEAKER_03: But can we drive the brain into more normative,
38:34
SPEAKER_03: healthy brain states through closed-loop feedback?
38:36
SPEAKER_01: Which,
38:36
SPEAKER_01: okay,
38:37
SPEAKER_01: and this is gonna open up a whole other can of questions,
38:39
SPEAKER_01: but we've had things like deep brain stimulation,
38:42
SPEAKER_01: you've had,
38:42
SPEAKER_01: like,
38:43
SPEAKER_01: electroshock therapy.
38:44
SPEAKER_01: I mean, that goes back... Way back.
38:46
SPEAKER_01: So people didn't understand why electroshock therapy worked basically,
38:50
SPEAKER_01: like,
38:51
SPEAKER_01: at all,
38:51
SPEAKER_01: from my understanding,
38:52
SPEAKER_01: but we knew.
38:53
So somebody did some experiment somewhere where they started
38:56
shocking people,
38:58
SPEAKER_01: and something decent happened.
39:00
SPEAKER_03: Yeah.
39:01
SPEAKER_03: Yeah.
39:02
SPEAKER_03: I mean,
39:02
I think deep brain stimulators in particular are really
39:04
interesting because that's
39:05
SPEAKER_03: like really,
39:06
SPEAKER_03: it looks like neurotech,
39:08
SPEAKER_03: right?
39:08
SPEAKER_03: There's a needle that's going down into the brain.
39:10
SPEAKER_03: It's stimulating a part of the brain.
39:12
It's alleviating complex systems despite being really pretty
39:16
like crude as a
39:17
SPEAKER_03: device,
39:17
SPEAKER_03: just delivering voltage somewhere in the brain.
39:20
SPEAKER_03: And that goes all the way back into, like, 50s and 60s.
39:24
Robert Heath at the Tulane University has,
39:27
like,
39:27
this crazy round of experiments.
39:29
SPEAKER_03: This is,
39:29
SPEAKER_03: like,
39:30
SPEAKER_03: pre-FDA,
39:30
SPEAKER_03: so,
39:31
SPEAKER_03: like,
39:31
SPEAKER_03: a lot of this was,
39:32
SPEAKER_03: like,
39:32
SPEAKER_03: borderline unethical,
39:33
SPEAKER_03: just to be clear.
39:36
Or he's just,
39:36
like,
39:36
testing this in,
39:37
like,
39:37
every...
39:38
You could do that,
39:39
by the way.
39:39
SPEAKER_03: Yeah, you kind of could.
39:41
SPEAKER_03: You could test it just like every psychological,
39:45
SPEAKER_03: psychiatric indication you could possibly imagine.
39:48
SPEAKER_03: Depression, OCD, bipolar disorder, all of them.
39:51
SPEAKER_03: And he's finding ways that this is going to work.
39:52
SPEAKER_03: And so, in some ways, he is the control system, right?
39:55
SPEAKER_03: He's like,
39:56
I'm going to stick this thing in and I'm just going to zap the
39:58
hell out of the
39:59
SPEAKER_03: brain and this particular person that I think of as a hopeless case,
40:03
SPEAKER_03: If it works, it works.
40:04
SPEAKER_03: If it doesn't,
40:04
SPEAKER_03: oh,
40:04
SPEAKER_03: well,
40:05
we're just going to try somewhere else and I'll poke it in
40:07
again somewhere else.
40:09
It was very fly by the seat of your pants compared to how we do
40:12
things today.
40:14
But made an incredible amount of progress and deep brain
40:16
stimulation may have never
40:17
SPEAKER_03: come to fully exist the way it does.
40:19
I think there's 200,000-ish people currently implanted with
40:22
deep brain stimulators.
40:24
SPEAKER_03: Primarily for Parkinsonian tremor.
40:28
SPEAKER_03: So it's for motor dyskinesia.
40:30
SPEAKER_03: So people that have Parkinson's,
40:32
SPEAKER_03: you kind of see this sort of like...
40:33
SPEAKER_01: Yeah,
40:33
SPEAKER_01: you see those videos where they flick it on.
40:36
SPEAKER_03: Yeah, and they'll literally be going like this and they just stop.
40:38
SPEAKER_03: It's like magic.
40:40
SPEAKER_01: It's super cool.
40:40
SPEAKER_01: And like epilepsy, I think.
40:42
SPEAKER_03: Yeah,
40:43
so NeuroPace really kind of innovated a lot there where they
40:47
were doing closed-loop
40:47
SPEAKER_03: for epilepsy,
40:48
SPEAKER_03: so they're actually looking for a seizure,
40:50
SPEAKER_03: and when they see it,
40:50
SPEAKER_03: they try to interrupt it.
40:52
SPEAKER_03: In fact,
40:53
SPEAKER_03: Frank Fisher's on our impact committee at Forrest,
40:55
SPEAKER_03: so he was one of the co-founders and CEO of NeuroPace,
40:59
which I think is still the only company to go all the way from
41:02
a concept of a
41:03
neurotech device all the way to full commercialization in
41:06
humans without getting
41:08
SPEAKER_03: killed or acquired.
41:09
SPEAKER_03: Okay.
41:11
SPEAKER_01: With badge of honor.
41:12
SPEAKER_01: It is, yeah.
41:14
SPEAKER_01: But in severe mental health cases as well, right?
41:18
SPEAKER_03: It's being investigated in deep brain stimulators for depression.
41:22
SPEAKER_03: It's already approved for OCD.
41:24
SPEAKER_03: It's being investigated for depression.
41:26
People are interested in using it for substance use disorder
41:28
and other disorders of compulsion.
41:31
So I think there's still quite a lot of alpha to be found in
41:34
deep brain stimulation.
41:36
And many neurotech companies are aware of this and interested
41:40
in going after it.
41:42
And in some ways,
41:42
you could even say that Forrest is a deep brain stimulator.
41:46
SPEAKER_03: It's just not the traditional one.
41:48
We can deliver doses of energy deep into the brain by focusing
41:52
the ultrasound beam,
41:53
but we're not invasive in the sense that we're not actually
41:56
going into the brain
41:57
SPEAKER_03: parenchyma,
41:57
SPEAKER_03: into the brain tissue itself.
41:59
We're sitting completely outside the brain and also completely
42:02
outside the dura
42:03
SPEAKER_03: mater,
42:03
which is this protective membrane that goes around your brain
42:06
and really separates
42:07
SPEAKER_03: the outside world from the central nervous system.
42:10
SPEAKER_03: Central nervous system kind of lives on its own immune system,
42:12
SPEAKER_03: and you really don't want to mess with that.
42:13
SPEAKER_03: That's where the risk from brain surgery comes from.
42:16
We sit outside that,
42:17
which means we're almost more like a plastic surgery than brain
42:22
surgery.
42:23
For people outside of brain land,
42:24
I mean,
42:25
I hear this come up all the time.
42:26
SPEAKER_01: It's like you're either on top of the dura or underneath.
42:30
And if you're underneath,
42:31
you're in a whole different world of opportunity and problems.
42:36
SPEAKER_01: And if you're outside, considered much, much safer.
42:39
SPEAKER_01: So you're drilling a hole in the skull,
42:44
SPEAKER_01: the implant's going in,
42:45
SPEAKER_01: and then you're essentially,
42:46
SPEAKER_01: like,
42:46
sitting on top of these protective layers before you have to
42:50
start dealing with as
42:51
SPEAKER_01: much,
42:51
SPEAKER_01: um,
42:52
SPEAKER_01: scarring, doing damage to brain tissue, all of that stuff.
42:55
SPEAKER_01: And then when you're...
42:56
SPEAKER_01: So when you're talking about your form of stimulation,
42:59
SPEAKER_01: it's just you're sending in ultrasound pulses.
43:03
SPEAKER_01: That's right.
43:03
SPEAKER_01: And you can... You aim them with what?
43:07
SPEAKER_01: Like software on this chip?
43:09
SPEAKER_01: Software on the chip.
43:09
SPEAKER_03: Okay.
43:10
SPEAKER_03: So this is what's called a phase array.
43:11
So in the current instantiation of the chip,
43:14
we have about 9,000,
43:14
8,960 elements.
43:15
SPEAKER_03: Each one of those can vibrate at ultrasonic frequencies.
43:20
And so you can take this big 2D array and I can energize one
43:24
side and the other
43:25
side at the same time and then stair-step my way into the
43:27
middle and that will
43:28
SPEAKER_03: create a focus beam.
43:29
SPEAKER_03: If I do it the opposite,
43:30
I can create a diverging wave or a plane wave or any shape that
43:33
I want and I can
43:34
holographically construct deep into the brain whatever pattern
43:38
I really need to.
43:40
SPEAKER_01: And so the idea is,
43:40
SPEAKER_01: like,
43:41
SPEAKER_01: if you're depressed,
43:42
SPEAKER_01: you have maybe some neural pathways that are overactive or dormant,
43:48
SPEAKER_01: and this is helping break up that sort of fixation of these neurons.
43:56
SPEAKER_01: I don't know.
43:57
SPEAKER_01: I'm sure I did a terrible job.
43:58
SPEAKER_03: No, no, no, that's fine.
44:00
SPEAKER_03: Yeah,
44:00
SPEAKER_03: so actually,
44:00
SPEAKER_03: maybe just backing up a little bit too,
44:02
SPEAKER_03: I think one of the things that really stood out to myself and Tyson,
44:08
SPEAKER_03: my co-founder,
44:08
and the entire Forest team and people that are kind of aware of
44:12
this is like,
44:13
SPEAKER_03: these electrodes are amazing.
44:15
SPEAKER_03: You put them in, the person can control a robotic arm.
44:19
But there's this really tricky problem where a patient will
44:22
come to you and say,
44:23
SPEAKER_03: you know,
44:24
SPEAKER_03: hey,
44:24
I really love coming in and working with you guys and
44:26
controlling the robotic arm.
44:27
SPEAKER_03: It's fun.
44:28
SPEAKER_03: But I'm also just depressed, man.
44:31
SPEAKER_03: I'm sad.
44:32
SPEAKER_03: And that's really common among people with spinal cord injury.
44:36
SPEAKER_03: And he's like, can the implant do anything for me?
44:39
SPEAKER_03: And it's like, well, no.
44:41
You'd either need to have a whole other surgery with a totally
44:43
different type of
44:44
SPEAKER_03: investigative device,
44:45
SPEAKER_03: probably a deep brain stimulator.
44:47
SPEAKER_03: And, you know, that's an added risk, totally different thing.
44:51
And so there's just not a transferability or scalability of
44:55
those types of implants.
44:56
SPEAKER_03: And so,
44:57
SPEAKER_03: kind of going back to the drawing board,
45:00
SPEAKER_03: when I was finishing up my PhD,
45:02
SPEAKER_03: I was sort of disillusioned.
45:04
SPEAKER_03: I'd use everything under the sun.
45:05
SPEAKER_03: I'd worked with patients.
45:06
SPEAKER_03: And when you work with patients,
45:07
SPEAKER_03: unless you're sociopathic,
45:08
SPEAKER_03: you're like,
45:09
SPEAKER_03: man,
45:09
SPEAKER_03: you become really motivated to like,
45:11
SPEAKER_03: I want something that works.
45:13
SPEAKER_03: And I became disillusioned with EG,
45:15
SPEAKER_03: MEG,
45:16
SPEAKER_03: SEG,
45:17
SPEAKER_03: MRI,
45:18
SPEAKER_03: electrode arrays,
45:19
SPEAKER_03: and I was just like,
45:20
SPEAKER_03: what would work?
45:21
SPEAKER_03: And ultrasound physics are just kind of cool.
45:24
SPEAKER_03: They go through soft tissue.
45:25
SPEAKER_03: Ultrasound waves go through soft tissue really well,
45:28
SPEAKER_03: and this is why we can use it to look at an unborn baby,
45:30
SPEAKER_03: right?
45:32
And it also turns out that their resolution is
45:36
diffraction-limited by the size of
45:37
SPEAKER_03: the wave to about hundreds of microns,
45:40
SPEAKER_03: which is 100 microns is about the width of a human hair.
45:42
SPEAKER_03: It's incredibly precise.
45:45
And it really just hadn't been used in the brain because it
45:47
doesn't go through bone.
45:49
SPEAKER_03: But if you're willing to remove the bone,
45:51
SPEAKER_03: which the neurotech industry is,
45:53
SPEAKER_03: then you have access to this whole space.
45:56
SPEAKER_03: And all of these things that are big societal problems,
46:01
SPEAKER_03: like we mentioned,
46:01
SPEAKER_03: depression,
46:02
SPEAKER_03: OCD,
46:02
SPEAKER_03: bipolar disorder,
46:04
It's something like 21 million people that have severe and
46:08
treatment-resistant,
46:09
SPEAKER_03: meaning they failed three or more drugs,
46:11
SPEAKER_03: forms of those.
46:12
SPEAKER_03: So these are debilitating just, like, people out of the workforce.
46:16
SPEAKER_03: They're depressed.
46:16
They're,
46:17
like,
46:17
borderline suicidal or so compulsive that they just can't live
46:20
a normal life.
46:22
SPEAKER_03: And, like, we're just, like, failing them.
46:24
SPEAKER_03: Every day, we don't create something like this.
46:26
And ultrasound felt like such a good match,
46:28
and there was just a lack of research in that area.
46:33
So the idea is if you can bring ultrasound in,
46:36
we know that it can monitor.
46:38
And that was really the work I did at Caltech was showing that
46:40
it had enough
46:40
SPEAKER_03: sensitivity to monitor function of the brain.
46:43
SPEAKER_03: And then going back to like 50s and 60s,
46:45
we know that you can then focus that beam to actually
46:49
upregulate or downregulate
46:50
SPEAKER_03: neural activity.
46:51
SPEAKER_03: And it's more of like a bump rather than like a precise thing today.
46:55
SPEAKER_03: And Forrest wants to take that further.
46:59
SPEAKER_03: And so,
46:59
SPEAKER_03: like you said,
47:00
SPEAKER_03: going back to depression,
47:01
SPEAKER_03: your original question,
47:02
if you see a biomarker that is suggestive of or correlated with
47:07
that person's
47:07
SPEAKER_03: depressive state,
47:09
SPEAKER_03: can you then stimulate parts of the brain,
47:11
SPEAKER_03: much of which we're learning from deep brain stimulators,
47:13
SPEAKER_03: the electrode kind,
47:15
SPEAKER_03: by the way,
47:15
SPEAKER_03: as a good starting point.
47:17
Can we do that in a less invasive way and a way where we can
47:22
retarget every day?
47:23
So then if that person that was implanted originally for
47:25
epilepsy is like,
47:27
SPEAKER_03: hey,
47:27
SPEAKER_03: you know,
47:27
SPEAKER_03: I'm also depressed,
47:28
SPEAKER_03: we go,
47:28
SPEAKER_03: oh,
47:29
SPEAKER_03: we can solve that for you.
47:30
SPEAKER_03: It's a change in software.
47:31
SPEAKER_03: We don't need any change to the hardware device itself.
47:34
SPEAKER_03: So it's really a paradigm shift
47:36
SPEAKER_03: where one device can do many, many things.
47:40
SPEAKER_03: And that's, like, actually the change that Forrest cares about.
47:43
SPEAKER_03: Like, ultrasound, whatever, those are all just tools.
47:45
SPEAKER_03: What we want is the one device that can do many things.
47:47
SPEAKER_03: I see.
47:48
SPEAKER_03: Okay.
47:49
SPEAKER_01: And as far as, like,
47:52
SPEAKER_01: Learning about this and proving it out,
47:54
SPEAKER_01: um,
47:55
SPEAKER_01: you and I talk about things outside of this podcast,
47:58
SPEAKER_01: so I don't wanna,
47:58
I don't wanna overstep on some things that we've talked about
48:01
in the past.
48:02
SPEAKER_01: What,
48:02
SPEAKER_01: what,
48:02
SPEAKER_01: what can we say about what stage you guys are at,
48:06
SPEAKER_01: um,
48:06
SPEAKER_01: with,
48:07
SPEAKER_01: with,
48:07
SPEAKER_01: with trials and,
48:08
SPEAKER_01: and where,
48:09
SPEAKER_01: what you've learned?
48:10
SPEAKER_03: Yeah, so we have two lines of hardware stood up.
48:15
SPEAKER_03: Actually, much of which was created in this building.
48:17
SPEAKER_01: Yeah, we'll give Playground Global a shout-out.
48:21
SPEAKER_01: So my podcast studio is in Playground Global's offices,
48:24
SPEAKER_01: and they've been a backer of the company?
48:29
SPEAKER_01: No,
48:30
SPEAKER_03: they are also a venture capital firm,
48:32
SPEAKER_03: but we are actually a non-profit,
48:34
SPEAKER_03: so we only accept donations.
48:36
SPEAKER_03: That being said,
48:37
SPEAKER_03: one of our anchor donors,
48:38
SPEAKER_03: Eric Schmidt,
48:40
SPEAKER_03: is also kind of part of the playground-iverse,
48:43
SPEAKER_03: let's call it.
48:44
And so they've been very helpful and kind and really helped us
48:47
out a lot in getting
48:48
SPEAKER_03: lab space set up and treated us almost like a portfolio company.
48:51
SPEAKER_01: Okay,
48:52
SPEAKER_01: okay,
48:52
SPEAKER_01: so yes,
48:53
SPEAKER_01: so some of the Playgrounds,
48:54
SPEAKER_01: some of the Force teams,
48:55
SPEAKER_01: that's not too far away from me.
48:57
SPEAKER_01: Yeah, okay, so there's two devices.
49:00
SPEAKER_03: Yep, two devices.
49:01
SPEAKER_03: And I think we have five approved non-significant risk IRB.
49:07
SPEAKER_03: So these are early feasibility studies in human beings.
49:10
SPEAKER_03: So I'm not sure how many hours we're up to,
49:13
SPEAKER_03: but,
49:13
SPEAKER_03: you know,
49:14
SPEAKER_03: several,
49:14
SPEAKER_03: probably a couple hundred hours worth of data recording now.
49:19
SPEAKER_03: And yeah, I can kind of jump into whatever level you want.
49:24
SPEAKER_01: Well, I mean, what have we learned?
49:26
SPEAKER_01: What have we seen?
49:27
SPEAKER_01: Is it working?
49:29
SPEAKER_02: It does work.
49:31
SPEAKER_03: Yeah,
49:32
SPEAKER_03: I won't name names,
49:33
SPEAKER_03: but actually someone at the company,
49:35
SPEAKER_03: they're right at the end of the pipeline,
49:37
SPEAKER_03: so that means they're the ones that see the data first.
49:40
SPEAKER_03: I actually wasn't there,
49:41
SPEAKER_03: which I'm really sad,
49:41
SPEAKER_03: but apparently she was just crying at her desk one day.
49:44
SPEAKER_03: And people are like, what's going on?
49:45
It was the very first data we'd ever collected from humans and
49:49
it just came out and
49:50
SPEAKER_03: it was like,
49:50
SPEAKER_03: boom,
49:51
SPEAKER_03: it works.
49:52
SPEAKER_03: And just to give you a sense there,
49:54
SPEAKER_03: when we started this company,
49:56
SPEAKER_03: which again is a nonprofit research org,
49:57
SPEAKER_03: we can talk about that.
50:00
SPEAKER_03: I would say the majority of people just said,
50:03
SPEAKER_03: like,
50:04
SPEAKER_03: hey,
50:04
SPEAKER_03: you know,
50:04
SPEAKER_03: it's a cool concept,
50:05
but there's no way you're ever gonna get this miniaturized
50:08
ultrasound on chip to
50:09
SPEAKER_03: get that level of sensitivity.
50:10
SPEAKER_03: Because up until that time,
50:12
SPEAKER_03: like,
50:12
SPEAKER_03: the stuff we used at Caltech,
50:13
SPEAKER_03: it's like two server racks full of equipment,
50:17
SPEAKER_03: you know,
50:18
SPEAKER_03: a million-dollar scanner,
50:19
SPEAKER_03: and you get out of that 128 channels,
50:21
SPEAKER_03: but it's really sensitive and very programmable.
50:24
SPEAKER_03: And everyone's like,
50:25
SPEAKER_03: yeah,
50:25
SPEAKER_03: that's great,
50:26
SPEAKER_03: but if you try to put that in this tiny little thing,
50:28
SPEAKER_03: it's really hard to make stuff small,
50:30
SPEAKER_03: right?
50:30
SPEAKER_03: Yeah, because you have to put ultrasound in this tiny device, yeah.
50:35
SPEAKER_03: Yeah,
50:35
SPEAKER_03: and we're like,
50:35
SPEAKER_03: we're gonna do a 9,000-channel array,
50:37
SPEAKER_03: it's gonna be the size of your thumb,
50:39
and it's gonna be more sensitive than any research scanner
50:41
you've ever seen.
50:42
SPEAKER_03: People were like,
50:43
SPEAKER_03: Okay, bro.
50:45
SPEAKER_03: I'll believe it when I see it.
50:47
SPEAKER_03: And so,
50:48
SPEAKER_03: for some of the very first images we got to just work,
50:51
SPEAKER_03: and they were by no means optimal.
50:52
SPEAKER_03: We've improved immensely since then.
50:55
SPEAKER_03: But yeah, we've detected functional signal in the brain.
50:58
SPEAKER_03: We've talked earlier about corticothalamic loops.
51:01
We've detected activity deep in the brain and on the surface of
51:05
the brain in concert.
51:07
SPEAKER_03: And then we've had clinically meaningful stuff as well.
51:10
SPEAKER_03: So I'll maybe just like tease this a little bit,
51:12
but one of the things that we'll want to share with the world
51:15
pretty soon was
51:16
SPEAKER_03: actually looking at the brain states of people who are in a coma.
51:21
SPEAKER_03: And we were able to detect this phenomenon called covert consciousness,
51:26
SPEAKER_03: where a person is functionally unable to move,
51:28
SPEAKER_03: disconnected from the outside world.
51:29
They're in a coma,
51:30
so their family's asking the kind of usual questions like,
51:32
hey,
51:33
SPEAKER_03: Are they going to wake up?
51:34
SPEAKER_03: Can they hear me?
51:35
SPEAKER_03: Should I sit here and talk to them?
51:36
SPEAKER_03: Is it worth my time?
51:38
SPEAKER_03: And we know from fMRI,
51:40
SPEAKER_03: so credit where credit's due,
51:43
SPEAKER_03: fMRI literature tells us that in some of those people,
51:46
you can actually detect brain states that indicate that they
51:49
can actually hear you
51:50
and they're interpreting what you're saying and in some cases
51:52
can even respond
51:54
SPEAKER_03: through simple brain activity.
51:57
SPEAKER_03: So you say like, hey, the first study was actually like this.
52:00
SPEAKER_03: It was like, if your name is Susan, then please imagine playing tennis.
52:05
Like,
52:06
try to hit a tennis ball as hard as you can,
52:08
because she was a tennis player.
52:10
And they would see this big spike in activity in the part of
52:12
your brain that would
52:13
be active if you were trying to move your right arm or whatever
52:16
her dominant arm
52:16
SPEAKER_03: was.
52:17
Um,
52:17
and so you can create rudimentary communication with someone
52:20
who's literally in a coma.
52:21
SPEAKER_03: How long has that been going on?
52:24
SPEAKER_03: I want to say that... I don't remember exactly.
52:26
SPEAKER_03: It's like 2004, 2014.
52:27
SPEAKER_03: Okay.
52:27
SPEAKER_01: I feel like this is like one of these big questions.
52:30
SPEAKER_01: This is like everybody's nightmare, right?
52:32
SPEAKER_01: Is that you're in a coma, and you can't communicate out.
52:35
And I think the reason that you don't hear as much about this
52:39
is because it's
52:40
SPEAKER_03: amazing,
52:41
SPEAKER_03: but it's not super actionable.
52:43
And the reason why is because a person who's in a coma might be
52:45
in a neuro ICU,
52:47
taking them off life support or wheeling their life support
52:51
somehow into an MRI
52:52
SPEAKER_03: scanner is so difficult.
52:55
SPEAKER_03: And running that study, it's functional MRI.
52:57
SPEAKER_03: This isn't like your normal neighborhood doctor's going to do it.
53:00
This is like you need a specialized team that's going to do it,
53:03
analyze the data.
53:04
SPEAKER_03: Really, it's a science project.
53:07
SPEAKER_03: And so,
53:07
SPEAKER_03: although there's proofs of concept,
53:09
SPEAKER_03: it's super far away from being kind of standardized clinical care.
53:13
SPEAKER_03: But here now,
53:14
SPEAKER_03: for us,
53:16
we have this little tiny device that we can just stick on the
53:19
head bedside.
53:20
SPEAKER_03: It's a wearable, but again, only a wearable if you're missing your skull.
53:24
But for many people who are in a coma,
53:26
if they've had a traumatic brain injury,
53:28
SPEAKER_03: that's put them in the coma in the first place,
53:30
they'll have a decompressive craniectomy where they actually
53:32
remove a part of your
53:33
SPEAKER_03: skull to allow for swelling to happen safely.
53:38
SPEAKER_03: And during that time,
53:39
there's an unobstructed view into the brain that's just through
53:42
soft skin,
53:43
SPEAKER_03: scalp,
53:44
SPEAKER_03: and right into the brain.
53:45
SPEAKER_03: So we can actually monitor the brain that way.
53:48
SPEAKER_03: And so,
53:49
SPEAKER_03: we were actually able to detect in,
53:51
SPEAKER_03: I think,
53:52
SPEAKER_03: one of the very first patients we had,
53:55
SPEAKER_03: not only the kind of audio listening parts of her brain,
53:59
but the parts that actually interpret speech active whenever we
54:03
would talk to her.
54:04
SPEAKER_03: And it was not us talking, it was recording.
54:06
SPEAKER_03: And then interestingly,
54:07
SPEAKER_03: if you take that same recording and play it backward,
54:10
SPEAKER_03: so now it's kind of like it's the same volume,
54:12
SPEAKER_03: same waveform,
54:13
SPEAKER_03: so mathematically it's kind of the same thing,
54:15
SPEAKER_03: but it's unintelligible,
54:16
because you can't really understand when something's being
54:18
played backwards.
54:20
SPEAKER_03: Then we'd see that signal disappear,
54:22
so we could actually watch these speech interpretation areas
54:25
ramp up every time she
54:26
SPEAKER_03: was hearing spoken language.
54:29
SPEAKER_01: Yeah.
54:31
SPEAKER_01: Okay.
54:31
And that means,
54:32
like,
54:34
so she was an active participant on some level then.
54:38
SPEAKER_01: Yeah, yeah.
54:38
SPEAKER_01: I mean, it's not like... I didn't ask that question very well.
54:41
SPEAKER_01: I mean, if you're...
54:45
If you're playing it backwards and it's sort of nonsense and
54:47
your brain's not
54:48
SPEAKER_01: responding,
54:48
SPEAKER_01: I mean,
54:49
SPEAKER_01: I would still hear something.
54:50
SPEAKER_01: Wouldn't I be trying to, like, decode it?
54:53
SPEAKER_03: You might, yeah, but it depends on what area of the brain you're in.
54:56
SPEAKER_03: So,
54:57
your kind of auditory cortex is just gonna be going crazy
55:00
because it's still
55:01
SPEAKER_03: hearing sound and doing its thing.
55:03
SPEAKER_03: But your language interpretation areas are kind of like,
55:06
SPEAKER_03: they're downstream of that,
55:07
SPEAKER_03: and they don't really have anything to interpret.
55:09
So,
55:10
you know,
55:10
somewhere in this processing pipeline,
55:13
things are getting cut down.
55:14
So I think that's what I was trying to fish out,
55:16
if it was passive or active.
55:20
SPEAKER_03: I think that might turn into a philosophical question.
55:23
SPEAKER_03: Somewhere in between those, yeah.
55:25
SPEAKER_03: I mean,
55:25
SPEAKER_03: I think what it suggests is that on some level,
55:28
SPEAKER_03: functionally speaking,
55:29
SPEAKER_03: she was hearing and interpreting the people around her.
55:32
SPEAKER_03: And therefore,
55:33
SPEAKER_03: actionably,
55:36
SPEAKER_03: for her family,
55:37
SPEAKER_03: it might make more sense to spend time chatting with her.
55:40
SPEAKER_03: And not only that,
55:41
SPEAKER_03: but people who...
55:43
SPEAKER_03: We know this from these great fMRI studies,
55:45
and there was a great one in New England Journal of Medicine
55:47
just last year that
55:48
SPEAKER_03: kind of concatenated all of them and said,
55:49
SPEAKER_03: oh,
55:50
SPEAKER_03: across these 400 or so people that have been in a study like this,
55:54
SPEAKER_03: those that had cognitive motor dissociation,
55:56
SPEAKER_03: this kind of covert consciousness effect,
56:00
SPEAKER_03: were,
56:00
SPEAKER_03: I forget the exact number,
56:01
SPEAKER_03: but much more likely to wake up,
56:04
SPEAKER_03: and their post-wake-up recovery was much more substantial.
56:08
So these are...
56:09
It tells you something actionable about the course of action.
56:12
I mean,
56:12
the scary part about this,
56:13
like you said,
56:14
the nightmarish part about this,
56:16
is the idea that someone's gonna pull the plug while you're
56:18
sitting there listening
56:19
SPEAKER_03: to them talk about it,
56:21
SPEAKER_03: right?
56:21
SPEAKER_03: And that's the number one cause of death in these folks,
56:23
because about 25% of people in a coma do have cognitive-motor
56:28
dissociation,
56:28
SPEAKER_03: we think.
56:29
SPEAKER_01: Okay,
56:29
SPEAKER_01: that's what I was gonna ask,
56:31
SPEAKER_01: is if you haven't sustained some kind of tremendous brain damage,
56:37
SPEAKER_01: like,
56:37
SPEAKER_01: do we think that most people in a coma then are...
56:42
SPEAKER_01: are alert to all this, yeah.
56:45
SPEAKER_03: Not most, but 25% is a good chunk.
56:48
You're talking thousands of people every year that are at risk
56:51
of getting a plug
56:52
SPEAKER_03: pulled that shouldn't.
56:53
SPEAKER_01: Like,
56:53
SPEAKER_01: the positive seems to be that you could flag this to the family,
56:57
SPEAKER_01: you could have people then spending more time talking to you,
57:00
SPEAKER_01: reading to you.
57:02
SPEAKER_01: Obviously,
57:02
the horrible downside is that you know this and you still can't
57:08
kind of get things
57:09
SPEAKER_01: out.
57:10
SPEAKER_03: Yeah, maybe I'm oversharing, but whatever.
57:17
We've been internally calling this project Orpheus because
57:20
Orpheus in Greek
57:23
SPEAKER_03: mythology goes down into the underworld to rescue his wife who had been
57:28
SPEAKER_03: had died.
57:29
SPEAKER_03: And so he's communicating with someone in the underworld.
57:32
SPEAKER_03: And so that's kind of like where that's coming from.
57:34
And so we want to eventually get to the point that we can
57:36
actually allow them to
57:37
SPEAKER_03: communicate back to us from their coma state.
57:40
SPEAKER_03: And then later down the road,
57:42
SPEAKER_03: which is,
57:42
SPEAKER_03: you know,
57:43
SPEAKER_03: we haven't really started on this yet,
57:44
SPEAKER_03: but maybe jokingly,
57:47
SPEAKER_03: maybe not called Project Lazarus,
57:48
SPEAKER_03: which is like,
57:48
SPEAKER_03: can you actually stimulate the brain?
57:51
SPEAKER_03: in a way that would bring that person back.
57:53
SPEAKER_03: And because now you have this closed-loop feedback,
57:55
SPEAKER_03: can you listen to the brain,
57:57
monitor these levels of consciousness on some of our biomarkers
58:01
of consciousness,
58:02
SPEAKER_03: and then try to stimulate them to come higher and higher and higher?
58:06
SPEAKER_03: And there's some early evidence to suggest that that may be possible.
58:09
SPEAKER_03: Oh, my God.
58:10
SPEAKER_01: Well, do you remember the name of that podcast I sent you about on?
58:13
SPEAKER_03: Yeah, yeah, I listened to, I'm like halfway through.
58:16
SPEAKER_01: Okay, wait, vamp for a second.
58:17
SPEAKER_01: I'm gonna look this up.
58:20
So you're doing these studies,
58:24
and...
58:26
Now that I feel on the spot,
58:27
I have nothing to say.
58:29
SPEAKER_01: You gotta carry me here.
58:32
SPEAKER_03: Yeah, so I mean, this is really just... Blink, blink.
58:36
SPEAKER_01: Blink, that's it, yeah.
58:39
SPEAKER_01: I don't think it's like,
58:41
SPEAKER_01: when I first recommended it to you,
58:43
SPEAKER_01: now that I've listened to the whole thing,
58:45
SPEAKER_01: I'm not sure it's like fully apples to apples,
58:47
SPEAKER_01: because this guy,
58:47
SPEAKER_01: he basically has this...
58:51
SPEAKER_01: I'm not gonna divulge too much of the story,
58:53
SPEAKER_01: because it's crazy,
58:54
SPEAKER_01: and it has a juicy ending.
58:56
SPEAKER_01: But he has this terrible reaction to smoking heroin,
59:00
SPEAKER_01: and so it's like this chemical-induced coma.
59:04
SPEAKER_01: But...
59:05
SPEAKER_01: basically he's in this coma for a really long time.
59:08
SPEAKER_01: And then he comes out and he was like,
59:10
SPEAKER_01: yeah,
59:10
SPEAKER_01: I was awake and I saw like everything you guys were doing to me.
59:14
SPEAKER_01: And I just heard people fighting all around me and then,
59:17
SPEAKER_01: and,
59:18
SPEAKER_01: um,
59:18
and so I feel like his is a little bit different because of the
59:21
circumstances,
59:22
SPEAKER_01: but yeah,
59:22
SPEAKER_01: I mean,
59:23
SPEAKER_01: I was just like,
59:23
SPEAKER_01: oh my God,
59:24
SPEAKER_01: if we could,
59:25
SPEAKER_01: um,
59:27
SPEAKER_01: if you made progress on this front,
59:28
SPEAKER_01: it seems like such a fascinating scientific avenue.
59:32
SPEAKER_03: Yeah,
59:32
SPEAKER_03: and I think,
59:32
SPEAKER_03: you know,
59:33
SPEAKER_03: just to highlight too,
59:34
SPEAKER_03: because we're this focused research organization,
59:37
SPEAKER_03: a non-profit group,
59:39
SPEAKER_03: and I mentioned earlier,
59:40
SPEAKER_03: like,
59:40
SPEAKER_03: the real thrust here is one device,
59:42
SPEAKER_03: many applications.
59:44
SPEAKER_03: It's not that we're fundamentally only interested in this,
59:47
SPEAKER_03: but it was pretty exciting that one of the first indications that we were
59:50
SPEAKER_03: interested in,
59:51
SPEAKER_03: like,
59:52
SPEAKER_03: boom,
59:52
SPEAKER_03: right off the bat,
59:53
SPEAKER_03: just showed something so promising.
59:55
And so I hope that ultrasound-based neurotechnology can help in
59:59
that space.
1:00:00
SPEAKER_03: And at the same time,
1:00:01
SPEAKER_03: we're learning,
1:00:02
SPEAKER_03: you know,
1:00:03
we just partnered with ARIA and a group in the UK where we're
1:00:07
kind of starting to
1:00:08
go down the route for depression to understand and eventually
1:00:11
treat depression.
1:00:11
SPEAKER_03: So that's a neurostimulation study.
1:00:13
SPEAKER_03: And we're interested in a bunch of other indications as well.
1:00:16
SPEAKER_03: I mentioned we have five of these IRBs.
1:00:17
Each one's a little different in how we use this with really
1:00:20
the goal to show like,
1:00:22
SPEAKER_03: yeah,
1:00:22
SPEAKER_03: not only can it do this thing in coma,
1:00:24
SPEAKER_03: but it can do this thing in depression and do this other thing.
1:00:26
SPEAKER_03: Okay.
1:00:27
SPEAKER_01: Okay.
1:00:28
I'm going to ask you what's like maybe an unfair question
1:00:31
because I don't think it
1:00:32
SPEAKER_01: has a black and white answer and I sort of wanted to,
1:00:34
SPEAKER_01: but,
1:00:35
SPEAKER_01: um,
1:00:36
SPEAKER_01: Okay, we've got Neuralink doing surgery into the brain.
1:00:40
SPEAKER_01: We've got companies like Precision Neuro doing this layer of electrodes.
1:00:46
I think they are under the dura,
1:00:49
but they're not going like into the tissue.
1:00:52
SPEAKER_01: They're kind of like sitting on the brain.
1:00:54
Synchron goes through a blood vessel in your neck,
1:00:58
almost like a stent into your brain.
1:01:01
SPEAKER_01: It kind of hangs out in the blood vessel.
1:01:04
SPEAKER_01: Downside is the resolutions, not
1:01:07
SPEAKER_01: Like,
1:01:08
SPEAKER_01: everything has this trade-off,
1:01:09
SPEAKER_01: everything...
1:01:10
We're also in this spot where almost all these devices right
1:01:14
now are being used for
1:01:16
SPEAKER_01: people in quite dire straits who might have ALS,
1:01:20
SPEAKER_01: or they might be paralyzed.
1:01:21
SPEAKER_01: You know, we're seen quite far from, like, the consumer.
1:01:26
SPEAKER_01: We're merging with the machine as a general populace era.
1:01:31
SPEAKER_01: And...
1:01:32
And so I get bombarded with all these different approaches and
1:01:36
everybody makes a
1:01:36
SPEAKER_01: pretty strong case to me on some level about why,
1:01:42
SPEAKER_01: like,
1:01:42
it feels like we're heading into this world where we're
1:01:44
treating people who are
1:01:46
SPEAKER_01: sick in very specific ways where you have,
1:01:50
SPEAKER_01: you've got this one implant that
1:01:53
SPEAKER_01: It's like,
1:01:53
SPEAKER_01: okay,
1:01:53
SPEAKER_01: well,
1:01:53
SPEAKER_01: for your condition,
1:01:55
SPEAKER_01: you kind of want the less invasive one,
1:01:57
SPEAKER_01: you know,
1:01:57
SPEAKER_01: and then for this one,
1:01:58
SPEAKER_01: yours a little more serious.
1:01:59
SPEAKER_01: We're going to go a little bit deeper,
1:02:01
SPEAKER_01: but that also feels like we have almost like too many options or,
1:02:05
SPEAKER_01: and I know it's early days.
1:02:06
And so this is all shaking out,
1:02:08
but I was just so curious,
1:02:09
like how you think this plays out.
1:02:11
SPEAKER_01: I mean,
1:02:12
do you end up with one device that does everything or do we end
1:02:17
up with like six
1:02:18
SPEAKER_01: different implants for,
1:02:19
SPEAKER_01: for 20 different conditions?
1:02:23
SPEAKER_03: Yeah, I think we end up with one device that can do everything.
1:02:27
SPEAKER_03: But it's not tomorrow.
1:02:28
SPEAKER_03: And that's why I think all of these devices you just mentioned,
1:02:32
SPEAKER_03: every one of them has a place right now.
1:02:36
You mentioned also that there's a trade-off between
1:02:39
invasiveness and performance.
1:02:40
SPEAKER_03: The fact is,
1:02:41
SPEAKER_03: the closer you get to the neuron today with an electrode,
1:02:44
SPEAKER_03: when you're talking about electricity,
1:02:46
SPEAKER_03: the closer you get to the neuron,
1:02:47
the more precise the signal is because you're getting closer to
1:02:49
the origin of the
1:02:49
SPEAKER_03: signal itself.
1:02:51
SPEAKER_03: The further you get away,
1:02:52
SPEAKER_03: so you move out of the deep brain,
1:02:53
SPEAKER_03: you move into the cortex,
1:02:54
SPEAKER_03: eventually move on top of the cortex like precision,
1:02:56
SPEAKER_03: or you move epidural,
1:02:59
SPEAKER_03: or you move into the vasculature like Synchron is doing,
1:03:02
SPEAKER_03: you get kind of further and further away.
1:03:04
So performance degrades,
1:03:05
but you get this invasiveness trade-off or less invasive
1:03:09
trade-off.
1:03:11
And I think that if you are beholden to that electrophysiology
1:03:16
phenomenon,
1:03:17
SPEAKER_03: that trade-off is pretty hard to break.
1:03:20
SPEAKER_03: You're not going to jump off of that line very easily.
1:03:23
SPEAKER_03: So I've been calling these like generation one or level one interfaces.
1:03:28
These are the first generation of commercialized neurotech that
1:03:32
I think are really
1:03:33
SPEAKER_03: going to make an impact on people's lives.
1:03:35
SPEAKER_03: But at the same time,
1:03:36
SPEAKER_03: what you're seeing is,
1:03:38
SPEAKER_03: if you look a little further down the line,
1:03:41
SPEAKER_03: entirely new types of physics and approaches being used,
1:03:44
SPEAKER_03: ultrasound being one of them,
1:03:46
SPEAKER_03: that start to break that mold a little bit and go,
1:03:49
we're actually going to put aside the biophysics of electricity
1:03:52
for a second,
1:03:53
SPEAKER_03: and let's look at other mechanisms by which we can detect neurofunction.
1:03:57
SPEAKER_03: And now we can do it with remote sensing.
1:03:59
SPEAKER_03: And I think that whether it's ultrasound or it's light or it's magnetism,
1:04:05
SPEAKER_03: remote sensing is kind of the way forward.
1:04:08
SPEAKER_03: There's this great concept by Marcus Meister called the butcher number,
1:04:11
SPEAKER_03: which is kind of a brutal way to say it,
1:04:13
SPEAKER_03: but it's the number of neurons that you have to kill to record from one.
1:04:19
And so the Utah array,
1:04:20
which we talked a lot about,
1:04:21
has a butcher number of something like 200.
1:04:23
So you killed 200 neurons putting this spike into the brain in
1:04:26
your cord for one at the bottom.
1:04:28
SPEAKER_03: So it's fundamentally not scalable.
1:04:30
SPEAKER_03: You're going to end up damaging a lot of brain to put a lot of these in.
1:04:32
SPEAKER_01: Definitely not elective or something like that.
1:04:34
SPEAKER_03: Yeah.
1:04:36
SPEAKER_03: And then,
1:04:36
SPEAKER_03: you know,
1:04:37
SPEAKER_03: Neuralink brings that down and Neuropixels,
1:04:39
SPEAKER_03: which is another more kind of R&D technology,
1:04:41
SPEAKER_03: brings that down to probably single digits.
1:04:43
SPEAKER_03: So it's like three, four, something like that.
1:04:45
SPEAKER_03: Neuropixels is.
1:04:46
SPEAKER_03: Neuralink's never published theirs, but...
1:04:48
SPEAKER_01: Yeah, I think they're a little secretive on that.
1:04:51
I don't want to butcher this because I'll get in trouble if I
1:04:57
get it wrong.
1:04:57
SPEAKER_01: But,
1:04:58
SPEAKER_01: you know,
1:04:58
SPEAKER_01: I hear them cite like 99.999% are untouched,
1:05:05
SPEAKER_01: you know,
1:05:06
SPEAKER_01: of the neurons as you're going through.
1:05:07
SPEAKER_01: But I don't know how that translates into the butcher number.
1:05:10
SPEAKER_03: Yeah,
1:05:11
SPEAKER_03: I mean,
1:05:11
SPEAKER_03: yeah,
1:05:12
SPEAKER_03: I also will get in trouble if I...
1:05:15
SPEAKER_03: So we'll just say that it's good,
1:05:17
SPEAKER_03: but it's non-zero.
1:05:19
SPEAKER_03: And fundamentally,
1:05:20
SPEAKER_03: if you need something to scale to whole-brain structures,
1:05:24
SPEAKER_03: then you need a butcher number of zero.
1:05:27
SPEAKER_03: And,
1:05:28
SPEAKER_03: you know,
1:05:28
SPEAKER_03: there are ways of,
1:05:29
SPEAKER_03: like you said,
1:05:30
SPEAKER_03: like sliding something on top of the brain,
1:05:32
SPEAKER_03: but kind of like underneath the dura,
1:05:35
SPEAKER_03: but there's complications and risk of infection and everything else.
1:05:38
SPEAKER_03: And so,
1:05:39
SPEAKER_03: for me,
1:05:40
SPEAKER_03: the long-term,
1:05:41
SPEAKER_03: this kind of generation two interfaces are a departure from traditional
1:05:45
SPEAKER_03: electrophysiology.
1:05:46
SPEAKER_01: Okay, like, because when you're saying remote sensing, you mean what?
1:05:49
SPEAKER_01: You mean...
1:05:50
I mean something that sits on top of the brain or outside of
1:05:53
the brain tissue itself.
1:05:54
SPEAKER_03: So you're not actually touching the brain tissue.
1:05:57
SPEAKER_03: And you're sending energy into the brain tissue.
1:06:00
SPEAKER_03: And that energy can be light, magnetism, sound, et cetera.
1:06:04
SPEAKER_01: Okay,
1:06:05
SPEAKER_01: so you think if we use these different approaches,
1:06:06
we will get enough of the information we need to not need to
1:06:14
sit right next to a
1:06:15
SPEAKER_01: neuron to do amazing things.
1:06:18
SPEAKER_03: That's right, yeah.
1:06:19
And so I mentioned earlier there's these 21 million people with
1:06:22
psychiatric and
1:06:22
SPEAKER_03: cognitive disorders,
1:06:23
SPEAKER_03: and that's just your treatment-resistant severe cases.
1:06:26
SPEAKER_03: That's like the worst of the worst.
1:06:28
SPEAKER_03: Well,
1:06:28
21 million is a lot bigger than the roughly 1.5 million people
1:06:32
that are living with
1:06:33
SPEAKER_03: severe forms of paralysis.
1:06:35
So just the opportunity size to help people grows tremendously
1:06:39
as you move beyond
1:06:40
SPEAKER_03: motor neuroprosthetics.
1:06:42
But you need whole brain coverage,
1:06:44
or not whole,
1:06:45
but a lot of the brain to do this.
1:06:47
SPEAKER_01: So multiple devices.
1:06:49
SPEAKER_03: So either you have a ton of Neuralinks all over the place,
1:06:52
SPEAKER_03: and actually,
1:06:53
SPEAKER_03: I'm not even sure that would work because you have to go deep,
1:06:55
SPEAKER_03: although there are efforts working on that too.
1:06:58
SPEAKER_03: I'll be careful again.
1:07:00
SPEAKER_02: Go on.
1:07:04
SPEAKER_03: But you're going to turn the skull into Swiss cheese trying to do that.
1:07:07
SPEAKER_03: I think the Neuralink,
1:07:08
SPEAKER_03: we back of the napkin right now,
1:07:11
SPEAKER_03: despite being a technological tour de force,
1:07:13
SPEAKER_03: getting nothing to take away from them,
1:07:14
SPEAKER_03: but covers about a sixth of 1% of surface cortex.
1:07:19
That's to say nothing of all these sulcal folds or deep brain
1:07:22
structures that we
1:07:22
SPEAKER_03: know are so intimately involved in all of these psychiatric disorders.
1:07:27
SPEAKER_03: So if you want to get to whole brain,
1:07:29
SPEAKER_03: you need something that is stable over long periods of time,
1:07:31
SPEAKER_03: covers huge amounts of the brain,
1:07:33
and the physics will tell you that electricity just doesn't get
1:07:36
you there unless
1:07:37
SPEAKER_03: you're literally,
1:07:38
SPEAKER_03: you know,
1:07:39
SPEAKER_03: just implanting all over the place.
1:07:41
SPEAKER_01: Okay, so is... And this looks like a helmet or a...
1:07:47
SPEAKER_01: Or it's just multiple devices.
1:07:51
SPEAKER_03: Well, with remote sensing, that's what's cool, right?
1:07:53
SPEAKER_03: So with ultrasound,
1:07:54
SPEAKER_03: we can put one device on one part of the skull here,
1:07:58
SPEAKER_03: but we can send out a beam just about anywhere in the skull.
1:08:01
SPEAKER_03: And then we can electronically steer this thing and sweep it around.
1:08:05
SPEAKER_03: And I mentioned earlier,
1:08:06
SPEAKER_03: with one device,
1:08:06
SPEAKER_03: we're getting today something like 20% of the human brain by volume.
1:08:11
SPEAKER_03: So,
1:08:11
SPEAKER_03: if you wanted to cover,
1:08:12
SPEAKER_03: you know,
1:08:13
SPEAKER_03: let's call it 75%,
1:08:15
SPEAKER_03: covering 100 is always going to be hard,
1:08:17
SPEAKER_03: because they're just kind of edge cases and funny geometry.
1:08:20
But something like four devices is going to get you awfully
1:08:22
close to everything you
1:08:23
SPEAKER_03: could possibly need.
1:08:24
SPEAKER_03: Okay.
1:08:26
SPEAKER_01: So,
1:08:26
SPEAKER_01: kind of,
1:08:26
SPEAKER_01: I mean,
1:08:26
SPEAKER_01: I don't think it fully worked the way intended,
1:08:29
SPEAKER_01: but,
1:08:30
SPEAKER_01: like,
1:08:30
SPEAKER_01: Kernel's kind of like,
1:08:33
SPEAKER_01: I mean,
1:08:34
SPEAKER_01: conceptually in the ballpark of where some of that might be heading.
1:08:39
SPEAKER_03: Yeah.
1:08:42
SPEAKER_03: So,
1:08:42
SPEAKER_03: for those that don't know,
1:08:43
SPEAKER_03: the kernel headset is non-invasive,
1:08:45
SPEAKER_03: sits on top of the head entirely,
1:08:47
SPEAKER_03: and uses functional near-infrared spectroscopy.
1:08:50
SPEAKER_03: So, what does that mean?
1:08:51
SPEAKER_03: You're sending light down into the brain.
1:08:53
SPEAKER_03: And so,
1:08:54
SPEAKER_03: again,
1:08:54
remote sensing is kind of going down through the skull and
1:08:58
scalp and through the
1:08:59
SPEAKER_03: dura and into the brain,
1:09:00
SPEAKER_03: and it's scattering all over the place.
1:09:02
SPEAKER_03: But near-infrared scatters less than other wavelengths.
1:09:05
SPEAKER_03: And so you have what's called an optode.
1:09:07
So you have a source that's sending that laser light down in,
1:09:10
and then you have a detector.
1:09:13
SPEAKER_03: And so most of those photons scatter all and kind of disappear,
1:09:16
SPEAKER_03: but some of them happen to scatter back to the detector in this kind of
1:09:19
SPEAKER_03: probabilistic,
1:09:20
SPEAKER_03: almost banana-shaped curve.
1:09:23
And so you can say that summed activity of everything in that
1:09:26
curve will tell me
1:09:27
something about the neural activity because the laser light
1:09:31
transversing through
1:09:32
SPEAKER_03: the brain is being absorbed by hemoglobin,
1:09:34
SPEAKER_03: which carries oxygen in your bloodstream,
1:09:36
SPEAKER_03: right?
1:09:37
And you send in two wavelengths,
1:09:40
and then you use this concept called the Beer-Lambert law,
1:09:43
SPEAKER_03: where one wavelength absorbs hemoglobin more or less than the other one,
1:09:48
SPEAKER_03: and so you kind of get the differential between these,
1:09:50
SPEAKER_03: and you can go,
1:09:51
SPEAKER_03: oh,
1:09:51
there's more or less oxygenated blood in this banana-shaped
1:09:55
curve over time,
1:09:57
SPEAKER_03: and you can watch this signal change.
1:09:59
SPEAKER_03: Now, why do we care about blood?
1:10:01
SPEAKER_03: As neurons are active,
1:10:03
SPEAKER_03: they use metabolic resources like oxygen,
1:10:05
SPEAKER_03: and that has to be resupplied in the bloodstream.
1:10:07
SPEAKER_03: So you get this neural correlate that's sort of temporally blurred,
1:10:11
SPEAKER_03: but in this case,
1:10:13
SPEAKER_03: spatially blurred,
1:10:15
SPEAKER_03: and that will recover some correlate of human brain function.
1:10:19
SPEAKER_03: But when you're doing it noninvasively, it's pretty noisy.
1:10:22
SPEAKER_03: Not good enough, yeah.
1:10:23
SPEAKER_01: Okay,
1:10:25
SPEAKER_01: in this future that you're talking about,
1:10:28
we can use these techniques possibly to help millions and
1:10:31
millions of people in
1:10:33
SPEAKER_01: different conditions.
1:10:34
If we're talking about a future where we are heading towards a
1:10:37
more consumer thing,
1:10:38
SPEAKER_01: this is an elective surgery,
1:10:40
SPEAKER_01: we're trying to merge with the machines,
1:10:41
SPEAKER_01: we're trying to keep up with the AI same thing,
1:10:44
SPEAKER_01: or that needs to be implanted in the brain?
1:10:47
SPEAKER_03: It's going to need to be implanted first.
1:10:50
SPEAKER_03: Now,
1:10:51
SPEAKER_03: there's a leap of faith that I'm happy to take and say,
1:10:53
SPEAKER_03: like,
1:10:55
SPEAKER_03: every naysayer that says,
1:10:56
SPEAKER_03: oh,
1:10:57
SPEAKER_03: that will never happen is just proved wrong by history,
1:10:59
SPEAKER_03: so I'm not going to be that guy.
1:11:01
But there is not an obvious path to a non-invasive
1:11:05
high-bandwidth brain-computer
1:11:07
SPEAKER_03: interface today.
1:11:08
And the physics that we have available to us right now just
1:11:13
suggests that we're
1:11:14
SPEAKER_03: very,
1:11:14
SPEAKER_03: very far away from it.
1:11:16
I'm very confident that the first high-bandwidth interface will
1:11:19
be an invasive one.
1:11:20
SPEAKER_03: That's not to say it won't translate eventually,
1:11:22
SPEAKER_03: but you're going to see that high performance first in an invasive one.
1:11:24
SPEAKER_03: Okay.
1:11:25
SPEAKER_01: If you would say never,
1:11:27
SPEAKER_01: and then we could replay this on YouTube a bunch when you're wrong,
1:11:31
SPEAKER_01: that would be good for maybe.
1:11:34
SPEAKER_03: Not gonna do it.
1:11:36
SPEAKER_03: Not gonna do it.
1:11:38
SPEAKER_03: And I think that there are certain tricks that can get you closer,
1:11:40
SPEAKER_03: but ultimately,
1:11:41
SPEAKER_03: the highest performance interface is gonna be the one that's there,
1:11:45
SPEAKER_03: invasive and closest to the neurons.
1:11:47
SPEAKER_03: Okay.
1:11:50
So I kind of talked about type one electrophysiology in there,
1:11:54
touching the brain.
1:11:55
Type II interfaces are these remote sensing that buy you whole
1:11:58
brain coverage,
1:11:59
SPEAKER_03: but they make a sacrifice,
1:12:01
SPEAKER_03: right?
1:12:01
They're not touching the neuron anymore,
1:12:03
and so they lose some performance.
1:12:05
So in Forrest's case,
1:12:06
we also use hemodynamics,
1:12:08
so we lose the temporal precision.
1:12:10
SPEAKER_03: So we're not a very good motor neuroprosthetic.
1:12:13
We're great at neuropsychiatric and detecting these things like
1:12:17
disorders of
1:12:17
SPEAKER_03: consciousness,
1:12:18
SPEAKER_03: and so we kind of serve that 21 million population,
1:12:21
SPEAKER_03: but we don't really serve that 1.5.
1:12:24
SPEAKER_03: And then in the future,
1:12:25
SPEAKER_03: I'll just tease this a little bit,
1:12:26
like what I'm most excited about is where you actually start to
1:12:30
combine the hard
1:12:31
SPEAKER_03: tech,
1:12:31
SPEAKER_03: this remote sensing new modalities of physics with biologics.
1:12:35
And so you start to actually modify the neurons themselves or
1:12:39
grow new neurons into the brain.
1:12:40
And we're seeing these sorts of things start to come back and
1:12:44
proliferate.
1:12:45
SPEAKER_03: I mean, like efforts in the space that are doing this.
1:12:48
SPEAKER_03: But once you combine biologics and hard tech,
1:12:50
you can take the precision of type one interfaces and the
1:12:53
coverage of type two
1:12:54
SPEAKER_03: interfaces and you can put them together.
1:12:56
And now you have like the ultimate high band with brain
1:13:00
computer interface.
1:13:00
SPEAKER_03: And that's actually to me the like,
1:13:03
SPEAKER_03: you know,
1:13:03
SPEAKER_03: decades away,
1:13:05
decade away solution that actually gets you to the like fully
1:13:11
Ray Kurzweilian
1:13:11
SPEAKER_03: technology.
1:13:12
SPEAKER_01: So,
1:13:13
SPEAKER_01: okay,
1:13:13
SPEAKER_01: so Science Corporation,
1:13:14
SPEAKER_01: Max Hodak,
1:13:15
SPEAKER_01: I mean,
1:13:15
they're working on growing these neurons that are sitting
1:13:19
outside of the brain,
1:13:20
SPEAKER_01: and then you kind of fuse them into your brain,
1:13:23
SPEAKER_01: or they grow axons and dendrites into your tissue and give you this extra
1:13:28
SPEAKER_01: horsepower.
1:13:28
SPEAKER_01: That's one...
1:13:30
SPEAKER_01: That would be,
1:13:31
SPEAKER_03: in my mind,
1:13:32
SPEAKER_03: so that bio-hybrid interface is,
1:13:35
SPEAKER_03: to me,
1:13:35
SPEAKER_03: one of the first real,
1:13:37
SPEAKER_03: like,
1:13:38
SPEAKER_03: big attempts at creating a Gen 3 or Type 3 interface,
1:13:43
SPEAKER_03: as I've been calling them.
1:13:44
SPEAKER_03: Okay.
1:13:45
SPEAKER_03: I keep thinking there's been... But it's old.
1:13:46
SPEAKER_03: You know that, right?
1:13:47
SPEAKER_01: No, when you said people have been doing this before, I was...
1:13:51
Okay,
1:13:53
so the very first,
1:13:55
it might be the first human implanted BCI was a bio-hybrid
1:13:59
interface.
1:14:01
SPEAKER_03: Do you know the story of Phil Kennedy?
1:14:02
SPEAKER_03: No.
1:14:03
SPEAKER_03: Dude, you got to find Phil Kennedy.
1:14:05
SPEAKER_03: I'm dying to meet this guy.
1:14:07
SPEAKER_03: Okay, so Phil Kennedy.
1:14:08
SPEAKER_03: Find Phil.
1:14:09
SPEAKER_03: You got to find, hashtag find Phil.
1:14:11
SPEAKER_03: I actually think I know where he is.
1:14:15
SPEAKER_03: So you take,
1:14:16
SPEAKER_03: in this case,
1:14:17
SPEAKER_03: a glass cone,
1:14:19
SPEAKER_03: and you have a gold strip running down into it,
1:14:23
SPEAKER_03: and then you put these...
1:14:24
SPEAKER_03: You actually grow neurons in the cone itself,
1:14:27
SPEAKER_03: and the neurons actually grow out into the brain tissue itself,
1:14:32
SPEAKER_03: and then you basically have a link,
1:14:34
SPEAKER_03: an organic link between the normal brain,
1:14:38
SPEAKER_03: these kind of lab-grown neurons,
1:14:40
SPEAKER_03: and then it's touching these electrode contacts.
1:14:43
So the electrodes never actually go into the brain,
1:14:45
you never penetrate the brain.
1:14:46
SPEAKER_03: So butcher number zero, probably.
1:14:49
SPEAKER_03: And so Phil Kennedy actually came up with this,
1:14:52
SPEAKER_03: and I don't know if he was actually the inventor,
1:14:54
SPEAKER_03: but he was one of the pioneers in this space,
1:14:56
SPEAKER_03: and this is back in,
1:14:57
SPEAKER_03: I think,
1:14:57
SPEAKER_03: the 90s.
1:14:59
SPEAKER_03: And then he wants to test it on humans,
1:15:01
SPEAKER_03: running into multiple kind of roadblocks and hurdles,
1:15:04
SPEAKER_03: and eventually decides to actually implant himself.
1:15:07
SPEAKER_03: So it goes like full rogue scientist,
1:15:09
SPEAKER_03: flies to Mexico,
1:15:10
SPEAKER_03: gets this thing implanted,
1:15:13
and then actually comes home and gets like 10 days of recording
1:15:16
or something like
1:15:17
SPEAKER_03: that before it gets infected.
1:15:18
SPEAKER_03: So the U.S.
1:15:19
SPEAKER_03: healthcare system, no doctor in the U.S.
1:15:20
SPEAKER_03: was willing to put the thing in.
1:15:21
SPEAKER_03: That's why you had to go to Mexico.
1:15:22
SPEAKER_03: Which is why I had to go to Mexico.
1:15:24
But then they were like,
1:15:26
well,
1:15:26
it's life-threatening now,
1:15:27
so we'll take it out.
1:15:30
And so,
1:15:30
like,
1:15:30
I think he got away with,
1:15:31
like,
1:15:31
minimal neurological damage.
1:15:33
SPEAKER_03: But,
1:15:33
SPEAKER_03: like,
1:15:34
SPEAKER_03: just the craziest story of,
1:15:35
SPEAKER_03: like,
1:15:35
SPEAKER_03: rogue scientist implanting himself with,
1:15:37
SPEAKER_03: like,
1:15:37
SPEAKER_03: a biohybrid interface.
1:15:39
SPEAKER_01: And he hasn't really talked about this since?
1:15:42
SPEAKER_03: He has.
1:15:42
SPEAKER_03: Okay.
1:15:43
SPEAKER_03: But, like, he's not an active voice on the PCI field.
1:15:46
SPEAKER_01: How are they making synthetic neurons?
1:15:50
SPEAKER_01: Were they from an animal or...?
1:15:53
In his case,
1:15:55
they might have been stem cell derived,
1:15:58
but I think they were.
1:15:59
SPEAKER_01: Because it was like pre-IPSC, isn't it?
1:16:01
SPEAKER_01: Yeah, exactly.
1:16:03
SPEAKER_01: Okay.
1:16:04
SPEAKER_01: Now my mind is racing.
1:16:06
SPEAKER_01: So where are you on the, you think...
1:16:15
SPEAKER_01: Okay, I guess two questions.
1:16:18
SPEAKER_01: So it sounds like you do think at some point in the future,
1:16:21
SPEAKER_01: humans will not only have the option to have these kinds of implants,
1:16:27
SPEAKER_01: but will elect to do them.
1:16:30
SPEAKER_01: You do?
1:16:31
SPEAKER_03: Yeah,
1:16:32
it just seems actually really obvious that the Overton window
1:16:35
on this will move
1:16:36
SPEAKER_03: when the benefit is there.
1:16:38
SPEAKER_03: Today, it seems totally crazy because it is totally crazy.
1:16:41
Why would you have a hole drilled in your skull so you can move
1:16:43
a cursor on a screen?
1:16:44
SPEAKER_03: I can do that with my mouse and pretty good at it.
1:16:47
SPEAKER_03: Yeah.
1:16:48
SPEAKER_03: Of course, it seems crazy.
1:16:50
SPEAKER_03: But,
1:16:51
SPEAKER_03: I mean,
1:16:51
SPEAKER_03: I don't remember exactly what the numbers are,
1:16:53
but the number of people that have elective plastic surgery for
1:16:57
what is ultimately
1:16:58
SPEAKER_03: a psychological benefit,
1:16:59
SPEAKER_03: right?
1:17:00
SPEAKER_03: And in most cases,
1:17:01
SPEAKER_03: it's like a temporary bump,
1:17:03
and then you kind of level back off to where you started and
1:17:05
hedonic treadmill
1:17:06
SPEAKER_03: kicks in.
1:17:08
SPEAKER_03: But plastic surgery is a huge industry, single-payer driven.
1:17:12
So people are paying out of pocket for surgeries that give them
1:17:14
a minor psychological bump.
1:17:17
SPEAKER_03: So,
1:17:17
if you have a device that can be implanted one time and not
1:17:20
only provide you a
1:17:20
SPEAKER_03: minor psychological bump,
1:17:21
SPEAKER_03: but a permanent knob on a tunable mood,
1:17:25
SPEAKER_03: better sleep,
1:17:27
SPEAKER_03: improved depressive or anxious symptoms,
1:17:30
SPEAKER_03: name it,
1:17:31
SPEAKER_03: right?
1:17:32
SPEAKER_03: All of a sudden, that seems to actually be totally reasonable.
1:17:37
SPEAKER_03: The same thing happened with pacemakers.
1:17:39
SPEAKER_03: I love the pacemaker story.
1:17:41
SPEAKER_03: In the 1950s,
1:17:42
SPEAKER_03: like,
1:17:42
SPEAKER_03: the very first pacemaker that was put in a human was sort of done sort of
1:17:45
SPEAKER_03: fly-by-night in,
1:17:46
SPEAKER_03: I think,
1:17:46
SPEAKER_03: Sweden.
1:17:47
SPEAKER_03: And it was in a... They literally cast it in a, like, a shoe polish tin.
1:17:53
SPEAKER_03: Just to give you a sense of, like, how advanced this was.
1:17:55
SPEAKER_03: And this is a device that beats a person's heart.
1:17:58
SPEAKER_03: If it stops working, that person dies.
1:18:00
SPEAKER_03: Yeah.
1:18:01
SPEAKER_03: And then eventually, over 20 years or so, these become nuclear-powered.
1:18:05
SPEAKER_03: So you're talking about in the 50s and 60s,
1:18:06
SPEAKER_03: it's like,
1:18:07
SPEAKER_03: hey,
1:18:07
I'm gonna put in a nuclear device in your chest that's gonna
1:18:10
beat your heart for
1:18:10
SPEAKER_03: you.
1:18:11
SPEAKER_03: That must have also sounded crazy.
1:18:13
SPEAKER_03: We don't even think about it now.
1:18:14
SPEAKER_03: It's just like, well, of course.
1:18:16
SPEAKER_01: Yeah.
1:18:17
I mean,
1:18:17
there's a lot that comes once we start messing with our mind,
1:18:20
right?
1:18:21
SPEAKER_01: I mean,
1:18:22
SPEAKER_01: well,
1:18:23
I'm not gonna break new ground here on where this science
1:18:26
fiction future could go,
1:18:28
SPEAKER_01: but,
1:18:29
SPEAKER_01: you know,
1:18:30
SPEAKER_01: just us sitting around,
1:18:31
SPEAKER_01: pleasuring ourselves all day,
1:18:33
SPEAKER_01: sitting in a chair with happiness.
1:18:35
Flowing over us and or...
1:18:37
This is actually one of the things that makes me kind of glad.
1:18:40
I mentioned earlier that I really do believe that the
1:18:43
high-performance interfaces
1:18:44
SPEAKER_03: start as invasive,
1:18:45
SPEAKER_03: right?
1:18:46
SPEAKER_03: Which means they do,
1:18:47
SPEAKER_03: unless you're like the one-off rogue Phil Kennedy of the world,
1:18:50
need to be prescribed by a surgeon that needs to go through the
1:18:54
FDA's most rigorous
1:18:56
SPEAKER_03: pre-market authorization approvals.
1:18:58
SPEAKER_03: It's a Class III device.
1:18:59
So there's already actually regulation and ethical boundaries
1:19:04
set up around this
1:19:04
through the medical system and regulatory system that mean that
1:19:08
we'll actually have
1:19:08
a little bit of a moratorium even if we do end up creating
1:19:12
non-invasive consumer
1:19:13
type devices where we'll learn what these devices are capable
1:19:16
of and we'll learn
1:19:17
where the downsides and the risks are with like regulation in
1:19:21
place today.
1:19:22
So I think when people get like really nervous about like,
1:19:24
where's the regulation on this?
1:19:25
SPEAKER_03: It's like, dude, there's a lot.
1:19:27
SPEAKER_01: I don't know,
1:19:27
SPEAKER_01: man,
1:19:28
SPEAKER_01: I think Elon,
1:19:28
SPEAKER_01: they're gonna pop one,
1:19:29
SPEAKER_01: Elon's just gonna chuck that in his head,
1:19:32
SPEAKER_01: and he'll be like,
1:19:33
you thought I could do a lot of stuff before all of a sudden
1:19:36
he's gonna be like 65,
1:19:37
SPEAKER_01: and just supercharged.
1:19:40
SPEAKER_01: I think he honestly would just play more Polytopia.
1:19:42
SPEAKER_01: I mean,
1:19:43
SPEAKER_01: you do...
1:19:43
SPEAKER_01: I have noticed,
1:19:44
SPEAKER_01: like,
1:19:45
SPEAKER_01: just about everybody who gets one of these Neuralink implants,
1:19:49
SPEAKER_01: they end up playing just video games.
1:19:51
SPEAKER_01: It's like... Well, that's because it's what it's good at.
1:19:52
SPEAKER_01: Yeah, right.
1:19:53
SPEAKER_03: It's a cursor on a screen, click, and off you go.
1:19:56
SPEAKER_01: Yeah, yeah.
1:19:56
SPEAKER_01: And where are you on the...
1:19:59
humans need to do this anyway to compete with AI and have a
1:20:05
place in this universe.
1:20:07
SPEAKER_03: I'm not like a full doomer in the Silicon Valley or SF scene.
1:20:13
SPEAKER_03: But what I would say is that, like, AI is gonna be transformative.
1:20:16
SPEAKER_03: Like,
1:20:16
SPEAKER_03: we're at the elbow on another,
1:20:19
SPEAKER_03: you know,
1:20:20
SPEAKER_03: transformational curve in human history,
1:20:22
SPEAKER_03: the agricultural revolution,
1:20:23
SPEAKER_03: industrial revolution,
1:20:25
SPEAKER_03: computer revolution,
1:20:26
SPEAKER_03: now the intelligence revolution.
1:20:28
So it's going to have a lot of churn and turmoil,
1:20:31
and jobs will be changed.
1:20:34
SPEAKER_03: I'm not gonna say lost, right?
1:20:35
It's like,
1:20:36
it's always a temporary,
1:20:36
like,
1:20:37
well,
1:20:37
I don't do this anymore,
1:20:38
now I do this.
1:20:40
And I think that it's like,
1:20:41
we're in for a very fascinating,
1:20:44
not boring decade.
1:20:46
SPEAKER_01: Well, yeah, because it's so bizarre, because we're...
1:20:49
advancing these artificial minds at this moment in time when
1:20:52
we're learning far
1:20:54
more about the human mind than we've ever known because all of
1:20:57
a sudden we have
1:20:58
SPEAKER_01: tons of computing going into brains and all this money.
1:21:03
SPEAKER_01: Which curves these two things go on is super fascinating.
1:21:08
SPEAKER_03: Yeah,
1:21:08
SPEAKER_03: unfortunately,
1:21:09
SPEAKER_03: I think that BCIs,
1:21:11
SPEAKER_03: because we build these things in atoms,
1:21:13
SPEAKER_03: it's hardware,
1:21:14
SPEAKER_03: and it's slow,
1:21:15
SPEAKER_03: and it's regulated,
1:21:17
SPEAKER_03: is not on the same runaway curve that AI is.
1:21:20
SPEAKER_03: So when people say,
1:21:22
SPEAKER_03: like,
1:21:22
SPEAKER_03: oh,
1:21:22
SPEAKER_03: this is gonna be the solution to alignment,
1:21:26
SPEAKER_03: I think that just raises a lot of questions for me.
1:21:29
SPEAKER_03: What do you mean by that?
1:21:30
SPEAKER_03: How does that actually align AI?
1:21:31
SPEAKER_03: And I hosted workshops on this in the past and thought a lot about it.
1:21:35
SPEAKER_03: And I do think,
1:21:37
to stay relevant as a species in a world where intelligence is
1:21:42
no longer only
1:21:44
SPEAKER_03: human,
1:21:45
SPEAKER_03: BCIs have a place.
1:21:46
SPEAKER_03: But they're not your one-stop solution.
1:21:49
SPEAKER_01: Do you think that there's all, oh, this is a very hard thing to predict.
1:21:53
SPEAKER_01: One thing I,
1:21:54
this is like some intuitive thing just I have from having
1:21:58
covered companies for
1:22:00
SPEAKER_01: many,
1:22:00
many years in different markets is I look at the amount of
1:22:03
money all these BCI
1:22:05
startups have raised and the market that they have to play in
1:22:10
for the foreseeable
1:22:10
SPEAKER_01: future is this medical market.
1:22:12
You pointed out that in some cases that the number of patients
1:22:17
is actually relatively,
1:22:19
SPEAKER_01: Small,
1:22:20
SPEAKER_01: I wonder,
1:22:21
SPEAKER_01: I feel like there's going to be,
1:22:23
SPEAKER_01: um,
1:22:23
there's gonna be a natural thing where a lot of these startups
1:22:26
go away because they
1:22:27
SPEAKER_01: can't,
1:22:28
SPEAKER_01: they can't advance their technology and find,
1:22:30
SPEAKER_01: find a big enough market for it.
1:22:31
SPEAKER_01: I even think though,
1:22:33
SPEAKER_01: um,
1:22:33
SPEAKER_01: there's gonna be,
1:22:35
SPEAKER_01: you know,
1:22:35
SPEAKER_01: that even for the most successful companies,
1:22:38
that could be a protracted period of time where you're trying
1:22:40
to make money off of
1:22:42
SPEAKER_01: people who are sick and it has cost you an awful lot to get to this spot.
1:22:50
I guess maybe this technology will always be interesting enough
1:22:53
at this point that
1:22:54
SPEAKER_01: somebody will invest in the next thing.
1:22:56
I just wonder if we're gonna have like a little bit of a valley
1:23:00
of despair when it
1:23:02
SPEAKER_01: takes longer than people think to find a multi-billion dollar
1:23:09
SPEAKER_01: at the end of this, yeah.
1:23:11
SPEAKER_03: Yeah, this is increasingly the interesting space, though, right?
1:23:14
SPEAKER_03: Is deep tech investments,
1:23:17
SPEAKER_03: which is what we're,
1:23:18
SPEAKER_03: you know,
1:23:18
SPEAKER_03: VCs would call this,
1:23:20
SPEAKER_03: and others would call it,
1:23:20
SPEAKER_03: is they're fundamentally interesting because they're hard.
1:23:24
SPEAKER_03: They have long timelines, big capital requirements.
1:23:27
SPEAKER_03: But if you can make it to the end,
1:23:30
SPEAKER_03: or even some,
1:23:30
SPEAKER_03: like,
1:23:30
SPEAKER_03: modicum of the end,
1:23:32
SPEAKER_03: you're interacting with every facet of the human experience,
1:23:35
SPEAKER_03: what it literally means to be human,
1:23:36
SPEAKER_03: to interact with the brain.
1:23:38
SPEAKER_03: So even if you don't capture the whole thing,
1:23:41
SPEAKER_03: it's like,
1:23:42
SPEAKER_03: yeah,
1:23:42
SPEAKER_03: we shot for the full augmented human merge with AI,
1:23:47
SPEAKER_03: and at the end,
1:23:48
SPEAKER_03: all we got was like,
1:23:50
SPEAKER_03: oh,
1:23:50
SPEAKER_03: we solved depression.
1:23:52
SPEAKER_03: That's actually not so bad.
1:23:54
SPEAKER_03: That's a great outcome.
1:23:56
SPEAKER_03: I had a professor that used to talk about failure modes.
1:23:59
SPEAKER_03: In engineering, we think a lot about failure modes.
1:24:02
SPEAKER_03: A plane falling out of the sky is a bad failure mode,
1:24:04
SPEAKER_03: which is why we spend a lot of time building redundancy into the system.
1:24:08
SPEAKER_03: The flip side of that is like an escalator.
1:24:10
SPEAKER_03: Escalators have the best failure mode.
1:24:11
SPEAKER_03: They just become stairs.
1:24:13
SPEAKER_03: And that's actually kind of the world that I see for BCIs,
1:24:17
SPEAKER_03: where like,
1:24:18
SPEAKER_03: yes,
1:24:18
SPEAKER_03: big capital investments,
1:24:20
SPEAKER_03: but if you never try,
1:24:21
SPEAKER_03: you'll never know.
1:24:22
SPEAKER_03: If we had never scaled large language models or transformers,
1:24:26
we would have possibly never figured out the AI revolution that
1:24:30
we're experiencing
1:24:31
SPEAKER_03: now.
1:24:31
SPEAKER_03: You have to try it to know.
1:24:33
And the failure mode is we solve some of the most debilitating
1:24:36
diseases on Earth.
1:24:38
SPEAKER_03: That, to me, is fundamentally good for humanity.
1:24:40
SPEAKER_03: That seems okay.
1:24:41
SPEAKER_01: All right, last question.
1:24:45
SPEAKER_01: Downloading brains.
1:24:47
SPEAKER_01: Do we download our brains?
1:24:48
SPEAKER_01: Do we live eternally in a hard drive in the sky?
1:24:51
SPEAKER_03: Oh, I'm gonna get myself in so much trouble.
1:24:54
SPEAKER_03: Why did this have to be the last question?
1:25:00
SPEAKER_03: I really love the show Pantheon, by the way.
1:25:02
SPEAKER_01: Yes.
1:25:02
SPEAKER_01: Pantheon's amazing.
1:25:03
SPEAKER_01: It's awesome.
1:25:04
SPEAKER_01: I watch it with my kids and...
1:25:06
SPEAKER_01: It takes place in Palo Alto.
1:25:08
SPEAKER_01: Pretty much in this office is the headquarters of the...
1:25:12
SPEAKER_01: If you look at where it probably is,
1:25:14
SPEAKER_01: it's right here.
1:25:15
SPEAKER_03: Credit where credit's due.
1:25:16
Max Hodak told me to watch that a while back before it made it
1:25:20
onto Netflix,
1:25:20
SPEAKER_03: and I was dying to find it.
1:25:21
SPEAKER_03: I was watching some bootleg sketchy copy,
1:25:23
SPEAKER_03: and then in the middle of watching it,
1:25:25
SPEAKER_03: it came out on Netflix,
1:25:26
SPEAKER_03: and the resolution's much better.
1:25:27
SPEAKER_01: However, have you done season two?
1:25:28
SPEAKER_01: Yeah.
1:25:29
SPEAKER_01: You have?
1:25:30
SPEAKER_01: Okay.
1:25:30
SPEAKER_01: So, most people I know, I have not.
1:25:32
SPEAKER_01: And it's like, season one's amazing.
1:25:34
SPEAKER_01: And for some reason,
1:25:35
SPEAKER_01: because of these weird rights issues,
1:25:37
SPEAKER_01: everyone sort of ditched season two,
1:25:38
SPEAKER_01: and now it's on Netflix.
1:25:40
SPEAKER_01: Anyway, okay, that's good.
1:25:41
SPEAKER_01: All right.
1:25:42
SPEAKER_01: Season two is good?
1:25:43
SPEAKER_03: It's not as good as one until the end.
1:25:47
SPEAKER_01: Okay.
1:25:47
SPEAKER_01: And then it gets really good.
1:25:48
okay but in in relation to this question there's some brain
1:25:52
downloading all right
1:25:53
so you're gonna make me answer the question yes yes i will um
1:25:56
all right i have a
1:25:57
contrarian to all my friends take um i fundamentally find brain
1:26:04
uploading to be
1:26:06
problematic because a brain in a vat is banal it's not
1:26:11
interesting like a brain
1:26:14
SPEAKER_03: it's
1:26:15
Its information content is created by the entropy of the world
1:26:19
around it.
1:26:19
SPEAKER_03: It is its experience, it is creating connections, it is learning.
1:26:23
SPEAKER_03: That is kind of what makes you a human.
1:26:26
And so just uploading the weights of a brain onto a hard drive
1:26:29
I think is missing something.
1:26:32
SPEAKER_03: it needs a couple of things.
1:26:33
SPEAKER_03: One, I do believe it needs embodiment.
1:26:35
SPEAKER_03: Now that can be silicon-based embodiment,
1:26:37
SPEAKER_03: which means that you take that,
1:26:39
SPEAKER_03: it's not just the brain,
1:26:40
it's the whole central nervous system that you have to emulate
1:26:42
and you have to then
1:26:43
SPEAKER_03: connect that to some kind of input-output state.
1:26:47
SPEAKER_03: It may not be like a humanoid-type
1:26:50
SPEAKER_03: virtual body,
1:26:51
SPEAKER_03: it could be just about anything,
1:26:52
SPEAKER_03: but it needs to be wired up to something in a way.
1:26:55
And that thing needs to live in a physical construct that has
1:26:59
enough consistency
1:27:00
SPEAKER_03: that it can be modeled.
1:27:01
SPEAKER_03: So you have to create a world model.
1:27:03
SPEAKER_03: So you need the brain, the body, and the world model.
1:27:06
SPEAKER_03: And also,
1:27:06
SPEAKER_03: by the way,
1:27:07
SPEAKER_03: we're kind of social beings,
1:27:08
SPEAKER_03: so we should probably put other agents in there,
1:27:10
SPEAKER_03: whether artificial or biologically uploaded intelligences.
1:27:14
SPEAKER_03: And now you've basically just reconstructed the world Atom for Atom.
1:27:17
And for me,
1:27:18
that is like,
1:27:20
sure,
1:27:21
but so many of the time...
1:27:24
I just feel like nuance is lost here.
1:27:25
SPEAKER_03: So many of the timelines you hear is like,
1:27:27
SPEAKER_03: well,
1:27:27
SPEAKER_03: we uploaded a fly brain,
1:27:28
SPEAKER_03: we uploaded a worm brain,
1:27:30
SPEAKER_03: we uploaded whatever,
1:27:30
SPEAKER_03: or uploaded...
1:27:32
SPEAKER_03: Many of these aren't even functional connectomics,
1:27:34
SPEAKER_03: right?
1:27:34
SPEAKER_03: But let's say that's... Like, mapped it.
1:27:36
SPEAKER_03: Yeah, so we've mapped it.
1:27:38
SPEAKER_03: And then we're gonna get to mapping
1:27:40
SPEAKER_03: Um,
1:27:41
SPEAKER_03: the mouse brain in the next few years,
1:27:43
SPEAKER_03: I think,
1:27:44
SPEAKER_03: really,
1:27:44
SPEAKER_03: with some of the efforts in the space,
1:27:46
SPEAKER_03: many of which I know you're aware of.
1:27:47
SPEAKER_03: And then, human brains are, like, tangibly there, but far away still.
1:27:53
Probably,
1:27:54
you know,
1:27:55
tens of billions of dollars in a decade might get you there.
1:27:57
SPEAKER_01: Just to map a brain in a snapshot of time.
1:28:01
SPEAKER_03: Exactly.
1:28:02
SPEAKER_03: And that doesn't take into account the functional aspect.
1:28:05
SPEAKER_03: Now you need to actually simulate the connections between those neurons.
1:28:09
SPEAKER_03: And then there's kind of something missing there,
1:28:11
SPEAKER_03: too,
1:28:12
SPEAKER_03: where maybe you actually need more resolution than that.
1:28:14
SPEAKER_03: You need to know what's happening at every synapse.
1:28:16
You need to know the synaptic cleft and every connection,
1:28:19
and that's where plasticity happens.
1:28:20
So there's probably memory formation that's happening at the
1:28:23
molecular level.
1:28:24
SPEAKER_03: Do you need to simulate at the molecular level, the atomic level?
1:28:27
SPEAKER_03: Like, how deep does the rabbit hole go?
1:28:28
SPEAKER_03: We don't know.
1:28:30
SPEAKER_03: But let's say that you don't,
1:28:31
SPEAKER_03: and you just have the snapshot,
1:28:33
SPEAKER_03: and then you need to make the snapshot functional,
1:28:35
SPEAKER_03: and then you need the central nervous system,
1:28:36
SPEAKER_03: and then you need the multi-agent and the physics simulation.
1:28:38
SPEAKER_03: I think it's just like... Too big.
1:28:42
SPEAKER_03: It's big.
1:28:43
SPEAKER_03: It's a really,
1:28:44
SPEAKER_03: really big problem that we should definitely keep pushing forward on,
1:28:47
SPEAKER_03: and I hope I'm wrong.
1:28:49
But maybe it's just because it's not my primary focus,
1:28:53
but it feels really hard.
1:28:54
SPEAKER_01: AI, bro.
1:28:55
SPEAKER_01: AI, man.
1:28:56
SPEAKER_03: Something, something, AI.
1:28:57
SPEAKER_03: Something, something, profit.
1:29:00
SPEAKER_01: Um, okay.
1:29:01
SPEAKER_01: Well, this is fascinating.
1:29:02
SPEAKER_01: My brain, like, actually physically hurts.
1:29:06
SPEAKER_01: Um, you, you... That was awesome.
1:29:09
SPEAKER_01: Thank you so much.
1:29:09
SPEAKER_01: Um, thank you so much.
1:29:12
SPEAKER_03: Yeah, it was great to be here.
1:29:13
SPEAKER_03: Thanks, Ashton.
1:29:14
Um,
1:29:15
and...
1:29:16
Well,
1:29:17
I'm sure we'll have you on again,
1:29:18
if you will,
1:29:19
if you'll...
1:29:20
SPEAKER_03: I'll come back anytime.
1:29:21
SPEAKER_03: I've got endless brain thoughts.
1:29:22
SPEAKER_01: I want you to be a resident brain expert.
1:29:25
SPEAKER_03: That sounds fun.
1:29:26
SPEAKER_03: Can I get a title, though?
1:29:27
SPEAKER_03: I need something for my LinkedIn.
1:29:29
SPEAKER_03: Yeah, I mean, what do you want?
1:29:32
SPEAKER_03: Brain in residence.
1:29:33
SPEAKER_01: Core memories, like memory... Chief Memory Officer or something.
1:29:39
SPEAKER_01: Ooh.
1:29:40
SPEAKER_01: I just got upgraded.
1:29:40
SPEAKER_01: Now we're getting somewhere.
1:29:42
SPEAKER_01: Sumner, thank you so much.
1:29:44
SPEAKER_01: We will put links to all your endeavors,
1:29:47
SPEAKER_01: and then,
1:29:47
SPEAKER_01: yeah,
1:29:47
SPEAKER_01: I think we should...
1:29:48
I'll try to go extra on this one and put some links to some of
1:29:51
this history in so
1:29:53
SPEAKER_01: people can deep dive on some of these studies and different things.
1:29:57
SPEAKER_01: Awesome.
1:29:58
SPEAKER_01: Love that.
1:29:58
SPEAKER_00: And do you have a YouTube presentation?
1:30:04
SPEAKER_03: There's a few out there.
1:30:05
SPEAKER_03: Yeah, yeah, yeah.
1:30:06
Foresight Institute hosted me for a few different talks and
1:30:09
protocol labs.
1:30:10
SPEAKER_01: Where you walk through some of this, right?
1:30:12
SPEAKER_03: Yeah, I go through some of the history stuff.
1:30:14
SPEAKER_03: I think I mentioned bits and pieces of that today.
1:30:16
SPEAKER_03: But yeah,
1:30:16
I find it can be very easy to get lost in the swirl of all the
1:30:20
things happening
1:30:21
SPEAKER_03: when you're trying to build a company or organization that's looking 30,
1:30:25
SPEAKER_03: 50 years in the future.
1:30:27
SPEAKER_03: And it helps to just look back 30 or 50 years.
1:30:29
SPEAKER_03: And so it's my hobby.
1:30:31
SPEAKER_03: No, I think it's awesome.
1:30:33
SPEAKER_01: All right.
1:30:33
SPEAKER_01: Thank you so much, man.
1:30:34
SPEAKER_01: Appreciate it.
1:30:35
SPEAKER_01: And we'll speak again soon.
1:30:37
SPEAKER_01: Sounds good.
1:30:37
SPEAKER_01: Later.
1:30:41
core memory podcast is hosted by me ashley vance it is produced
1:30:46
by david nicholson
1:30:48
and me our theme song is by james mercer and john sortland and
1:30:54
the show is edited
1:30:56
SPEAKER_01: by the john sortland
1:31:00
SPEAKER_01: Thanks, as always, to Brex and E1 Ventures for making this possible.
1:31:06
Please visit our Substack YouTube and podcast channels to get
1:31:11
more of what Core Memory makes.
1:31:13
SPEAKER_01: Thanks, y'all.