The Alien in the Room

Summary

---

Radiolab: "The Alien in the Room"

Date: December 12, 2025
Hosts: Lulu Miller, Latif Nasser
Guest Contributors: Simon Adler (reporter/producer), Stephen Cave, Terry Sejnowski, Grant Sanderson, Fan Hui, Tom Mullaney

---

Episode Overview

In this emotionally resonant and intellectually playful episode of Radiolab, the team asks: "What is Artificial Intelligence, really?" Instead of getting lost in speculation or hype, hosts and guests break down the mechanics of AI, tracing its evolution and exploring why modern machine intelligence is so profoundly alien to us. Using vivid metaphors (octopuses! lightbulb grids!), historical anecdotes, and deeply human stories, the episode equips listeners to understand not just how AIs work, but how their minds diverge from ours—and what that means for the future of humanity.

---

Key Discussion Points & Insights

1. Setting the Stage: The Alien Among Us

• AI is everywhere and urgent, but deeply misunderstood. (03:01)
• Two camps: "AI will outsmart and destroy us" vs. "AI is just mimicking, nothing to worry about." (02:30)
• The show’s goal: get under the hood and actually explain what AI is, not just what it does.

[Notable Quote]

"Something that everybody's talking about, but nobody seems to actually understand."
– Simon Adler (02:06)

---

2. What Even Is Artificial Intelligence?

• Stephen Cave: Director, Leverhulme Centre for the Future of Intelligence (03:10).
• His team runs the "Animal AI Olympics," pitting AIs against tasks from animal psychology to see where AI fits in the "cognitive tree of life."
• Finding: AIs do not possess even animal "common sense"; tasks like manipulating objects or understanding gravity stump them.
• Moravec’s Paradox: For AI, “easy things are hard, hard things are easy.” Advanced math is simple, but simple perception is hard.

[Memorable Exchange]

"These systems don't have the common sense of a mouse..."
– Stephen Cave (05:12)

• Alien Metaphor: AI is less like a human, more like an octopus—intelligent but in a fundamentally different, distributed, "alien" way (06:06).

---

3. Tracing the Evolution of Machine Learning

a) Early Neural Nets

• Terry Sejnowski: Early neural network pioneer. Initially, AI was all “rules,” incapable of dealing with real-world complexity. (09:02)
• The breakthrough: Let the machine learn instead of manually programming every rule.
• Experiment: Teaching a neural network to pronounce English using child speech data—no rules, just pattern learning via feedback (11:46–14:19).
• Learning Process: Parallels with human brain development—random initial connections, gradually pruned and strengthened via feedback.

[Quote]

"What we didn't appreciate back then was that NetTalk was a little bit of 21st-century AI in the 20th-century..."
– Terry Sejnowski (14:19)

b) How Neural Networks Actually Work

• Guided by Grant Sanderson (ThreeBlueOneBrown), the show walks through a "mental image" of a neural network using circles and pixels (16:25–26:47).
  • Diagram: layers of "lightbulbs" (neurons), connected by wires (synapses).
  • Learning as a process of math and feedback, using calculus to minimize error.
• Key Idea: The middle layers find “clues” or features, but we (still) don’t know exactly what those clues are. This is the “black box.” (24:48)
• Scaling up: Add more layers to go from recognizing circles to recognizing faces, text, and beyond.

[Notable Quote]

"It's figuring that out itself...without anyone labeling any of those intermediate [features]."
– Latif Nasser (27:34)

---

4. From Recognition to Generation

• Prediction powers everything: Whether it's identifying a circle or predicting what word comes next, it's about guessing the next piece (33:29–36:05).
• The move from rule-based chatbots to predictive language models: They train on huge numbers of texts, using math to predict each next word/token.
• Scale blows up: Words are represented as vectors of thousands of numbers. Modern language models (like GPT-3) have 175 billion+ parameters. (43:34)
• Transformer models (such as those described in Google's "Attention Is All You Need" paper) solved long-range attention and context problems, enabling today's impressive AI feats. (41:08–44:17)

[Memorable Moment]

"It's so weird that...that's the simpler version for it."
– Latif Nasser (36:39) (on encoding words as thousands of numbers)

---

5. An Explosion of "Alien Minds"

• The transformer breakthrough led to the rapid proliferation of specialized AIs—text, image, video generators, etc.
• Fundamentally, it's still the same underlying math-driven prediction, just on a much grander scale and in different "mediums." (46:13)
• Even "creativity" in AI is just a controlled statistical drift (via "temperature" settings), not genuine inspiration. (47:01–47:59)

---

6. Human Reactions: Confronting the Alien

a) Fan Hui & AlphaGo: Losing to a Machine

• Fan Hui, three-time European Go champion, recounts losing to AlphaGo, Google’s AI (50:00).
  • Go is immensely complex, thought to be uniquely human in mastery.
  • Losing to AlphaGo shook his sense of self and confronts what it means to be human.
  • Profound take-away: The AI had no "mind" to read. It was a mirror, but a blank one, forcing Fan to confront only himself.
• AI as relentless, errorless math— winning with no emotion, no strategy to intuit.

[Notable Quotes]

"I like fight. But AlphaGo don't fight with me. If I want something, AlphaGo gives me very easily."
– Fan Hui (54:16)

"When you think about this, the confidence is crash. It's Crash. All crash."
– Fan Hui (55:19)

"AlphaGo teaches me that our life—we will always lost, lost, lost...It's our life. I think this is human. This is important for us."
– Fan Hui (56:30)

---

b) What Will This Mean For Us?

• Tom Mullaney (Stanford historian): No matter how smart AI gets, it cannot experience life, suffering, joy, or death as a human does. It will always be alien—just as an octopus is to us. (58:10–59:36)

[Notable Quote]

"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 humans can..."
– Tom Mullaney (58:26)

• Yet: “This is gonna get weird down to the fabric.” But—fundamentally—future human life will still involve humans “rejoicing, suffering.”

---

Notable Quotes & Memorable Moments (with Timestamps)

• Stephen Cave: "Intelligence can be so profoundly different—alien." (07:22)
• Stephen Cave: "These systems don't have the common sense of a mouse..." (05:12)
• Grant Sanderson: "This is what calculus is all about. Like, Newton, if he was rising from the grave, would just be like showing fireworks right now..." (23:15)
• Fan Hui (on losing to AlphaGo): "When you think about this, the confidence is crash. It's Crash. All crash." (55:19)
• Tom Mullaney: "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...solved that issue." (60:45)

---

Key Timestamps

• [03:10-05:51] – Animal AI Olympics and Moravec's Paradox
• [09:02-14:19] – Early neural nets, text-to-speech training
• [16:25-26:47] – Neural net "lightbulb" analogy and black box learning
• [33:29-36:05] – "Prediction" as the foundation of generative AI
• [41:08-44:17] – GPUs, transformers, and the data explosion
• [50:00-57:05] – Fan Hui’s story: Defeat by AlphaGo, existential reflection
• [58:10-60:45] – Tom Mullaney: The immutable strangeness and resilience of humanity

---

Tone, Language & Style

The hosts maintain their signature mix of curiosity, humility, and humor, balancing rigorous explanation with relatable analogies and deeply human moments. Jargon is broken down (often with self-deprecating laughter), and technical experts are encouraged to tell stories that resonate emotionally and intellectually.

---

Takeaways & Themes

• “Alien” isn’t just a metaphor: True machine intelligence is legitimately strange—powerful, but unfathomable in the ways that matter most to humans.
• AI as “math with no mistakes”—sterile, relentless, sometimes dazzling, but ultimately without intention or feeling.
• Confronting AI’s abilities can be humbling—forcing us to ask what’s unique (and valuable) about being human.
• The future will be strange and disruptive, but human life (with its messy, felt experience) continues—and the project of “how to be human” remains unsolved.

---

This episode is a must-listen for anyone seeking to peer behind the curtain of AI, not just how it works, but why it is—and will likely remain—an alien intelligence in the room with us.

Transcript

[Latif Nasser]

Radiolab is supported by Apple TV. It's 1972. A young British family is attempting to sail around the world when disaster strikes. Their boat is hit by killer whales and it sinks in seconds. All they have left is a life raft and each other. How will they survive? The true story of a family's fight for survival, hosted by Becky Milligan. This is Adrift, an Apple original podcast produced by Blanchard House. Apple TV subscribers get special early access to the entire season. Follow and listen on Apple Podcasts.

[Lulu Miller]

Radiolab is supported by Story Publishing. Presenting the universe in 15 portals to wonder Through Science and Poetry. An illustrated collection of short essays and accompanying poems by Maria Popova of the Marginalian. Radiolab creator Jad Abumrad writes, this book is a wonder. Science writing so often negates the inherent wonder of science, which comes so brilliantly alive here. Something bursts open in the mind. And let me say this. Maria Popova has the rare gift of starting a sentence and leaving you entranced by its end. I'd read anything she writes. Available wherever books are sold.

[Latif Nasser]

Oh, wait, you're listening.

[Lulu Miller]

Okay.

[Stephen Cave]

All right.

[Latif Nasser]

Okay.

[Stephen Cave]

All right.

[Simon Adler]

You're listening to Radiolab Radio Lab from wny.

[Lulu Miller]

See?

[Latif Nasser]

Yep.

[Simon Adler]

Okay. After all of that, it is time to finally discuss. Let's if the question.

[Latif Nasser]

Yeah.

[Simon Adler]

The topic.

[Latif Nasser]

Okay.

[Simon Adler]

The theme of the moment, perhaps.

[Latif Nasser]

Climate change.

[Simon Adler]

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

[Stephen Cave]

Simon.

[Latif Nasser]

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. We're here today to talk about a different huge, overwhelming thing that we're all in the middle of.

[Simon Adler]

I mean, I don't wanna put words in your mouth, but what I have been feeling is a general sense of frustration.

[Latif Nasser]

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

[Simon Adler]

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

[Latif Nasser]

That's right. Which is, of course, artificial intelligence.

[Simon Adler]

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.

[Latif Nasser]

All. Right?

[Simon Adler]

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

[Latif Nasser]

Right.

[Simon Adler]

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

[Stephen Cave]

Yeah.

[Simon Adler]

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.

[Stephen Cave]

Yeah, because we don't know.

[Terry Sejnowski]

Right.

[Stephen Cave]

This is one of the most extraordinary things about machine learning. AI is that we don't really know what they are.

[Simon Adler]

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.

[Stephen Cave]

Stephen Cave. I'm the director of the Leverholme Centre for the Future of Intelligence.

[Simon Adler]

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

[Stephen Cave]

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, tests designed to.

[Simon Adler]

Measure how well a mouse can problem.

[Stephen Cave]

Solve 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?

[Simon Adler]

And they've actually turned these tests into a sort of competition that they call the Animal AI Olympics.

[Stephen Cave]

Yes, indeed.

[Latif Nasser]

Okay. Oh, that just sounds fun, right?

[Simon Adler]

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

[Latif Nasser]

Okay.

[Simon Adler]

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

[Latif Nasser]

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

[Simon Adler]

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

[Latif Nasser]

Okay.

[Simon Adler]

And then they take these AI which are running on basically the same kind of engine that power 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 this 3D world and say, there is food in here, find it.

[Latif Nasser]

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

[Simon Adler]

Really food, it's this green orb that they're looking for.

[Latif Nasser]

Okay.

[Simon Adler]

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

[Latif Nasser]

But it's like physical world problem solving.

[Simon Adler]

Absolutely. And I mean, while this is the sort of task that mice or pigeons can pull off pretty easily for these.

[Stephen Cave]

AI agents, things like manipulating objects and understanding gravity, it's a real challenge.

[Simon Adler]

Like they struggle to press a lever or perceive an edge, which any animal.

[Stephen Cave]

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.

[Latif Nasser]

That's the Moravec's paradox. Right. Like, it's like easy things are hard and hard things are easy.

[Simon Adler]

Exactly, yeah. 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 they have a.

[Stephen Cave]

Completely different profile of capabilities and skills than any animal.

[Simon Adler]

They are not like us.

[Stephen Cave]

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

[Simon Adler]

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

[Stephen Cave]

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 for lots of wonderful octopus escape stories.

[Simon Adler]

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

[Stephen Cave]

Um, no.

[Simon Adler]

Then why this metaphor?

[Stephen Cave]

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

[Simon Adler]

Okay?

[Stephen Cave]

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

[Simon Adler]

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.

[Stephen Cave]

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

[Simon Adler]

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.

[Latif Nasser]

Right?

[Simon Adler]

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, 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. So this is great.

[Latif Nasser]

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

[Simon Adler]

Great. Fill your glass because here we go. Hey, you guys can hear me?

[Terry Sejnowski]

Yes, I can hear you, Simon.

[Simon Adler]

Hello, Terry. How are you?

[Tom Mullaney]

Very good, thank you.

[Simon Adler]

Sorry for the slight delayed start here. Some classic technical difficulties, you know. So there are a lot of different first contacts we could point to with this Alien species. But the most fun place to start that I found is with this guy.

[Terry Sejnowski]

Terry Sinofsky, professor at the Salk Institute for Biological Studies.

[Simon Adler]

Yeah, sort of like the midwife of AI. Is that a helpful way to think of you or no?

[Terry Sejnowski]

Yes, yes, actually. Well, it's obviously more complicated than that, but that's not a bad analogy.

[Simon Adler]

Terry trained as a neurobot. He came up poking probes in monkeys.

[Terry Sejnowski]

Heads to try to understand how the brain works.

[Simon Adler]

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.

[Terry Sejnowski]

But it was going nowhere.

[Simon Adler]

Okay.

[Terry Sejnowski]

Because everything was based on rules at the time.

[Simon Adler]

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, this sort of web of logic, which when it comes to recognizing sounds or pictures, was a problem because for each rule there are.

[Terry Sejnowski]

You know, tens of thousands, a hundred thousand exceptions.

[Simon Adler]

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

[Terry Sejnowski]

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, he started to wonder.

[Simon Adler]

If there was a different way to tackle this learning.

[Terry Sejnowski]

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

[Simon Adler]

They set out to build a machine that could learn. And one of the first things they tried to teach it was how to.

[Terry Sejnowski]

Pronounce English, you know, text to speech in computer science.

[Latif Nasser]

And amazingly, Demonstration of network Learning by Terry Sinowski and Charles Rosenberg.

[Simon Adler]

They have recordings from these early training sessions.

[Terry Sejnowski]

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

[Simon Adler]

And so. So, you ready?

[Levon]

Ooh.

[Lulu Miller]

Ah.

[Levon]

Bleh.

[Simon Adler]

Okay. Sometimes, Yvonne, look at me. They took a transcript of a kid talking, a transcript I had my friend and neighbor Levon reenact.

[Levon]

When we walk from school, I go to my grandmother's house because he gives us candy.

[Latif Nasser]

Nice.

[Simon Adler]

That's perfect. You ready for the next one? 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 first.

[Latif Nasser]

Recording de novo learning.

[Simon Adler]

And here it is. No, no, no, no, no, no, no.

[Levon]

No, no, no guns.

[Latif Nasser]

Wow. Right?

[Simon Adler]

So it has no idea what it's doing?

[Latif Nasser]

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

[Simon Adler]

It's chaos, right? It's like noise. Effectively, yeah. But then, as it continued quizzing itself, comparing its output to what it should.

[Levon]

Have said, When I go to my cousins, I play badminton.

[Terry Sejnowski]

All that slowly, we could actually hear the learning. You could hear it figuring out the difference between vowels and consonants. And then it would start pronouncing small.

[Levon]

Words, you know, oh, we gotta go to Micah. Can't get one mile.

[Lulu Miller]

No.

[Levon]

Yum, yum. Come dino it.

[Terry Sejnowski]

We sleep and, you know, it only took a couple of days when we.

[Levon]

Walked home from school. I like to go to my grandmother's house. Well, because she gave us candy and.

[Simon Adler]

It was acing it.

[Levon]

And we eat there sometimes. We eat there sometimes. Sometimes we sleep overnight there. Sometimes we sleep overnight there. Sometime when I go to go to my cousins, I get up late. Family to all that.

[Simon Adler]

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

[Levon]

He won't stop jumping around in the bathtub.

[Simon Adler]

It pronounced those too.

[Levon]

He beats jumping around, gets tired when he got stuck. But when he finally gets to sleep, it was phenomenal. Sometimes I get to go to bed at 12:30 sometimes, but most of the times I don't.

[Terry Sejnowski]

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

[Levon]

Are we done?

[Latif Nasser]

We're done.

[Simon Adler]

Thank you so much.

[Latif Nasser]

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

[Simon Adler]

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.

[Latif Nasser]

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

[Simon Adler]

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 know that this is a square. Or if you poke a cat, you get scratched.

[Terry Sejnowski]

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.

[Simon Adler]

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

[Terry Sejnowski]

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

[Simon Adler]

Okay, but how did you. How did you do that? Like, what is going on under the hood here that allows it to do this?

[Terry Sejnowski]

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

[Simon Adler]

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

[Latif Nasser]

Okay.

[Grant Sanderson]

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 at all?

[Simon Adler]

With the help of this guy, Grant Sanderson.

[Grant Sanderson]

Yeah. I run a YouTube channel that's named Three Blue, One Brown. I often talk about math, but math adjacent things as well.

[Latif Nasser]

Great.

[Simon Adler]

We're just going to draw like a mental image of what one of these networks looks like.

[Latif Nasser]

Okay, let's go.

[Simon Adler]

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 shapes. What shape is this?

[Levon]

A circle.

[Grant Sanderson]

Can we get a computer to see a circle?

[Simon Adler]

How about this?

[Levon]

A circle.

[Simon Adler]

A very childlike task.

[Latif Nasser]

Yeah, sure.

[Simon Adler]

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.

[Latif Nasser]

Okay.

[Simon Adler]

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.

[Latif Nasser]

Okay.

[Simon Adler]

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

[Latif Nasser]

Okay.

[Simon Adler]

I can see it from there for reasons that'll make sense in a minute. Below that, we're gonna add a smaller grid of 10 light bulbs. And then below that, just one bottom.

[Latif Nasser]

Bulb at the top, 100 light bulbs. And then another layer, 10 light bulbs, another layer, one bulb.

[Simon Adler]

Exactly. And that final bulb, that is just the answer. The output that when it turns on says.

[Levon]

Yes, Circles.

[Fan Hui]

Circles.

[Simon Adler]

That's right. There's a circle here.

[Latif Nasser]

Okay.

[Simon Adler]

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. 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.

[Grant Sanderson]

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 visual cortex to try to visualize all this. But setting aside all of the technical.

[Simon Adler]

Terminology, there's one last thing to do. We have to wire all of these bulbs together so that electricity can 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, and then go and connect every bulb in the middle 10 to that final bulb.

[Latif Nasser]

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

[Simon Adler]

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

[Latif Nasser]

Got it.

[Simon Adler]

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

[Levon]

Hey, why are you using a microphone again?

[Simon Adler]

To record your voice. Lighting up some of the bulbs in that top. What shape is this? The electricity passes down through these random connections from the top to the middle down to the bottom. And in all likelihood, I don't want rectangle.

[Grant Sanderson]

It's completely wrong on this.

[Simon Adler]

That final bulb might be a little lit up or half lit up or just completely off. Okay, 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.

[Latif Nasser]

Right, right.

[Simon Adler]

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.

[Latif Nasser]

Right, right, right.

[Simon Adler]

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

[Latif Nasser]

Yeah.

[Simon Adler]

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.

[Latif Nasser]

Right.

[Simon Adler]

And so instead, we're going to 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, There is a 12.2%.

[Levon]

Likelihood of a circle in this image.

[Simon Adler]

We're going to say, bad robot. There is absolutely a circle in this image. Try again.

[Levon]

Okay, I will try again.

[Simon Adler]

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. Uh, okay.

[Grant Sanderson]

It's not that complicated. It's mostly multiplication.

[Latif Nasser]

All right. Okay, let's go.

[Simon Adler]

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

[Levon]

1, 2, 3, 4, 5.

[Simon Adler]

And the wires, you can really just think of them as variables that multiply these numbers x times 2 as they.

[Levon]

Pass through them, y times point 3.

[Simon Adler]

A good wire multiplies the electricity by 5 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.

[Grant Sanderson]

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.

[Simon Adler]

Bit of mathematical trickery, this equation can represent the difference between the output it is giving.

[Levon]

There is a 12.2% likelihood of a.

[Simon Adler]

Circle, and the output we want it.

[Levon]

To give, There is a 100% likelihood of a circle.

[Grant Sanderson]

And if we think, hey, I've got this function, and I want to find a minimum of that.

[Simon Adler]

Like, minimize the difference between your output and the output we want.

[Grant Sanderson]

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.

[Latif Nasser]

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

[Simon Adler]

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. Okay, so 12% likelihood after the first wrong answer, the equation says, no. Machine tries again.

[Levon]

25% likelihood.

[Simon Adler]

And the equation says closer. And the machine tries again.

[Levon]

77% likelihood.

[Simon Adler]

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

[Latif Nasser]

Bulbs getting it closer and closer to right.

[Simon Adler]

Exactly. And what happens over time is 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.

[Latif Nasser]

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

[Simon Adler]

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

[Latif Nasser]

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

[Simon Adler]

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, circles, circles, but with tens, circles, 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. Just average it all together. All of the wire strength you got from all the examples for wire one get averaged down to one value. All the wire strengths 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.

[Latif Nasser]

Holy cow.

[Simon Adler]

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.

[Latif Nasser]

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

[Simon Adler]

Yes.

[Latif Nasser]

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

[Simon Adler]

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.

[Stephen Cave]

Great.

[Latif Nasser]

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 and, like, crazy fast.

[Simon Adler]

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.

[Latif Nasser]

But, like, a circle. Recognizing a circle is one thing. And, like, now we're talking, like, actually making. Like, making, you know, a sonnet as if Shakespeare wrote it. That seems like a very wide gulf. It seems like there's still a lot of place to go, for sure.

[Simon Adler]

And our little alien is going to have to evolve here. But in terms of its architecture, of how it does this, it's basically the exact same. The only real difference is we're shifting its focus from recognizing to a slightly different skill, and we're gonna get to that. You want to predict what I'm gonna say next right after a quick break. Exactly right after a quick break.

[Lulu Miller]

Radiolab is supported by bilt. Nobody wants to pay rent, but if you have to, BILT works to make it more worthwhile by paying rent. Through Bilt, you can earn flexible points that can be redeemed toward hundreds of hotels and airlines, a future rent payment, your next Lyft ride, and more. But it doesn't stop there. You can dine out at your favorite local restaurants and earn additional points, get VIP treatment at certain fitness studios, and enjoy exclusive experiences just for BILT members. Every month, earn points on rent and around your neighborhood, wherever you call home, by going to joinbuilt.com Radiolab. That's J-O-I-N-B-I-L-T.com Radiolab. Hi, I'm Lulu Miller, and this episode is sponsored by BetterHelp the holidays are a time of old traditions. Lighting candles in the window, singing centuries old songs. But why stay wed to the past? I've heard all kinds of stories lately of people inventing new traditions. Like my friends Lena and Jamie, who celebrated what they called trash giving last year by picking up trash in the morning and then eating trashy food together at night. Or my friend Terry, who skipped out on Thanksgiving altogether last year for a solo camping retreat of replenishment. Holidays can bring up the feels, good feels and hard feels. It's natural, and a great way to navigate all of them is through the regular practice of therapy. Why not try something new and different, like giving yourself the gift of therapy? And one of the easiest ways to get you going on this journey is BetterHelp. BetterHelp does the initial matching for you so you can focus on your therapy goals. A short questionnaire helps identify your needs and preferences, and their 12 plus years of experience means they typically get it right the first time. And if you aren't happy with your match, you can switch to a different therapist at any time from a list of recommendations tailored to your needs. This December, start a new tradition. Treat yourself by taking care of you. Our listeners get 10% [e-mail] Radiolab that's betterhelp.com Radiolab Radiolab is supported by Capital One with no fees or minimums on checking accounts. It's no wonder the Capital One bank guy is so passionate about banking with Capital One. If he were here, he wouldn't just tell you about no fees or minimums. He'd also talk about how most Capital One cafes are open seven days a week to assist with your banking needs. Yep, even on weekends it's pretty much all he talks about in a good way. What's in your wallet? Terms apply. See capitalone.combank Capital One NA member FDIC.

[Latif Nasser]

Radiolab is supported by Rippling Finance teams often spend weeks chasing receipts, reconciling spreadsheets and fixing errors across disconnected spend tools. This can be frustrating. And that's not software as a service. That's sad software as a disservice. If you've been thinking about replacing stitched together tech stacks with one platform for all departments, Rippling can help. Rippling is a unified platform for global hr, payroll, IT and finance, helping people replace their mess of cobbled together tools with one system. Designed to help give leaders clarity, speed and control. By uniting employees, teams and departments in one system, Rippling works to remove the bottlenecks, busy work and silos in business software With Rippling, you can choose to run hr, IT and finance operations as one, or pick and choose the products that best fill the gaps. Right now you can get 6 months free when you go to rippling.com Radiolab learn more at r I p p l-I n g.com Radiolab terms and conditions apply.

[Simon Adler]

Latif Simon, Radiolab so you asked this question to me before the break, like, how did this thing evolve from being able to recognize shapes to generate stuff?

[Latif Nasser]

Yeah.

[Simon Adler]

And I posed that very question to Grant Sanderson.

[Latif Nasser]

Okay. Yeah.

[Grant Sanderson]

Okay. So I would say there's many different ideas at play here.

[Simon Adler]

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.

[Grant Sanderson]

Like with the circle example, there's two.

[Simon Adler]

Things that it could predict, circle or not.

[Latif Nasser]

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.

[Lulu Miller]

Right?

[Simon Adler]

But it becomes important when we're talking about generative stuff.

[Latif Nasser]

Okay.

[Simon Adler]

Like in the case of language, predict.

[Grant Sanderson]

What word comes next.

[Simon Adler]

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.

[Stephen Cave]

There, how are you today?

[Simon Adler]

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 bulbs, those layers it's passing through, they haven't been trained to categorize a sentence. 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.

[Grant Sanderson]

Just one word. Just one. It's not even a word. Also, there's a nuance here between the notion of words and tokens, but excessive nuance.

[Latif Nasser]

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 that right answer. Right, but like, in a sentence that could go any million number of ways, how can it ever Have a right answer to train back to.

[Simon Adler]

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 right answer was the most likely word to follow the preceding words.

[Latif Nasser]

So 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 in this exact scenario?

[Simon Adler]

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.

[Latif Nasser]

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

[Simon Adler]

Yeah. And so, like in the way that a pixel value in the circle example was like, basically a 0 or a 1, it's like every word is this list of 13,000 numbers.

[Latif Nasser]

It's so weird that it.

[Simon Adler]

Like that.

[Latif Nasser]

That's the simpler version for it.

[Simon Adler]

I know.

[Latif Nasser]

Let me turn it into this, like.

[Simon Adler]

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

[Latif Nasser]

Yeah, they're really not us.

[Simon Adler]

Not at all.

[Stephen Cave]

Wow.

[Simon Adler]

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 takes all the words in your question plus the first word it predicted, sends all that through the network again.

[Grant Sanderson]

And then it just predicts the next.

[Simon Adler]

Word after that, sends that through those bulbs again.

[Grant Sanderson]

And then the next word after that.

[Simon Adler]

Does the whole thing again and plays.

[Grant Sanderson]

The same game over and over and over. And one of the words in its vocabulary is the end conversation token. So it has some notion of when to stop, but the act of stopping is itself just one more prediction. It's one more probability in that big list of things that should happen next.

[Simon Adler]

And as I said, this is how they were doing it all the way back in the 80s. 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.

[Latif Nasser]

Really?

[Simon Adler]

Do you have any recollection of when you first came in contact with one? Oh, God.

[Latif Nasser]

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

[Simon Adler]

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

[Latif Nasser]

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

[Simon Adler]

And a big part of why they were bad was they had difficulty dealing with longer stretches of text. This is Steven Levy, editor at large at Wired. He's been covering this stuff for. Yeah, yeah, yeah. I mean, a long time. I published a book in 1992 called Artificial Life. I was two years old, by the way. Thanks for that. Sorry. Yeah, yeah, thanks. 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.

[Terry Sejnowski]

Exactly.

[Latif Nasser]

Huh.

[Simon Adler]

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?

[Latif Nasser]

It's like, I can see why that would be confusing, right?

[Simon Adler]

Like, you have to somehow know that in that sentence, dog is really the operative term here. The important noun. It's not I or door.

[Latif Nasser]

Right, right, right.

[Simon Adler]

And so in 2017, this guy, you know, Oscar Eep, who worked at Google, set out to solve this dog door problem. 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. And now the question becomes like, how the heck does one go about doing that? And what they figured out was, 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 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. But then just, like, set that aside, because you're gonna take in all that information again, and then we're gonna 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.

[Latif Nasser]

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

[Simon Adler]

Yeah, and like, I. I've looked at this stuff for months, and I still don't totally understand exactly how a machine does this, but. Well, I mean, something like that. And also, you know, you can say no. You can tell me no, I mean, in the raw sense. Yeah, that's the idea.

[Latif Nasser]

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.

[Simon Adler]

Totally. 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.

[Latif Nasser]

GPUs components that go into data centers.

[Simon Adler]

Or the company maker, Nvidia. Nvidia, the most valuable company in history that makes these things. Its story, of course, wrapped up in.

[Lulu Miller]

The frenzy around the future of artificial intelligence.

[Simon Adler]

These things in this company have been at the center of the conflict between China and the US when it comes to export controls. The idea here is for the US.

[Latif Nasser]

To kind of limit the ability for.

[Fan Hui]

China to catch up when it comes to AI.

[Simon Adler]

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

[Grant Sanderson]

Computer games, video games, things like that.

[Simon Adler]

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

[Grant Sanderson]

Exactly. 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.

[Simon Adler]

And 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.

[Latif Nasser]

Sure, sure, sure.

[Simon Adler]

And so with these gpu' 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.

[Latif Nasser]

All right.

[Simon Adler]

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

[Grant Sanderson]

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.

[Simon Adler]

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, in our.

[Grant Sanderson]

Smaller example with the circle, there's something.

[Simon Adler]

Like a thousand and some odd parameters.

[Latif Nasser]

Right.

[Simon Adler]

A thousand or so of those wires.

[Grant Sanderson]

GPT3, which was kind of dumb by today's standards, but it came out, it had 175 billion parameters.

[Simon Adler]

175 billion things that could be tweaked.

[Latif Nasser]

Yeah.

[Grant Sanderson]

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

[Simon Adler]

And as they fed basically all the things we humans have ever said on the Internet into this thing, throwing way.

[Grant Sanderson]

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

[Simon Adler]

Slowly they started to notice that with.

[Grant Sanderson]

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.

[Simon Adler]

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 there was a passage in there.

[Terry Sejnowski]

Saying, we think this can work for images and video. And indeed that turns out to work.

[Simon Adler]

That same basic model of massive parallel processing with tons of input that could predict the next part of an image or sound.

[Latif Nasser]

The moment civilization was transformed.

[Simon Adler]

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.

[Stephen Cave]

New artificial intelligence systems, machines that can.

[Simon Adler]

Teach themselves superhuman skills.

[Latif Nasser]

ChatGPT3, GPT4, Apple Intelligence, Dall? E, an app called Lensa. Bard.

[Lulu Miller]

Bard.

[Simon Adler]

It's called midjourney Text to video art generated by. It's crazy.

[Latif Nasser]

Look at this. So I don't know what AI it is they're using.

[Simon Adler]

Yes. It feels like an episode of Black Mirror.

[Latif Nasser]

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.

[Simon Adler]

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.

[Latif Nasser]

It does send you into a spiral. Because it's like. It's like. Is what I do any different from that? Just spewing out, you know, some iteration of everything else? I've seen before this.

[Terry Sejnowski]

Yeah.

[Simon Adler]

But first of all, you're not. You're not pulling from the whole Internet.

[Lulu Miller]

Right.

[Simon Adler]

Like, you have to depend on just the limited things you've experienced or can even maybe remember.

[Latif Nasser]

That's fair.

[Simon Adler]

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

[Latif Nasser]

Yeah.

[Simon Adler]

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.

[Stephen Cave]

I think it's.

[Simon Adler]

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.

[Tom Mullaney]

No.

[Latif Nasser]

So you can control, like, how precise you want the math. Like, you can. You can say, I want it a little stanky.

[Fan Hui]

Yeah.

[Simon Adler]

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 chart turned up on, like, every third word so that there's this almost spontaneous feeling, serendipitous creation, act of creation that comes out of this rigid math.

[Latif Nasser]

Like, it's like something startlingly creative might.

[Simon Adler]

Just be a less right answer.

[Latif Nasser]

A less right answer. Wow.

[Simon Adler]

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

[Latif Nasser]

Yeah.

[Simon Adler]

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

[Latif Nasser]

Really? I didn't hear about that.

[Simon Adler]

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.

[Latif Nasser]

To me, I think the thing that makes me realize is when you see under the hood, what you see is less like something spooky and ethereal. Like, there are times when it gets spooky when there'll be a time I'll be listening to an AI generated podcast, and then one of the hosts breathes, and I'm like, wait, that's so weird. 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 much less eerie. Cause you can see where it got it from, right?

[Simon Adler]

But well, okay, 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.

[Fan Hui]

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.

[Grant Sanderson]

No.

[Fan Hui]

So yeah, it's me.

[Simon Adler]

And so I got in touch with.

[Fan Hui]

This guy, Fan Hui. I'm a professional Go player, three time European champion.

[Simon Adler]

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.

[Fan Hui]

It's very simple idea, but it's difficult.

[Simon Adler]

Because with such simple rules, there are just this crazy number of ways 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.

[Fan Hui]

Yes.

[Simon Adler]

Anyhow, back to Fon.

[Fan Hui]

I remember I discovered a goal age like 6 in my school in Xi'. An and I feel something. Oh, this game I can play. And I progressed very quick. One year after I learned Go for my school, I'm number one.

[Simon Adler]

Three years after that, I'm in the.

[Fan Hui]

Best team in the province.

[Simon Adler]

And not long after that I stopped my school.

[Fan Hui]

I only learn Go game, I mean for years, every day, only thing you do is just to play go game. 12 hours.

[Simon Adler]

12 hours of playing, yeah, it's no joke. Around age 15, he went pro. And somewhere along the way he says he noticed this almost magical quality of the game.

[Fan Hui]

Go for me is like a mirror. A mirror mirror, yeah, because when you play, you can see your mind on the board.

[Simon Adler]

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.

[Fan Hui]

It's like print the mind, print and your opponent's mind.

[Simon Adler]

He says it's printed there too.

[Fan Hui]

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.

[Simon Adler]

But this mirror of his, well, it.

[Fan Hui]

Was about to get shattered 2015 that Mr. Hasabis researcher at Google sent 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?

[Simon Adler]

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?

[Fan Hui]

So I tell him, 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.

[Grant Sanderson]

Never.

[Fan Hui]

It's like 0% chance to win with me. 0%.

[Simon Adler]

And why? Why were you so confident?

[Fan Hui]

Because I know the best program this moment I can give. Six stone handicap, handicap, handicap, handicap game.

[Simon Adler]

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

[Fan Hui]

So how you can possibly make the technique make the huge difference? Just months impossible.

[Simon Adler]

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

[Fan Hui]

That first game, all the game, I feel good. I think I will win. But end of the game, with just.

[Simon Adler]

A few stones left to play.

[Fan Hui]

Oh, how I stupid. I made some mistake and I lost my first game.

[Simon Adler]

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

[Fan Hui]

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

[Simon Adler]

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, I feel something.

[Fan Hui]

Really difficult, very difficult.

[Simon Adler]

Because I like fight.

[Fan Hui]

But Afaku don't fight with me. And if I want something, AFU give me very easily.

[Simon Adler]

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

[Latif Nasser]

No.

[Simon Adler]

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

[Fan Hui]

I see Africa want to do this. Africa won't do that. But why he won't do this? You cannot fund it. You can't.

[Simon Adler]

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

[Fan Hui]

Good move, Bad move. Good move. What mean bad move? What mean? Good mood? What sink my teacher? The good mood. What sink my student, Everybody, all my friends.

[Simon Adler]

And he realized that with all these emotional pushes and pulls that eventually I will make mistake.

[Fan Hui]

But Avago, no, never. When you think about this, the confidence is crash. It's Crash. All crash. And I lost again. Very, very badly. And I lost again for third, fourth and the last one.

[Latif Nasser]

Yeah.

[Simon Adler]

Damn.

[Fan Hui]

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

[Simon Adler]

Really? You think AlphaGo taught you to be more myself?

[Fan Hui]

Yeah, I think this alphago teaches me about that.

[Simon Adler]

And why?

[Fan Hui]

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.

[Simon Adler]

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.

[Latif Nasser]

Right.

[Simon Adler]

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.

[Fan Hui]

Today I teach the GO in China with the students.

[Simon Adler]

Wait, why are you teaching go? The computer will always win.

[Fan Hui]

Yes, yes, but be careful because I think all you experiment to learn is still useful. So don't worry, it will be coming. You can do nothing, accept it and just learn.

[Tom Mullaney]

Yeah, I get that.

[Simon Adler]

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

[Tom Mullaney]

I'm professor of Modern Chinese history at Stanford University.

[Simon Adler]

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

[Tom Mullaney]

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

[Simon Adler]

So how would you respond?

[Tom Mullaney]

Well, 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 humans can, in the same way that we cannot live and die and suffer and comprehend and feel the way an octopus can, 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.

[Simon Adler]

Well, that was very beautiful. And while I love that, I'm still like, but this is gonna mess everything up so badly. I don't know if it's. Oh, no, I agree. Okay.

[Tom Mullaney]

It's gonna. No, it's. I mean, this is gonna get weird down to the fabric.

[Latif Nasser]

But.

[Tom Mullaney]

But Fast forward this 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 rejoicing, suffering. It will still be that condition. And so it's kind of a. It's kind of a liberatory for me. It's like a little bit of a liberatory time. It's 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.

[Simon Adler]

You.

[Stephen Cave]

It.

[Latif Nasser]

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

[Simon Adler]

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 Sandbach is our executive director.

[Latif Nasser]

Our managing editor is Pat Walters.

[Simon Adler]

Dylan Keefe is our director of Sound design. Our staff includes Simon Adler, Jeremy Bloom.

[Latif Nasser]

W. Harry Fortuna, David Gable, Maria Paz Gutierrez, Sindhu Nyanasambandan, Matt Kielty, Mona Maggavkar.

[Simon Adler]

Annie McEwen, Alex Neeson, Sara Khari, Anissa Vitza, Arianne Wack, Molly Webster, and Jessica Young, with help from Rebecca Rand. Our fact checkers are Diane Kelly, Emily.

[Latif Nasser]

Krieger, Ana Pujol Mazzini, and Natalie Middleton.

[Simon Adler]

Leadership support for Radiolab 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.

[Latif Nasser]

Radiolab is supported by the National Forest Foundation, a nonprofit transforming America's love of nature into action for our forests. Did you know that national forests provide clean drinking water to 1 in 3Americans? And when forests struggle, so do we. The National Forest foundation creates lasting impact by restoring forests and watersheds, strengthening wildfire resilience and expanding recreation access for all. Last year they planted 5.3 million trees and led over 300 projects to protect nature and communities nationwide. Learn more at nationalforests.org Radiolab.

[Lulu Miller]

Radiolab is supported by John Duvall, author of the children's book the Great Spruce Presents a new heartwarming Tale. The young crow, Alec and Rosa, are the best of friends when they find a fledgling crow that has fallen out of its nest, who they name Blue. They do their best to take care of it in any way they can before it's ready to venture out on its own. Before spring ends, will Blue be able to fly? And how will Alec and Rosa feel if Blue does gain the strength to flap his wings? Available now on Amazon.