Why Silicon Valley Is Hiring Bird Experts
Decrypted is brought to you by Hewlett Packard enterprise. Introducing HP green lake a new way to consume IT as a service, visit H P, E dot com slash green lake. To learn more. A few years ago around the time of Twitter's IPO. I noticed what I thought was an odd coincidence. It's chief engineer had studied birds specifically he'd studied, the auditory cortex of zebra finches, I thought that was pretty funny for an engineer, especially because Twitter's mascot, is that little blue bird. I forgot about it until a couple of years ago when I noticed, another bird brain scholar in the top echelons of tech. This person had been hired by Elon Musk the entrepreneur behind, tesla and SpaceX to join his new company, neural link nearly Inc is a very secretive futuristic company, which is trying to supercharge, the human brain. Let's all sounds pretty obscure, you'd think studying bird brains wouldn't be to relevant to studying human brains or figuring out social media rate. That's why remembered it, it just felt so random. But now here we're too. So one day when I was a little bored. I. Tried typing zebra, Finch and the names of several big tech companies into Google. And what did you find? I found quite a few employees who knew a lot about zebra finches. That was that companies like Intel and apple and Google, too. I was surprised. Okay. Hang on. Sara, I'm googling this does. Leaper Finch is the most common stranded, Finch of central Australia and ranges over most of the continent. Avoiding only the cool, moist south and tropical far north zebra finches are loud and boisterous singers. So what's the connection here? Well, that's where I jumped into the reporting process goes to figure out why companies are hiring bird brain experts in this little quest has taken us to college campuses around the country. We've is it. Several university labs some smelled better than others. We heard a lot of bird saw, so this guy, and so they're, they're very late. Super active. How big would you say that thing is like, so they're about fifteen grams total weights? They're mostly feathers. That's Tim Achi a neuroscientist and assistant professor at Boston University showing us zebra Finch in has bird lab and telling us about his research on bird, brains. He's running some pretty extraordinary experiments own over the the the implant. So this is actually just a lens. Okay. And so if you look at the top see that little black thing, that's coming up. So that's a lens if you look down at you could see the brain, but there's probably not enough like getting down there that you could actually see it. It turns out what Tim's learning about the brains of these tiny birds of great interest to the world's largest tech companies. I'm Brad stone and Sarah McBride that I'm Ashley Vance. And you're listening to decrypt it. So guys, tell me a little bit more about Tim Timoci and his, I'm sure wonderfully smelling lab. Out in Boston. Yeah. We'll Tim's lab was one of the key stops on our trip. Tim took over a Boston University lab from Tim Gardner, the guy who left to work for Alon musket and early Inc. Tim Achi is the forty year old guy with black frame glasses who throws in plenty of references to Portland area and other TV shows when he chats on his office wall. He's got print of voltage, chases called five seconds of Donkey Kong. Now we're talking about something I'm familiar with. So what what's the connection here? It's hers out that, that old eighties video game became the subject of a famous. Neuro science research, paper, there's one other thing you should know about Tim his a giant zebra Finch tattoo on his right forearm. Brad, how familiar are you is Seabra? Well, Ashley other than my very brief Wikipedia search just now other than small birds that sing a lot, I would say, not very familiar. They are very cute. Little birds. Indeed. They're about four inches long, the males have orange cheeks, and black and white stripe. Feathers across their chests. Hence the name the super easy to breed. And they chirp allot his Tim's interpretation. So how did Tim get into studying bird brains? He took a circuitous route into neuroscience. One thing I found interesting. Is he started as an engineer working at a company that helped factories to automate, his job was teaching robots how to sort stuff, everything from car parts to gizmos for circuit boards? It was just a stoning to me, how difficult it was to get things to do this. And these were tasks that, you know, children. Do it really putting my mind the idea that, you know, a lot of the, the things that, that children can do almost effortlessly, and without almost any training are incredibly impossible to get artificial systems to do or take an enormous amount of thought. Okay. So I think I understand Tim was curious about why a certain task is so easy for child, but so difficult for a robot. Yeah, that's exactly it. So after a detour to Tim ended up studying neuroscience at Harvard and that's where he discovered zebra finches. Now, he's teaching at BU and doing his own research. One thing Tim focuses on is how zebra finches learned to sing ten described his study of zebra Finch is more of a means to an end. I don't really think of myself or or care, too much about, you know. Songbird neuroscience. Specifically, I take it as a as a way to investigate general principles and mechanisms in neuro science, and how brains function, generally. So Tim is saying that he studies these teeny tiny zebra Finch brains because he thinks it will give us insights into the way human brains function. That's right. Researchers study all different kinds of animals for all different kinds of purposes. But in this case, they're trying to learn more about the human brain seeing how birds learn to sing, for example can provide insights into how we learn things. So I think that, you know, the songbird is one of those systems that we probably understand the best in terms of the different brain regions that are involved in terms of the roles of those different brain regions. And so I think that we can ask very, very precise questions in the songbird about the interaction between brain activity and behavior, a Sarah, you started out by saying that studying the brains of zebra, finches or somehow interesting for. Tech companies. So if Tim studies the birds, because their connections between bird brains and human brains are there also connections between bird brains and computers. Absolutely. But to understand it any to explain a little bit more about Tim's research when we met him in Boston before we went into the room with the real life zebra, finches, he played us when at his best songbird clips. And so that's what the song sounds like. Okay. So that's one. Yeah. So that's what the, that's what the this particular zebra, Finch, sings, all of them have slightly different songs. Thank a cartoon. Yeah, it sounded a little like, Woody woodpecker, right? But that, but that is actually that is actually what it sounds like. Yeah. Yeah. We listen to these all day long. Tim told us while male and female zebra, finches, conserve only males finches sing, even if they're song is so short, it doesn't sound like much of a song to us. Tim studies. The brains of the baby birds is they learn here's the baby's zebra Finch trying to imitate his dad's song. He makes an early effort and then a better one a month later shoes. The father. This month when this one. Well, okay. Then we can do that with the middle on where he's not quite got it, right. Yeah. So Ashley, I'm kind of praying that in these experiments. He's not hurting these beautiful teeny tiny birds. Every basically what it does is they take these birds that have been injected with a benign virus. The virus makes their brains produce. A type of protein, the causes individual neurons to light up when they fire they glow green and red to see them in action Tim in the grad students at his lab, performed very delicate surgery on the zebra finches to implant tiny tiny microscopes in their brain. Tim showed us one bird, who had gone through this procedure male zebra Finch. And so I believe on his head. He's got one of these windows. I was telling you about. And so this guy has already been through the procedure in which we inject the viruses into the brain. And then we've put a a a, a window or a lens on top so that we can actually image through it. He will eventually get a microscope attached. Sarah. That sounds completely crazy. Give us a better sense, for what this looks like. So there are all these little birds happing around with a tiny bit of their skull missing. And instead, they have a maker scope there and a little hole where you can look in and see what's going on in their brain. The microscope sits there for days, weeks months at a time looking at their neurons. Fire in, in real time and the birds are in their little cages. And then when you go into their lab, it's actually it's pretty cool. There's all these wires going off these cages straight until, like a data center where they store all of this information. And so, basically, you just get to get to watch this bird's brain behave in real time, and then go back and look through all information. Sarah, should we feel sorry for these birds? I didn't they seemed really happy. They were hopping around the chirping. They were acting normal, as far as they could tell based on the ones I saw without microscopes in their head. I really didn't seem to be too much difference. Okay. So how is this useful for the neuro-scientists? They look deep inside each birds brain, and they look to see which neurons are firing. And for how long when say the bird is learning to sing, and they can make hypotheses on the relationship between different neurons. So that's actually a technique used pretty widely in science now on all kinds of animals. While we revisiting Tim in Boston. We also stopped by mouse lab that does something similar and official AB these are the labs the neuroscientists are studying other things as well. Like how animals move what happens in their brains? When they make decisions for example at the Rowland institute in Cambridge. We saw a couple of video games that mice play using tiny joysticks size for mouse PA a wow. So how do they how do the mice get the quarters in the video game machine? The highly trained is e and the video games tells us what it tells us how they're making decisions. Exactly. So the, the mice are a lot like the birds they have bits of their skull of been removed and were watching their brains in real time again. And so you watch them play these video games. And you see how they adapt. Sometimes they get different rules. And, and you see how the the mouse learns the rules of the game, and they move this choice, stick around in one of them to find the edges of a box. And if the mouse's successful, it gets a little bit of sugar water and all this time. The scientists are sitting there seeing which parts of the brain, light up and how the mouse reacts to these different situations. So what is it about this research that tech companies find so interesting when reason is that the scientists are working with tons, and tons of data and that's something that every tech company needs to do? And also, it's to do with artificial intelligence the field that has computer systems mastering. Tasks that require human traits, like visual perception or decision making, maybe how do I dent cat? That sounds simple. But it's actually way more nuanced than a traditional computer task and very hard for a computer to master. So there's this one school of thought that moving forward should loosely be modeled after the human brain they I systems we have today, but still basically number crunching systems. They're doing tons of statistical calculations. And if we want to get to this, this future that Sarah's talking about where you actually have decision, making, and much more sophisticated thought the, the ideas that we could borrow from the human brain, and maybe have something that's way more flexible than what a computer could do. And so presumably since we can't cut open human, skulls, and make people play video games against their will the zebra Finch brands first step. Yeah. Exactly. Human brains are a little too big and complicated to study in the. Kind of detail we can get from animals right now. They're smaller easier to study, and obviously the ethics studying living human brain. Very tricky. So I'm so curious about why zebra finches become the bird to study. And what exactly tech companies are doing with all this? Let's get to that after the break. Business leaders are increasingly demanding the outcomes. They want when they want them and paying only for what they use at Hewlett Packard enterprise. We're bringing that flexibility and control to IT with the introduction of HP green lake pay per use outcomes on your terms, come to HP green lake and see the future of IT get there at HP dot com slash green lake. Okay. Sarah nationally, so, before the break you explained, how neuroscience is starting to inform the way tech companies designed AI systems. So I guess this means it's a very good time to be a neuroscientist. Yes. So these these people that used to be academia, or working at universities for their whole career are now finding tons of job opportunities in Silicon Valley companies like Google and apple and Amazon are all snatching them up, including a lot of these zebra Finch experts that we've been talking about. It's how the company that probably makes the chips in your laptop won't say exactly how many people it has working in a I, but it's a lot and many of them have this expertise Intel helps its customers soup up machines to get them to behave smarter. And in more human lake waves. That could mean may be working with a self driving car maker, and that carmaker needs its vehicles to make lightning fast decisions on the road. And they have to be good decisions promising way to do that is to build computer systems that imitate how the human. Brain works in the human brain. The systems are called sign ups is in pathways in a machine. They're called neural networks. We talked to Amir customers Shahi the chief technology officer for Intel's AI products division. Since around two thousand eleven. In a while. I don't know why. Vision. Speech, navigation reinforcement learning things that are related to narrow science because is also Catholic humans do pretty well. Amir's, the computational neuroscientist by training. He got his gauge at UC Berkeley. And he ended up hiring another Berkeley PHD, grad Tyler Lee to work at Intel and helping virtual assistance understand, human speech, Tyler spends a lot of time thinking about how speech works in different environments, like cars, knowing, context, simple, things like if the speakers the driver or the passenger makes it a lot easier to understand what they're saying, for example, drivers more likely to ask about directions, some of the work on context, does actually reminds him a lot of his PHD work studying. You'd guessed it zebra finches Newburgh as recognize what type of call is being being emitted by by the wanna tearing, and then then it can go in recognize who that Vert is, is that is that a family member mate? You know, is it another ranger a different type of bird, should I be concerned? The vocal identification is one thing where it's, it's. Context Pacific and recognizing the context. Let you better understand. We'll go signature mostly Tyler says his studies helped him with big picture stuff neuroscience teaches you how to think about complex problems of signal processing, where I take some something from the world that comes in. It's an image. It's it's sound and it's it's noisy, and it's high very, very high dimensional, and I have to break it down into into features that I can then use to, to, like, do something with solve a task with, that's what the brain does all the time. And that's sort of the abstract level Tyler's boss. Amir says he's no zebra Finch chauvinist. He's got people on his team, who studied flies rats locus, even worms these very simple organisms. Exhibit really complex behaviors that are still a challenge for us to, to simulate and silicon New Zealander networks and machine learning. So even a simple warm inchworm is a really complicated. Robotic machine. That's really macula us. So we look for inspiration from simple to humans. So, I guess, Intel and all these other companies must be developing products, which incorporate a I developed with the help of the zebra Finch experts besides commands to self driving cars. Is there any area where knowing a lot about sound itself is helpful, and I guess, went went to all these years of studying zebra Finch, finally pay off? You could think about features on your phone, or computer that let you unlock the device with your voice or stuff, like noise reduction and phone calls. And on video calls I can't wait till I can sing a little zebra Finch song and my phone unlocks but anything beyond gadgets. Yeah, absolutely animal. Neuroscience connects to a lot of fields, especially health somewhere could be relevant for Parkinson's research because animals help. Researchers figure out how to stop tremors. They also help in areas like how to handle limbs outside of medicine, the work might grow, even more futuristic. This is where we. We get into dystopia and scenarios. I suspect what, what else exactly like neural link a mentioned that company earlier because Elon Musk had hired zebra Finch scholar. And it's one of the companies that we believe is working on very futuristic technology. You must keep dropping hints on Twitter that the company's about to announce a big breakthrough. Nobody knows exactly what, but it's going to have something to do with brain machine interfaces. So I'm trying hard not to think about some Star Trek episodes on this topic, which all ended quite badly. But help me understand Ashley, what a brain machine interfaces about of it at its most basic level as idea that you have a two way, interplay between humans and computers where you could actually funnel, information back and forth. We already have examples of stuff like this with implants that help people here or stop Parkinson's tremors. In this case, I think people are looking at much more futuristic applications where you might even have like a mesh that's attached to your brain. And you could full on. Download your brain to a machine or learn Japanese and five seconds. There's another company called kernel. That's in the same field and like neural Inc. It's also very mysterious. And in some ways, that's kind of the best part, when we don't know exactly what they're doing. We can imagine all kinds of crazy stuff going back to ULA Muskie. He's been talking about where nearly could go maybe allowing people to have this kind of super human cognition where you could you could think on par with a machine, or certainly much better than we do today. That means basically stuff, like you could download an entire for language directly into your brain. Or maybe instantly grabbing encyclopedia people like to macho, you've been thinking about exactly these scenarios for years and can really nerd out on the possibilities trick trivia, would be over. Jeopardy would not be thing anymore. Alex trebek's would be out of a job. He has some more serious thoughts on the topic to find the idea that we could, you know, pretend one day in maybe distant future. You know, really right information directly into the brain that we could actually have a high band with way to get really Sifi about it, kind of a matrix. Like, I think that would be amazing. We are nowhere near knowing anything about how to get there we can barely even scratched the surface of what that would be like. But I you know, in terms of fantasy, what would I like to do one day, I would love to be able to contribute, even a small way to figuring out how we can have this sort of bidirectional interface with the brain? Oh my God. So, yeah, this is wild evoke ation of the matrix does not make me feel more comfortable about this. What are the skeptics say about about at all? Well, there are plenty of skeptics out there. They're checked in with one scholar at the university of Chicago din. Margaglio sh the idea that we're going to reverse engineers britainop, reverse engineer forward engineer, the, the, the human brain. So we can download tons of material into it very rapidly. And I don't know what pick up a language overnight or something I it's, it's the way people make progresses to dream. And so I'm, I'm a scientist. I'm one hundred percent for that. But that's really sounds more fantastical than realistic. Okay. Did ignores ignores the remarkable ways we learn and ignores our evolutionary history? So, I would, I would it'll be interesting to see what progress they make for sure. So guys, the possibility of super human cognition. Sounds appealing, you know, if it were up to you to and, and somebody was offering to put a chip or an apparatus in the year brains like they're doing the to the poor little zebra Finch. Would you do it? Substance. We're already doing this stuff today. If you have an implant to help you here or things to stop Parkinson's tremors. Yeah. If I had one of those conditions I would absolutely get one of these implants, and supercharge myself when you start going into this, this next wave of stuff, it gets far more philosophical and complicated. Because you're talking about changing humans from what they are some sort of weird. Next step of evolution where we're kind of half man half machine. You know, in some ways, there's people, I talked to like the guys at Colonel who argue that this is the only way, humans will be able to keep up with machines. And we always hear about losing jobs start official intelligence and seeing what humans can do going away. And so, you know, if you're having half you can, you can keep up but maybe keep some of your humanness as well. Sarah, what about cyborgs, Sarah McBride will be ever ever see that? Hate me cell phone. So I doubt it. But what about this idea of keeping up with robots and, and, yeah, there are people who think that AI could help us solve problems like climate change? I think that's too optimistic. The funny borrowed will do the story for be was that, that in the AI cab, you tend to have I feel like people who think the technology is really far along you. If you're talking about specifically computer scientists in the Silicon Valley kind of crew, they're very impressed. What they've come up with over the last few years, where we went to talk to all these brain. Researchers the ones who are down in the box down at the the neurons. They seemed on the whole to me, much more skeptical about when we would see huge break Thursdays seem to think that a lot of the stuff was years and years away, and I felt like they had this sense of how complicated the brain really isn't that a locking its secrets is is gonna take a long. But given the fact that they are taking the first steps to what will ultimately be very transformative and challenging controversial technology Sarah, did you get the sense that they were wrestling with the ethical complications of their work, you actually turned out to have studied philosophy at some point in their careers, which I thought was pretty interesting. And yeah. They talk about the decisions that a carmaker might have to make it more important to preserve the life of a passenger, or pedestrian stuff like that. So, yeah, they're thinking about these big problems. That doesn't mean they know how to answer them though anymore than we would. But they say that means these systems will be very human one day. Do they understand that? They have now found themselves right at the center of this next wave in, in computing. They to degree that definitely excited. I mean, just in very crass terms, a lot of these people get much better job offers than they would have in the past, you know, there used to be far less neuro science graduates. It wasn't that a. Feeling of a field. We didn't know much about the brain all these promised breakthroughs. Weren't happening at all in, in now you can go to a university do this amazing work, or if you kinda get tired of that, or you want to poke around somewhere else that you can go work for one of these tech companies and get paid. I don't know like ten times what you make at one of these labs ought to bring this all the way back to the beginning, Sarah, is it too early when I talked to Alexa, or Siri or Google voice to thank the little zebra Finch. Can we can we see any of the zebra Finch in that research in today's AI, we see some of it already in things like voice recognition and making better audio quality? But I think some of the biggest stuff is yet to come, and we can always check in with Tim Achi. And that's it for this week's episode of decrypted. Thanks for listening lake to know what you think of the show, you can write to us at decrypted at Bloomberg dot net poor. I'm on Twitter at McBride s g marmot, Brad stone, and I'm at valley hack please help us spread the word about our show by leaving us a rating or review in your favorite podcast app. This episode was produced by he get Cari and Lindsey Crowell. Our story editor was Akihito. Thank you, also, to end, Vander may and Emily BSO, Francesca Levy is head of Bloomberg podcasts. We'll see you next week. Decrypted is brought to you by Hewlett Packard enterprise. Introducing HP green lake a new way to consume IT as a service, visit HP dot com slash green lake. To learn more.