The Alien in the Room

It’s faster than a speeding bullet. It’s smarter than a polymath genius. It’s everywhere but it’s invisible. It’s artificial intelligence. But what actually is it? Today we ask this simple question and explore why it’s so damn hard to answer. Special thanks to Stephanie Yin and the New York Institute of Go for teaching us the game. Mark, Daria and Levon Hoover Brauner for helping bring NETtalk to life. And a huge thank you to Grant Sanderson for his unending patience explaining the math of neural nets to us. To learn more about how these 'thinking machines' actually think, we highly recommend his wonderful youtube channel 3Blue1Brown (https://www.youtube.com/watch?v=aircAruvnKk). EPISODE CREDITS: Reported by - Simon AdlerProduced by - Simon AdlerOriginal music from - Simon AdlerSound design contributed by - Simon AdlerFact-checking by - Anna Pujol-Mazzini Sign up for our newsletter!! It includes short essays, recommendations, and details about other ways to interact with the show. Signup (https://radiolab.org/newsletter)! Radiolab is supported by listeners like you. Support Radiolab by becoming a member of The Lab (https://members.radiolab.org/) today. Follow our show on Instagram, Twitter and Facebook @radiolab, and share your thoughts with us by emailing [email protected]. Leadership support for Radiolab’s science programming is provided by the Gordon and Betty.

• 0:01Speaker 00

  Wait, you're listening.

• 0:03Speaker 01

  Okay.

• 0:04Speaker 00

  All right.

• 0:05Speaker 01

  Okay.

• 0:07Speaker 00

  All right.

• 0:08Speaker 01

  You're listening to Radiolab.

• 0:11Speaker 00

  Radiolab.

• 0:11Speaker 01

  From...

• 0:12Speaker 00

  WNYC. See? Yep.

• 0:18Speaker 02

  Okay, after all of that, it is time to finally discuss Latif's question.

• 0:24Speaker 03

  Yeah.

• 0:24Speaker 02

  The topic.

• 0:25Speaker 03

  Okay.

• 0:25Speaker 02

  The theme of the moment, perhaps.

• 0:28Speaker 03

  Climate change.

• 0:30Speaker 02

  No, that, nobody cares about climate change, man. Come on.

• 0:33Speaker 03

  Simon! Hey, I'm Latif Nasser. This is Radiolab, where despite what reporter-producer Simon Adler just said, we here at the show, including Simon, do care about climate change. But we're here today to talk about a different, huge, overwhelming thing that we're all in the middle of.

• 0:51Speaker 02

  I mean, I don't want to put words in your mouth, but what I have been feeling...

• 0:54Speaker 03

  Yeah.

• 0:54Speaker 02

  Is a general sense of frustration.

• 0:56Speaker 03

  Yeah, yeah. Something that everybody's talking about, but nobody seems to actually understand.

• 1:02Speaker 02

  You and I have even done interviews together with people on this stuff.

• 1:04Speaker 03

  That's right.

• 1:05Speaker 02

  Which is, of course, artificial intelligence. So much of the coverage about this stuff right now is like, this running debate, right, where you've got people on one side saying, "These AI, you know, they think they are intelligent, and eventually they'll outsmart and destroy us all."

• 1:21Speaker 03

  Right.

• 1:21Speaker 02

  And then on the other side, you've got people being like, "No, they, they, they aren't actually intelligent. They're, they're just mimicking us. And it's not as big a deal as everyone says."

• 1:30Speaker 03

  Right.

• 1:31Speaker 02

  And I, I don't actually know who to believe.

• 1:34Speaker 03

  Yeah.

• 1:34Speaker 02

  And I think it's because, like, I don't know what AI is. Like, I don't know how it does what it does under the hood.

• 1:45Speaker 04

  Yeah, because we don't know, right? This is one of the most extraordinary things about, you know, machine learning, AI, is that we don't really know what they are.

• 1:52Speaker 02

  But after reading countless articles, talking to tech people and scientists, I finally felt like I was getting at that question when I talked to this guy.

• 2:00Speaker 04

  Stephen Cave. I'm the director of the Leverhulme Center for the Future of Intelligence.

• 2:04Speaker 02

  He leads this sort of think tank at the University of Cambridge.

• 2:07Speaker 04

  And there's about 50 of us now. Trying to understand these systems using a really wide range of methods, including tests taken from animal psychology.

• 2:18Speaker 02

  Tests designed to measure how well a mouse can problem solve.

• 2:21Speaker 04

  And applying them to AI agents in order to understand, well, where are we in the kind of evolutionary cognitive tree of life of AI.

• 2:30Speaker 02

  And they've actually turned these tests into a sort of competition.

• 2:33Speaker 03

  Huh.

• 2:34Speaker 02

  That they call the Animal AI Olympics.

• 2:37Speaker 04

  Yes, indeed.

• 2:39Speaker 03

  Okay, well, that just sounds fun.

• 2:41Speaker 02

  Right, yeah, exactly.

• 2:43Speaker 03

  Yeah.

• 2:43Speaker 02

  So to do this, they've created a slightly lower resolution Toy Story looking digital world.

• 2:52Speaker 03

  Okay.

• 2:52Speaker 02

  Or maybe even more accurately, like if you know the game Minecraft.

• 2:55Speaker 03

  Oh, yeah, yeah, yeah. Sure, sure, sure.

• 2:57Speaker 02

  It it looks like that. It's this three dimensional space filled with all these different bright primary colored objects.

• 3:05Speaker 03

  Okay.

• 3:06Speaker 02

  And then they take these AI, which are running on basically the same kind of engine that powered ChatGPT. And they give these things a little avatar, like a hedgehog or a pig or a panda. And then they just sort of place them in in this 3D world and say, there is food in here, find it.

• 3:26Speaker 03

  So it has to like navigate the digital world to find, I mean, like, I assume it's not really food.

• 3:32Speaker 02

  It's this green orb that they're looking for.

• 3:34Speaker 03

  Okay.

• 3:35Speaker 02

  And I mean, there are walls that they have to like figure out how to get around. There are transparent walls.

• 3:41Speaker 03

  But it's like, it's like, it's like physical world problem solving.

• 3:44Speaker 02

  Absolutely. And I mean, well, this is the sort of task that mice or pigeons can pull off pretty easily. For these AI agents.

• 3:51Speaker 04

  Things like manipulating objects and understanding gravity. It's a real challenge.

• 3:58Speaker 02

  Like they struggle to press a lever or perceive an edge.

• 4:02Speaker 04

  Which any animal can do, or at least, you know, any mammal, say. And so effectively, these systems don't have the common sense of a mouse. Whereas higher reasoning, maths and so on, they can do a hell of a lot better than humans can.

• 4:17Speaker 03

  That's the, that's the more of X paradox, right? Right? Like it's like easy things are hard and hard things are easy.

• 4:22Speaker 02

  Uh, exactly.

• 4:23Speaker 03

  Yeah.

• 4:23Speaker 02

  And like, we've known this for a long time and it's pretty obvious at this point. But after running all of these AIs through this thing dozens, hundreds of times, what Stephen has seen over and over is that.

• 4:36Speaker 04

  They have a completely different profile of capabilities and skills than any animal.

• 4:42Speaker 02

  They are not like us.

• 4:44Speaker 04

  Uh, no. I mean, one of its capabilities might be convincing us it's human-like, but it isn't.

• 4:49Speaker 02

  Well, okay, so then what is it like? I mean, uh, is the AI little tadpoles or what, what is it?

• 4:56Speaker 04

  Well, there is one metaphor that some people like to use, and that's the octopus. You know, what's wonderful about the octopus is they are phenomenally smart. They can use tools, for example, without being taught. They develop sophisticated tactics of all kinds, where lots of wonderful octopus escape stories.

• 5:15Speaker 02

  Well, wait, cause that doesn't sound like AI at all.

• 5:20Speaker 04

  No.

• 5:22Speaker 02

  Then why this metaphor?

• 5:24Speaker 04

  Well, it's helpful, not because AIs are like them, but because in a way it really shows how different intelligence can be.

• 5:32Speaker 02

  Okay.

• 5:33Speaker 04

  I mean, octopuses, their intelligence is distributed through their tentacles.

• 5:37Speaker 02

  He says, you know, we and all mammals have this one central brain. But octopuses, they have nine little brains, one in the center and then one in each limb.

• 5:48Speaker 04

  So their tentacles can function much more independently, uh, which is how they managed to have eight of them, all doing like clever things all at once. And, you know, this kind of intelligence is fundamentally alien to us. And that's a good way of looking at AI. Alien.

• 6:09Speaker 04

  Profoundly alien.

• 6:12Speaker 02

  Which, on the one hand, makes this thing feel sort of unknowable, impossible to understand. But then, on the other, while it is alien, it did not evolve in some far-off galaxy, or even the depths of the ocean. Right. Like, this is an alien we created, year by year, transistor by transistor. And so this is what we're doing today. We are going to trace the evolution of this alien in our midst, this alien that we designed, in the hopes, at least, of like, coming to some deeper understanding of what it actually is today. And then, maybe, if we're lucky,

• 6:55Speaker 02

  that will give us some insight into this thing we are all, almost certainly, going to have to face off with at some point or another.

• 7:06Speaker 02

  So.

• 7:07Speaker 03

  This is great. Like, I feel like we all need this. We all need this explainer.

• 7:12Speaker 02

  Great. Fill your glass, because here we go.

• 7:27Speaker 02

  Hey, you guys can hear me?

• 7:28Speaker 05

  Yes, I can hear you, Simon.

• 7:29Speaker 02

  Hello, Terry. How are you?

• 7:31Speaker 05

  Very good, thank you.

• 7:32Speaker 02

  Uh, sorry for the slight, delayed start here. Some classic, technical difficulties, you know? So, there are a lot of different first contacts. Yeah. We could point to with this alien species. But the most fun place to start that I've found is with this guy, Terry Sinofsky.

• 7:49Speaker 05

  Professor at the Salk Institute for Biological Studies.

• 7:52Speaker 02

  Who, uh, yeah, sort of like the midwife of AI. Is that the helpful way to think of you, or no?

• 7:58Speaker 05

  Yes, yes, actually. Uh, well, it's obviously, uh, more complicated than that.

• 8:04Speaker 02

  Right, right.

• 8:06Speaker 05

  But that's not a bad analogy.

• 8:08Speaker 02

  Terry trained as a neurobiologist. He came up poking probes in monkeys' heads to try to understand how the brain works. But then, in the mid-'80s, he teamed up with some computer scientists, trying to make computers do animal brain Like things, like hear and recognize sounds or visuals.

• 8:25Speaker 05

  But but it was going nowhere.

• 8:28Speaker 02

  OK.

• 8:29Speaker 05

  Because everything was based on rules at the time.

• 8:32Speaker 02

  Like all computer programming at this point, it was this incredibly complicated set of like if this, then that statement. So if you see this and you see that, but you don't see that, then that means this.

• 8:48Speaker 05

  Right.

• 8:48Speaker 02

  This sort of web of logic.

• 8:50Speaker 05

  Right.

• 8:51Speaker 02

  Which, when it comes to recognizing sounds or pictures, was a problem because.

• 8:55Speaker 05

  For each rule, there are tens of thousands, 100,000 exceptions.

• 8:59Speaker 02

  Just too many nuances in the rules to hard code in.

• 9:02Speaker 05

  And so it was clear that this approach, this way of doing it through rules, was really hopeless. And so, together with my friend and collaborator, Geoffrey Hinton.

• 9:13Speaker 02

  He started to wonder if there was a different way to tackle this.

• 9:16Speaker 05

  Learning.

• 9:19Speaker 05

  And so with a small group, with computers that were puny by today's standards.

• 9:24Speaker 02

  They set out to build a machine that could learn.

• 9:28Speaker 02

  And one of the first things they tried to teach it was how to pronounce English.

• 9:33Speaker 05

  You know, text to speech in computer science.

• 9:37Speaker 02

  And amazingly.

• 9:39Speaker 06

  Demonstration of network learning by Terry Sinofsky and Charles Rosenberg.

• 9:44Speaker 02

  They have recordings from these early training sessions.

• 9:48Speaker 05

  Now, if you want to learn from experience, you have to have lots of data.

• 9:53Speaker 02

  And so.

• 9:54Speaker 05

  So.

• 9:54Speaker 02

  You ready?

• 9:55Speaker 07

  Ooh, ah, bad. OK, sometimes, Ivan, look at me.

• 10:00Speaker 02

  They took a transcript of a kid talking, a transcript I had my friend and neighbor, Levon, reenact.

• 10:06Speaker 07

  When we walk from school, I go to my grandmother's house because she gives us candy. Nice, that's perfect.

• 10:14Speaker 05

  You ready for the next one?

• 10:16Speaker 02

  And then what Terry did was give the computer this text and then also gave it the exact phonemes, like the symbols for the proper pronunciation for those words. No rules, just actual pronunciations. And then said to the computer, "Quiz yourself." Like, "Go ahead and try, and then compare what you tried to the correct pronunciation."

• 10:36Speaker 06

  First recording, de novo learning.

• 10:38Speaker 02

  And here it is.

• 11:15Speaker 02

  Wow.

• 11:16Speaker 05

  Right, so it has no idea what it's doing, right?

• 11:18Speaker 03

  Yeah, not even close.

• 11:19Speaker 05

  No, no.

• 11:20Speaker 03

  Doesn't sound like a baby either. Like, that just sounds, like, glitched out.

• 11:24Speaker 05

  It's chaos, right? It's like, noise, effectively.

• 11:27Speaker 03

  Yeah.

• 11:28Speaker 02

  But then,

• 11:30Speaker 02

  as it continued quizzing itself, comparing its output

• 11:36Speaker 02

  to what it should have said.

• 11:37Speaker 07

  When I go to my cousins, I play badminton and all that.

• 11:41Speaker 02

  Slowly, we could actually hear the learning.

• 11:50Speaker 05

  You could hear it figuring out the difference between vowels and consonants.

• 11:58Speaker 05

  And then it would start pronouncing small words, you know, oh.

• 12:11Speaker 05

  You know, it only took a couple of days.

• 12:14Speaker 07

  When we are home from school, I like to go to my grandmother's house. Well, because she gives us candy.

• 12:22Speaker 02

  And it was acing it.

• 12:25Speaker 07

  Sometimes. And we there sometimes. Sometimes we sleep over night there. Sometimes we sleep over night there. Sometime, when I go to go to my cousins, I get to play. Family and all that.

• 12:40Speaker 02

  But the really astonishing thing is that when they gave the program new words and new sentences that it had never seen before.

• 12:49Speaker 07

  He won't stop jumping or running the bathtub.

• 12:53Speaker 02

  It pronounced those too.

• 12:54Speaker 07

  He keeps jumping around gets tired when he goes to bed when he finally gets to sleep.

• 13:00Speaker 05

  It was phenomenal.

• 13:02Speaker 07

  Sometimes I get to go to bed at 12:30. Sometimes, but most of our times I don't.

• 13:09Speaker 05

  What we didn't appreciate back then was that Net Talk was a little bit of 21st century AI in the 20th century. That this process of learning was the future.

• 13:27Speaker 07

  Are we done?

• 13:28Speaker 05

  We're done. Thank you so much.

• 13:33Speaker 03

  Well, OK, but like what actually happened there? What is it doing? How do you get a machine to learn that?

• 13:39Speaker 02

  Well, take a baby human. You know, it's born with this clump of gray stuff in its head, which is really a bunch of neurons that are all connected in like a random messy way.

• 13:49Speaker 03

  Oh, they are connected? I just imagined a baby brain was like, nothing was connected. It was a blank slate.

• 13:55Speaker 02

  No, when the baby emerges, the neurons are all connected. They're just not connected in ways that make sense in terms of the world they've just popped into. But then, when it gets some input, like it touches something hot, gets yelled at, gets cuddled, it starts to strengthen some of these connections and prune others back. Until you have this just unbelievably complicated network of connections that can recognize patterns in the world around it and, you know, know that this is a square or if you poke a cat, you get scratched.

• 14:28Speaker 05

  That's right. In the brain, you adapt to your world that you happen to be in by changing the strengths of connections. Between neurons.

• 14:36Speaker 02

  And so basically Terry and others wanted to create some version of that in a machine.

• 14:41Speaker 05

  Yeah, that you you you hit it. The models we were developing, these neural network models, were based on very simplified versions of brain circuits.

• 14:49Speaker 02

  OK, but how did you? How did you do that?

• 14:52Speaker 05

  Uh.

• 14:53Speaker 02

  Like, what is what is going on under the hood here that allows it to to do this?

• 14:59Speaker 05

  Well, we understand mathematically how they work, and we're making progress now with trying to translate the mathematics into something that humans understand.

• 15:10Speaker 02

  And so, Latif, here is my best attempt to translate this for us humans.

• 15:15Speaker 03

  OK.

• 15:16Speaker 02

  I mean, so just setting aside all of the technical setup on like, how is it even interpreting the data or what are you inputting? Uh, with the help of this guy, Grant Sanderson. Yeah, I I run a YouTube channel that's named Three Blue and Brown. I often talk about math, but math-adjacent things as well. Great. We're just going to draw like a mental image of what one of these networks looks like.

• 15:36Speaker 03

  OK, let's go.

• 15:37Speaker 02

  Now, as we all know, these neural nets can do crazily complex things, but for now, we are going to give one a very simple problem. I'm going to draw a couple of shapes. What shape is this? A circle. Can we get a computer to see a circle? How about this? A circle. A very childlike task.

• 15:57Speaker 03

  Yeah, sure.

• 15:58Speaker 02

  Now, first things first, to get an image into the computer, we're going to chop it up into a bunch of pixels, like a 10 by 10 grid of them.

• 16:06Speaker 03

  OK.

• 16:07Speaker 02

  And we're going to imagine those pixels as 100 light bulbs. One light bulb for every pixel and light bulbs that will be on if their corresponding pixel is filled in with ink and off if their pixel is empty.

• 16:24Speaker 03

  OK.

• 16:24Speaker 02

  So you've got this circle of illuminated bulbs in this grid of bulbs that are off.

• 16:29Speaker 03

  OK, I can see it.

• 16:30Speaker 02

  From there, for reasons that'll make sense in a minute, below that, we're going to add a smaller grid of 10 light bulbs, and then below that, just one bottom bulb.

• 16:44Speaker 03

  At the top, 100 light bulbs, and then another layer, 10 light bulbs, another layer, one bulb.

• 16:50Speaker 02

  Exactly, and that final bulb, that is just the answer. The output that when it turns on says yes. Circles? Circles, that's right. There's a circle here.

• 17:00Speaker 03

  Okay.

• 17:00Speaker 02

  But this last bulb, it's a little bit special. It's not like the other bulbs in that it's actually on a dimmer. So it can also answer like, maybe a circle, because it could be a square if it's kind of bright. Or I'm pretty sure if it's pretty bright. Or if it's all the way on, that means this is definitely a circle.

• 17:23Speaker 03

  As a side note, yeah, this feels like quite the challenge where we're torturing the poor audience members here, probably like on their drive and not able to allocate their like visual cortex to like try to visualize all this. But setting aside all of the technical terminology.

• 17:38Speaker 02

  There's one last thing to do. We have to wire all of these bulbs together so that electricity can like flow from that top grid through that middle grid down to that last bulb, which will hopefully turn it on. So we call up an electrician. We tell him, go and connect every bulb in the top 100 to every bulb in the middle 10.

• 18:03Speaker 02

  And then go and connect every bulb in the middle 10 to that final bulb.

• 18:08Speaker 03

  So literally every bulb is connected to every other bulb, basically.

• 18:12Speaker 02

  Exactly. So that electricity can flow down from any bulb that's lit up and kind of cascade through all of them.

• 18:21Speaker 03

  Got it.

• 18:22Speaker 02

  And so the electrician starts pulling the wires, soldering, and they say, I'm done. But the thing about this electrician is they're shit. Like, they just do a terrible job. Some of the wires that they put in are like a strong copper. Others are just twine, so they can't even carry electricity. And so when this is all said and done, this network we get is kind of like a fresh baby brain with just random neurons clumped together.

• 18:58Speaker 03

  Got it.

• 18:58Speaker 02

  And so when we do send an image of a circle into it, into the machine.

• 19:04Speaker 07

  Microphone again.

• 19:05Speaker 02

  To record your voice. Lighting up some of the bulbs in that top grid. What shape is this?

• 19:11Speaker 07

  Uh.

• 19:12Speaker 02

  The electricity passes down through these random connections from the top to the middle, down to the bottom, and in all likelihood.

• 19:19Speaker 07

  I don't, rectangle.

• 19:22Speaker 02

  It's completely wrong on this. That final bulb might be a little lit up, or half lit up, or just completely off.

• 19:29Speaker 03

  Okay.

• 19:30Speaker 02

  Now, when a child gets something wrong. No, what is that? And, like a parent scolds them, that is altering the connections between the neurons in the brain, strengthening some, pruning others back, right?

• 19:42Speaker 03

  Right.

• 19:43Speaker 02

  And that, that is what we want to do with this machine. We want to mess with those wires, the strengths of those connections between the bulbs.

• 19:52Speaker 03

  Right, right, right.

• 19:53Speaker 02

  Now, we could just go in there and rewire this thing by hand.

• 19:57Speaker 03

  Yeah.

• 19:57Speaker 02

  We could pick out the important bulbs, because we know which ones are lit up for a circle, and direct their current through the middle bulbs to that final bulb, but, you know, that would take just as long as hard coding it.

• 20:10Speaker 03

  Right.

• 20:10Speaker 02

  And so instead, we're gonna give this thing the chance to learn all this, to learn what the connections should be. So, when it gives us that first random wrong answer.

• 20:23Speaker 07

  There is a 12.2% likelihood of a circle in this image.

• 20:27Speaker 02

  We're gonna say, "Bad robot, there is absolutely a circle in this image. Try again."

• 20:32Speaker 07

  Okay, I will try again.

• 20:34Speaker 02

  But then, after that first try, instead of us standing there saying yes or no, we are going to set it up to learn all on its own. We're going to step away and let math be its babysitter, be its teacher. And so, this is the moment where we have to dive into the math a bit.

• 20:51Speaker 03

  Uh, okay.

• 20:53Speaker 02

  It's not that complicated. Mostly multiplication.

• 20:55Speaker 03

  All right. Okay, let's go.

• 20:57Speaker 02

  First of all, these bulbs in the computer, they're really just numbers.

• 21:01Speaker 07

  One, two, three, four, five.

• 21:03Speaker 02

  And the wires, you can really just think of them as variables that multiply these numbers.

• 21:08Speaker 07

  X times two.

• 21:10Speaker 02

  As they pass through them.

• 21:11Speaker 07

  Y times point three.

• 21:15Speaker 02

  A good wire multiplies the electricity by five or whatever. A bad one divides it in half or even zeroes it out. And that means we can just take this entire array of bulbs and wires and turn it into a giant equation. You know, A times B plus C times D plus E times F. There's some other math strewn in there very artfully and deliberately. But the key here is with a bit of mathematical trickery, this equation can represent the difference between the output it is giving.

• 21:48Speaker 07

  There is a 12.2% likelihood of a circle.

• 21:51Speaker 02

  And the output we want it to give.

• 21:53Speaker 07

  There is a 100% likelihood of a circle.

• 21:56Speaker 02

  And if we think, hey, I've got this function and I want to like find a minimum of that. Like minimize the difference between your output and the output we want. There's a whole field of math that is just built, ready to do exactly this kind of thing. This is what calculus is all about. Like Newton, if he was rising from the grave, would just be like showing fireworks right now, saying, hey, I got this. I know how to do this one.

• 22:21Speaker 03

  So somehow the calculus tells you in math equation form, if you're getting closer to the right answer?

• 22:28Speaker 02

  Yeah. And now, don't worry. We're not going to go into the calculus other than to say, we walk away and the calculus becomes the teacher.

• 22:35Speaker 03

  Okay.

• 22:36Speaker 02

  So.

• 22:36Speaker 07

  12% likelihood.

• 22:38Speaker 02

  After the first wrong answer, the equation says, no, machine tries again.

• 22:42Speaker 07

  25% likelihood.

• 22:44Speaker 02

  And the equation says. Closer, and the machine tries again.

• 22:48Speaker 07

  77% likelihood.

• 22:50Speaker 02

  And each time it tries, it messes with the wiring, the weight of the connections between the bulbs.

• 22:56Speaker 03

  Getting it closer and closer to right.

• 22:58Speaker 02

  Exactly. And what happens over time is that that middle grid of 10 bulbs, their connections back to that top grid are getting tweaked in such a way that it's like they're starting to pick up clues. Like, maybe it's getting stronger signals from bulbs that are part of a curve. Or maybe it figures out that the corner bulb can't be on for it to be a circle. And like, the thing is, we actually don't know. I mean, when people talk about these things being a black box, this is what they mean. It's this middle grid. It's all automated by math. It's picking up something.

• 23:39Speaker 03

  We don't know what the clues are. We just know that they're right, that the clues are...

• 23:45Speaker 02

  Yeah, it's finding some signal that tells it, there is a circle-ish thing here. And as it keeps giving answers and the equation keeps telling it whether it's right or not, or closer or further away, eventually, each of those middle bulbs is receiving the right electricity from the right top bulbs to know if these characteristics of a circle are there. And if they are, they pass that along to the final light bulb, which will light up if enough of those characteristics are present.

• 24:23Speaker 02

  And at that point, yeah, our little network here has learned to recognize this circle, which...

• 24:29Speaker 03

  That's actually kind of astonishing. That's pretty amazing.

• 24:32Speaker 02

  It is, but it's only this one circle. And so the important thing is that if you do this process, not with just this one circle, Circle.

• 24:42Speaker 07

  Circle.

• 24:44Speaker 02

  But with tens, hundreds, thousands of examples. You know, big circles, little circles, messy circles, circles drawn by you and me. And you have the machine tweak all those different wires for all those different examples. You can then take all of that and do one final, actually very simple bit of math.

• 25:07Speaker 02

  Just average it all together.

• 25:11Speaker 02

  All of the wire strength you got from all the examples for wire one get averaged down to one value. All the wire strength that you got for wire two get averaged down to one value. And if you've done this right, you can then send in any of the drawings it's seen before or new drawings it's never seen. Circles drawn by a two-year-old or a picture of an orange. And it will say, yes, there is a circle there.

• 25:38Speaker 03

  Holy cow.

• 25:40Speaker 02

  Now, that process we just went through can recognize way more sophisticated things than just a shape, like cats or dogs. And I mean, the only real difference in the model is instead of these three grids we just used, these three layers, you know, an input, a middle, and an output, you just add more layers of bulbs in the middle. These multiple middle layers allow the computer to recognize progressively more complicated components of the picture. So like, the first layer might just find the edges. The second might find textures. The third, forms. The fourth, maybe eyeballs.

• 26:19Speaker 03

  Because it's like, because everything is made up of building blocks of the layer before it.

• 26:24Speaker 02

  Yes.

• 26:24Speaker 03

  Without, crucially, without anyone labeling any of those intermediate... Like, it's figuring that out itself.

• 26:33Speaker 02

  Exactly. And then, using the same mathematical reinforcement, it can tune and tweak to get shit right. Okay, wow. Well, I need a drink after all of this to sort of let all this settle in. Great.

• 26:48Speaker 03

  Okay, like this, I'm... I wish my kids could learn like this. Like, the way they learn is so physical, so emotional. It matters who's saying it. It matters how they're saying it. It matters the tone. It matters all these different things. Like, this is so clean.

• 27:03Speaker 02

  And like, crazy fast. I mean, what just took us 10, 15 minutes to explain, that all happens in seconds. So it can learn the circle thing at basically lightning speed.

• 27:15Speaker 03

  But like, a circle, recognizing a circle is one thing.

• 27:18Speaker 02

  Sure.

• 27:18Speaker 03

  And like, now we're talking like, actually making...

• 27:23Speaker 03

  Like making a, you know, a sonnet as if Shakespeare wrote it. That seems like, that seems like a very wide gulf. It seems like there's still a lot of place to go.

• 27:33Speaker 02

  For sure. And our little alien is going to have to evolve here.

• 27:37Speaker 03

  Yeah.

• 27:38Speaker 02

  But in terms of its, its architecture, of how it does this, it's basically the exact same.

• 27:45Speaker 03

  Huh.

• 27:45Speaker 02

  The, the only real difference is we're shifting its, its focus from recognizing to a slightly different skill. And we're going to get to that. You want to predict what I'm going to say next?

• 27:59Speaker 03

  Right after a quick break.

• 28:01Speaker 02

  Exactly. Right after a quick break.

• 28:24Speaker 01

  After some actual fake polls came to light, pollsters are in the hot seat. But they say, don't be too quick to dismiss the entire industry. People want to raise taxes on the rich. I can tell you that definitively because we know it from polling. And if we lose trust in polls, we have no way to understand what Americans want to be done. On this week's On The Media from WNYC. Find On The Media wherever you get your podcasts.

• 28:58Speaker 03

  Latif.

• 28:58Speaker 02

  Simon.

• 28:59Speaker 03

  Radiolab.

• 29:00Speaker 02

  So you, you asked this question to me before the break. Like, how did this thing evolve from being able to recognize shapes to, to generate stuff?

• 29:10Speaker 03

  Yeah.

• 29:11Speaker 02

  And I posed that very question to Grant Sanderson.

• 29:15Speaker 03

  Okay. Yeah, okay. So I would say.

• 29:18Speaker 02

  There's, there's many different ideas at play here. Who, again, Youtuber, has thought a hell of a lot about this stuff. And he says, "The important next step is to realize that, yes, you could think of what we did with those circles as having the machine recognize them, or you could say we were asking the machine to predict the answer we wanted." Like with the circle example, there's, there's two things that it could predict. Circle or not.

• 29:43Speaker 03

  Okay, so but that, so that's, so it's not anything meaningfully different. It's just like, let's just call everything a prediction.

• 29:50Speaker 02

  Right, but it becomes important when we're talking about generative stuff.

• 29:53Speaker 03

  Okay.

• 29:54Speaker 02

  Like, in the case of language. Predict what word comes next. So, to explain, going all the way back to the 80s, IBM began playing around with these chatbots that you could type to and it would respond. Hello there, how are you today? And the way it would do what it was doing was it would take every word that you typed in as your question, turn those words into numbers. We're not going to go into how, because that would take an hour in and of itself. But turn those words into numbers, send it through this multi-layered set of bulbs. But in this case, those, those bulbs, those layers it's passing through, they haven't been trained to categorize a sentence.

• 30:37Speaker 02

  Like, we don't want it to say, "That was a question." Instead, it has been trained to spit out the word that is most likely to come next, to predict the most likely next word. Just one word, just one, not even a word. Also, there, there's a nuance here between the notion of words and tokens, but excessive nuance.

• 30:56Speaker 03

  Yeah, but it's like, what is it even basing? Like, how is it predicting that? With a circle, you know it's a circle. We know the right answer, we're giving it the right answer. It's calculating back to, to that right answer.

• 31:08Speaker 02

  Right.

• 31:08Speaker 03

  But like, uh, in a sentence that could go any million number of ways, how, how can it ever have a right answer to train back to?

• 31:15Speaker 02

  Well, so what IBM was doing was giving it a bunch of texts, books, transcripts, conversations, feeding that into this machine. And so then the, the right answer was the most likely word to follow the preceding word or words.

• 31:32Speaker 03

  So it's like, it's just like, here's a giant stack of human talking, and in this giant stack, what's the most likely thing that would have been said next?

• 31:45Speaker 03

  In this exact scenario.

• 31:46Speaker 02

  Exactly. That's right. And just one brief aside, because it's sort of fun. I think I have this right, that a word is a big, long list of like 13,000 numbers.

• 31:59Speaker 03

  What? A computer has to turn a word, one word, just like a one word into 13,000 numbers?

• 32:06Speaker 02

  Yeah. And so like, in the way that a pixel value, in the circle example, was like, basically a zero or a one, it's like every word is this list of 13,000 numbers.

• 32:18Speaker 03

  It's so weird that it like, that that's, that's the simpler version for it.

• 32:26Speaker 02

  I know.

• 32:26Speaker 03

  Let me turn it into this, like, phone book of numbers.

• 32:30Speaker 02

  Which is again, like, which, which points to how these things are so not us.

• 32:35Speaker 03

  Yeah, they're really not us.

• 32:37Speaker 02

  Not at all.

• 32:38Speaker 03

  Wow.

• 32:39Speaker 02

  But they're using us though, right? Like, it's our talk that's getting turned into numbers. And it literally does it one word at a time. So after it's written, the first word of its response, it just does the whole process over again. It, it takes all the words in your question, plus the first word it predicted, sends all that through the network again. And then it just predicts the next word after that. Sends that through those bulbs again. And then the next word after that. Does the whole thing again. And plays the same game over and over and over. And one of the words in its vocabulary is the like, end conversation token. So it like, it has some notion of when to stop, but the act of stopping is itself just one more prediction. It's, it's one more probability in that big list of things that should happen next.

• 33:21Speaker 02

  And as I said, this is how they were doing it all the way back in the eighties. And, I mean, if you interacted with a chatbot, even in like, the 20 teens, this is the way they were doing it as well.

• 33:32Speaker 03

  Really?

• 33:33Speaker 02

  Do you have any recollection of when you first came in contact with one?

• 33:36Speaker 03

  Ah, God. Um, I feel like it must have been one of those like customer service bots on a website kind of thing.

• 33:43Speaker 02

  And I'm sure, not just because it's a customer service experience, but because it was an early chatbot experience, it, it wasn't very good.

• 33:51Speaker 03

  No, no, no, terrible. No, terrible.

• 33:53Speaker 02

  And a big part of why they were bad was. They had difficulty dealing with.

• 33:57Speaker 00

  Longer stretches of text.

• 34:00Speaker 02

  This is Stephen Levy.

• 34:01Speaker 00

  Editor at large at Wired.

• 34:03Speaker 02

  He's been covering this stuff for

• 34:04Speaker 00

  Yeah, I mean.

• 34:05Speaker 02

  A long time.

• 34:06Speaker 00

  I published a book in 1992 called Artificial Life.

• 34:09Speaker 02

  I was two years old, by the way.

• 34:12Speaker 00

  Thanks for that.

• 34:13Speaker 02

  Sorry.

• 34:14Speaker 00

  Yeah, yeah, thanks.

• 34:15Speaker 02

  And he says, because it predicted words one at a time and one after the other, the longer the question or the longer the answer, the more likely it was to miss or lose the larger meaning. And so eventually predict a word that just doesn't make sense or is out of place.

• 34:32Speaker 00

  Exactly.

• 34:33Speaker 02

  Huh. And so just to give one very concrete example to illustrate it, like the sentence, what sound does my dog make when I slam the door? It's like

• 34:46Speaker 03

  That's so, I can see why that would be confusing.

• 34:49Speaker 02

  Right, like you have to somehow know that in that sentence, dog is really the operative term here.

• 34:55Speaker 03

  Right, right.

• 34:56Speaker 02

  The important noun, it's not I or door.

• 34:59Speaker 03

  Right, right, right.

• 35:00Speaker 02

  And so in 2017.

• 35:02Speaker 00

  This guy, you know, Oscar Reed, who worked at Google.

• 35:06Speaker 02

  Set out to solve this dog door problem.

• 35:08Speaker 00

  He thought that the thing should be able to figure out, oh, this is the most important part of the sentence. This is what I should pay attention to.

• 35:14Speaker 02

  And now the question becomes like, how the heck does one go about doing that? And what they figured out was the, the problem here is we're giving it one word at a time and we're having it predict one word at a time. And what we need to be able to do instead is have it somehow process the sentence as a whole so that, you know, something at the end of the sentence can sort of feed back on the weight or meaning it gives to something at the beginning of the sentence. And one way that you can just imagine it doing this is that instead of just making a prediction and giving an answer, you need to take in all the information, make a prediction,

• 35:56Speaker 02

  but then just like set that aside because you're going to take in all that information again. And then we're going to send it through again and again and again, each time focusing on a different word in the sentence, generating a different possible prediction before landing on some final prediction, which, God willing. Would be bark.

• 36:13Speaker 03

  It's like the computer simultaneously lives in the multiverse of that sentence, where each word in that sentence is the most important.

• 36:24Speaker 02

  Yeah, and like I, I've, I've, I've looked at this stuff for months and I still don't totally understand exactly how a machine does this, but.

• 36:32Speaker 00

  Well, I mean, something like that. And also, you know.

• 36:35Speaker 02

  You can say no, you can tell me.

• 36:37Speaker 00

  Well, I mean, I mean, I, I, in, in the, in the raw sense, yeah, that's, that's, that's the idea.

• 36:42Speaker 03

  Like the complexity here, you can see it's going through the roof here, like where you're like, oh God, this is so much more computing you need to do.

• 36:49Speaker 02

  Totally. And, and this was a big barrier for a long time. I mean, that's why these chatbots were almost as bad in the early 2000s as they were in the 1980s. And this is where we get to the next step in the evolution of our little alien friend here, which, as many evolutionary leaps are, was mostly a hardware upgrade. I mean, if you have been following the news about AI at all, you've probably heard this term GPU.

• 37:16Speaker 03

  GPUs, components that go into data centers.

• 37:19Speaker 02

  Or the company. Computer chipmaker Nvidia. Nvidia. The most valuable company in history. That makes these things. Its story, of course, wrapped up in the frenzy around the future of artificial intelligence. These things and this company have been at the center of the conflict between China and the US when it comes to export controls.

• 37:36Speaker 07

  The idea here is for the US to kind of limit the ability for China to catch up when it comes to AI.

• 37:42Speaker 02

  And interestingly, what these GPUs, these graphical processing units were originally designed for

• 37:49Speaker 02

  was computer games.

• 37:50Speaker 00

  Video games, things like that.

• 37:51Speaker 02

  And what they're really good at is just doing a bunch of different math problems all at once.

• 37:57Speaker 00

  Exactly. It's just, it's just all about multiplying and adding numbers as fast as you can. There's some other things, but like by and large, like. Just do those two things and we're off to the races.

• 38:06Speaker 02

  In doing these math problems all at once, which is called parallel processing, that's exactly what these learning machines needed to do some version of that super complicated multiverse prediction thing we discussed.

• 38:17Speaker 03

  Sure, sure, sure.

• 38:18Speaker 02

  And so with these GPUs and this new parallelized architecture that Google named a transformer, all of a sudden, they could get a machine to parse those longer sentences and give at least reasonable answers to more complicated questions.

• 38:34Speaker 03

  All right.

• 38:35Speaker 02

  But what really sent these AI chatbots into the stratosphere was a kind of knock on effect of this parallel processing.

• 38:41Speaker 00

  Because when you can process everything at the same time in parallel, you can actually train on a lot more material in the same amount of time.

• 38:49Speaker 02

  And so eventually, they just gave it basically the entire internet, almost everything we humans have ever said on the internet, as its training material and started sending that through this network of light bulbs and wires that was just unimaginably big. Like, to get a sense.

• 39:06Speaker 00

  In our smaller example with the circle.

• 39:08Speaker 02

  There's something like a thousand and some odd parameters.

• 39:12Speaker 03

  Right.

• 39:13Speaker 02

  A thousand or so of those wires.

• 39:15Speaker 00

  GPT-3, which was kind of dumb by today's standards, but it came out, had 175 billion parameters.

• 39:21Speaker 02

  175 billion things that could be tweaked.

• 39:26Speaker 00

  Yeah, and many of the ones that we have now, they're trillions of parameters.

• 39:31Speaker 02

  And as they fed basically all the things we humans have ever said on the internet into this thing.

• 39:37Speaker 00

  Throwing way more training examples and way more compute than anyone would reasonably think to do.

• 39:41Speaker 02

  Slowly, they started to notice.

• 39:43Speaker 00

  That with a sufficiently large amount of data on a sufficiently large model, run with sufficiently many cycles of training, these new computers do seemingly intelligent things.

• 39:58Speaker 02

  Now, a lot of what I just described was written up in a paper called Attention is All You Need. And these findings are really what unlocked these large language models like ChatGPT. And that's all it was really intended for, but.

• 40:15Speaker 00

  There was a passage in there saying, we think this can work for images and video. And indeed, that turns out to work.

• 40:29Speaker 02

  That same basic model of massive parallel processing with tons of input, that could predict the next part of an image or sound. The moment civilization was transformed. And that moment, that realization, is really what triggered the explosive proliferation of artificial intelligences, different kinds, practically different species of AIs that we are living amongst today. New artificial intelligence systems. Machines that can teach themselves superhuman skills.

• 41:03Speaker 03

  Chat GPT-3. GPT-4. They're doing things Apple Intelligence. DALL-E. An app called Lensa.

• 41:10Speaker 01

  Bard. Bard. It's called Midjourney. Text-to-video art-generated.

• 41:13Speaker 03

  It's crazy. Look at this.

• 41:14Speaker 01

  So I don't know what AI it is they're using. Yes, it feels like an episode of Black Mirror.

• 41:23Speaker 03

  So it's like, it's like all of these different apps doing all of these different things in all these different mediums. They're taking in a huge amount of examples. And then they're using fancy math to basically predict the next word, the next pixel, the next note. And from that, it's like generating this whole huge diversity of new stuff.

• 41:55Speaker 02

  Yeah, basically. And I mean, it, it's also, just as we described, doing something that I don't totally understand, that's more holistic than just looking at the thing that happens next. But it is drawing on the examples it's been given to decide what should happen next, which suddenly sounds not so simple.

• 42:13Speaker 03

  No, it, it does send you into a spiral because it's like, it's like, is, is what I do any different from that? Just spewing out, you know, some iteration of everything else I've seen before this?

• 42:27Speaker 02

  Yeah, but first of all, you're not, you're not pulling from the whole internet, right? Like, you have to depend on just the, the limited things you've experienced or can even maybe remember.

• 42:36Speaker 03

  That's fair.

• 42:37Speaker 02

  And you're like, math. It's also just way sloppier. It's not as accurate.

• 42:41Speaker 03

  Yeah.

• 42:42Speaker 02

  And to that point, and maybe we shouldn't even go here, but there's this one other thing that you can control in these models, which is called the temperature, which is like this final knob you get to tweak on the thing. And so if you have, I think it's if you have the temperature all the way down, it will give you the most likely thing to come next. If you turn the temperature up a little bit, though, it then is going to pick like the second or third most likely thing to come next.

• 43:09Speaker 03

  No, so you can control like how precise you want the math? Like you can, you can say, I want it a little stanky?

• 43:15Speaker 02

  Yeah, like there's a little bit of randomness in it then, that it's then acting upon in what it does next. So maybe you just want the temperature turned up on like every third word. So that there's this almost spontaneous feeling, serendipitous creation, act of creation that that comes out of this rigid math.

• 43:32Speaker 03

  Like it's like something startlingly creative might just be.

• 43:39Speaker 02

  A less right answer.

• 43:40Speaker 03

  A less right answer. Wow.

• 43:41Speaker 02

  Yep. And just by doing that, it's going to keep doing stuff that we are going to get increasingly uncomfortable with.

• 43:50Speaker 03

  Yeah.

• 43:51Speaker 02

  Like right, right now, there is an AI-generated song on the Billboard country charts.

• 43:56Speaker 03

  Really? I didn't hear about that.

• 43:58Speaker 02

  But like, if that's the case, I see no way that eventually a fully AI-generated film won't hit the box office. Like that, that's just going to happen. But when it happens, it will be only because of all of this math.

• 44:15Speaker 03

  To me, I think the thing that makes me, it makes me realize is when you see under the hood, what you see is less like something spooky and ethereal.

• 44:29Speaker 02

  Yeah.

• 44:29Speaker 03

  Like there are times when it gets spooky, when like, like there'll be a time, like I'll be listening to like an AI-generated podcast, and then one of the hosts breathes. And I'm like, wait, that's so weird. Like it doesn't even need oxygen. Why is it breathing? And now it's like, oh, because you know that like, that's just the next statistical thing that would come in that sentence is a breath. That to me, that to me is like, it's, it's much less eerie because you can see where it got it from.

• 44:55Speaker 02

  Right. But.

• 44:58Speaker 02

  Well, OK, I do have one bit more for you. Because, I don't know, I still found myself wondering how it will feel as these things get better and better. And in particular, what it'll feel like in the moments we sit across from it and it is better than us at something we have spent our lives working on. That it is better than us at something we truly love. Yeah, many, many people, also all my friends tell me like, wow, you are the first professional Go player. Be famous because you lost the game. No. So, yeah, it's me. And so I got in touch with this guy.

• 45:42Speaker 02

  I'm a professional Go player, three-time European champion. So, real quick, Go. It is an ancient Chinese game, considered probably the most complicated board game in the world, to teach a computer to play. Because of just how open-ended it is. All you really need to know is you are trying to control as much of the board as possible. You go back and forth with your opponent, placing one stone at a time. And you control portions of the board or territory by either like cordoning off sections of it or encircling your opponent's stones. It's very simple idea, but it's difficult. Because with such simple rules, there are just this crazy number of ways

• 46:27Speaker 02

  the game can play out. In fact, folks like to say that there are more possible ways for a Go game to go than there are atoms in the known universe. Yes. Anyhow, back to Fong. I remember I discovered Go, aged like six, in my school in Xi'an. And I feel something, oh.

• 46:50Speaker 07

  This game I can play and I progress very quick. One year after I learned Go for my school, I'm number one.

• 46:58Speaker 02

  Three years after that.

• 46:59Speaker 07

  I'm in the best team in the province.

• 47:02Speaker 02

  And not long after that.

• 47:03Speaker 07

  I stopped my school. I only learn Go game.

• 47:06Speaker 02

  I mean, for years.

• 47:08Speaker 07

  Every day, only only thing you do is just to play Go game. Twelve, twelve hours.

• 47:12Speaker 02

  Twelve hours of playing.

• 47:14Speaker 07

  Yeah, it's no joke.

• 47:16Speaker 02

  Around age fifteen, he went pro. And somewhere along the way, he says, he noticed this almost magical quality of the game.

• 47:26Speaker 07

  Go for me, it's like mirror.

• 47:29Speaker 02

  A mirror.

• 47:31Speaker 07

  Mirror, yeah, because when you play, you can see your mind on the board.

• 47:36Speaker 02

  He says, all the choices you make, whether you're aggressive in attack or are patient and waiting. You know, in a sense, how you think stares back up at you.

• 47:46Speaker 07

  It's like print, mind print.

• 47:49Speaker 02

  And your opponent's mind, he says, it's printed there, too.

• 47:52Speaker 07

  So I play with someone, I don't know him, I never talk with him. I play one game, I know him. This is magical.

• 48:00Speaker 02

  But this mirror of his, well, it was about to get shattered.

• 48:05Speaker 07

  Two thousand fifteen, that was the husband.

• 48:07Speaker 02

  Researcher at Google.

• 48:08Speaker 07

  Send me an email, like, we have some very exciting Go project. Can you go to our office, visit? We will show you our project. I tell, yes, okay, why not?

• 48:20Speaker 02

  And what they showed him was this thing called AlphaGo. It was a computer that had learned how to play the game. And they asked him, like, will will you play against this?

• 48:30Speaker 07

  So I took it, okay, we can play together because I will win. It's just a program, just a program. What can you do? You can win with me? Never. It's like zero percent chance to win this game, zero percent.

• 48:43Speaker 02

  And why? Why were you so confident?

• 48:46Speaker 07

  Because I know the best program this moment. I can gave six stone handicap.

• 48:52Speaker 02

  Handicap.

• 48:52Speaker 07

  Handicap, handicap game.

• 48:53Speaker 02

  Got it, got it, got it, got it.

• 48:54Speaker 07

  So how you can possible make the technique make the huge difference?

• 48:59Speaker 05

  Just a month is impossible.

• 49:03Speaker 02

  And so, a month later, in this windowless office room, Fawn faced off with this computer and its human stone-placing helper in a best-of-five-game match.

• 49:16Speaker 05

  The first game, all the game, I feel good. I think I will win. But, end of game.

• 49:22Speaker 02

  With just a few stones left to play.

• 49:24Speaker 05

  Ah, oh, how stupid. I make some mistake. And I lost my first game, okay?

• 49:30Speaker 02

  But, you know, he's thinking, I was sort of arrogant going into this. I was overly confident.

• 49:36Speaker 05

  So next game, I will be careful. I will play more seriously. I will win the game.

• 49:43Speaker 02

  So, the next day, next game, he sits down at the board, starts carefully placing his stones. And it's looking good on the board. But inside his head.

• 49:55Speaker 05

  I feel something really difficult, very difficult, very difficult.

• 50:00Speaker 02

  Because.

• 50:01Speaker 05

  I like fight, but AlphaGo don't fight with me. And if I want to something, Alpha give me very easily.

• 50:08Speaker 02

  Looking down at the board, he was not able to see his opponent's mind in the way he always had.

• 50:16Speaker 05

  No.

• 50:18Speaker 02

  There was no bravery, there was no subterfuge that he could sense.

• 50:24Speaker 05

  I see AlphaGo won't do this, AlphaGo won't do that. But why he want to this? You cannot find it, you can't.

• 50:32Speaker 02

  And so, he didn't know how to respond to it. His mind started to race.

• 50:39Speaker 05

  Good move, bad move. Good move, what mean? Bad move, what mean? Good move, what think my teacher? The good move, what think my student? Everybody, all my friends.

• 50:49Speaker 02

  And he realized that with all these emotional pushes and pulls that eventually.

• 50:55Speaker 05

  I will make mistake.

• 51:00Speaker 05

  But AlphaGo, no. Never. When you think about this,

• 51:09Speaker 05

  the confidence is crush, is.

• 51:13Speaker 05

  Crush, all crush. And I lost again, very, very badly.

• 51:20Speaker 05

  And I lost again for third, fourth, and the last one.

• 51:24Speaker 02

  Yeah, damn.

• 51:55Speaker 05

  But you know, this experiment is really good for me. This is, this is the moment I really see myself.

• 52:04Speaker 02

  Really? You think AlphaGo taught you to be more?

• 52:07Speaker 05

  Myself, yeah. I think this AlphaGo teaching me about that.

• 52:10Speaker 02

  And why?

• 52:11Speaker 05

  Because I see myself. So it's like AlphaGo teach me that our life, we will always lost, lost, lost, lost. Sorry, it's real life. It's our life. I think this is a human. This is important for us.

• 52:30Speaker 02

  I think what he saw in that game as he was losing was kind of what you were saying about seeing under the hood making AI less spooky. Like he could see it wasn't magic. It was math with no mistakes.

• 52:46Speaker 03

  Right.

• 52:47Speaker 02

  And when he saw himself, you know, like not being the perfect Go player in any given moment or in every given moment, like that's what makes him a person. A person who could love something but still lose at it. Maybe feel bad about that and then use that feeling to figure out what to do next.

• 53:13Speaker 05

  Today I teaching the Go in China with the student.

• 53:15Speaker 02

  Right, why are you teaching Go? The computer will always win.

• 53:20Speaker 05

  Yes, yes, but be careful because I think all you experiment to learn is still useful. So don't worry.

• 53:28Speaker 02

  It will be coming. You can do nothing. Accept it and just learn.

• 53:37Speaker 02

  Yeah, I get that. Before we wrap this thing up, I wanted to put all of this in front of someone. And not, not an AI person, but somebody with a really wide scope on technology and history. And so I went to this guy.

• 53:51Speaker 05

  Tom Malaney, I'm professor of modern Chinese history at Stanford University.

• 53:55Speaker 02

  I worked with him years back on a story about typing in Chinese. And he's, he's just one of the most thoughtful and informed people I know.

• 54:02Speaker 05

  That means a lot. That means a lot to me.

• 54:04Speaker 02

  So how, how would you respond?

• 54:07Speaker 05

  Well, I, I mean, everyday life is at its core a study of this awful, amazing, horrifying, never-ending surprise of what it means to be born and live and die as a human. And even if, at the end of the day, an AI is orders of magnitude smarter, AI, just by definition, cannot suffer and rejoice and live and die in quite the same way that, that humans can, in the same way that we cannot live and die and suffer and comprehend and feel the way an octopus can.

• 54:53Speaker 05

  I mean, the only thing an AGI will be able to do is contemplate, my goodness, what does it mean to be an AI? And so I am not worried at all about what AI means with regard to meaning, human identity, what it means to be human, or any of that.

• 55:17Speaker 02

  Well, that was very beautiful.

• 55:20Speaker 02

  And while I love that, I'm still like, but this is gonna mess everything up so badly. I don't know if-

• 55:28Speaker 05

  No, I agree. It's gonna, no, it's, I mean, this is gonna get weirded down to the fabric. But fast forward this, you know, 20, 30 years, if we're still around at a sort of climate change level, when another future human is sitting in this fabric-altered world, it will still be a group of humans.

• 55:52Speaker 05

  Rejoicing, suffering. Like, it will still be that condition. And so, it's kind of a, it's kind of a liber-- for me, it's like a little bit of a liberatory time. It's, it's, it's a, a great, maybe we'd get to free up a little bit more space to get back to work thinking about how to be human. Because we have not, we have not even come close to solving that issue.

• 57:46Speaker 03

  Special thanks to Stephanie Yin and the New York Institute of Go for teaching us the game, to Mark, Daria, and Levon, to Barbara Svenich, and of course, thank you to Grant Sanderson for his unending patience explaining the math of neural nets to us. Grant is kind of like your favorite math nerd's favorite math nerd. His YouTube channel is Three Blue, One Brown. Check it out. This story was reported and produced by Simon Adler, with original scoring and sound design by Simon Adler. Which brings me to the last unsavory thing I have to say, which is goodbye to Simon Adler, who happens to be one of our best reporter producers here at the show,

• 58:31Speaker 03

  and also a friend. He's going off to, among other things, pursue his music career. And this was his last episode on staff with us. Chances are, if you list out your favorite episodes from the last 11 years at the show, more than a couple will be his. Could be some of the tech stories he did. He did stories about drones in Ukraine, about content moderation on Facebook. Could be some of the international stories he did. He reported about the hunt for an endangered rhino in Namibia. He did a story about a species of raccoon in the Caribbean islands of Guadalupe. He did a lot of stories about democracy as well. Covered a town, Seneca, Nebraska, that voted itself out of existence.

• 59:15Speaker 03

  He did a story back in 2017 about a New York City council race where the campaign manager was a little-known guy named Zoran Mamdani. Besides being a killer reporter, not to mention composer and interviewer, Simon has also spent so many hours coaching an entire generation, and staffers, and interns. He's so generous with his expertise and his time. Really someone who makes everyone around him better. Anyway, we have been so lucky to have him as part of our nerdy band for 11 years. Check out his band, Windstar Enterprises, on Instagram. That's Simon and another fellow former radio lover,

• 59:58Speaker 03

  Alex Overington. We just, we already miss you, Simon. And good luck out there.

• 1:00:14Speaker 01

  Oh, you want me to say this? Oh, that's fun. Hi, I'm Cordelia and I'm from New York City. And here are the staff credits. Radiolab is hosted by Lulu Miller and Latif Nasser. Soren Wheeler is our executive editor. Sarah Sandback is our executive director. Our managing editor is Pat Walters. Dylan Keefe is our director of sound design. Our staff includes Simon Adler, Jeremy Blum, W. Harry Fortuna, David Gable, Maria Paz Gutierrez, Sindhu Nyan Nasa Bandhan, Matt Keelty, Mona Magavkar, Annie McEwen, Alex Neeson, Sarah Kari, Anisa Vitza, Ariane Wack,

• 1:00:57Speaker 01

  Molly Webster, and Jessica Young. With help from Rebecca Rand. Our fact checkers are Diane Kelly, Emily Krieger, Anna Pujol Mazzini, and Natalie Middleton.

• 1:01:10Speaker 01

  Leadership support for Radiolab's science programming is provided by the Simons Foundation and the John Templeton Foundation. Foundational support for Radiolab was provided by the Alfred P. Sloan Foundation.