Latest / The Meaning of Health Podcast / Episode 150 – Grow Your Own Intelligence Part 2: Biological Computers
Transcript
- 0:00SPEAKER_00: Thanks everyone for joining us for another part of our Science Week episode, which is part two.
- 0:07SPEAKER_00: This week we sat down for a conversation with Dr. Brett Kagan.
- 0:13SPEAKER_00: Is that a doctor?
- 0:14SPEAKER_00: Yeah, I don't even know.
- 0:15SPEAKER_00: I think so.
- 0:15SPEAKER_00: Yeah.
- 0:15SPEAKER_01: I don't know.
- 0:18SPEAKER_01: I think he does mention it in the episode, so you can find out for us.
- 0:20SPEAKER_00: Yeah, okay.
- 0:23SPEAKER_00: I'll just start again.
- 0:25SPEAKER_01: Do you want me to confirm?
- 0:27SPEAKER_01: I can...
- 0:28SPEAKER_01: I think I've got it somewhere.
- 0:31SPEAKER_01: I think he is a doctor because
- 0:34SPEAKER_01: he's got
- 0:37SPEAKER_00: a principal.
- 0:38SPEAKER_00: Yeah, Dr. Brett Kagan, I can see.
- 0:41SPEAKER_01: Or you can introduce him as chief scientific officer as well for Cortical Labs.
- 0:47SPEAKER_00: Yep.
- 0:49SPEAKER_00: Thanks everyone for joining us for part two of our National Science Week episodes.
- 0:54SPEAKER_00: This week we've got Dr Brett Kagan, who is the Chief Scientific Officer for Corticolabs.
- 1:00SPEAKER_00: And Brett took some time between flights,
- 1:03SPEAKER_00: I believe,
- 1:04SPEAKER_00: back from America to speak to us when he was traveling through Los Angeles airport.
- 1:10SPEAKER_00: There was a little bit of background noise as a result because he was in
- 1:12SPEAKER_00: the departure lounge, I think.
- 1:15SPEAKER_00: But yeah, Courtney, what can you tell us about Brett?
- 1:17SPEAKER_01: Yeah, so Brett made some time for us to have this conversation,
- 1:22SPEAKER_01: which is really, really exciting.
- 1:23SPEAKER_01: And I wanted to briefly introduce the idea of Cortical Labs, the company that Brett's a part of.
- 1:29SPEAKER_01: And I'm going to actually read the section on the website about Cortical Labs
- 1:34SPEAKER_01: because I think it's just a really nice summary.
- 1:37SPEAKER_01: So Cortical Labs,
- 1:40SPEAKER_01: their little summary says,
- 1:41SPEAKER_01: our technology merges biology with traditional computing to create
- 1:45SPEAKER_01: the ultimate learning machine,
- 1:47SPEAKER_01: creating what others only imagined, the world's first code deployments.
- 1:53SPEAKER_01: Now, if that doesn't spark creativity, I don't know what does.
- 1:57SPEAKER_01: I think that's really, really exciting.
- 1:59SPEAKER_01: And that is exactly why we decided that BRETT and Cortical Labs would be part two
- 2:05SPEAKER_01: of our National Science Week.
- 2:07SPEAKER_01: Now just a little bit about National Science Week.
- 2:10SPEAKER_01: So this is one of the largest science festivals in the world.
- 2:14SPEAKER_01: It's occurring between the 15th and 23rd of August.
- 2:17SPEAKER_01: So this would be coming towards the end of the week and there will be over two and a half
- 2:28SPEAKER_01: So you can find out a little bit more and about events near you at scienceweek.net.au.
- 2:35SPEAKER_01: So hopefully part two of Grow Your Own Intelligence,
- 2:39SPEAKER_01: our little part series for National Science Week spikes your curiosity
- 2:45SPEAKER_01: and you can enjoy this conversation.
- 3:21SPEAKER_00: Thanks for joining us for another episode of The Meaning of Health.
- 3:24SPEAKER_00: We have another part in our Science Week special looking at brains
- 3:28SPEAKER_00: and how we can use machines and brains together.
- 3:31SPEAKER_00: Courtney is joining me today.
- 3:33SPEAKER_00: Hello.
- 3:34SPEAKER_00: Do you want to tell us a bit about today's episode and our guest?
- 3:38SPEAKER_01: Yeah, yeah, so this is a part two in a series for National Science Week.
- 3:43SPEAKER_01: It's very, very exciting to have Brett here from Cortical Labs.
- 3:48SPEAKER_01: I think this is part two of an episode of a series and we're kind of exploring
- 3:56SPEAKER_01: the use of brain cells and how we can use it in terms of technology and
- 4:01SPEAKER_01: the first episode was around art.
- 4:04SPEAKER_01: Brett, hello, welcome to the podcast.
- 4:07SPEAKER_01: It's lovely to have you here.
- 4:09SPEAKER_01: Do you want to talk a little bit about who you are and what you do?
- 4:15SPEAKER_03: Sure.
- 4:15SPEAKER_03: Thanks very much for having me, Courtney and Craig.
- 4:17SPEAKER_03: Pleasure to be here, virtually, as it were, of course.
- 4:21SPEAKER_03: So I'm Brett.
- 4:23SPEAKER_03: I sit as Chief Scientific Officer and Chief Operating Officer at Cortical Labs,
- 4:27SPEAKER_03: which is actually an Australian-based in Melbourne.
- 4:29SPEAKER_03: Mm-hmm.
- 4:56SPEAKER_01: Very interesting and I guess the first thing that I kind of want to talk about
- 5:00SPEAKER_01: with you is actually your background and how you even got to where you are right now
- 5:06SPEAKER_01: because I feel like there'd be a very interesting pathway
- 5:09SPEAKER_01: and you probably didn't think back originally that you were going to be doing
- 5:14SPEAKER_01: a startup around brown cells.
- 5:18SPEAKER_03: Well, I mean, you know, it's interesting, the pathways that we go through in life.
- 5:23SPEAKER_03: I've always had a strong interest in this area going back for my whole adult life.
- 5:30SPEAKER_03: Admittedly, I never thought that there would be funding to do a startup like this.
- 5:34SPEAKER_03: But what's great,
- 5:35SPEAKER_03: and maybe what's a little scary in some ways,
- 5:37SPEAKER_03: is that the startup community and investors,
- 5:40SPEAKER_03: they're actually more willing to take a risk to try and find something wonderful
- 5:43SPEAKER_03: than Unfortunately,
- 5:45SPEAKER_03: these days,
- 5:45SPEAKER_03: our government grants would allow,
- 5:47SPEAKER_03: and it really should be the other way around, but that's the world we live in.
- 5:52SPEAKER_03: So my background is I started with a degree in psychology,
- 5:57SPEAKER_03: did some law,
- 5:59SPEAKER_03: realized I didn't want to be a psychologist or a lawyer,
- 6:01SPEAKER_03: started to go down into the neuro route,
- 6:04SPEAKER_03: did a masters in neuroscience,
- 6:06SPEAKER_03: and then followed it up with a PhD in neuro,
- 6:09SPEAKER_03: but focusing on stem cell therapy for neonatal brain injury.
- 6:13SPEAKER_03: Developed allergies to rodents and also realized
- 6:16SPEAKER_03: that strictly pre-clinical translational section wasn't really for me
- 6:20SPEAKER_03: because there wasn't that much intellectual creativity for the most part.
- 6:25SPEAKER_03: There are exceptions,
- 6:26SPEAKER_03: but for the most part Because typically what you're doing at
- 6:29SPEAKER_03: that translational level,
- 6:31SPEAKER_03: preclinical level,
- 6:32SPEAKER_03: is figuring out how to optimize, make things safe, make it scalable.
- 6:36SPEAKER_03: And those are challenging endeavors.
- 6:39SPEAKER_03: But I really wanted something where there was some more guidance.
- 6:43SPEAKER_03: Laying of the foundations.
- 6:46SPEAKER_03: So I moved a bit into bioinformatics and then did about a year postdoc in that
- 6:51SPEAKER_03: before I realized that academia really wasn't going to be the pathway for me
- 6:55SPEAKER_03: to move forward.
- 6:58SPEAKER_03: It's a challenging career and it takes a very specific sort of person
- 7:02SPEAKER_03: and I really always just wanted to move a little quicker than what it facilitated
- 7:05SPEAKER_03: and so when I got the opportunity to come in on the ground floor of Cortical Labs,
- 7:11SPEAKER_03: I jumped at it.
- 7:12SPEAKER_03: It was a great opportunity and we've been working on that now for nearly seven years.
- 7:18SPEAKER_01: Yeah, wow.
- 7:19SPEAKER_01: That's, yeah, very interesting.
- 7:20SPEAKER_01: It's good to hear that you basically,
- 7:23SPEAKER_01: although you had this strong desire,
- 7:26SPEAKER_01: there was lots of different pathways you tried and basically was like,
- 7:30SPEAKER_01: no, not for me, try something else.
- 7:32SPEAKER_01: No, not for me, try something else.
- 7:34SPEAKER_01: I think that's really useful and I think a lot of people kind of think
- 7:38SPEAKER_01: that they're stuck in one pathway.
- 7:40SPEAKER_01: But do you think all of that experience that you've had previously in psychology
- 7:44SPEAKER_01: and law and all that kind of thing has helped you Your position now?
- 7:50SPEAKER_03: Absolutely.
- 7:51SPEAKER_03: I mean, there's a strong old idea,
- 7:55SPEAKER_03: I didn't come up with it, called failing forwards and failing fast.
- 8:00SPEAKER_03: And I think that's kind of where I see myself.
- 8:02SPEAKER_03: And I, you know,
- 8:03SPEAKER_03: when I do talk about career,
- 8:05SPEAKER_03: you know,
- 8:05SPEAKER_03: especially to PhD students and the like,
- 8:07SPEAKER_03: you know,
- 8:07SPEAKER_03: I often say to them like, hey,
- 8:09SPEAKER_03: if you're doing your PhD and things are going really great and you love it,
- 8:12SPEAKER_03: this talk isn't going to be for you.
- 8:14SPEAKER_03: But if you're there struggling and you feel like everything you're doing is failing
- 8:17SPEAKER_03: and failing and failing, here's where things can go.
- 8:20SPEAKER_03: And you never know the outcome of something.
- 8:24SPEAKER_03: You never.
- 8:24SPEAKER_03: I remember when I went to Vienna to do a postdoc and I got there and
- 8:29SPEAKER_03: the postdoc was not what was discussed.
- 8:32SPEAKER_03: I didn't want to do it.
- 8:33SPEAKER_03: I wasn't set out to do it.
- 8:35SPEAKER_03: And I was devastated because I'd turned down other things
- 8:38SPEAKER_03: and it's very hard finishing your PhD to then jump between multiple positions,
- 8:42SPEAKER_03: especially when you need salaries and other things.
- 8:45SPEAKER_03: And I thought, well, that's it.
- 8:47SPEAKER_03: You know, I've wasted my life.
- 8:49SPEAKER_03: I've spent...
- 8:50SPEAKER_03: Five years,
- 8:51SPEAKER_03: master's, PhD,
- 8:52SPEAKER_03: and now I've taken a bad postdoc, turned down other options, and I was devastated.
- 8:57SPEAKER_03: But for that,
- 8:59SPEAKER_03: I would not have had the opportunity to do what I'm doing now,
- 9:01SPEAKER_03: which I love more than anything I could have imagined.
- 9:04SPEAKER_03: And I could pick a dozen other examples of things that I thought were failures at
- 9:08SPEAKER_03: the time, but laid the foundation to be able to do exactly what we ended up doing.
- 9:15SPEAKER_00: I think you've tapped into something there that goes all the way back
- 9:18SPEAKER_00: to Thomas Edison and before,
- 9:19SPEAKER_00: that And,
- 9:21SPEAKER_00: you know,
- 9:22SPEAKER_00: science really is about finding ways that don't work so that you can find the way
- 9:25SPEAKER_00: that does work, right?
- 9:26SPEAKER_00: So I think he said, I didn't fail.
- 9:28SPEAKER_00: You know, I haven't failed.
- 9:30SPEAKER_00: I just found 10,000 ways not to do something.
- 9:33SPEAKER_03: Yeah, and I think that's true.
- 9:34SPEAKER_03: And building the resilience to,
- 9:37SPEAKER_03: you know,
- 9:37SPEAKER_03: I had said this just yesterday talking to someone,
- 9:40SPEAKER_03: you know, I said it's not about not getting knocked down.
- 9:44SPEAKER_03: It's getting up one more time than you've been knocked down.
- 9:48SPEAKER_03: And that makes a difference,
- 9:49SPEAKER_03: you know,
- 9:49SPEAKER_03: most people,
- 9:50SPEAKER_03: and I understand why,
- 9:51SPEAKER_03: but they get knocked down,
- 9:53SPEAKER_03: and there's just, there's a lot going on, and I get it, right?
- 9:56SPEAKER_03: Life is hard, life is messy, and they say, okay, that's it.
- 10:01SPEAKER_03: But if it matters to you,
- 10:02SPEAKER_03: and not everything
- 10:04SPEAKER_02: should matter to you,
- 10:04SPEAKER_02: I'm not glorifying the get up and do it no matter what,
- 10:07SPEAKER_02: not everything
- 10:08SPEAKER_03: is worthwhile to do that for.
- 10:10SPEAKER_03: But if it is, if it is worthwhile enough to you, you just keep going until, until you get there.
- 10:18SPEAKER_01: All right.
- 10:19SPEAKER_01: Yeah, very, very interesting.
- 10:20SPEAKER_01: I wonder what part of,
- 10:23SPEAKER_01: and we'll get into what Cortical Labs is in a moment and the work that you do,
- 10:26SPEAKER_01: but what makes your work worthwhile and passionate for you right now?
- 10:32SPEAKER_03: I've always been really interested in these fundamental science questions.
- 10:36SPEAKER_03: What gives rise to intelligence at the most fundamental level of what I like
- 10:41SPEAKER_03: to call information physics?
- 10:43SPEAKER_03: And I think this is the part that a lot of people who are exploring
- 10:46SPEAKER_03: in this space are a little nervous about getting involved with
- 10:49SPEAKER_03: because these questions of how does intelligence arise
- 10:53SPEAKER_03: from these base principles is one of the most challenging questions that faces us as
- 10:58SPEAKER_03: a species, never mind us as a company.
- 11:02SPEAKER_03: But that's my personal interest and where it gets excited
- 11:04SPEAKER_03: and building these platforms and getting other people involved,
- 11:08SPEAKER_03: customers, collaborators, whatever the case is, is all a way to achieve that goal.
- 11:12SPEAKER_03: And I've always been the belief that if you build something great,
- 11:17SPEAKER_03: then the other details will work itself out.
- 11:20SPEAKER_03: You know, I've never been the one to say, well, I want to start a business.
- 11:23SPEAKER_03: I'm going to start a business or I want to make money.
- 11:26SPEAKER_03: So it's let's do something great.
- 11:30SPEAKER_03: And then what happens from that?
- 11:32SPEAKER_03: Everything else is just added bonus.
- 11:34SPEAKER_01: I think that's a great leeway to lead into what Cortical Labs is.
- 11:41SPEAKER_01: So do you mind giving a brief description of what the company is and what you're actually doing?
- 11:48SPEAKER_03: Yeah, so Cortical Labs, we're at about 36 people now.
- 11:52SPEAKER_03: Our main operations are in Melbourne, Australia.
- 11:54SPEAKER_03: We have a couple of people in Malaysia,
- 11:56SPEAKER_03: and we're looking to grow at a few other spots in the world soon.
- 11:59SPEAKER_03: And our focus is really this fundamental question of,
- 12:02SPEAKER_03: Brain cells we know can process information, they can compute.
- 12:07SPEAKER_03: What if we could use these capabilities but outside the body?
- 12:11SPEAKER_03: And so what we do is we grow biological neurons through synthetic biology.
- 12:16SPEAKER_03: We plate them down on these things called multi-electrode array or microelectrode arrays, MEA.
- 12:23SPEAKER_03: And we then interact with them through electricity
- 12:26SPEAKER_03: because electricity is this shared language between the silicon systems we use
- 12:31SPEAKER_03: to interact with them and the biological systems that do the information processing.
- 12:36SPEAKER_03: And then to take that a step further,
- 12:38SPEAKER_03: we've now built devices to allow anybody to do this,
- 12:41SPEAKER_03: either by purchasing devices or even over the cloud.
- 12:47SPEAKER_01: Wow.
- 12:48SPEAKER_01: Yeah, that's really exciting.
- 12:51SPEAKER_01: Yeah, so who...
- 12:53SPEAKER_01: Who came up with this idea and how did they get there?
- 12:57SPEAKER_03: I mean, the initial idea to start the company was our CEO, Dr. Hongwen Chong.
- 13:06SPEAKER_03: But the idea of brain cells in a dish or in a vat goes all the way back to,
- 13:12SPEAKER_03: I think, 1929, nearly 100 years ago, is the first recorded mention I could find.
- 13:17SPEAKER_03: There might be an older one.
- 13:20SPEAKER_03: J.C. Burnell was the author who basically reasoned the thing
- 13:24SPEAKER_03: that makes humans special is our brains.
- 13:27SPEAKER_03: So he is quite a quirky guy,
- 13:30SPEAKER_03: but he wrote it's inevitable that we'll end up preserving that part of humans.
- 13:37SPEAKER_03: But, yeah, the idea goes back a long time.
- 13:38SPEAKER_03: Then it was adopted by H.P.
- 13:40SPEAKER_03: Lovecraft, J.C.
- 13:41SPEAKER_03: Vanell influenced Ursula Le Guin and a range of other sci-fi authors,
- 13:46SPEAKER_03: Arthur C. Clarke, who represent it in certain ways.
- 13:51SPEAKER_03: But it might go back even before that.
- 13:55SPEAKER_00: I'm curious about what the main applications for this technology would be.
- 14:01SPEAKER_00: When you guys were...
- 14:03SPEAKER_00: Envisaging this and developing the idea.
- 14:07SPEAKER_00: Where do you see this going in terms of use cases?
- 14:12SPEAKER_03: So it's always been a platform technology and it has to be because as I was saying
- 14:16SPEAKER_03: before, figuring this out is daunting, right?
- 14:20SPEAKER_03: There's four key pillars.
- 14:21SPEAKER_03: I talked about the computational neuroscience information physics part.
- 14:25SPEAKER_03: There's the wetware,
- 14:26SPEAKER_03: the growing the biological cells,
- 14:28SPEAKER_03: there's the hardware interacting with the cells,
- 14:30SPEAKER_03: keeping them alive,
- 14:31SPEAKER_03: and then there's the software,
- 14:33SPEAKER_03: which is being able to program and interface with them in an accessible way.
- 14:38SPEAKER_03: And for us, the applications have always been buried in many, and that's why it's a platform.
- 14:43SPEAKER_03: So to be like concrete, not just vague, you know, you can look at the drug testing, right?
- 14:49SPEAKER_03: We've published and we've shown that diseased cells,
- 14:52SPEAKER_03: or disease models anyway,
- 14:54SPEAKER_03: tend not to perform very well in these, what we call embodied environments.
- 14:58SPEAKER_03: You want them to learn the game pong,
- 14:59SPEAKER_03: if they're an epilepsy model, typically can't learn very well.
- 15:03SPEAKER_03: You apply anti-epileptics on them,
- 15:05SPEAKER_03: And you find that some,
- 15:06SPEAKER_03: but not others,
- 15:07SPEAKER_03: will restore learning,
- 15:08SPEAKER_03: some will change modularity, and you get all this rich information you don't get.
- 15:12SPEAKER_03: You can take that a step further, and you can actually grow cells.
- 15:15SPEAKER_03: This is what we do.
- 15:16SPEAKER_03: We grow stem cells from people's donations of blood.
- 15:20SPEAKER_03: Not too many at the moment,
- 15:21SPEAKER_03: but this could be scaled, and you could do this in a personalized approach.
- 15:25SPEAKER_03: But then you can pivot and go to the other end of the spectrum,
- 15:27SPEAKER_03: and you can look at computing, general intelligence, robotics, physical, and AI.
- 15:33SPEAKER_03: We know biology can move objects around in the real world and adapt quickly.
- 15:38SPEAKER_03: We are the proof of it.
- 15:39SPEAKER_03: This isn't a, well, maybe this could happen one day.
- 15:42SPEAKER_03: No, we know it happens.
- 15:44SPEAKER_03: So the question isn't if, but how.
- 15:47SPEAKER_03: And so that doesn't mean how is easy, but it means it is possible.
- 15:52SPEAKER_03: And that's the other end of the spectrum.
- 15:54SPEAKER_03: And plenty of stuff in between.
- 15:58SPEAKER_00: Yeah, it's really interesting.
- 16:00SPEAKER_00: So...
- 16:01SPEAKER_00: As simply as you're able to,
- 16:03SPEAKER_00: can you explain for us how these biological computers,
- 16:08SPEAKER_00: for want of a better term, work?
- 16:12SPEAKER_00: What are the components and how are they interacting?
- 16:15SPEAKER_03: Yeah, sure.
- 16:16SPEAKER_03: So in short,
- 16:18SPEAKER_03: basically we grow neural cells,
- 16:20SPEAKER_03: we plate them on these electrodes that allow us to interface them with electricity.
- 16:24SPEAKER_03: We then have software that basically can work in real time.
- 16:28SPEAKER_03: So it has very, very tight loops.
- 16:30SPEAKER_03: And what it does is it can read the information or the electrical activity from
- 16:34SPEAKER_03: the cells as they're active with their action potentials and other forms
- 16:38SPEAKER_03: of electrical activity that are present.
- 16:41SPEAKER_03: We can then take that information,
- 16:43SPEAKER_03: do something with it,
- 16:44SPEAKER_03: and that something depends on the tasks,
- 16:46SPEAKER_03: and then we can put in some electrical information to update these cells about
- 16:51SPEAKER_03: the outcome of their actions or anything else that you might do.
- 16:57SPEAKER_03: And I'm a bit vague on that because there's a lot of different approaches how
- 17:00SPEAKER_03: to use this technology.
- 17:02SPEAKER_03: You can use it for classification,
- 17:03SPEAKER_03: you can use it for robotics control,
- 17:05SPEAKER_03: you can use it for, as I said, a drug discovery.
- 17:07SPEAKER_03: And all of these tasks are a little different.
- 17:09SPEAKER_03: And that's why we built this really flexible platform that allows you to log on
- 17:14SPEAKER_03: If you're using it remotely or access it directly if you have a unit yourself.
- 17:18SPEAKER_03: And program it just using very, very simple code blocks.
- 17:22SPEAKER_03: While previously you would have to be able to program down
- 17:24SPEAKER_03: on this low level hardware, which only a few people can do.
- 17:29SPEAKER_03: And it's very slow to do it even if you have that talent.
- 17:53SPEAKER_01: So this is probably me not quite understanding.
- 17:58SPEAKER_01: So the neurons themselves,
- 18:03SPEAKER_01: the clusters of cells that you're using,
- 18:06SPEAKER_01: you're inputting in electrical signals
- 18:09SPEAKER_01: and they're outputting out other electrical signals to then control things.
- 18:16SPEAKER_01: Is it acting more like a battery or is it,
- 18:23SPEAKER_01: Actually understanding feedback that you're then giving it to then learn.
- 18:30SPEAKER_01: So how does that learning process
- 18:32SPEAKER_02: work, I guess?
- 18:34SPEAKER_02: Great question.
- 18:37SPEAKER_02: The detailed answers, we're still working at ourselves.
- 18:39SPEAKER_03: We have made a few compelling inroads that have been well received in the scientific fields.
- 18:49SPEAKER_03: I think like your reference I think that question of is it like a battery?
- 18:54SPEAKER_03: I'd say no,
- 18:54SPEAKER_03: it's not like a battery because what drive these are very complex nonlinear dynamics
- 19:00SPEAKER_03: and we've published
- 19:01SPEAKER_03: and shown actually it's not even just simple connections changing either.
- 19:05SPEAKER_03: You actually get network-wide reorganization very rapidly and in complex ways.
- 19:13SPEAKER_02: Some of the features that drive these systems are things like
- 19:20SPEAKER_03: being able to predict,
- 19:21SPEAKER_03: it appears at this very basic rudimentary level,
- 19:24SPEAKER_03: the outcomes of their actions or their environment,
- 19:27SPEAKER_03: maintain themselves in these states often called critical dynamics.
- 19:32SPEAKER_02: But there are very likely other features that influence them as well,
- 19:35SPEAKER_02: and these are some of what we're exploring now still.
- 19:38SPEAKER_02: So that's a bit of a non-answer,
- 19:40SPEAKER_02: if only because the detailed technical answer is a solid hour computational lecture.
- 19:47SPEAKER_02: But what we can do,
- 19:48SPEAKER_02: in short,
- 19:49SPEAKER_02: is to take inspiration,
- 19:51SPEAKER_02: not necessarily try to copy,
- 19:53SPEAKER_02: that's another discussion,
- 19:55SPEAKER_02: what happens in neural systems,
- 19:58SPEAKER_02: and then leverage this inspiration to try and drive changes in our system.
- 20:02SPEAKER_02: And there's a spectrum here of physiologically relevant Try to copy the brain
- 20:09SPEAKER_02: and engineered,
- 20:11SPEAKER_02: try to leverage these physical principles,
- 20:13SPEAKER_02: but in a different way, that it really needs to be designed on a per task basis.
- 20:20SPEAKER_02: And I'll use an
- 20:20SPEAKER_03: analogy here, what I mean about why that makes sense.
- 20:24SPEAKER_03: A bird can fly and a plane can fly.
- 20:27SPEAKER_03: They both use aerodynamics, but they don't do it in the exact
- 20:31SPEAKER_02: same way.
- 20:32SPEAKER_02: So this is the same idea here.
- 20:36SPEAKER_01: Right, okay,
- 20:37SPEAKER_01: so you're leveraging the cells to do a similar output but just in
- 20:41SPEAKER_01: a different mechanism or a different way to do it.
- 20:44SPEAKER_03: Yeah, yeah,
- 20:45SPEAKER_03: I mean the fundamental physical principles must almost certainly be quite similar
- 20:51SPEAKER_03: but the level of how that's applied can differ and that gives you huge flexibility.
- 20:58SPEAKER_00: And do the cells and the neurons that you guys are creating and maintaining,
- 21:05SPEAKER_00: can the input that they get back from the computer change those cells as well?
- 21:09SPEAKER_00: Can they evolve?
- 21:12SPEAKER_03: Yeah, absolutely.
- 21:13SPEAKER_03: And we see that not only can the cells change their activity, but they do so very quickly.
- 21:18SPEAKER_03: So we've got papers where we've shown that these change will happen,
- 21:22SPEAKER_03: depending on the paper, between about one to two minutes up to about 10.
- 21:26SPEAKER_03: The sweet spot seeming to be about five minutes,
- 21:28SPEAKER_03: generally, to start to see some really dynamic network changes.
- 21:32SPEAKER_03: Depending on your culture,
- 21:33SPEAKER_03: composition,
- 21:34SPEAKER_03: the task,
- 21:35SPEAKER_03: and a whole range of other things,
- 21:37SPEAKER_03: there's variability there that we're starting to be able to explain.
- 21:40SPEAKER_03: There's also variability that we're still trying to explain.
- 21:43SPEAKER_03: That's the fun thing
- 21:44SPEAKER_03: of doing these cutting-edge things is we don't necessarily know exactly how all
- 21:50SPEAKER_03: of it works,
- 21:50SPEAKER_03: but what we
- 21:52SPEAKER_02: do find is that we can iteratively explore and understand things
- 21:55SPEAKER_02: that previously you just had no way of even exploring.
- 22:00SPEAKER_01: Yeah, I think that kind of leads on to my next question about unpredictability
- 22:05SPEAKER_01: because I feel like computers as they are now and the way that we work with them
- 22:10SPEAKER_01: and how they function,
- 22:11SPEAKER_01: it's all very predictable and it's something that we understand relatively well as
- 22:17SPEAKER_01: a species, I guess.
- 22:18SPEAKER_01: Yeah.
- 22:19SPEAKER_01: But when you're introducing something that is almost living,
- 22:25SPEAKER_01: there is a nature or there is a level of unpredictability about what they will learn
- 22:29SPEAKER_01: from it or how it works because we don't know yet.
- 22:32SPEAKER_01: So how do you as a company deal with that unpredictability of the thing that you're selling?
- 22:41SPEAKER_03: Yeah, that's a great question.
- 22:45SPEAKER_03: And I think part of it is trying to make as much of it controllable and predictable as possible.
- 22:51SPEAKER_03: So, for example,
- 22:53SPEAKER_03: one of the big weaknesses of the approaches people,
- 22:57SPEAKER_03: including ourselves,
- 22:58SPEAKER_03: did before and other approaches that people are currently still doing is Is
- 23:03SPEAKER_03: that they have a lot of jitter,
- 23:05SPEAKER_03: a lot of variability in the hardware and the software itself, right?
- 23:08SPEAKER_03: So we focused and the team worked incredibly hard,
- 23:13SPEAKER_03: incredibly talented people work incredibly hard to make sure that that jitter,
- 23:19SPEAKER_03: that variability was all reduced down to as little as physically possible with
- 23:25SPEAKER_03: the systems that exist.
- 23:27SPEAKER_03: And so I think if we can at least control what we can,
- 23:31SPEAKER_03: that allows us to better cope with what we can't control.
- 23:35SPEAKER_03: And at the moment,
- 23:35SPEAKER_03: this uncertainty and predictability is actually one of the draw cards
- 23:39SPEAKER_03: of what makes this technology interesting.
- 23:41SPEAKER_03: These are non-deterministic systems.
- 23:43SPEAKER_03: Understanding where you can use non-deterministic systems is part of the challenge
- 23:48SPEAKER_03: and what's great and Gratifying to us is that there are so many people
- 23:53SPEAKER_03: who are excited to use the tools we've built,
- 23:55SPEAKER_03: work on these problems with us together.
- 23:58SPEAKER_03: It's a very rare case where you can have customers come and help you solve
- 24:02SPEAKER_03: the technology and then want to build their own technology off it.
- 24:05SPEAKER_02: So we're very fortunate with the community that's really rapidly growing in this space.
- 24:11SPEAKER_00: Yes, as someone with a background in law, Brett,
- 24:14SPEAKER_00: obviously any research that involves humans or components of humans and
- 24:18SPEAKER_00: where humans are likely to be interacting with technology can carry some risks.
- 24:23SPEAKER_00: So I'm just wondering what your company's protocols
- 24:26SPEAKER_00: and rules around managing risks are,
- 24:29SPEAKER_00: particularly when you start to get sort of unpredictable results from something
- 24:33SPEAKER_00: that you're doing.
- 24:35SPEAKER_03: Yeah, well,
- 24:36SPEAKER_03: I think that that's, I mean,
- 24:37SPEAKER_03: especially when you say any technology that humans are interacting with,
- 24:40SPEAKER_03: that's pretty much all technology.
- 24:43SPEAKER_03: I guess there's maybe some parts that aren't anymore with some
- 24:46SPEAKER_03: the new agentic workflows,
- 24:47SPEAKER_03: but even so, maybe that carries even more risks to humans.
- 24:52SPEAKER_03: But I think it's really important to break it down into,
- 24:55SPEAKER_03: there's a regulatory component,
- 24:58SPEAKER_03: And then there's the ethical component as well, right?
- 25:00SPEAKER_03: And what is risk?
- 25:01SPEAKER_03: Because there's a risk of causing physical, clear,
- 25:05SPEAKER_03: obvious harm,
- 25:07SPEAKER_03: but there is also the risk to,
- 25:11SPEAKER_03: let's say, the character or the morality of the humans, right?
- 25:17SPEAKER_03: And so broadly we've identified three key areas,
- 25:19SPEAKER_03: because we've worked with independent ethicists, we've published many, many papers,
- 25:26SPEAKER_03: on these issues because we take it seriously and we want this research
- 25:29SPEAKER_03: to be developed and applied responsibly.
- 25:32SPEAKER_03: And so there's probably three areas.
- 25:35SPEAKER_03: There's one, which is the sourcing of the material, right?
- 25:39SPEAKER_02: That's probably the one you're thinking of when you're saying interacting with humans, right?
- 25:44SPEAKER_02: We take some material.
- 25:45SPEAKER_02: Now that material is normally blood as what we use.
- 25:49SPEAKER_02: You can use skin.
- 25:49SPEAKER_02: Some people have recently used urine
- 25:51SPEAKER_02: and they developed these pluripotent stem cells,
- 25:54SPEAKER_02: which are of the stem cell groups by far the least controversial.
- 26:00SPEAKER_03: Whether that makes sense or not,
- 26:01SPEAKER_03: I think it does,
- 26:02SPEAKER_03: but I'll leave that separately,
- 26:04SPEAKER_03: but at the very least, undisputably, it's the least controversial approach.
- 26:08SPEAKER_03: Even fairly conservative groups like the Vatican Church are broadly not opposed
- 26:14SPEAKER_03: to these, if you look into that literature.
- 26:15SPEAKER_03: Embryonic stuff, yes.
- 26:17SPEAKER_03: Induced polypotent stem cells, less so.
- 26:20SPEAKER_03: And so that's one.
- 26:22SPEAKER_03: The second one is the applications, right?
- 26:25SPEAKER_03: What are you going to use it for?
- 26:27SPEAKER_03: And as many applications as I said,
- 26:28SPEAKER_03: generally when you look at drug screening,
- 26:30SPEAKER_03: disease modeling, people
- 26:32SPEAKER_02: are broadly very supportive of that.
- 26:36SPEAKER_02: You get more speculative questions when
- 26:38SPEAKER_03: you start discussing things like automation
- 26:40SPEAKER_03: and robotics and general intelligence.
- 26:44SPEAKER_03: Now what's interesting is there's some recent work
- 26:46SPEAKER_03: and generally people are actually becoming more and more positive about this
- 26:50SPEAKER_03: as it gets developed.
- 26:53SPEAKER_03: Do hope that part of that is due to the approach that we've been taking
- 26:57SPEAKER_03: of really engaging with the ethicists.
- 26:59SPEAKER_03: But this is the other thing, the applications, right?
- 27:03SPEAKER_02: The final one,
- 27:03SPEAKER_02: which is the most quirky and
- 27:06SPEAKER_03: the most important perhaps,
- 27:08SPEAKER_03: but also the most difficult to address,
- 27:11SPEAKER_03: is what's the moral status of these devices themselves?
- 27:16SPEAKER_03: And you refer to it as part of human behavior.
- 27:18SPEAKER_03: And it is in the same way that blood donation is a part of human.
- 27:26SPEAKER_02: But that doesn't mean that it should be imbued with humanity.
- 27:29SPEAKER_03: We could dive into the depths of the Latin descriptions of these systems
- 27:34SPEAKER_03: and go back even further to Aristotle, Plato, and talk about definitions of life.
- 27:38SPEAKER_03: But in short, we think that there's a very scientific way to approach these.
- 27:43SPEAKER_03: And essentially, it's one, define the language.
- 27:46SPEAKER_03: If you want to say, oh, I'm worried it's conscious, what do you mean by conscious?
- 27:50SPEAKER_03: Because I've just attended two different conferences.
- 27:52SPEAKER_03: One's a stem cell one and one is a computational one.
- 27:56SPEAKER_03: And they're very different definitions two days apart.
- 28:01SPEAKER_03: Once we've defined what the language is, we need to define what metrics we're using.
- 28:05SPEAKER_03: If we're saying, well,
- 28:06SPEAKER_03: this is conscious or this has agency,
- 28:08SPEAKER_03: we need to be able to quantify and say, well, this is a marker of agency.
- 28:13SPEAKER_03: This is a marker of consciousness.
- 28:15SPEAKER_03: And maybe there are 12 different things we need to look for,
- 28:18SPEAKER_03: but we need to actually put the work in and do that.
- 28:21SPEAKER_03: And it's not easy, but it's right.
- 28:24SPEAKER_03: And then finally,
- 28:25SPEAKER_03: we can start to say,
- 28:25SPEAKER_03: okay, When is it right or wrong to build a system
- 28:30SPEAKER_02: that might have some of these?
- 28:32SPEAKER_02: Now I'll say right now,
- 28:33SPEAKER_02: I do not believe any of the systems we build now
- 28:36SPEAKER_02: or even could build would show anything like consciousness.
- 28:39SPEAKER_02: But I would also say that it's wrong to think that just
- 28:42SPEAKER_02: because it may have some basic level of consciousness, we should never test on it.
- 28:47SPEAKER_02: Why?
- 28:47SPEAKER_02: Because we already accept that we test on things that I feel certain are conscious.
- 28:52SPEAKER_02: There's animals that are tested,
- 28:53SPEAKER_02: there's mice,
- 28:54SPEAKER_02: there's monkeys, there's even people when we need to do clinical trials.
- 28:58SPEAKER_02: So we accept that there is a degree where we're prepared to test on something
- 29:01SPEAKER_02: with amoral status.
- 29:03SPEAKER_02: And so if this could replace or reduce the need to use animals to
- 29:07SPEAKER_02: a meaningful extent,
- 29:09SPEAKER_02: Even in the incredibly unlikely chance that one day it does have some sort
- 29:13SPEAKER_02: of moral status,
- 29:15SPEAKER_02: it's not immediately obvious that you would just stop working on it.
- 29:18SPEAKER_02: Because you might be actually saving things with far more consciousness
- 29:22SPEAKER_02: and more capability of suffering.
- 29:25SPEAKER_00: Yeah, and I think that is a potentially really useful thing.
- 29:30SPEAKER_00: Use case for this technology is we could have the same advancements in science
- 29:37SPEAKER_00: and knowledge without some of the harms that we currently accept but still do
- 29:44SPEAKER_00: to living creatures.
- 29:47SPEAKER_00: Yeah, I'm just curious,
- 29:48SPEAKER_00: are you guys subject to governance from any ethics committees anywhere,
- 29:51SPEAKER_00: like a university ethics committee?
- 29:54SPEAKER_02: We do.
- 29:55SPEAKER_02: We have ethical approvals for whenever we generate these stem cell lines.
- 30:02SPEAKER_03: There's not a lot of other formal ethical components.
- 30:06SPEAKER_03: There are, of course, regulatory components about how you run a PC2 lab or BSL2 lab.
- 30:12SPEAKER_03: If you're not in Australia, that's the other common one.
- 30:16SPEAKER_03: And we follow all of those very strictly.
- 30:19SPEAKER_02: The biggest regulatory push is actually self-regulation.
- 30:25SPEAKER_02: And as I said,
- 30:25SPEAKER_02: we actually engage with independent ethicists and we're really leading,
- 30:31SPEAKER_02: I think it's fair to say from an industry perspective,
- 30:33SPEAKER_02: we are leading the field in pushing for responsible developments of these classes
- 30:40SPEAKER_03: of technology.
- 30:41SPEAKER_03: And we're doing that,
- 30:42SPEAKER_03: as I said,
- 30:43SPEAKER_03: those different breakdowns I gave you before,
- 30:46SPEAKER_02: like we are pushing through all of that.
- 30:48SPEAKER_02: This isn't just a matter of sort of ethics washing or saying,
- 30:51SPEAKER_02: you know, hey, hey, we're really ethical, but don't look too closely.
- 30:54SPEAKER_02: When you look closely, you'll see we are actually walking what we talk about.
- 31:00SPEAKER_01: Yeah, that's super interesting.
- 31:04SPEAKER_01: I think one of the questions that I have is around the blood donation.
- 31:08SPEAKER_01: So who do you get these blood donations from?
- 31:11SPEAKER_01: How frequently?
- 31:12SPEAKER_01: And do you think the genetics of the person impacts the stem cells
- 31:17SPEAKER_01: that you eventually grow from those donations?
- 31:20SPEAKER_01: Yeah.
- 31:21SPEAKER_02: So naturally, I can't tell you who we get them from because of confidentiality.
- 31:26SPEAKER_02: But we get them from people who are aware and able to provide informed consent,
- 31:31SPEAKER_02: broadly speaking.
- 31:35SPEAKER_02: How often would they?
- 31:35SPEAKER_02: We've only ever made five.
- 31:36SPEAKER_02: We have five in-house lines we've made.
- 31:38SPEAKER_02: And contrary to what a lot of people think,
- 31:40SPEAKER_02: we don't have to make more for the perspective of biocomputing.
- 31:43SPEAKER_02: The only reason we would want to scale this up is if and when we want to go down
- 31:48SPEAKER_02: the personalized medicine route.
- 31:50SPEAKER_02: But in that case, the donation would be in that person's bespoke interest.
- 31:57SPEAKER_02: Now, regarding do we see differences from different people?
- 32:00SPEAKER_02: Yes, we do.
- 32:04SPEAKER_02: They tend to be subtle on the performance side, but you can see some differences.
- 32:10SPEAKER_02: Getting it to be really reproducible in a certain standard is
- 32:13SPEAKER_02: beyond what we've been able to test for now because that is a very large type
- 32:16SPEAKER_02: of study.
- 32:18SPEAKER_02: We'd like to do it one day, maybe partner up with people who specialize in this sort of work.
- 32:22SPEAKER_02: But in the meantime, I can say we do see differences.
- 32:26SPEAKER_02: But again,
- 32:27SPEAKER_02: this isn't sort of,
- 32:28SPEAKER_02: we wouldn't for computation look to be generating hundreds of these lines.
- 32:31SPEAKER_02: What you really want is one stable, reproducible line for the computation.
- 32:36SPEAKER_02: Because again, you asked before, how do we deal with this system that's inherently variable?
- 32:41SPEAKER_02: By minimizing every bit of variability
- 32:43SPEAKER_03: we can.
- 32:44SPEAKER_03: So that's not really,
- 32:46SPEAKER_03: you know,
- 32:46SPEAKER_03: when we started out,
- 32:47SPEAKER_03: a lot of people sort of threw their hands up in the air saying,
- 32:49SPEAKER_03: think of the children,
- 32:51SPEAKER_03: they're going to be taking advantage and trying
- 32:53SPEAKER_02: to pull out people's brains.
- 32:54SPEAKER_02: And it's like brush notes,
- 32:56SPEAKER_02: like one donor of a small bit of blood,
- 32:59SPEAKER_02: like you'd give to any doctor for testing,
- 33:02SPEAKER_02: was all we ever needed from a computation perspective.
- 33:05SPEAKER_02: Everything else is just a biomedical perspective of how do you help
- 33:08SPEAKER_00: people?
- 33:11SPEAKER_00: Yeah, so...
- 33:13SPEAKER_00: Is that sort of a key focus then is the biomedicine and health,
- 33:19SPEAKER_00: improvements in health,
- 33:20SPEAKER_00: was that sort of an underlying kind of motivator for you guys originally?
- 33:25SPEAKER_03: That's always been one end of the spectrum.
- 33:30SPEAKER_03: A lot of the team's background come from that preclinical
- 33:34SPEAKER_03: or even sometimes clinical perspective.
- 33:38SPEAKER_02: It's important.
- 33:39SPEAKER_02: It's between 93 to 99.9% of clinical trials in neurological psychiatric areas fail.
- 33:49SPEAKER_02: And why is that?
- 33:51SPEAKER_02: Because the preclinical testing to get there, for the most part, is obviously not sufficient.
- 33:57SPEAKER_02: And we've already published and shown that you can get a broader range of results
- 34:02SPEAKER_02: with these approaches.
- 34:04SPEAKER_02: And of course you can, because...
- 34:07SPEAKER_02: When you're testing something before it goes to clinic, you want to test the function, right?
- 34:13SPEAKER_02: But like, let me ask you this, what is the function of the brain?
- 34:20SPEAKER_00: Well, I guess it's essentially to control Yeah, it's multifunctional.
- 34:27SPEAKER_00: But one of the primary functions is to control the body and to regulate
- 34:32SPEAKER_00: the body's organs, I think.
- 34:34SPEAKER_00: And then you get into sort of the consciousness sort of intellectual functions as well.
- 34:40SPEAKER_03: So it's to take information either from organs or from outside the world and
- 34:44SPEAKER_03: to do something with it.
- 34:45SPEAKER_02: I noticed you didn't say to have lots of spiking electrical activity.
- 34:53SPEAKER_02: If you read these pre-clinical papers,
- 34:56SPEAKER_02: they say,
- 35:00SPEAKER_02: and yet if you look at any pre-clinical paper,
- 35:07SPEAKER_02: they'll say we did functional testing,
- 35:09SPEAKER_02: but what they mean is we look to see if the activity went up and down.
- 35:13SPEAKER_02: And more activity is not always better.
- 35:15SPEAKER_02: If you think of the patient with epilepsy,
- 35:17SPEAKER_02: their brains can be very active at certain points,
- 35:21SPEAKER_02: not terribly functional when they're having a seizure and have
- 35:25SPEAKER_02: that heightened electrical activity, hyper-excitability.
- 35:28SPEAKER_02: So function is the ability to process and to do something with that information.
- 35:33SPEAKER_02: That something depends on the area.
- 35:35SPEAKER_02: So if you're not testing that function in the dish,
- 35:38SPEAKER_02: how are you going to accurately predict the outcome in a clinical trial?
- 35:41SPEAKER_02: And if you're testing We know there's differences between mouse and human.
- 35:44SPEAKER_02: We know there's differences between monkey and human, quite big ones.
- 35:48SPEAKER_02: And so if you're not testing the right material or the right function,
- 35:52SPEAKER_02: it's kind of like, of course, there's going to be troubles translating that.
- 35:55SPEAKER_02: And we offer a way to try and make that better.
- 35:59SPEAKER_01: Yeah, that is super interesting.
- 36:01SPEAKER_01: And something that because I did an undergrad in psychology and I think...
- 36:07SPEAKER_01: It was something that I never quite understood why you're looking at peaks
- 36:12SPEAKER_01: of activity and then suddenly associating that with the thought or the process
- 36:18SPEAKER_01: of what you're doing.
- 36:18SPEAKER_01: That correlation or causation that they're suggesting has never happened.
- 36:24SPEAKER_01: Quite understood.
- 36:25SPEAKER_01: I've not been able to understand exactly how they've gotten to those outcomes.
- 36:30SPEAKER_01: So that itself is really interesting.
- 36:33SPEAKER_01: And I guess on the personalized medical kind of pathway,
- 36:39SPEAKER_01: how would you see the technology
- 36:43SPEAKER_01: that you're creating here actually impact personalized medicine?
- 36:50SPEAKER_03: So I'll give an example.
- 36:52SPEAKER_03: My brother-in-law has both non-verbal autism and epilepsy.
- 36:58SPEAKER_03: And when he developed epilepsy, he was
- 36:59SPEAKER_02: already 18 or so.
- 37:03SPEAKER_02: And because he doesn't communicate efficiently,
- 37:07SPEAKER_02: the only way for them to test which anti-seizure medications worked best for him was
- 37:12SPEAKER_02: to give him different drugs at different doses and wait to see if he would have
- 37:16SPEAKER_02: a seizure.
- 37:17SPEAKER_02: Every time he had a seizure, or nearly every time, he had to be hospitalized.
- 37:21SPEAKER_02: It has taken years and years,
- 37:22SPEAKER_02: and they're still working to try and figure out the right approach.
- 37:25SPEAKER_02: What if we could have taken a small sample of blood,
- 37:28SPEAKER_02: set up 30 what we call CL1 units,
- 37:33SPEAKER_02: Grown cultures and tested a range of different drugs on them across three
- 37:37SPEAKER_03: to six months and came down and said,
- 37:39SPEAKER_03: all right, these are the top three candidates out of 30 options.
- 37:45SPEAKER_03: Cut it down to the top three.
- 37:48SPEAKER_01: Okay, that makes total sense.
- 37:50SPEAKER_01: It's like basically bacterial infections in that category.
- 37:55SPEAKER_01: You're just coming up with the new drugs that are specific for the bacterial line
- 37:58SPEAKER_01: or whatever it is, but instead it's humans.
- 38:02SPEAKER_01: So we're coming up with specific drugs that are for the specific human and the issue
- 38:07SPEAKER_01: that the human is facing.
- 38:09SPEAKER_02: And the genetic component is large.
- 38:11SPEAKER_02: And the thing is,
- 38:11SPEAKER_02: from a public health perspective,
- 38:14SPEAKER_02: because let's go there for your audience, This actually makes economic sense.
- 38:22SPEAKER_02: The cost of having people continue to come back to specialists,
- 38:27SPEAKER_02: continue to go back to hospital is huge.
- 38:30SPEAKER_02: The cost of doing these pre-clinical testing that we can do
- 38:34SPEAKER_02: at scale is actually pretty moderate.
- 38:38SPEAKER_02: And it's not just epilepsy, right?
- 38:42SPEAKER_02: It's mental health, depression, anxiety affects at least one in five people.
- 38:47SPEAKER_02: I think the numbers are going higher now.
- 38:50SPEAKER_02: But being able to help people who struggle,
- 38:55SPEAKER_02: I mean these diseases are debilitating to To the people who experience them,
- 39:03SPEAKER_02: to their friends and family.
- 39:05SPEAKER_02: And again,
- 39:05SPEAKER_02: the current way that people are treated when they go down
- 39:08SPEAKER_02: to pharmaceutical approach is as guinea pigs.
- 39:12SPEAKER_02: And I don't blame clinicians for this.
- 39:13SPEAKER_02: There has been no better method.
- 39:16SPEAKER_02: And we don't know for sure that our method, we have to have that and say we don't know for sure.
- 39:21SPEAKER_02: But what we can do is with reasonable little funding relative to what is spent, test it.
- 39:27SPEAKER_02: And we're at the start now where we've built the platform
- 39:31SPEAKER_02: that makes testing this affordable and scalable.
- 39:51SPEAKER_01: So I think what I was going to ask was about we've got the health of the person
- 39:59SPEAKER_01: and I wanted to ask about the health of the stem cells.
- 40:03SPEAKER_01: And so we've kind of touched on how these stem cells are not Really,
- 40:09SPEAKER_01: like they're living, but they're not living at the same time.
- 40:12SPEAKER_01: And, you know,
- 40:14SPEAKER_01: they don't have a conscious and they don't have morals and all that kind of stuff.
- 40:16SPEAKER_01: And that's a possibility for super future long term ideas.
- 40:23SPEAKER_01: I guess in terms of the health of the cells itself and how
- 40:27SPEAKER_01: that impacts your technology,
- 40:30SPEAKER_01: the one question I have is around cancer
- 40:35SPEAKER_01: because cancer is obviously how these cells kind of live and then deform basically.
- 40:46SPEAKER_01: Can your cells that you use develop cancer?
- 40:51SPEAKER_01: What would you do in that case?
- 40:53SPEAKER_01: And how do you think that would impact the technology?
- 40:58SPEAKER_03: Okay.
- 40:59SPEAKER_03: Yeah.
- 41:00SPEAKER_03: This is a new question.
- 41:01SPEAKER_03: I haven't had this one before.
- 41:02SPEAKER_01: It was the first thing I thought of when I actually looked up the company.
- 41:07SPEAKER_01: And like I have,
- 41:08SPEAKER_01: like there's a personal connection with cancer,
- 41:10SPEAKER_01: which is probably why I've thought of this question.
- 41:14SPEAKER_01: But yeah, I'm just so curious about it.
- 41:18SPEAKER_02: So people...
- 41:20SPEAKER_02: Do have different models with neural cultures.
- 41:26SPEAKER_02: I know there are glioblastoma models and a range of others.
- 41:31SPEAKER_02: I'm not an expert in this by any means.
- 41:34SPEAKER_02: It would be possible to introduce these types of models
- 41:38SPEAKER_02: into otherwise healthy cultures and see how that might propagate and disrupt
- 41:42SPEAKER_02: the network.
- 41:45SPEAKER_02: Cancer, I think,
- 41:46SPEAKER_02: would be, well, yeah,
- 41:48SPEAKER_02: it's not one we've looked at before,
- 41:50SPEAKER_02: but it's one that I think you could use these systems to explore.
- 41:53SPEAKER_02: Having said that, it's not our expertise.
- 41:55SPEAKER_02: And this is actually the exact reason why we wanted to make a platform technology
- 42:00SPEAKER_02: that was accessible to everyone.
- 42:01SPEAKER_02: Because there's too many important questions.
- 42:03SPEAKER_02: Like what you raise now, such a vitally important question.
- 42:08SPEAKER_02: There's no one group in the world,
- 42:10SPEAKER_02: no matter how good they are, Who can answer all of these questions?
- 42:14SPEAKER_02: The focus would be too broad, too destructive.
- 42:17SPEAKER_02: So our focus is let's provide the tooling and let's allow other people
- 42:24SPEAKER_02: who have those expertise to use the tooling to be able to answer these questions.
- 42:31SPEAKER_01: Good answer that's also a non-answer.
- 42:35SPEAKER_02: Well, we don't do it.
- 42:36SPEAKER_02: So anything I could give you would be...
- 42:39SPEAKER_02: You know, would just be making it up on the spot, which I can do.
- 42:43SPEAKER_02: But, you know, again, as I say, it's too important.
- 42:46SPEAKER_02: So, like,
- 42:47SPEAKER_02: rather what I'd rather do is put the tool into the hands of someone who does have
- 42:51SPEAKER_02: that deep technical knowledge and say, well, I know X, Y, and Z.
- 42:55SPEAKER_02: Here's how I can explore it.
- 42:56SPEAKER_02: And there are people who are interested in doing that.
- 43:01SPEAKER_01: Awesome.
- 43:01SPEAKER_01: And how do they access that platform?
- 43:06SPEAKER_02: So we actually have three key markets that we can actually provide the technology through.
- 43:14SPEAKER_02: One is the standard,
- 43:15SPEAKER_02: we can sell devices, we can sell you server racks of devices pretty scalably.
- 43:20SPEAKER_02: And we're hoping to see some support come through.
- 43:23SPEAKER_02: Actually, I know you're based in Western Australia.
- 43:25SPEAKER_02: We're hoping to see some support come through and get a few
- 43:28SPEAKER_02: of these server racks there for people in Western Australia to use.
- 43:32SPEAKER_02: There's the cloud.
- 43:35SPEAKER_02: So if you don't grow your own cells,
- 43:37SPEAKER_02: because most people out there don't grow neural cells in their garage,
- 43:41SPEAKER_02: I hope not anyway.
- 43:43SPEAKER_02: And most people then don't have bio labs to do it in.
- 43:46SPEAKER_02: You can actually access it over the cloud.
- 43:47SPEAKER_02: So we'll grow the cells for you.
- 43:49SPEAKER_02: You can access them, program them, interact with them through very simple code that we use.
- 43:53SPEAKER_02: And this is called Cortical Cloud.
- 43:56SPEAKER_02: And then the final way is a service business where essentially you come to us
- 44:01SPEAKER_02: and you give us a question and we'll go away and do that exploration and come back
- 44:07SPEAKER_02: with the answers for you.
- 44:08SPEAKER_02: And there's a bit of options with all three, but those are the three options broadly.
- 44:14SPEAKER_01: So, really,
- 44:15SPEAKER_01: that allows you to facilitate all of the kind of research
- 44:18SPEAKER_01: and answering questions like the one I asked,
- 44:20SPEAKER_01: where people have those individual questions,
- 44:22SPEAKER_01: they can come to you and be like, hey, what do you think the best way to do this is?
- 44:26SPEAKER_01: Am I able to access the cells myself?
- 44:28SPEAKER_01: Blah, blah, blah.
- 44:29SPEAKER_01: And you kind of help facilitate answering those research questions.
- 44:34SPEAKER_02: Exactly.
- 44:35SPEAKER_02: So we always do that.
- 44:36SPEAKER_02: And our goal really is, you know, of the opinion here.
- 44:40SPEAKER_02: And I think this is quite a different approach to the classic sort
- 44:44SPEAKER_02: of Silicon Valley approach of like take as much as you can and,
- 44:49SPEAKER_02: you know, kind of winner takes all mentality.
- 44:52SPEAKER_02: We think that this is going to be a big field.
- 44:54SPEAKER_02: And the best way to make it bigger is to give people a piece of it, let them share.
- 44:58SPEAKER_02: So if you use our technology, you...
- 45:03SPEAKER_02: Own what you develop.
- 45:05SPEAKER_02: Oh, awesome.
- 45:06SPEAKER_02: So if you're using the cloud, we're not trying to stamp that and you actually own it.
- 45:11SPEAKER_02: That's yours.
- 45:11SPEAKER_02: We actually have people now who have startups spun out of using our technology.
- 45:16SPEAKER_02: Right.
- 45:17SPEAKER_02: And that's great.
- 45:18SPEAKER_02: Because that's good for us.
- 45:19SPEAKER_02: And really,
- 45:19SPEAKER_02: as a business model,
- 45:21SPEAKER_02: and this sounds,
- 45:22SPEAKER_02: we've been saying this long before,
- 45:23SPEAKER_02: they were the most valuable company in the world,
- 45:25SPEAKER_02: but we really want to be the NVIDIA of biological computing.
- 45:29SPEAKER_02: And if you think about it, NVIDIA, the big...
- 45:34SPEAKER_02: Areas have made them so valuable.
- 45:35SPEAKER_02: They didn't develop those applications.
- 45:37SPEAKER_02: They didn't develop cryptocurrency.
- 45:39SPEAKER_02: They didn't develop AI.
- 45:40SPEAKER_02: They didn't really develop gaming.
- 45:42SPEAKER_02: They developed chips for gaming.
- 45:43SPEAKER_02: They developed chips for AI.
- 45:45SPEAKER_02: But what they did was had a multipurpose platform that Allowed anybody to access it.
- 45:51SPEAKER_02: They built CUDA,
- 45:52SPEAKER_02: which is their software,
- 45:54SPEAKER_02: if you're not familiar with that, that allows people to use them.
- 45:57SPEAKER_02: And that's actually the reason.
- 45:58SPEAKER_02: The fact that they built a successful platform is the reason they are
- 46:01SPEAKER_02: the most valuable company in the world.
- 46:03SPEAKER_02: Not that they make GPUs.
- 46:04SPEAKER_02: Other people make GPUs.
- 46:06SPEAKER_02: They didn't make GPUs or only GPUs.
- 46:09SPEAKER_02: They made an ecosystem.
- 46:10SPEAKER_02: And we want to try to do the same thing.
- 46:12SPEAKER_02: We grow it by giving it away.
- 46:17SPEAKER_01: That's, yeah.
- 46:18SPEAKER_00: Yeah, that's really fascinating.
- 46:20SPEAKER_00: And, yeah,
- 46:21SPEAKER_00: I think an interesting comparison with NVIDIA there,
- 46:24SPEAKER_00: I think that's, yeah, really telling.
- 46:27SPEAKER_02: These days,
- 46:28SPEAKER_02: everyone wants to be the NVIDIA for X.
- 46:29SPEAKER_02: I feel like an old school hipster where I'm like,
- 46:32SPEAKER_02: hey, we said that before it was cool.
- 46:34SPEAKER_02: But it really was the basis of why we took this approach all the way back since 2019.
- 46:40SPEAKER_00: Yeah.
- 46:42SPEAKER_00: Okay.
- 46:43SPEAKER_00: Look, and...
- 46:44SPEAKER_00: In the grand scheme of things, you guys have moved very quickly.
- 46:47SPEAKER_00: You know, 2019 isn't that long ago.
- 46:49SPEAKER_00: That might feel like it.
- 46:52SPEAKER_02: It was a while ago,
- 46:53SPEAKER_02: but, you know, the great thing about it now is that, like, the team's growing.
- 46:58SPEAKER_02: We're getting more and more amazing people across the world.
- 47:01SPEAKER_02: We're getting more and more,
- 47:03SPEAKER_02: not just a team of people employed by Cortical Labs,
- 47:05SPEAKER_02: but the sort of Synthetic biological intelligence team across the world is growing.
- 47:11SPEAKER_02: We have amazing collaborators.
- 47:13SPEAKER_02: We have amazing customers and users who are doing really cool things.
- 47:18SPEAKER_02: And it's getting to the weird point now where I'm seeing things online
- 47:21SPEAKER_02: and I'm going, oh, that's cool.
- 47:23SPEAKER_02: And I'll scroll through it and I'm, oh, that's an AL technology and I didn't even know it.
- 47:28SPEAKER_02: I loved them the other day and I saw someone said, brain cells playing Mario Kart.
- 47:33SPEAKER_02: Oh, that's AL systems.
- 47:35SPEAKER_02: Like that's such an amazing feeling to know
- 47:36SPEAKER_02: that you have enabled this cutting edge research
- 47:41SPEAKER_02: and it's just growing now organically, biologically.
- 47:45SPEAKER_02: Amazing.
- 47:47SPEAKER_01: How cool.
- 47:47SPEAKER_01: What's the next step for the company?
- 47:50SPEAKER_02: Well, we're currently raising now and then the plan is to grow,
- 47:53SPEAKER_02: to build the next levels of the technology so we have our CL1
- 47:57SPEAKER_02: and provide more access.
- 47:59SPEAKER_02: I mean, unfortunately at the moment,
- 48:02SPEAKER_02: And look, as a company this is a nice problem to have, but it's still a problem.
- 48:05SPEAKER_02: But we have more demand than we can supply.
- 48:09SPEAKER_02: And so we need some,
- 48:10SPEAKER_02: we're currently gathering some capital and we're talking with some great investors
- 48:15SPEAKER_02: to grow the company.
- 48:16SPEAKER_02: And that's really going to unlock the next level.
- 48:20SPEAKER_02: And then beyond that, improve the technology.
- 48:23SPEAKER_02: And there's lots of ways we can do that.
- 48:25SPEAKER_02: Lots of pathways forward.
- 48:26SPEAKER_02: And this is only the beginning.
- 48:28SPEAKER_02: And I think that's what also makes it really exciting to me.
- 48:32SPEAKER_02: The key thing is that we, again, reiterate what I said at the start.
- 48:37SPEAKER_02: We have a ground truth that biology can achieve so many amazing things.
- 48:42SPEAKER_02: I remember in the early days talking to a rather large name in the tech sphere,
- 48:47SPEAKER_02: I won't say who,
- 48:48SPEAKER_02: and they said to me, oh,
- 48:48SPEAKER_02: but biology,
- 48:50SPEAKER_02: you get a lot of stuff for cheap,
- 48:52SPEAKER_02: but it's so limited, you know, with what you can do with it.
- 48:56SPEAKER_02: And I said to them, limited?
- 48:59SPEAKER_02: Look at the world.
- 49:01SPEAKER_02: This was built through biological intelligence.
- 49:03SPEAKER_02: Yes, we made machines,
- 49:04SPEAKER_02: we've done everything, but it's built with biological intelligence of the core.
- 49:09SPEAKER_02: And so we can figure out how to leverage that better.
- 49:14SPEAKER_02: The things we could unlock,
- 49:15SPEAKER_02: I think,
- 49:15SPEAKER_02: really can help not just make an industry, but hopefully make our species better.
- 49:22SPEAKER_01: And I think this might be the last question unless Craig's got another one at the end.
- 49:27SPEAKER_01: What about the next steps for you specifically?
- 49:32SPEAKER_02: Well, you know, I'm committed to the mission.
- 49:35SPEAKER_02: So, you know, you kind of have to be pretty driven to continue to pursue these lines.
- 49:41SPEAKER_02: A little obsessive.
- 49:43SPEAKER_02: And so we want to keep pursuing and figure out.
- 49:45SPEAKER_02: And we've got a few exciting things on the horizon I haven't been able to talk about.
- 49:49SPEAKER_02: But hopefully they'll be out soon enough and people can sort of see the directions we're going.
- 49:55SPEAKER_00: Great.
- 49:56SPEAKER_00: Yeah.
- 49:56SPEAKER_00: No, there's nothing further from me.
- 49:57SPEAKER_00: I think you've given us a great overview of you and your company's activities,
- 50:02SPEAKER_00: and it's been really fascinating.
- 50:03SPEAKER_00: Thanks for your time, Brett.
- 50:05SPEAKER_00: It's been a pleasure to chat with both of you.
- 50:07SPEAKER_01: Wonderful.
- 50:08SPEAKER_01: Thanks, Brett.
- 50:11SPEAKER_00: So that was our conversation with Dr Brett Kagan, who was in LA Airport.
- 50:17SPEAKER_01: It was a little noisy.
- 50:18SPEAKER_00: Yeah, it was.
- 50:19SPEAKER_00: But amongst all the noise,
- 50:21SPEAKER_00: there was some really interesting conversations about not only the technology,
- 50:24SPEAKER_00: but how it's sort of evolving quite quickly,
- 50:28SPEAKER_00: the sort of risks with that as well as the benefits.
- 50:31SPEAKER_00: So yeah, I thought it was really fascinating.
- 50:32SPEAKER_00: And the fact that I think Brett originally had a law background obviously resonates with me too.
- 50:37SPEAKER_00: So yeah, that was another little bonus.
- 50:40SPEAKER_01: Yeah, that means that we expect some big things from you, Craig.
- 50:43SPEAKER_01: Are you going to invent your own computer or invent your own intelligence?
- 50:47SPEAKER_00: Yeah.
- 50:50SPEAKER_00: I mean, I have got my own intelligence, but I don't know how innovative it is.
- 50:56SPEAKER_01: Yeah, that's fair.
- 50:59SPEAKER_01: I feel like there's a lot we can learn from the conversations we've had with Brett.
- 51:05SPEAKER_01: It was just, yeah, it was really, really fascinating.
- 51:08SPEAKER_01: And I love also the fact that he wasn't able to answer all the questions that I had.
- 51:14SPEAKER_01: And it is about kind of integrating and allowing research with the infrastructure
- 51:20SPEAKER_01: that they create.
- 51:21SPEAKER_01: I think it's such a cool idea.
- 51:23SPEAKER_01: And it's also slightly scary,
- 51:24SPEAKER_01: but that's okay because I think it'll just keep going forward anyway.
- 51:28SPEAKER_00: Yeah.
- 51:29SPEAKER_00: No, I think the model they've got where they do allow other people
- 51:32SPEAKER_00: to leverage their platform to create some innovative science is really
- 51:38SPEAKER_00: a really good part of their business structure.
- 51:42SPEAKER_00: So I'm looking forward to seeing what happens next.
- 51:44SPEAKER_01: Yeah, me too.
- 51:45SPEAKER_01: And if you guys want to find out anything more about Cortical Labs and also how they
- 51:49SPEAKER_01: or how you can use their computers for research and any of
- 51:53SPEAKER_01: the other infrastructure they've created or the clouds,
- 51:57SPEAKER_01: the cloud drives that they've started,
- 51:59SPEAKER_01: you can have a look at our show notes and we'll have some links there available
- 52:03SPEAKER_01: for you.
- 52:04SPEAKER_01: But thanks for listening and I hope you enjoyed the episode.
- 52:10SPEAKER_01: Meaning of Health is produced at the University of Western Australia with
- 52:13SPEAKER_01: the support of the School of Population and Global Health,
- 52:17SPEAKER_01: the Digital Development and Enhancement Team and the Infrastructure
- 52:21SPEAKER_01: and Platforms Team.
- 52:23SPEAKER_01: The podcast is produced and hosted by Craig Cumming,
- 52:26SPEAKER_01: Courtney Webber, Leah Roberts, Ingrid Stacey, Dorinda Tartt.
- 52:31SPEAKER_01: Music, mixing and post-production is by Craig Cumming.
- 52:35SPEAKER_01: You can find more information about us at our website, meaningofhealth.com.au.
- 52:40SPEAKER_01: If you'd like to get in touch with us,
- 52:42SPEAKER_01: you can email us at meaningofhealthatoutlook.com and contact us on Instagram
- 52:48SPEAKER_01: or bluesky at healthmeanswhat or find us on Facebook or LinkedIn by searching for
- 52:54SPEAKER_01: The Meaning of Health.