Introduction to Fargo by Wells Fargo
00:00:00
Speaker
Fargo, the new virtual assistant from Wells Fargo, makes banking faster and easier. Like this. Fargo, what's my checking account routing number? And this. Fargo, turn off my debit card. And this. Fargo, what did I spend on groceries last month? And that's just the beginning. Do you Fargo? You can. In the Wells Fargo mobile app. Learn more at wellsfargo.com slash Get Fargo. Terms and conditions apply. Your mobile carrier's availability and message and data rates may apply. Wells Fargo Bank and a member of DIC.
00:00:31
Speaker
Hello, saver. Whether you're saving for that trip to the tropics or saving for an emergency, now is the time to take advantage of Wells Fargo's savings options. Wells Fargo offers savings accounts that can help you save towards your goals. So, what are you saving for? Visit a Wells Fargo branch or wellsfargo.com slash save to open a savings account today. Wells Fargo Bank N.A. Member FDIC.
Human Brain Organoids in Robotics Experimentation
00:01:06
Speaker
We integrated the tiny human brains we grow in jars into robots. Something that literally every science fiction series said will end terribly. Really, really terribly. A lab out of UC San Diego, the Maturi lab, decided to let a little brain organoid control a robot spider, in an effort to let it start to integrate with its environment. Initially they found that unstimulated brain organoids fired around 3,000 times per minute, but when they stimulated it and allowed it to learn the environment,
00:01:33
Speaker
it ended up firing 300,000 times. This is a lot more similar to what we see in newborn babies. We already know that the brain organoids, even without stimulation, still end up giving delta brainwaves, which is the kind of brainwaves you see in a pre-term baby, or somebody who's sleeping. The brain organoids can form every structure in the human brain during development.
00:01:52
Speaker
One of the large issues with it is integrating it with a circulatory system, which we might be able to do inside of robots. Maybe one day my AI robot can actually love me. But in other horrifying dystopian news, we're actually able to use the genomes that we found from Neanderthals to make Neanderthal brain organoids to help us
Ethical Implications of Neanderthal Brain Research
00:02:10
Speaker
try and figure out how they differed from us or how similar they were, which leads me to an ethical question.
00:02:16
Speaker
We are allowed to use human brain organoids right now. They seem to have the developmental brainwaves of a preterm baby. However, when we're integrating them into robots, giving them material that they can actually interact with seems to get them beyond that developmental stage, at least in what we're seeing right now. Would it be more ethical to use Neanderthal brain organoids?
00:02:40
Speaker
What you just saw was a video by the content creator known as GT on TikTok, a video all about the use of human brain organoids in controlling robots. Now let me take a minute here and unpack that. Essentially an organoid is sort of like an organ. It's made from human stem cells and they can essentially be teased and harvested to create any organ in the human body. That means you can have a
00:03:05
Speaker
a heart organoid or a liver organoid. And yes, that does include a brain organoid. Now, there are differences between an organoid and an actual organ. And in the case of a brain organoid, if you go through any of the literature available or any of the videos on YouTube, the creators insist time and time again that they are not conscience and they don't experience things.
00:03:29
Speaker
Yet as a researcher myself, I haven't found a satisfying reason for that or a real reassuring explanation for their confidence in that. So there's a lot of unanswered and scary questions today.
00:03:45
Speaker
In this episode of the Breaking Math Podcast, we're gonna talk about a few things involving brain organoids and some other topics. We're gonna talk about what happens when stem cells are grown into brain organoids. We will talk about the difference between artificial neural networks in machines and the networks in brains and talk about why perhaps brains are more efficient and better at what they do.
GT on Brain Organoids and AI in Breaking Math Podcast
00:04:08
Speaker
And of course, we'll talk about the ethical implications of using brain organoids and what the future of that area
00:04:14
Speaker
might mean. I am extremely happy to have today on our show, as I said earlier, the content creator known as GT. GT, thank you so much for joining me today. Yeah. Can I talk a little bit about the difference between sentience and consciousness? Please, I'm sure our audience wants to know. I certainly do as well.
Exploring Sentience vs Consciousness in Organoids
00:04:36
Speaker
So, sentience is essentially the characteristic of being able to perceive your environment and then make decisions based on that perception. Consciousness is more the theory of mind, so the ability to decide that you are separate from other individuals and understanding that they have experiences that are different
00:04:55
Speaker
than yours. All life shares sentience because all life can experience their environment but not all life is conscious. So we've pretty definitively said that the brain organoids are sentient but we haven't decided if they're conscious or conscious yet.
00:05:10
Speaker
Yeah. And I was watching so many hours of videos on this. And in fact, I'm going to put these same videos in the show notes so our audience can watch them too, and also help us to hopefully better articulate the questions to ask about these things. One thing I looked at is the number of neurons. I didn't write it here on the outline, but I know the number of neurons counted I think was at least an order of magnitude, maybe two, less than even in a baby. Do you happen to know the neuron count?
00:05:37
Speaker
Off the top of my head, I think the largest brain organoids have a billion neurons, which is very much, much smaller than a human baby. Okay. Yeah. So that's good. And the other thing I looked at, and this gave me a little bit of a sigh of relief, a little bit is pain receptors. So the human brain does not have pain receptors. Those are in our central nervous system.
00:06:00
Speaker
Now that doesn't mean that we don't experience discomfort, but that does mean that we don't experience pain. So there's lots and lots of questions, and there'll be many more questions. And I think just to call it out, the most uncomfortable thing for me is to think, even if these brain organoids don't have any semblance of consciousness like we have, or perhaps they have a semblance of it, but it's not quite the same,
00:06:24
Speaker
What does make me a bit uncomfortable is just because they don't now does not mean that the science won't perfect the process and they may not develop that in a short time. That's where I have the greatest discomfort and that's also where I have not seen a whole lot of conversations happening. I don't know if you've caught any conversations on that.
00:06:43
Speaker
I would qualify that and say that we don't have pain receptors in our brains, but we do have pain receptors in our eyes and we can grow eyes on the brain organoids. We're also capable of experiencing an awful lot of psychological discomfort. When I talked to researchers about it, how they taught them, they used ordered signals, which the brain organoids seem to prefer, and to give them feedback when they're doing something quote unquote wrong, they would give them disordered signals, which seems to cause them a degree of discomfort. At least they don't seem to like it.
00:07:13
Speaker
The beginning of this podcast started off with the robot spider, but I want to go ahead and back it up a little bit. Can you tell us about some of your earlier videos when you first discovered them and what your earlier videos were about, such as the one that was playing Pong?
00:07:29
Speaker
Yeah, so when I first started reading about them, I did a bit of a deep dive. And I started talking to the researchers who actually did it. And I was shocked to find out that they considered them sentient. So I kind of started at the beginning of the research and worked my way forward to actually putting them in robots. For the Pong one, they found that they were able to learn better and faster playing Pong than even our best AI models. And they don't entirely have an explanation for that.
00:07:59
Speaker
Actually, I think we even chatted about that a little bit in the show notes. And again, for our listeners, forgive me, I know we're gonna jump from topic to topic because there's just so much rich material. I want to remind our listeners that our entire transcript will be available in the show notes, as well as all of the citations for the
00:08:15
Speaker
papers that we read or the videos that we watch. So if you want to see any more about a given topic, check out the show notes and you should be able to find plenty. Now you had just mentioned that the brain organoids learn how to play Pong more quickly than in an artificial neural net. Yes, I saw that and I think it was also about an order of magnitude or so.
00:08:35
Speaker
The number of trials was much faster in the brain organoids. It is worth mentioning that in the artificial neural net, it can do each trial much faster. So that's another consideration when we're talking about efficiency. But overall power efficiency, the energy used by the brain organoid is less than, I want to say, three orders of magnitude less than the total energy used by the neural net.
Energy Efficiency of Brain Organoids vs Neural Networks
00:09:00
Speaker
So they're just way more efficient.
00:09:02
Speaker
You might be interested to know, and I have a pre-print of the paper from Cortical Labs, they actually started getting the brain organoids to undergo something akin to sleep. So if they're allowed to sleep, they actually have the same signaling cascade that we do when we sleep, and they're able to learn to play the game faster.
00:09:19
Speaker
Oh, wow. No, I had no idea about that. I just read a paper recently about some sort of concept of sleep used in artificial neural nets. And I didn't even get as far as opening up that paper. But yeah, sleep itself isn't even well understood in its purpose. So that's absolutely fascinating. Now, I'm curious. When we're sleeping, I know that there's a certain brain wave pattern that we have.
00:09:43
Speaker
Is that similar or different than the typical brainwave when the brain organelles are being used? Yeah, so when human beings are asleep, they experience something called delta brainwaves. And preterm babies only experience delta brainwaves. The brain organoids, they found that they had delta very quickly upon having coordinated brain activity. It's a little bit odd that they're able to have the signaling cascade of sleep independent of
00:10:13
Speaker
having the same brainwaves for either the quote-unquote sleep or wake state.
00:10:18
Speaker
So it's almost like they're dreaming. Yeah, interesting. Oh, and I also I want to mention I'm hearing two different things depending on the source. And I wonder if you may have more information on this from certain sources. I watched it says that the brain organoids essentially have all the same components as as a human brain does or rather it could. But then I heard from other sources that they're trying to grow different components of a brain and bring them together. And that kind of threw me off a little bit. And I wasn't quite sure what they were referring to.
00:10:48
Speaker
So you can make a brain in organoid that has all the components of the human brain. That's essentially the standard brain organoid. But there's a problem because once they get too large, they start to become hypoxic because they don't have a circulatory system. So another option is giving them specific stimulating factors to turn into a certain part of the brain or even extract
00:11:09
Speaker
brain tissue from fetuses, aborted fetuses, unfortunately. And then they can grow that portion and then bring them together to create a larger coordinated brain. Okay. And again, I'm hoping, you know, fingers crossed that the ethics on that are, you know, slamming the brakes until we understand what these things really are. I think, and again, happening in the US, in the US brain organoids are made from skin tissue in other countries like Australia, China, and Japan, they're using aborted fetal brain tissue.
00:11:38
Speaker
OK, OK, wow. And then I'm still very curious about, even with the world over, I know there's different ethical oversight boards at the world over. I am quite curious what they'd consider a line that they don't want to cross in other places. And yes, that is a very discomforting thing to think about. But as a member of the human race in this life, these are questions that we run into as a society. And this is conversations that are worth having. So yeah.
00:12:07
Speaker
Here's a question that I also have. Our brains, the way I like to think about it is starting from being born, our brains are only ever used to the experience of being inside a human body. So if something like a brain organoid is grown in a lab or is put on a chip,
00:12:26
Speaker
and it has stimulation from the environment, it doesn't know anything except for what it's been exposed to. So I think of that, you know, this opens more questions than it answers, but I don't know. It doesn't necessarily know anything else, at least I assume so, unless there's, you know, embedded memories, you know, within the cells from what it's host experienced and what its host would experience.
00:12:53
Speaker
That is a very big bag of chips. So we can start by saying like, if someone is born without a hand, they really, really have to work on how to use an artificial hand. Our brain only has an experience of what our body looks like. It has a neocortex. It has a map of what we've experienced. That's why people who are born without hands can end up using their feet very efficiently.
00:13:17
Speaker
So I imagine they pretty effectively just think they are whatever they're in. But when we're talking about like our
00:13:26
Speaker
There have been studies that have found that if you expose mice to a negative stimulus, like the smell of lavender and associate that with being electrocuted, then their offspring will continue to be afraid of lavender for multiple generations. And there actually have been repeat studies on that because it sounds so fantastical and they found similar results. So there are genetic memories, but we don't understand them. And it's a little bit of a disturbing context.
00:13:54
Speaker
concept to think about whether or not a brain organoid has genetic memories of its past life. That's a thought.
00:14:01
Speaker
Yeah, now I know that this brings up another one of your videos. There's a whole series of videos. The first one we talked about Pong and I think the long and short of it is that it learned more efficiently than the AI. And actually, real quick, I don't know how much we talked about this, but I've been very curious about why human neural nets are more efficient. I'm sorry, learn certain things as good or better than deep learning neural nets. And the short answer is that it is unknown.
00:14:31
Speaker
But I do know that, for example, the human brain is shallow but wide, and that makes me wonder if in any capacity there's some parallel processing happening that somehow can come back together. I don't know. Maybe someone can educate me on that.
00:14:49
Speaker
I was trying to explain this to my 12-year-old daughter. And the only analogy I thought of is like a towel drying. Like if you've got two wet towels, if there's one that's all squashed together, but one kind of separate, one of them dries faster. And I just think, yeah, from a thermodynamic level, if something is happening distributedly, it can be more efficient than if it's deep. And that analogy may not even be great. I struggle to think of effective and useful scientific analogies. So I don't know. Do you have any thoughts on that?
00:15:19
Speaker
Our brains have evolved for millions of years to be efficient at learning. When we're talking about artificial models of intelligence, it's all nascent. We've just started figuring out how to put these things together. And it's also a different learning model. So when you think of being told a cell phone number just once you hear it, you're probably not going to be able to remember it, but a computer could.
00:15:44
Speaker
Conversely human brains are better at learning other things like how to play a game. Yeah Oh and actually I did run across a study not even a year ago. That's come up so many times specifically about sequential memories in chimpanzees You may have heard this study where I think with chimpanzees they had a computer screen where it was a game kind of like concentration where you had to See an order of things and then recall that order and chimpanzees were able to recall a sequence of over 30 with no effort out
00:16:14
Speaker
all. And with human beings, most of us really got much less efficient after about eight or so. So yeah, in terms of just sequential memory for sight, they have got us licked. The paper did go a little bit into the actual architecture, but that's not my expertise. And I was not really able to make heads or tails of what we have in our heads.
00:16:38
Speaker
uh that's different than chimpanzees and you know uh why are they better at that and why are we able to do things that as far as we know they cannot do um so yeah that's that's a question i'm very curious about taking me back to freshman college psychology right now um
00:16:54
Speaker
I would think we're just better at different things. It's a trade-off. Our brains are very hungry organs. They use around 30% of the calories we burn in a day. We probably just figured out how to focus our memory on what's important for our environment. Okay. Granted, I have ADHD and I can forget my phone number occasionally.
00:17:13
Speaker
Yeah, yeah. Same here. Same here, actually. That's a fun conversation. But I do enjoy the hyper-focusing, though. That's really fun, actually. I've been curious about the reward functions in a brain of somebody with ADHD who is hyper-focusing, because I can be gone for hours on a task and not get other practical things done that are very important. So I'm thankful that I have a wife who's very good at that and co-workers that are very good at that as well.
00:17:39
Speaker
In fact, I'll mention that we've actually have a good relationship with my coworkers and I because I can deep dive into a paper or a series of papers and obsess about them for days where they will be much better at all of the tasks that we have to do. And together, when the boss comes around, we show them all of what we've done. I'll talk about the theory and the extensions and they will show what they've made in the last few days. And he's very happy with all of us. So in my case, it works.
00:18:09
Speaker
Yeah. Yeah. Um, I'm the theory person in my lab. I'm the person who comes up with ideas of why things work the way they do.
00:18:17
Speaker
But I'm really bad at putting 500 things in, 600 other things. No, I got you. What's cool, though, is as I've learned about collaborations between scientists, that's not an uncommon thing. I mean, I just read the entire biography on Robert Oppenheimer. And there was so much discussion on who does what well. And Robert Oppenheimer was like a supremely good headhunter for talent. And he was a big theory guy.
00:18:44
Speaker
But Lawrence was an experimentalist, and he was very good at proving things through experiment. Through the two of them, they worked on the Manhattan Project. Of course, there were many other minds.
00:19:01
Speaker
as well. But a big topic was the way they described themselves was brick makers and brick layers. And yeah, it's fascinating when people with different but complementary thinking skill sets can work together. I think that's that's what's needed. And I'm definitely with you. I think I relate more to Oppenheimer with a theory, you know.
00:19:21
Speaker
Something that just comes up a lot in all of the genetic editing and brain organoid videos that I make is whether or not we could start getting rid of conditions like autism or ADHD. Being a scientist, I am aware that both of those are overrepresented in science.
00:19:40
Speaker
making big changes we may not know what we're losing and we may lose some of the most effective teams no one can do at all yes you're entirely right and that's also why you don't want to just decide that a certain characteristic is a bad thing because
00:19:56
Speaker
bad thing in what environment? Yeah, like there are all these ethical questions that we don't know. I've also read about that, like, and there's other conditions as well, you know, just like mutations that are not compatible with life and things like, oh, gosh, what did I read? All kinds of things that can happen in utero with a brain where we're doing studies with brain organoids and seeing if we can't better understand these conditions. So lots of research happening.
00:20:26
Speaker
I know it's a little bit of a tangent, but on that note, if there are mutations that are not compatible with life, sometimes it's because there's something wrong with the adult chromosomes. So even if they were to correct that one mutation, they may be more likely to have that mutation again down the line in subsequent generations.
00:20:46
Speaker
There's a lot of translocations that result in more translocations of the chromosomes happening more often. Yeah. Yeah. That's why I'm glad we've got experts doing this. What I should have done is before I brought that up, I should have had the examples of the research that is being done because I just kind of made that one up as a for example. But yeah, you're right. That's a very important one. A great example is that a Chinese guy who
00:21:13
Speaker
changed three babies without the mother's consent to be resistant to HIV. Yeah, just got out of prison. Oh, yes. Yeah, I saw that. And he's back researching again now. Isn't that right?
00:21:26
Speaker
Yep. Shockingly. Yeah. I'm completely against experimenting on people without their consent. Real quick, I wanted to move on to a few other of your videos, like ones with the rat bodies. I think you've got multiple videos that involve human brain organoids and rats. Can we talk about those a little bit?
Integration of Brain Organoids in Rat Vision Restoration
00:21:48
Speaker
A couple of different very spooky experiments with putting human brain organoids into rat bodies. One of them, which is very, very exciting for medicine, they cut out a portion of the rat's brain responsible for relay of vision to the rest of the brain. And then they input a human brain organoid into that site, and it integrated and restored sight in the rat.
00:22:10
Speaker
I don't know why they use the human brain organoid as opposed to a rat brain organoid, but it does show that we could repair portions of the brain that are damaged. It could be really, really important for people who have strokes or people who have congenital abnormalities that result in blindness. The other one, which is a lot more spooky, it's the air interface paper, they put a decapitated
00:22:34
Speaker
fetal mouse next to the brain organoid while it was growing. And it was able to autonomously make dendritic connections into the spine of the mouse and then actually took it over. They did a few little electrical simulations to show the brain that it could contract the muscles and it did it. It was able to move those muscles independently after it learned that it had control of them.
00:22:58
Speaker
That one is important for things like spinal damage, but it does sound a little bit spooky when you hear about it. Yes, it absolutely does. And you said a few things that I want to just say again real quick. So the first thing that was interesting was the paper says that the brain organoid was placed near the decapitated body. I didn't get a time frame for how long this process took, but I did hear that it grew dendrites into it. So that makes me very, very curious
00:23:27
Speaker
when a brain organoid creates dendrites? What does it have to sense in order to create dendrites? That's a question I have, of course, the timeframe as well. And then the other thing that you said that, of course, is very, very relevant is once the dendrites were established, a human operator had to induce a shock
00:23:48
Speaker
that went through both the organoid dendrite and the mouse body in order to teach it that this is how you move a muscle, I suppose. But then after that, it did so independently. So yeah, that's still scary, but I just wanted to clarify what was required in order for that to happen. Do you happen to know how long it took to grow the dendrites?
00:24:13
Speaker
Yeah, so apparently they had actually kept these things alive for months. They used an air interface to supply it with oxygen to kind of bypass the issue of the brain organoids becoming hypoxic. But there were neural co-factors that were able to just be sensed by the brain organoid. I mean, they're in the kind of fetal state where they want to make more connections. Interestingly,
00:24:43
Speaker
they were able to allow the brain organoid to heal. When it became damaged or there were neurons that had died, they're able to actually stimulate it to create more. But yeah, that's interesting that it kind of just found the spine.
00:24:59
Speaker
Yeah, no, that's absolutely fascinating. That's very fascinating. So yeah, so we may have to put a pin in this and just think there's a lot more to learn about this process, what they did, how they kept both the rat body as well as the organoid alive. I know you just read a little bit about that, but there's just so many questions about this process. Yeah, one of the questions, if I may, that I would ask, if we're taking a brain organoid that's not made from our stem cells, is made from somebody else's,
00:25:27
Speaker
I wonder if any of that ingrained genetic memory might transfer over. I would hope that if we are repairing brains, we're going to be using fetal cord blood from our birth, so there aren't issues with rejection.
00:25:40
Speaker
Yeah. Wow. That's actually a fascinating question. There's so many questions here, and I'm sure we'll learn a lot more about these. I'll mention real quick that many of these companies actually have quite the social media presence. There's one of them on Reddit, and I think it was Matri Labs, I believe, on Reddit. Oh, Critical Labs is on Reddit.
00:25:59
Speaker
Oh, that's right. They also have a very active LinkedIn. I spent a long time on their Reddit specifically in preparation of this episode. And there's just so many topics that I was going through that I was just absorbing it. I've still barely synthesized it. But for our listeners, I encourage you do check them out. Go through the Reddit, the LinkedIn. I'm sure they're available on any kind of socials.
00:26:38
Speaker
His lab is the one that has done a couple of things. They've used brain organoids to research a variety of mental illnesses, but also they're the ones that programmed the brain organoid to pilot a spider robot. And that's the one that we began this podcast with. Can you tell us a little bit more about that and where you discovered it?
00:26:45
Speaker
I saw a whole lot of videos by Motri Labs and I want to say it correctly. I was saying it wrong.
00:26:57
Speaker
Yeah, so I spend time looking on Google Scholar for new brain organoid research because I'm fascinated by it and I like to make videos on it. So I found that from the paper first. And I actually went and contacted Alison because I wanted to learn more. So what they did is they took a brain organoid and put it on a chip with some of the technology we already have to link up brain organoids to computers.
00:27:24
Speaker
And they gave it the ability to control the robot. They had to use some feedback at first. They had to
Brain Organoids Controlling Robotics
00:27:30
Speaker
give it electrical impulses to tell it that it's capable of moving it, and some directionality, some physical information to the brain to show it that it's moving. And it actually took off. It did it better, and it did it faster the more it did.
00:27:45
Speaker
So I did ask him about whether or not the brain organoid essentially thinks it is a spider robot because that's its body. That's what's giving it feedback. But he told me he can't say at this time, but you can expect the paper to come out this year if you want to read it. I sent you that question. I'm pretty proud of that one. I sent you that one before this interview.
00:28:06
Speaker
And this is the end of our first part of our interview with the content creator GT, where we're talking about brain organelles and artificial intelligence and ethics and what is intelligence. The entire transcript is available on our website at breakingmath.io, or if you just email us at breakingmathpodcast.com, we can send you a copy of the transcript. Join us next week for part two of our interview where we get more into artificial intelligence. And as always, we have a special section for those of you who want to become
00:28:36
Speaker
Science content creators we talk about our journey and what we went through in science content creation last thing I'll say is for those who would like a commercial free Experience on the podcast on our patreon for our five dollar a month donation We offer a commercial free version of the mp3 file for your enjoyment. See you next time. I've been Gabriel
00:29:03
Speaker
Fargo, the new virtual assistant from Wells Fargo, makes banking faster and easier. Like this. Fargo, what's my checking account routing number? And this. Fargo, turn off my debit card. And this. Fargo, what did I spend on groceries last month? And that's just the beginning. Do you Fargo? You can. In the Wells Fargo mobile app. Learn more at wellsfargo.com slash Get Fargo. Terms and conditions apply. Your mobile carrier's availability and message and data rates may apply. Wells Fargo Bank and a member of DIC.
00:29:33
Speaker
Fargo, the new virtual assistant from Wells Fargo, makes banking faster and easier. Like this. Fargo, what's my checking account routing number? And this. Fargo, turn off my debit card. And this. Fargo, what did I spend on groceries last month? And that's just the beginning. Do you Fargo? You can. In the Wells Fargo mobile app. Learn more at wellsfargo.com slash Get Fargo. Terms and conditions apply. Your mobile carrier's availability and message and data rates may apply. Wells Fargo Bank and a member of DIC.