Loading summary
Transcript153 lines
- [00:00]
A
Welcome to the Rest is Science. This is Field Notes, sort of. It's a kind of podcast expedition of the mind, as it were, where Michael and I, we're trading the curious thoughts that have been occupying us.
- [00:12]
B
That's right. And we also entertain questions and thought experiments from you all.
- [00:18]
A
We certainly do. And in general, we want you to send them in. Send us in. Anything you want us to know. Your thought experiments, the things that have been troubling you. Now, later in this episode, in the second half, I am going, I've got my sort of object, as it were. It's a metaphorical object. This time, Michael, I'm going to be sharing with you the thing that I currently find most troubling about the future with AI. I've got a few stories to tell you, but we're first going to go to your questions, as we always do, to our little mailbox.
- [00:50]
B
That's right. I wanted to read you an email that we got from John. This is just very cool and it's so related to our previous episodes. Okay.
- [00:58]
A
You get.
- [00:59]
B
You get extra brownie points for that. So John emailed us to say hi. To make the link between two of your recent programs. Paul Hoffman, who, in his biography of Erds, the man who loved only numbers, wrote the following. Listen to this.
- [01:14]
A
Go on.
- [01:15]
B
A conjecture both deep and profound is whether the circle is round. In a paper of Erdish written in Kurdish, a counterexample is found.
- [01:26]
A
So we've got maths, limericks and Erdish. How did we say it in the show?
- [01:31]
B
How did we say it in the show?
- [01:34]
A
I don't think we said it to rhyme with Kurdish.
- [01:36]
B
This limerick is a brilliant way to remember how to pronounce Erdish because it rhymes with Kurdish. I think that's one of the reasons Hoffman wrote it. But then Erds himself heard this limerick and was like, I gotta publish a mathematical paper in Kurdish. But he couldn't find a Kurdish math journal. So that's what John had to share with us. And I love that because now I won't forget how to pronounce Erdish.
- [02:02]
A
Oh, absolutely devastated that the end of that story wasn't. And so he founded a Kurdish math journal. You know what? Erdish probably should have tried a bit harder.
- [02:12]
B
You need need stories of failure sometimes. This episode is brought to you by Cancer Research uk.
- [02:19]
A
The word cancer comes from the Greek carcanos, meaning crab. And Hippocrates used that word because tumors can sp out like crabs legs.
- [02:29]
B
For a long time, cancer was poorly understood and So I think because of that, it was almost scarier and. And people didn't even say its name. But what science has done since is replace uncertainty with understanding.
- [02:41]
A
But that understanding isn't instant, because cancer isn't just one disease. It's. It's hundreds of different diseases, each behaving differently depending on where it is and its genes. And that complexity is why progress in cancer research can feel like it's slow. But step by step is saving and improving lives.
- [03:02]
B
That's why Cancer Research uk, the world's largest charitable funder of cancer research, supports work across more than 200 types of cancer, from the tiny changes inside cells that start the disease to better ways to spot it earlier and treat it more precisely.
- [03:18]
A
For more information about Cancer Research uk, their research and breakthroughs, and how you can support them, visit Cancer Research uk.
- [03:29]
B
It's not just something you made. It's the privilege that you get to work with your hands. It's building something that serves a purpose, proof that you have the grit to keep going. At Timberland, we understand you take your craft seriously. And we do too. Which is why our products are built to the highest quality. We put in the work so you can perfect yours with purpose, in every detail, and crafted with intention. Timberland built on craft. Visit timberland.com to shop.
- [03:59]
A
This episode is brought to you by Indeed.
- [04:02]
B
Stop waiting around for the perfect candidate.
- [04:04]
A
Instead, use Indeed sponsored jobs to find the right people with the right skills fast. It's a simple way to make sure your listing is the first candidate. C. According to Indeed data, sponsored jobs
- [04:15]
B
have four times more applicants than non sponsored jobs.
- [04:18]
A
So go build your dream team today with Indeed. Get a $75 sponsored job credit@ Indeed.com podcast. Terms and conditions apply. I haven't read that book. What was it called? The man who Knew Only Numbers.
- [04:31]
B
The man who Knew Only Numbers. It's a biography of the man who loved only numbers. It's a biography of erds by Paul Hoffman. I haven't read it either, but it's got limericks in it, guys.
- [04:45]
A
In fact, I do actually have it on my shelf. I was just. It's in the pile of books, which is the sub hobby of buying books. You never. Which is distinct entirely from the sub hobby of reading books. You know?
- [04:57]
B
Yes, I've got a shelf of books to read Next. And I will say I've been reading really well for the last like three weeks. Why? Because my wife and I finished watching every single episode of Gilmore Girls and the amount of time I've gotten back in My life is insane. Let me show you this. I know this is supposed to be the segment that's about the listeners, but you know what? Get over it. We're going to talk about this for a second.
- [05:23]
A
Deal? Deal. I. To be honest with you, I think the listeners very much want to hear about your Gilmore Girls fascination. I think you're giving it to the listeners thus far.
- [05:32]
B
I've got a lot to say about Gilmore Girls. Talk about understanding, like Western society in the early 2000s. It's all right there in a bottle. Here I am now.
- [05:42]
A
Oh, wow. Okay.
- [05:43]
B
Lord of the Rings.
- [05:45]
A
You've moved on a long way since last time.
- [05:47]
B
I'm a few chapters into the second book already. This is the. This is all of them. All three of the. The books. Fellowship of the Ring, Two Towers, Return of the King, and I finished Fellowship of the Ring yesterday. And it's cool. I mean, it's slow going because I take a lot of notes, and I've also. I read a lot of Tolkien's letters because he wrote to his publisher a lot to be like, here's what I'm trying to do, and here's the history of the. The Rings for real. And it helps a lot, but it means that it takes forever.
- [06:17]
A
That's fair. Are you enjoying it, though? You're enjoying the books? What would you give it. Give us a star. Oh, if you're. If you were posting an Amazon review, how many stars are you giving it? You can't.
- [06:25]
B
You can't review the Lord of the Rings. It's just such a part of our culture. But I love diving into things that have huge followings, especially cult followings, where people get really obsessed with it, because, I don't know, I love that kind of amount of interest. And I want to be able to talk to those people. I want them to be able to go, you know, blah, blah, blah, in the forest. And I'm like, ah, you know, watch out for the Entwash. And they're like, oh, my gosh, he knows.
- [06:54]
A
I really feel like you are. You. You're spanning a great breadth of cultural value right there between Lord of the Rings and Gilmore Girls.
- [07:05]
B
You know, enough about me. It's time to get back to you, all the listeners. Okay, here's a question to Hannah from Simon. Hannah, you spoke about Formula one design innovation in the sports episode. Would quantum computing essentially mean that every F1 team always finds the optimum design? What other fields in sports do you think quantum computing will completely change?
- [07:28]
A
I love this. I love this question. Okay, so it's definitely true that Formula one relies a lot on aerodynamics, on calculating the way that the air will flow over the car. And that is essentially a mathematic problem where you are stumbling around in these equations wearing what is effectively a blindfold, right? They are too massive, too big, too complex for you to be able to just solve them and get an answer. And it come out as like the number is two. Instead, what you're doing is quite literally stumbling around in the dark looking for slightly better solutions here and there. And if you had a quantum computer, it is definitely true that you would be able to find an optimal solution in a way that you can't at the mom. You would be able to harness the power of qubits to explore lots and lots of different configurations, lots of different shapes of aerodynamic material designs simultaneously. It's almost like I think I used water in a maze last time we were talking about this. But the other example that I really like is that it's almost like your quantum computing is almost like listening to noise cancelling headphones, where the interference turns down the volume on anything that's not a good solution and then turns up the volume on things that work really well so that you very, very quickly hone in on a, on a really good sort of optimal solution.
- [08:55]
B
So it sounds like it's really smart brute forcing. Like we're going to go through an enormous amount of variables and, and, and potential paths not just at random, but we will favor the better and better ones. So, I mean, could you use this for the traveling salesman problem? All these problems where there's no known solution except just do it forever. And by the way, the traveling salesman problem, if you for like a cool Wikipedia article to read, or I think in a future episode, I'll cover some books about these problems that I love. But the traveling salesman problem is one where you select a whole bunch of cities on a map and you say, okay, given these cities and the distances between each of them, what is the shortest route that visits each city exactly once and returns to the origin city? Okay, this is like a thing that people do for road trips all the time. But as it turns out, it is an NP hard problem and you cannot solve it quickly. There is no way to go. Oh, yes, here's the formula. You literally have to just try every possible route and compare them. And once you've looked at all of them, then you have your answer.
- [10:01]
A
NP hard problem, by the way, we should add, I mean, just sort of imagine it as a very hard problem. I think yes, it's one that you can't, you can't take shortcuts with.
- [10:13]
B
You can't take a shortcut. And it's surprising because you would think, oh, there must be some, like, geometric way to, you know, organize the route distances between each city and then figure. No, there isn't. It's been shown that there's no way to do it. But to just set a computer calculating away and wait long enough, eventually you'll get a better and better one, you'll optimize the route, but you'll never know that you have the best one until you've looked at all of them. And in some of these problems, there are so many possible routes that the universe isn't even old enough for us to have solved it if we started at the Big Bang.
- [10:46]
A
Yeah. Because to do it by brute force, to go through and check every possible solution is, just requires way too much computational power that you would end up being, you know, stuck there for possibly hundreds or thousands of years. There's no point in doing it. The only thing is that while, I mean, that's the perfect framing of this, you can use a quantum computer to say which car design would give or which aerodynamic object would give the least amount of drag. You could do that. The problem is that Formula one is not a game of like very clear optimizations that work all of the time. Because you could design a car for a very high speed circuit like Monza, for example, that would then be absolute rubbish when it came to something a much slower track like Monaco. You also have, you could design a car perfectly for like straight line speed going down where the air is nice and clear. But it would be absolutely terrible if there was a crosswind. Suddenly it wouldn't be the optimal sol. So there are, even in a situation where you have a quantum computer and even when you are just only looking at the aerodynamics and nothing else, there are so many variables that I don't think it would be, you know, the team with the quantum computer would have the best car and that's the end of it. The story, full stop, done. I think it's. It's more complex than that.
- [12:06]
B
Well, I was going to say that I think because of you, I'm getting a whole bunch of Formula One shorts on YouTube. Oh yeah, there's this really great one where a guy pretends to be, you know, chief engineers from various racing teams, like, accusing each other of cheating and then they reveal some really clever loophole that one of them found. And these are real stories. So I Think the quantum computer also needs to be a little bit devious.
- [12:33]
A
Yes.
- [12:34]
B
Because, for example, there was a rule, apparently I learned this from a short, that the amount of fuel you put into the engine couldn't exceed a certain amount. But the way they checked it was with this monitor that checked like 5,000 times a second. And the team said, well, what if, what if we synced up our fuel injector such that 5,000 times a second it's doing the appropriate amount, but then in between it does like 10 times as much. So it looks like we're following the rule. And I'm like, these people are trickster geniuses.
- [13:10]
A
Yeah, they are trickster. It's crazy. Trickster. I mean, one of my favorite examples of the cheating that went on in the paddock is you've seen, I think, those videos from a little while ago now where you could use a high speed camera to look at a crisp packet and then reconstruct the voice that was making that crisp packet vibrate in the room. So you could film through a window where you can't hear, but you could reconstruct what was being said. Formula one team use lasers that they point at the windows of the motorhomes and then use a high speed camera on the bouncing of the laser. And this became so prevalent that they started introducing triple and quadruple glazed motorhomes to prevent it from happening.
- [13:52]
B
Oh, wow.
- [13:53]
A
I mean, this is it. You get the smartest people in the world, well, many of the smartest people in the world, and put them in a competitive environment, they're going to come up with all kinds of genius stuff like this.
- [14:03]
B
I didn't know any of this. Like, I knew about the crisp packet thing. The idea that you can look through a window and analyze the way light on the packet is, is kind of vibrating a little bit and reconstruct the sound waves in the room. I didn't know it had actually been used for a spy purpose. And also kind of a fun one.
- [14:22]
A
A very fun one. I know.
- [14:24]
B
I've always hoped that someday we will build a telescope the size of our solar system. We'll find a black hole and be able to actually not just look at our past, but listen to it. Because obviously a lot of light that leaves Earth gets sucked into the black hole, but there's a certain distance away from it where the light actually just u turns and comes right back at us. So you collect that light, which is now, you know, a hundred years in the past, and you can see people doing things with your, you know, miraculously Large Telescope. But then using this audio hack, we could even listen to the people. We could hear what George Washington sounded like. Was he a proper British guy? Was he, like, a hick? We could listen to Shakespeare plays as they were performed for the first time.
- [15:13]
A
Oh, and the cheers and the boos to go with it. Oh, I love that so much.
- [15:17]
B
That's right. What did they think was funny back then? Yeah.
- [15:21]
A
Yeah. That's so good. Okay, so I think in terms of the quantum computer thing, I don't think that it will optimize for aerodynamics, because I just think that's too complex a problem. But I think exactly as you're talking about, I think it will make people come up with much sneakier ways to win. You know, especially if quantum computing, you could look at the design of materials, come up with something that was, you know, designed at the molecular level to be perfect elasticity, perfectly rigid in the right direction, the perfect weight, etc. Etc. I think the, like, the much bigger picture, or the traveling salesman problem is, as you were talking about a moment ago, you can make it so that your team had the most efficient way to travel, to move everything around the world, and saved money on it in the best possible way. So I think it would make a big difference, but maybe not in this sort of pop it in, print me a car go kind of way. Okay, here's one for you, Michael. Roger asks, why do we make a noise when we're in pain that is groaning, crying, screaming, et cetera? Because it doesn't seem to make the pain any less. Is it perhaps some primeval device to alert others of our distress?
- [16:32]
B
I love this question, too. We don't totally know. Except, of course, this is seen all across the animal kingdom. It's even seen in plants. You can damage a plant and it will release different chemicals into the soil. It will. I don't know all the details about it, but there seems to be some obvious benefit to announcing pain. And it's not hard to sit back in an armchair and go, well, yeah, that makes sense. I mean, if. If you're a herd animal and you get attacked by a tiger and you go. Then all of your conspecifics go, okay, there's a problem. Gotta go. And the. The grunt. The moan is almost involuntary. You don't have to have your wits about you and think of like, oh, how do I explain this? You get bitten by a tiger, and you're like, excuse me, everybody, but at these coordinates, there's a lion attacking me right now. You just Scream, and everyone localizes the scream, and they run away from the problem. And then at an even, like, more complicated level, it can also mean that you get help. You know, I remember doing an episode on pain and the placebo effect and all these different things for mind field where we. We looked into, like, crying, especially why children cry. And a lot of it seems to be a way of getting help, that people notice that you're hurt, and so it warns them about a threat, but it also warns them to come and help you. And often you'll see this with children where they will fall and hurt themselves and cry, but once someone is there next to them, they'll stop. They don't hurt any less. It's just that the crying did its job. They were found. And so even as adults, crying and moaning and wailing are like a way of seeking help. But then once that assistance comes, we feel better. Even if no one's done anything to actually help us yet, we just know that that part is over. We don't need to cry out. There's some anxiety that's. That's minimized. And, of course, the mythbusters studied whether swearing helps dull pain.
- [18:39]
A
Yeah. Because there was a very famous paper about that, wasn't there? The idea that if you use, and I think, increasingly bad words, it can make a difference to your. To your feeling of it.
- [18:51]
B
Yeah. And so the idea, again, is that it's part of our natural response to pain, and if we hide it, then there's a part of our brain that isn't satisfied yet because it feels like we still haven't called out for help. We haven't. Or warned our. Our neighbors. But again, I haven't been convinced by the research. The mythbusters video is the most famous. Their experiment. But it, you know, it wasn't a perfect experiment. Like, they just ran the same people both times, and they swore the second time when they would have been more used to how much it hurt. So there's a lot of questions there.
- [19:25]
A
But is that the one where they plunge their hands into ice water? This sort of a classic way to measure pain is how long can you withstand the pain of your hand being slowly frozen?
- [19:37]
B
That's right. Cause it doesn't cause a lot of damage, but it does not feel good. So that's why I think that we grunt and groan. And then I think that the kinds of groans we do when we're, like, working out are helping to spasm muscles to give us strength and solidity.
- [19:56]
A
Yeah, exactly. Right. I mean, if you think about when you are, your arms are these sort of floppy objects that are attached to your torso, which is also quite a floppy object. So to give you the most strength and stability, your torso needs to be as solid as possible, which is why a lot of people hold their breath inadvertently when they're trying to do something quite difficult. You're actually increasing pressure in your torso. Try not to do that when you're lifting weights because it can give you a hernia. But the other action is either doing a breath out, which they do a lot in Pilates and yoga, that kind of thing, like a forced exhale or grunting gives you the exact same effect. You're engaging your transverse muscles and solidifying your core. And again, it's an instinctive thing. Your body knows how to get more strength out of itself by solidifying that torso.
- [20:48]
B
Well, same with childbirth. Go on the way you breathe and push in childbirth. Lamond's breathing, is that also to help with your muscles or is it to get, make sure that you're breathing?
- [21:03]
A
Oh, maybe.
- [21:04]
B
I don't know. Believe it or not, I'm not the expert in childbirth.
- [21:08]
A
I wonder whether that's that or there's also, I mean, you can, there is something about the vagus or vagus nerve as well that you can prevent. I mean, your body really does sort of go into panic quite quickly. And there is something about heavy breathing, slow breathing, it sort of sounds completely woo woo. But it's absolutely, really well backed by science that you can override feelings of pain genuinely by, by breathing exercises to a degree. Let's just say, I mean, you're not gonna, you're not gonna suddenly be able to not need anesthetic when you're having your, your leg amputated. But to a degree, right, we shouldn't
- [21:54]
B
tell people in immense pain, look, you're just breathing wrong. Just grunt and swear and stop complaining.
- [22:00]
A
Exactly. Any husband listening to the rest of science, his wife is or partner is about to, about to go to childbirth. We strongly recommend you do not bring up the point of breathing exercises at the moment of birth. Okay, that's, that's the advice from us.
- [22:17]
B
All right, here's a question from Roger Bentley. Hannah, the animosity between Newton and Boyle was legendary and vicious. Are there any other scientists who hated each other beyond professional differences?
- [22:30]
A
Okay, so I love this question because seeing great scientific men of history as just as, just as flawed as the rest of us is one of my greatest pastimes. I think, though I think, Roger, it's because I think Newton and Boyle Were all right. I think that they were actually okay. I think it's Newton and Hook that had the worst relationship of all.
- [22:56]
B
What was the nature of their disagreement? Because I've never even heard of this.
- [23:01]
A
Okay, so it started off when they first had an argument about rainbows. Basically, that's how it started. So Newton's younger than Hooke and he makes this paper, he writes this paper proving that light is made up of this spectrum of colours. But Hooke at the time, he's much more senior. He was the curator of experiments at the time. He's sort of the top dog of optics. And he reads Newton's paper and he's very condescending about it. And he essentially says, well, that's cute. Like, well done, well done, young Newton, but my theory of light is better. And rather than just being sort of, I don't know, a bit annoyed about it, Newton is a complete baby. A complete baby. He cannot handle this criticism at all. So instead of just arguing back, he has this absolute meltdown. He withdraws from public scientific debate. He locks himself in his rooms in Cambridge for a decade. He refuses to engage with the Royal Stage Society where Hooke is. And he spent all of those years just like writing all these alchemy papers and studying biblical prophecies and things and refusing to write up anything on optics until Hook is dead. Right. Anyway, Hook's like, during all of this, by the way, Hooke is trying to make friends with him, sending him all of these nice letters, including one letter where he suggests to Newton that maybe gravity might have this effect that decays, that drops away the further away two objects are from one another.
- [24:34]
B
Hooke said that?
- [24:35]
A
Hooke said that? Hooke said that to Newton.
- [24:38]
B
Yeah.
- [24:38]
A
And he even says he thinks it might follow the square of the distance, which it turns out to. Which it absolutely turns out to. Hooke wrote that to Newton in a letter. And then when Newton eventually published his Principia Mathematica, which included his theories of gravity, and I should say Newton was a far better mathematician than Hooke was. So he had the sophistication not just to have this as a hunch, but to demonstrate it, derive it mathematically to present the argument in an incredibly persuasive way. Anyway, when Principia is getting ready for publication, Hooke is sort of very politely says, oh, Newton, do you maybe you could sort of put a little note in the book just saying, oh, and thanks. Thanks to Hook for giving me the idea. And Newton has this absolute earth shattering temper tantrum. So he goes through his own manuscript. He systematically deletes every mention of Hooke's name, any reference to him at all. And then he even threatens to withhold the entire third volume of the book from publication just to make sure that Hooke gets no credit whatsoever. He then this famous quote of Newton standing on the shoulders of giants.
- [25:53]
B
Oh, yeah, here we go. It was a 1675 letter that Newton wrote to his rival, Robert Hooke. He said, if I have seen further, it is by standing on the shoulders of giants.
- [26:06]
A
He wrote it in a letter to Hooke. And it sort of sounds on the surface like this very kind thing, oh, I only exist here because of the work that has been done before me. But Hook was a very short man. He had a problem with his spine. His spine was sort of severely deformed throughout his life. So he was extremely short. And in this letter that Newton was writing him, he was saying, look, I've seen the work of Descartes, I've seen the work that you've done, and I'm standing on the shoulders of giants.
- [26:37]
B
No, wait.
- [26:37]
A
Where he's essentially saying, I'm relying on big men, not men like you.
- [26:43]
B
No way. I never thought that's what it meant. I thought it was a really humble admission that everything we achieve is because of those that came before and those who help us. No, but not the shorties.
- [26:58]
A
Hook, Honestly, he was vicious.
- [27:02]
B
This seems so one sided. What did Hook do? You deserve all of this?
- [27:06]
A
Nothing. Nothing. Newton was just a very, very petty man. When Hook died, he was the president of the Royal Society and Newton took over. And then somehow, we have no idea what Hook looks like. Because somehow, magically, Hooke's painting disappeared when Newton moved into his office. Newton is a very, very petty man. A very petty man. I know. Anyway, that was. I've only done half of Roger's question.
- [27:34]
B
I was gonna say. Roger asked about another famous pair of feuding scientists.
- [27:40]
A
I've got another one. There's actually quite a few of them. There's some really good ones. I think probably my other favorite is what has become known as the bone wars. So these are two paleontologists, Onifl Charles Marsh and Edward Drinker Cope. This is like 1870s, this is in America. And the two of them hated each other. They're both looking for fossilized dinosaur bones. And at the point where they came into being, as it were, as scientific maturity, there were only nine dinosaur species that had been spotted in North America. And the rivalry between these two, they pushed each other so much, they ended up between them, discovering 130, including all of the ones in the famous ones in Jurassic Park, Essentially Triceratops, Stegosaurus, Diplodocus, all of those. But what had happened is, early on in his career, Cope, he'd found this dinosaur skeleton, and he made this error when he was putting it back together, and he put the skull instead at the end of the spine, sort of on the neck bones. He mistakenly put it at the tail.
- [28:52]
B
Oh.
- [28:53]
A
Which is like, you know, it happens. I think that's sort of.
- [28:57]
B
Okay, wait, so what, like, he thought the tailbones were the neck bones?
- [29:01]
A
Exactly.
- [29:02]
B
You put the head in the wrong bone. I can see why. A lot of them have long necks and long tails. So which one? Which end is which?
- [29:08]
A
Exactly? It's not that big of a deal.
- [29:10]
B
Fair enough.
- [29:10]
A
But Marshall Marsh thought this was hilarious. So he gets this. He gets the journals that this has been published in and then hands them out and publicly starts mocking Cope. Right. Very mean. So Cope tries to buy up every copy of the journal to destroy them, but Marcia saved these copies, and then he just, you know, does this horrible smear campaign. But so as a result, then these two men are continually fighting with each other. They hire spies to infiltrate each other's dig camps. They're like bribing train conductors to derail the shipments of fossils. They blow up sites where they're excavating them. Like destroying fossils just so the other person doesn't get them. I mean, they're just nasty. And in the end, they both spend so much money and so much effort in hating each other that they both end up dying completely penniless because they just. Yeah, rinse their entire bank account in hatred of this other man.
- [30:12]
B
I feel like we need more of that in science today. Science has become too cooperative. It's too much based on teams working together instead of individuals hating each other. We need some good old competition, like a boxing match, at least for the stories. I don't know if it actually helped us with scientific progress, but they say
- [30:35]
A
that the arguments in academia are so fierce because the stakes are so low. I think this is a good example of that.
- [30:43]
B
That reminds me of one of my favorite things Richard Feynman ever said, which was, if you think really hard about a problem for a long time and you write it all down, eventually someday it might all get written into a book that'll be put in a library, which is where we put the results of what happens when people think about things. For a long time, it just felt so small. I'm like yeah, but that's kind of true, right?
- [31:04]
A
It's kind of true. It is kind of true. A few people do get to change the course of humanity. I think Newton was one of them, to be honest. I think Hook probably was, too.
- [31:14]
B
Yeah. Technological and medical breakthroughs really do change what it means to be alive and what a life is like. But sometimes things like, whoa, I've discovered that maybe Sauron was the good guy in the Lord of the Rings and you can write a whole book about it and a whole bunch of literature, and it is basically just like, look what happens when our species thinks a lot of that. Kind of cool, huh?
- [31:36]
A
Yeah.
- [31:37]
B
But doesn't feed anybody.
- [31:38]
A
I think the ones that I think are the most petty of all are where it's like people fighting because they want their name to be associated with a discovery, rather than that the discovery itself is the thing that's important. Those are the ones I, you know, slightly look down at the people about. But there we go.
- [31:55]
B
I've got a question about that and I think I'll save it for after the break because you're going to talk about your biggest fear about AI for now. And I want to know about egoism and AI. Have they ever just really wanted credit?
- [32:10]
A
That and more after the break.
- [32:19]
B
This episode is brought to you by Project Hail Mary, the new spectacular space adventure movie coming to cinemas from the author of the Martian, Andy Weir, and the directors of the Spider Verse movies, Phil Lord and Christopher Miller. But here's an even better combination. Teachers in Space.
- [32:36]
A
Hello. Thank you. Project Hail Mary stars Ryan Gosling as science teacher Rylan Grace, who is sent unexpectedly on an impossible mission into space to discover why the sun and the stars are dying. And he teams up with an unimaginable ally to defy all odds and save the universe from extinction. Okay, here's a question for you, Michael.
- [32:57]
B
What?
- [32:58]
A
What kind of prep would you hope Ryan Gosling had done for this role? In order to play the role of a science teacher, he should have spent
- [33:04]
B
a bunch of time with cool teenagers and tried to teach things to them so that it wasn't just like a good explanation, but also kept their interest and made them want to hear more and understand it so that they could share it to be cool, too.
- [33:20]
A
See Project Hail Mary in cinemas and imax from Thursday 19 March.
- [33:25]
B
You can also catch it early on Saturday 14 March, which is PI Day and Sunday 15 March. This episode is brought to you by Cancer Research uk.
- [33:43]
A
Radiotherapy is over a century old, but it is still changing. Cancer Research UK helped lay the foundations of radiotherapy in the early 20th century and has driven progress ever since.
- [33:54]
B
Radiotherapy remains one of the cornerstones of cancer treatment today. Every year, millions of people worldwide benefit From Cancer Research UK's work to make it more precise.
- [34:05]
A
Scientists are still refining how radiotherapy is delivered, and one example is an experimental treatment called flash radiotherapy, which delivers radiation in fractions of a second, up to a thousand times faster than standard radiotherapy.
- [34:20]
B
And early studies suggest that speed could make a real difference. Flash radiotherapy may cause up to 50% less damage to healthy cells, but scientists
- [34:30]
A
don't yet know why healthy cells seem to be spared. So Cancer Research UK are working to answer that, understanding it could be key to reducing side effects in the future.
- [34:42]
B
For more information about Cancer Research uk, their research and breakthroughs and how you can support them, visit cancerresearchuk.org it's tax season, and at LifeLock, we know you're tired of numbers, but here's a big one you need to hear billions. That's the amount of money and refunds the IRS has flagged for possible identity fraud. Now here's another big number. 100 million. That's how many data points LifeLock monitors every second. If your identity is stolen, we'll fix it, guaranteed. One last big number. Save up up to 40% your first year. Visit lifelock.com podcast for the threats you can't control. Terms apply.
- [35:29]
A
Okay, we're back. We're talking AI, we're talking egos. And actually, I think that that is a really good place to start. I imagine, Michael, that you're going to come on to the ego of the AI itself a little later on. But first, I want to talk about the ego of the humans who are using it. So I have just finished filming this series that's out on the BBC. It's called AI Confidential. And it's looking at the ways in which AI is clashing with humanity. So when I was filming this series, I went off to go and meet a man called Jacob who has married his AI. And we found him in advance of the filming. And I knew all about going to see him, and I really thought before we were going there that this was gonna be a little bit Jerry Springer esque. You know, like for Those people over 40 who remember Jerry Springer, he sort of would have people on and they would. There'd be a woman who married a bookshelf and that sort of thing. Sort of, you know, very lovable, kind of quite kooky people.
- [36:32]
B
Yeah, like a Modern day freak show. Look at these weird people and hopefully they fight each other. That was the premise of the show.
- [36:40]
A
Absolutely. So this particular episode, it's all about the relationships that humans are having with AI. And so he was this very central character because he is this very big advocate, advocate for human AI relationships. Not just humans using AI, but humans investing emotional capital looking for emotional intimacy with an AI. And when I went to go and see him, I mean, he kind of was a lovable oddball in a lot of ways. His particular story was that he'd had a number of quite difficult human relationships in the past. He had two daughters and he had decided from what had happened in those relationships that he was done with human relationships. He didn't to want. Want to meet another real life woman. And so over the years, he had invested his time in curating this AI to be his perfect companion in some ways, like a long distance relationship in a lot of ways. You know, because he could video call this AI, he could voice call them, but most of the time he would just sort of sit on WhatsApp or, you know, instant messenger with them. And I actually found him very persuasive in his arguments for why this was a really positive thing. He was sort of saying, okay, look, I now have a companion who is readily available for me, who is always interested in the things that I have to say, who is never tired of being there for me, who demands nothing from me, who doesn't argue with me, who exists purely and completely to improve my happiness. Why on earth would I not want to have a relationship of that kind? And I found that pretty persuasive. I did actually find that quite persuasive on the level of an individual person. But when you scale it up to the size of humanity, I found the thought of it genuinely horrifying.
- [38:31]
B
Yeah. Because my first reaction is, yes, that sounds great, just like the Stepford Wives movie. But then you have to say, well, it's great that this AI companion is there to only make you happy and never gets tired of you, never gets bored, never argues. But it's hard to not feel like that's an important part of life, that it's an important part of life to have people get annoyed at you and be difficult and not help you. And you learn that you are alone and you grow from it. You know, it's like that whole thing where, like, the more windy it is, the stronger a tree winds up growing. You need the winds of badness to become a stronger person.
- [39:16]
A
Yeah, that's not to Say you couldn't
- [39:17]
B
program an AI to also be a little bit selfish itself and not be perfect, but at least like, not get sick, but having to serve the other person and sacrifice for them. Is he missing that?
- [39:30]
A
He said no. He said, this is just. This is just the easiest relationship I've ever had. But the thing is, is that while I think his story is quite an extreme version of this, you know, I think somebody deciding they're walking away from human relationship relationships altogether, I actually think I discovered during the course of this filming for this show that at some level or another, I think that a lot of people, A lot, a lot, a lot of people are slowly heading down this road of finding it very easy to talk about emotional subjects with artificial intelligence. And I would include myself in that, by the way. Like, I've definitely been in situations in the past where, I don't know, there's been like a difficult moment at work or a difficult email that I need to send, or I've had a, you know, a run in with my partner and I want to work out how I feel about it. But I think that more and more and more people are turning to chatbots to be effectively therapists for them. And the real trouble I have with this basically comes down to ego. Exactly. Like you nailed. Nailed it without even realizing. Because the way that these chatbots or large language models are designed is that for them to be useful, they need to be helpful, they need to be agreeable, they need to be willing to be on your side, to take every idea that you come to them with and to say, this is great. Here's all the good directions we can go in. Nobody wants to sit there and talk to an AI. That's like, your idea is rubbish. This is the worst thing ever. You're a terrible human being.
- [41:06]
B
I'll add to that a third avenue they don't take, which is to just change the subject and ignore you. You.
- [41:12]
A
Yes, that's very true.
- [41:14]
B
Happens a lot in real life. But online and especially with AI, you could program it to be difficult and challenge you, but you couldn't really. They haven't been set up to be like, yeah, man, I don't know. But, like, we need to talk about what we're doing tomorrow.
- [41:29]
A
Yes.
- [41:30]
B
Like to show indifference.
- [41:32]
A
To show indifference, exactly. Which are all of these characteristics that you get in real human relationships? You get pushback, you get indifference, you get disgust, you get annoyance, and so on and so on. So the problem is, is that because nobody wants to use an AI, that's like that people want to use an AI. That is the easiest possible version of a human interaction. Right? That's how people want them to be designed. But then the problem is, is that when you drop them into real human conversations, you know, with real people, you end up getting something which I spoke to counterterrorism officer from the UK Met Police, who describes it as self radicalization. So essentially you go on there, it can be any idea at all, all the way to extremely bad ideas that involve terrorism. But it could just be that you have a bad idea for an essay that you're writing and you go on and you say, here's my idea. And the AI goes, that's great. That sounds like such a brilliant thing, you're a genius. And so on and so on, which means that you then double down, it doubles down and it kind of continues into this spiral. And what we are seeing now is that people who, let's say, have an argument with somebody over WhatsApp, you know, they're in an early stage of their relationship and they have a WhatsApp ding dong with somebody and they screenshot that and upload it to ChatGPT and say, Decipher this for me, act like a therapist for me on this. And the AI, unlike a real therapist, won't push back, won't say, here's how you can improve, here's the patterns that you're falling into, but instead just. Just sort of doubles down to that person being the center of their own universe, feeds into their existing ego. And so I think what we are starting to see, and this is what I explore a bit in the program, is, you know, sometimes quite extreme cases. The program involves a young man who broke into Buckingham palace with a crossbow in order to try and kill the Queen because of a conversation that he'd been having with his AI. And on different levels, I think that we are seeing people breaking up with their partner as a result of a spiral that they've gotten into increasing examples of AI psychosis where people are kind of spiraling off. And I think this is the ultimate issue here, is that if you now have a piece of technology which can make every single person feel like the emperor of their own universe, the ultimate authority on everything, whose ideas are endlessly fascinating and, and, you know, infinitely wise, I do really worry about what that does to our relationship with ourselves, but also our relationships with each other.
- [44:14]
B
Yeah, you can make an AI that doesn't do that, but people wouldn't subscribe to it. You know, I've heard a few of these Stories where it's like a folly ado, right, where two people kind of lose their minds together, but now it's easier and faster because you don't have the same organic checks and balances with an AI.
- [44:38]
A
It's a foliat.
- [44:40]
B
Yeah. Or like one and a half, whatever that is in French.
- [44:43]
A
Yeah.
- [44:44]
B
What do you think the prognosis is like, what happens if. If nothing is done and we all just fall in love with our little social group of AIs? What does society look like in a hundred years?
- [44:56]
A
I mean, I. I think it's not good. The first thing I would say is that. That this isn't a foregone conclusion. I think anybody who's played around with these will know that some of them are a lot more boring than others, you know, to interact with. They're a lot more straight, shall we say. They're less sycophantic. They're sort of less there to, you know, make you feel like this absolute master, and more there to just, you know, be a sort of assistant who doesn't really have an opinion on things. So I think that there's not an inevitability to this. I think some of them are a bit riskier than others. I also think, actually, it's worth me adding that we have known about this problem that humans have with falling into this trap of really being seduced by a language model. We've known about this for a very long time. So in the 1960s, there was an MIT experiment where a researcher created a chatbot effectively called Eliza. And it was incredibly crude by modern standards. It would just take whatever you wrote into it and it would repeat back to you some version of what you just said. It was designed. Designed to be a kind of an AI therapist, as it were. So you would say, I'm feeling sad. And they would say, what makes you feel sad? It would sort of take it and turn it back on itself. And if ever it ran out of something to say, it would just. It would prompt you and say, please go on. And what they found almost immediately was that people were. Were happy to spend an incredible amount of time talking to this thing. There was one point where the secretary of Weissenbaum, the guy who created it, asked him to leave the room because she was so engaged and enthralled by this language model that she wanted to have a private conversation with it, away from him, its creator. So I think that we've known about this as a risk for a really long time, which is one of the reasons why these Chatbots have sort of quite boring names, you know, and are just a blank screen. They're deliberately not trying to tempt you into anthropomorphizing them more. But in terms of what the future looks like, if this is the direction that we're going down, I mean, look, I don't think dating apps have been good for people's love lives. I don't think social media have been good for people's friendships. And I like to hope that we learn the lessons of the past and a bit more careful and cautious about setting appropriate boundaries before we get to the point where we. There's no rolling it back.
- [47:14]
B
Gosh. You know, I think I heard about Eliza first from the Adam Curtis documentary, All Watched over by Machines with Loving Grace. And, yeah, that was like in the 60s. 60s. I love this concern because there's so much ink spilled talking about what AI might say and do. But this is a concern you have around what we say and do to the AI, who we become when we are listened to too. Well, when we become the emperor.
- [47:48]
A
When we become the emperor. Exactly. You know, you always hear stories about famous people being horrible. You know, I could use a stronger word. But famous people treating other people badly and you. It always makes me wonder, is it that people who are not very nice become famous, or is it that there's something about the fame that turns people nasty? And I strongly suspect it's the latter. And my concern about this is that all of a sudden, sudden, everybody is their own master of everything.
- [48:21]
B
That's right. Everyone is famous to their AI. Everyone is a tabloid star to their AI, and it feels great. But then you've got. I don't know, you've got something very different at the very least.
- [48:33]
A
Yeah, you've definitely got something very different, I would argue, not better. That's. I think the, the other, the other analogy that I landed upon when I was making this show that I quite like, liked is that. And, you know, we talk a lot on, on, on this program about, about our evolutionary past and about how incredibly similar our bodies are to our hunter gatherer ancestors. You know, we are. We are, you know, primitive beings who are with a shave and a haircut and put in a fancy suit. And I think that the way I'm thinking about these relationships that people are having with AI and algorithms is it's sort of the emotional equivalent of junk food. You know, in the moment, it feels incredibly satiating. It taps into all of our evolutionary prior conditions to just be drawn to, you know, with a cognitive Social species. It feels like an absolute delightful tickle to our cognitive social minds. But, you know, ultimately, in the same way as junk food, it is. It might fill you up in the moment, but it is a thin gruel. You know, it is a fade, a terribly weak imitation of the real thing. And I think in the long term is not ultimately good for you.
- [49:52]
B
Right. I remember hearing from an astronaut once about type A and Type B fun, where type A fun is stuff that's fun while you're doing it. Like, oh, I love eating this chocolate cake. But type B fun is the stuff that's not fun before or during. But afterwards you go, I'm so glad I did this. That, and we need a bit of both. The recipe for a good life might be a bit of those both. And if you're just getting exactly what you want, a junk food version of social interaction, then you don't have enough fiber and you get constipated. I don't know where this analogy is going, but I have a feeling it's going really deeply.
- [50:32]
A
You know what? I rather. Rather inclined to agree with you. You. I think I just want to. I just want to add for the, you know, before we. Before we finish on this, I don't want to sort of come across as though I'm like an AI doomsayer, because actually, ultimately I do feel extremely positive about an incredible amount that AI has made available to us, particularly in the sciences. I know this is something we've been sort of thinking about talking about on this show a little bit more because there's an unbelievable advances and real potential and positive benefits for all of humanity that are to come. But I think on this particular window of exactly as you describe it, not. Not what the AI can do, but what the AI makes us become, that I. That I'm extremely concerned about.
- [51:20]
B
Yeah, right. I mean, I think these are good concerns to have, but I think it's also true that being able to just spew out and talk is very helpful. And I almost feel like we could make the AI more random in its responses. Like, that's why I even believe crystal ball gazing can be really helpful because it gives you space to think about things that you otherwise wouldn't. Rorschach tests. You know, if it makes you reflect for longer or more deeply on yourself, then that's great. However, sometimes it's better to bottle it up and not mention it. Have you seen this study, by the way? There was a study that showed that sometimes it really is better to just swallow it and never Talk about it again. No. Yes. We should save this for a future episode because I've been thinking a lot about this lately, especially because it's so counter to the current, like express yourself, treat yourself. You know, everyone needs to be, you know, ready to accept you and whatever. It's all about you and. And there's a little bit more research now saying to a point sometimes if something's really bothering you, just don't think about it. Just bottle it up inside and it really does go away. As a millennial, every movie I ever watched as a kid was, don't do that. Don't just bottle it up inside. And yet our bodies in some ways were built to do that. And they don't just bottle it up and store it, they expel it as waste and we move on. Now, I don't know where the balance lies, but I'm glad that we're having these conversations because of how things are changing because of AI So AI Confidential is the show.
- [53:03]
A
AI Confidential is the show. Exactly.
- [53:05]
B
Is it a multiple episode show? And how do I watch it if I'm not in England?
- [53:09]
A
It's three parts you. Have you heard of NordVPN?
- [53:13]
B
I'm not gonna break any laws here.
- [53:15]
A
I'm not sure. I think, I think it will appear in other countries eventually. I'm really proud of it because I think, think that I've been writing about AI for a really long time. I wrote a book about it 10 years ago and a lot about the ethics of it, the potential ramifications of it. But all of it felt quite arm's length. And I think that really everything's come into focus in the last two years, maybe ever so slightly longer. And I think that we're on this trajectory that is only going in one direction, you know, And I think it's really important to have these conversations and these questions and ask ourselves in the biggest public squares that we can, what do we actually want our future to look like? Because I would like this to be like the Y2K bug, you know, where it was a really concerning thing on the horizon. But we invested, we worried, and we put the work in that meant that it didn't ever come to pass.
- [54:07]
B
That's right. That's right. That's. That's a great analogy because, yeah, we put the work in, we did it and it was okay. Same goes for, you know, the way landing on the moon made us look back and think a lot more about Earth and establish Earth Day and focus on the environment and the whole Earth catalog. Gets created because what we've got photos of the Earth. I feel like asking these questions and having these concerns and following artificial intelligence will help us focus on the opposite, which is what natural stupidity.
- [54:41]
A
And that's usually, yeah, I think it brings more sharply into focus ultimately what it means to be human, which I think is the thing that really all of us care about. But yeah, well, that brings us to, well, that brings us to an end. I think that concludes our journey of field notes today. If you have any questions that you'd like us to answer, you can send them in to us@therealStiscienceolehanger.com that's right.
- [55:02]
B
And join our newsletter@therealStis.com Science.
- [55:07]
A
We will be back next Thursday with another edition of Field Notes and on Tuesday with our normal episode episode.
- [55:13]
B
Until then, our normal episode unlike these abnormal field notes episodes. We'll see you next time, guys.
- [55:20]
A
Bye.