Transcript Generated for Free by youtranscripts.com ==========================================================

Title: Cory Doctorow On Why AI Won't Replace Workers, But Will Crash The Economy ================================================================================

Generated on: 7/29/2026, 2:17:11 PM

And when the AI bubble pops, we're going to see a massive economic vacuum, right? 35% of the stock market vaporized. We will likely see governments uh defaulting to austerity, which has the

well-known effect of driving support for fascism, right? When you lose the things that matter to you, your your doctor, your job, your school, uh you become a sucker for people who show up and say

the reason you've lost the things that matter is that undeserving others have taken it and you should embark on racist pilgrims, which you know we see taking place everywhere. We see austerity. So

we'll see we'll see more austerity. We'll see a breakdown in our systems. Um, and I'm absolutely delighted to welcome back journalist, blogger, author, and friend

of the show, Corey Doctor. A pleasure to be talking to you again, sir. >> Thank you very much. Pleasure to be back on. >> You recently appeared on a lot of really

high-profile programs, and the comments under these videos are pretty unanimous. Most people are saying, >> "This is the voice we need right now." Additionally, I have to confess much as

it's hard for a cold-hearted Eastern European like me that I've never received so much private correspondence urging me to have someone back on the pod than I did after our last

conversation. So, before we talk about your new book, I wanted to ask you why do you think your analysis of the internet and AI and and of big tech in general is resonating with so many folks

right now? >> Well, I've been at this for 25 years. Uh, so I work for the Electronic Frontier Foundation. It's a digital rights group based in America, but I was

their inaugural European director. And I worked on their behalf in more than 30 countries. And having done this in 30 countries for 25 years, uh, I like to think I have some context that rather

than treating all of these issues as though they occur in a vacuum, like we were just hit by a meteor called AI or, you know, a meteor hit Silicon Valley 10 years ago and created in shitification.

I can tell you where these things came from. I I can tell you how we can get rid of them. I I don't have to resort to mystical explanations like Mark Zuckerberg's the smartest person that

ever lived and you know succeeded where Rasputin and Mesmer and MK Ultra and pickup artist failed and he built a mind control ray and that's why everyone's acting so weird cuz Facebook's got a

mind control ray. I can ground this in material and economic phenomena. I can tell you why we have bubbles. Uh, I can tell you why my fellow progressives are wrong when they say endless growth is

the ideology of a tumor. The reason they want endless growth in Silicon Valley has nothing to do with ideology. It's about material factors. And I think that when you can ground the phenomena around

you that are making you angry and sad and scared in real material things that because they're real and material suggest ways that we can resolve them that even if you're still angry and sad

and and upset at the end of that at least you know what to do, right? At least you're not scared in the sense that you're that you're a prisoner of history. Maybe you're scared because now

you understand the forces arrayed against you and what needs to be done to deal with them. >> I mean, Elon Musk also has a lot of context, but not even his followers are

claiming that he's the voice of reason. [laughter] Well, I think Elon Musk makes people uh follow makes people excited about him because he deals exclusively in a kind

of adivism in in kind of libidinal emotions. And while there's something libidal and something addivistic about words like inchification and the minor license to vulgarity in moments where

everything seems to be going wrong, I like to think that I really um respect the people who take the time to listen to me enough to assume that they have the native intelligence to follow uh a

nuanced argument that doesn't assume that, you know, Jews are importing people of color to outbreed white people and end western civilization. or any of the other absurd things that

Elon Musk believes, including that US doesn't do anything. >> All right, you mentioned incidentification a couple of times now. Uh we talked about it last time and the

term really took off. It's everywhere. I can only imagine how hard some Google execs jaws must be clenching when they hear this word being applied to one of their products. Did you ever I'm just

wondering, did you ever get like a sassy email or anything from any of these people? I mean, I sometimes hear from them. I most mostly what's happened is uh I've had a couple of high-profile

speaking gigs canceled because one of the sponsors objected and no one's ever told me who the sponsor was, but you know, you look at the sponsor roles and it's all the usual suspects. So, um I I

I think that's the that's how that is being felt. Um, and you know, I I guess I it's always a mystery to me why someone would invite me to speak at a conference that's being sponsored by

firms that I have a long track record of uh saying um in tempmperate but true things about that make people angry at those companies. I I you know more power to them for inviting me. I'm grateful

for it. But I I think that there is this very uh foreseeable outcome. I think in many cases that those sponsors are going to object eventually you know um I think 20 years ago 25 years ago the sector was

diverse enough there were enough players in it that there was always uh a kind of balance of power within it when I look at institutions like the W3C the worldwide web consortium that designed

standards for the web and I look back in those days when there were you know hundreds of members of almost the same size and so the worst impulses of the members kind of died in their tracks

because there would be other members who had better impulses. And I look now where when there's a group within the W3C of very large firms that want to do something that's obviously bad for the

web, they win. It's not because the W3C is weak or foolish or evil. It's because when your, you know, budget depends on a dozen of your highest paying members and they all decide that they want one

thing, it's very hard to resist them. You've since written a new book, The Reverse Centaur's Guide to Life After AI, uh, how to think about artificial intelligence before it's too late. I

know you've had to explain the concept of a reverse centaur a gazillion times already, but for anyone who might be unfamiliar still, would you be so kind as to break it down again?

Yeah. So, it won't surprise you to know that this is not the first time that labor and automation have come into conflict with one another. And so there's a pretty good literature

explaining how different kinds of automation scenarios have played out in labor markets and for the people who rely on the outputs of labor. You know, the things that we get from workers who

use machines. And one of the um I guess iron laws or truisms or cornerstones of that theory is that when workers adopt automation, they typically do so in order to improve quality. But when

bosses drive automation adoption, it's typically in service to increasing throughput because you know if you are paying a monthly subscription or if you've acquired an asset that's now

depreciating, you want to maximize your use of it, right? That's that's how you uh maximize your profit, right? You've got an expense now. You want to you want to make the most of it. You want to, as

they say, sweat the asset. And you know, this is particularly true where you have situations where you have monopolies or cartels or duopolies uh that control a market because they can make worse

quality products uh and still sell them because they're the only game in town. And so this is what you see when you know the Dollar General that's replaced the grocery store in your town uh makes

everything smaller and more expensive. What are you going to do? take the bus 10 miles to the last remaining grocery store. Uh you know so in those conditions you would expect things to

get worse. And so this is where centators and reverse centur come in. Uh in automation theory a centaur is someone who is assisted by a machine. The analogy here is you know the human

head on the horse's body. The body being the body of a horse. It's tireless. It's strong. It's fast. But it doesn't have discernment or judgment. that comes from the human head. So the human head is

directing this strong, fast, enduring system, machine body. Um, so when you're riding a bicycle, you're a centaur, right? Even look like one. When you're using a spell checker, you're a

centator, too. You know, uh, there are lots of ways in which we can become centators. A reverse centaur, the the corollery, that's the reverse, right? That is the horse's head on the human

body. That's the machine driving the human. That's the machine making decisions about the human. And of course, the machine is still stronger than you, faster than you. It has more

endurance than you. There's just some part of the task that it can't do. And so, you are its peripheral and you are its help meat. And because you are slower and weaker and get tired more

easily, it is going to work you at the limit of your endurance and at the limit of your capability. And that means that you're not just going to get used by the machine, you're going to get used up by

the machine. And we see this in our kind of memes and tropes about automation and labor. You know, people who've seen I Love Lucy will know this classic scene where Lucy and Ethel are trying to put

chocolates in the chocolate box and the conveyor belts going so fast that uh the chocolates are going everywhere or that classic scene in modern times where Charlie Chaplan gets sucked in the

machine and ground up in the gears. You know, this is that logical consequence of uh capital-driven automation, an inferior product that makes the workers who produce its lives hell. And it's why

the most automated warehouses we've ever built, the Amazon warehouses, also have the highest rate of injury of any warehouse in the sector. That's not a coincidence. It's the consequence,

right? If you're spending seven or eight figures on automation, you're not going to run the machines slower than you need to. You will run them at the limit of the worker's capacity. And of course, if

you're working at your limit for long stretches, you will eventually make a mistake. And when you make that mistake in a warehouse, sometimes you get impaled on a forklift.

The last part of the title of the book is even scarier than what you were just talking about. I'm not going to lie. What do you mean by before it's too late?

Well, look, AI is a bubble. Um, which is not to say that AI technology isn't useful. If labor was driving automation adoption here, if labor was in charge of AI, I'm sure we'd see lots of people

finding useful things for it. In fact, when you meet programmers who say, "Oh, I write code with AI and it's the best code I've ever written." And then you meet other programmers who are writing

code with AI and they say, "I can't believe how bad this code is. You know, I work in avionics and I don't think anyone should ever get on an airplane because no one has ever produced tech

debt at the scale that we're producing it now with these AI tools." The way to understand that contradiction is not that the first group is lying or the second group is lying. It's that the

first group are centators and the second group are reverse centur. So there are useful things we can do with AI. But the bubble where they've raised now over $1.4 trillion on less than $50 billion a

year in revenue. That bubble demands that we fire as many workers as possible and replace them with defective chat bots. And even if we do that, we're still going to see a rupture. That

bubble is going to burst. 35% of the US stock market, the seven companies are AI companies, and they trade around the same hundred billion dollar IOU really fast, pretending it's in all their bank

accounts at once. Every generation of AI technology loses more money than the previous generation. Every new AI customer loses money for the business. Every time an AI customer touches the

AI, the business loses even more money. Uh and so here we are replacing critical functions in our society with technology that will eventually just be switched off. Um we are uh wiping out the kind of

hard one expertise. What's called sometimes process knowledge, right? All the things that workers know that can't really be written down. That tacet knowledge that's held between workers

within a firm across different divisions and across different firms in a sector and up and down a supply chain. All of that stuff's being wiped out because we're firing those workers and they're

becoming discouraged or they're retraining or they're retiring. And when the AI bubble pops, we're going to see a massive economic vacuum, right? 35% of the stock market vaporized. We will

likely see governments uh defaulting to austerity, which has the well-known effect of driving support for fascism, right? When you lose the things that matter to you, your your doctor, your

job, your school, uh you become a sucker for people who show up and say the reason you've lost the things that matter is that undeserving others have taken it and you should embark on racist

pilgrims, which you know we see taking place everywhere. We see austerity. So we'll see we'll see more austerity. We'll see a breakdown in our systems. Um, and

it's going to be really hard. And the sooner we burst the bubble, the better. And if we wait too long, some of these AI companies are going to unload on regular investors. Right now, most of

that AI money is is big dumb money. It's institutionals, it's uh, Gulf States, it's billionaires. But, you know, those people are really itching to get you to uh relieve them of their soontobe

worthless stock. We just saw this with Elon Musk and SpaceX where he did a sweetheart deal with the NASDAQ 100 to debut SpaceX in the index of top NASDAQ shares. That's never done. Normally, if

a company debuts on the stock market, it's ineligible for inclusion in an index until it's had several quarters of of sustained profitability because it might just be like a pump and dump. It

might be a swindle. And those indexes are held by everyday savers. That's what most people put their retirement savings into is an index fund. And if you manage an index fund, you are legally obliged

to buy stock from every company in the index. So if SpaceX is in the top 100, you have to go by SpaceX. And what Musk did when he floated SpaceX is he didn't just get this uh exemption to the

restriction on debuting in the index, he also restricted the shares that the company put in uh up on the market. So the company put almost no shares up for sale, all of the shares that were sold

to fund managers who were required to buy them uh at any price with your retirement savings. All those shares came from insiders who were not obliged to follow the normal legal requirement

of a lockout where insiders aren't allowed to sell for the first 90 days or 180 days again in case it's a pump and dump. And so, you know, we see these these companies already starting to

unload everything on ordinary savers to create basically bag holders. And again that like there it's not just the injustice although this is very important or the social problem of

people having their retirement savings wiped out. It's also what it does to the trusted institutions and to our social cohesion. It paves the way for authoritarian movements when people lose

everything. And so that's the that's the real risk that we face right now is the collapse of our important firms and institutions because we'll replace skilled workers with AI and then the AI

will be shut off when the companies run out of money. And that will make people furious uh because they will uh enter a period of austerity right after having all of their savings stolen by stock

swindlers who are also insiders. And that's going to be just catastrophic. What do you mean AI is just going to be switched off? Because after the.com bubble burst, the internet wasn't just

like switched off and a lot of people are saying >> Yeah. >> Yeah. No, that's very true. So the dotcoms had very different unit

economics, right? So the a lot of them were dumb, right? For sure. But as a sector, the web had great unit economics. Every new web customer made more money for web companies. Every time

you used the web, the web made more money. and every generation of the web made more money than the previous generation of the web. Uh that's not how AI works. As I've said, AI is the

opposite of that. And so, while there will be, and I this is actually quite important, there's going to be a ton of GPUs kicking around. There's going to be like data centers looking for buyers at

fire sale prices, there's going to be skilled workers who know how to do things with AI who don't have to answer to their stupid bosses anymore. And there's going to be these open source

models that have barely been optimized and that whenever anyone tries to optimize them, they get all kinds of of new efficiencies out of them. That stuff will be around, but like no one's going

to be keeping Open AI around in its current form. Open AAI is a business that sells $100 bills for a dollar each, right? And like even if you wipe out all of its capital expenditure overhang and

start over again through a bankruptcy, you can't if you continue to give away $100 bills for a dollar each, you will quickly end up exactly where you started. So, you know, it might be that

you can replace a worker who costs you $100,000 a year with a chatbot that costs you $1,000 a year and not really care that the outputs are half as good as the worker produced. But if the

chatbot costs $120,000 a year and the outputs are half as good, that's not the same situation at all. And that's what we would be in if we were going to continue to consume AI resources at the

rate we're consuming them now. but actually had to pay for them instead of having them defayed by, you know, Gulf State sovereign wealth funds and and Silicon Valley billionaires.

Yeah, I'm no economist, but selling a $100 bill for $1 doesn't sound like >> Oh, and it gets worse than that. You know, when the companies were trying to clean up their balance sheets because

they were all talking about their IPOs last month, they were like, "What if we start charging $5 for these $100 bills?" And like all the CEOs who' bought AI for their businesses and said everyone has

to use AI and if I catch you not using AI I'm going to fire you. They turned around they said uh these $100 bills are not worth $5. I'm sorry. Like we just don't get enough use out of them uh to

pay $5 for them. We are going to institute you know companywide bans on using AI limits on using AI. You're going to have to ask your manager for more tokens if you use up the tokens

that you've been given. Right? So like there it's not just that they're selling $100 bills for a dollar. It's that no one will pay more than a dollar for their $100 bills. It's really bad.

I want to talk about the ideology behind this AI push a little bit. And I totally believe that the engineers working on these LLMs have probably read too much science fiction and are working

earnestly towards creating some sort of benevolent super intelligence or whatever. However, when it comes to their bosses, the CEOs who are steering this thing, do you reckon that in their

minds this whole project is basically given what we're talking about right now, just like every capitalist's wet dream? So, so a campaign to simply eliminate human labor.

I think that there's two things going on here. I think for sure some of these guys definitely like would like to live in a world without people. I think I think there's a kind of solopsism that

goes with being a billionaire. You know, you can't really inflict the kind of pain at scale that you have to in order to make a billion dollars or hundreds of billions of dollars if you really

believe other people feel the way you feel are as real as you are. You know, and that's true of this class when they're face to face in in individual interactions. So, you know, the Epstein

class just couldn't have believed that the children who are on that island are as real as their own children. But it's also true especially when they deal with people as a statistical abstraction. You

know, Jeff Bezos and his army of of drivers and warehouse workers who don't even get pea brakes, right? Like he would never stipulate to those indignities, right? Like he gets a

catheter fitted when he goes up on his pl on his spaceship, right? He doesn't just like pee in a bottle, right? Uh like so even under the most extreme circumstances, like there's lots of

toilets on his yacht, right? the idea that you you don't get a pea break is something for other people and not for him. And so, you know, you get Elon Musk calling everyone who disagrees with him

an an NPC and so on. So, I think there's a lot of these people who just don't think other people are as real as they are or at least, you know, everyday people. And certainly like if you run a

social media business, I think it's probably really bad. Like Mark Zuckerberg deals with people entirely as statistical abstractions, you know, he's just like twisting dials in the back end

of Facebook to see if he can events human behaviors at scale, right? So that this is really a disindividuated way of confronting people. But then you have on top of that you have um what what Milton

Kane's called the the beauty contest. So Milton Kane said that um making money by investing is not a matter of picking the most beautiful person in the beauty contest, the company that's going to

make the most money. Making money as an investor is about picking the person the other judges will think is most beautiful because you are trying to pick the stock that other people are going to

want to buy irrespective of whether you think that stock is good. So, you know, there's people who invest in, you know, the supplements that Andrew Tate and uh um you know, uh uh those other Chud

podcasters, Joe Rogan and so on that they >> Oh, yeah. You know, there's people who invest in those in those um supplements who don't

believe that the supplements work, right? They know that it's just a way of getting extremely expensive urine, but they also know that there's like a an infinite army of vulnerable, desperate

young men who will pay everything they have for the supplements. You don't have to think the supplements will work to invest in the supplements. You don't have to think that, you know, Gwyneth

Paltro has a jade egg that will make your vagina better to think that women will buy the jade egg and risk their lives by putting it in their bodies, right? And so I think by the same token,

there's a lot of people who think, you know, bosses are really easy sells for a sales pitch about how you can replace workers with software cuz like workers are a pain in the ass. They demand moral

consideration and even more so they provoke these kind of ego shattering confrontations where like you have an idea, you want to do a thing and then you have to go out and talk to a worker

who knows how that thing works and they tell you that your idea is like stupid or illegal or impossible or it's going to kill people. And you know, eventually you start to have this suspicion that

while you're telling yourself you're in the driver's seat, you know, the fact is that if you didn't show up for work, everything would be fine. And if they didn't show up for work, everything

would stop. And maybe you're in the back seat with a FisherPrice steering wheel. And so AI is like really the potential of like wiring the steering wheel directly into the drivetrain, right? You

have an idea and it just gets [ __ ] out by the AI. And so, you know, you don't have to believe that AI can do that to believe that bosses will buy it if you promise them that. And, you know, you

don't have to be able you don't have to be fast enough to outrun the bear. You just have to be fast enough to outrun everyone else. So long as you get out of the AI position before everyone realizes

that the AI can be useful for workers who want to augment their work, but not particularly useful for replacing workers, and you get out ahead, then it doesn't matter that it it was a bad

investment. It doesn't matter that the company failed. You got out. Okay. Right. There's plenty of people who made a lot of money on NFTs. NFTTS were are and always will be stupid and a bunch of

people made a lot of money from NFTs. I mean, it's bizarre how they're not even trying to come up with a sensible justification for all this, are they? It's just a halfbaked version of a of a

trust me, bro. that they're saying we have to throw everything into their company so they get to AGI as fast as possible and then out of the goodness of their hearts they will reap the benefits

on the rest of humanity. Now call me Mr. cynical. But when in the history of the planet did did it ever happen that the guy at the top and and usually it was a guy said, "You know what? Now that I

robbed all of you blind, I think I'm going to divide the pie fairly. Let everyone have a nice piece." Or maybe I'm, you know, just not as well read on my history as I thought I was.

Well, you know, there's a connection between this and some of the more esoteric forms of billionaire mysticism. So, you know, the the Sam Bankman Freed and a bunch of his cronies practiced a

thing called effective altruism. I've heard. Yeah. >> Yeah. It has its roots in some pretty benign stuff like trying to empirically determine whether you save more lives

investing in like antimmalaria research or mosquito nets, right? Uh but it quickly became this kind of hyper utilitarian self-parodying venture where you um earn

to give. So you take the most destructive, harmful job you can that pays the most to earn as much money as possible, no matter how much pain you inflict on people around you in the name

of earning so much money that you can give it to causes that will make people's lives better. But of course, the way to improve the most lives is not to help people who are alive today, but

rather to invest all your money in improving the lives of 10 to the 53 artificial humans who will come into existence on Venus in 10,000 years. Because if you make even the smallest

improvement to the lives of 10 to the 53 people, it will swamp all the benefit you could possibly confer on a mere 8 billion people who are alive today. Right? This is like this is a very weird

idea right and like I I it is you know as they say uh the I think again this this is I can't remember if this is Keynes or Galbreth but you know it's the uh endless pursuit of uh man's oldest

ideological project to find a superior moral justification for selfishness right so much of of of this uh you know kind of cashes out there but um I I think that that um

it's worse worse than the idea that they're just handwaving and saying, "Trust me, bro." And like using the shitty trope from science fiction novels where computers get complicated enough

and then they wake up, which is like what science fiction writers do cuz we're lazy and we just want to get straight to the futuristic parable. We don't want to come up with a bunch of

ways in which computers can become self-aware. We just say like, "Oh, you add enough transistors and now it's a brain, right?" Uh it's it's worse than that. When you read accounts of the NIH

researchers who pleaded with the Doge Boys to not cancel their long-running cancer research, cancel their long-running cancer research. Sorry, it's hard to say. Um, what they what the

Doge Kids said was, "We don't need cancer research because general AI is around the corner, and when general AI arrives, it will cure cancer." So, it's it's worse than that. It's it's it's not

just that they're making false promises. It's that they're discontinuing the actual promising things. If you want a more immediate version of this, every time an AI bro turns on a gas turbine

and is asked about the climate impact, they say, "Don't worry. When general AI comes along, it'll solve the climate uh emergency." Right? And our job now is to outrun the second law of thermodynamics.

Right? before we melt the the the ice caps, we have to make the general AI that tells us what to do once the ice caps melt. Right? It's um it's an incredible gamble they are making with

the planet and their basis for asserting that a general AI is around the corner is is is deeply foolish. They're basically saying that if you teach enough words to the word guessing

program, it will wake up. which is like saying that if you breed horses to run fast enough, one of them will give birth to a locomotive, right? Like it doesn't matter how fast the horse runs, it's not

a train. And it doesn't matter how well the program guesses words, it's not conscious. Consciousness is not just guessing words. >> Okay. Um, since we're here, let's talk

about the state of of this technology today because for somebody like me, he usually uses it as some sort of like an expanded Google search. I honestly have no idea if this thing is like the

philosopher stone or just a really confident parrot. And the discourse at large is really not helping me. For example, when I opened my ex yesterday, God helped me. My feed was suddenly full

of folks saying that AI has cracked some old mathematical problem and apparently it's true. It's not [ __ ] On the other hand, immediately after that, I watched a video on the work of of of

Jennifer Dodna. She's a scientist who I I want to have on the podcast someday. She's a pioneer in gene editing and she said that chat bots are basically useless when it comes to innovation and

that she thinks it's going to stay that that way for a while. So which of the two is it? I'm I'm confused. You know, one of the things that AI uh and this isn't the first generation of AI we've

ever had. AI was coined at a conference in the 1950s and we've had many different kinds of AI emerge. Um, and you know, while they're all radically different and while AI has been a

marketing term for each of them, what they all share in common is that they were all a bid to replicate and possibly even replace some function that we think of as innately human. And what each

generation of AI has done is shown us that while there is a uniquely human way to do certain things, that there are other ways to do it, too. Right? The way that humans play chess is not the way

that AI plays chess. uh and uh while they may look the same from the outside, they're doing something different. And you can achieve the state of a chess game with AI, like

a successful chess game, a victorious chess game in a way that's different from the way people do it, which is interesting, super interesting, right? I mean for a long time you look at the

history of of uh um human flight and for a long time we were kind of become in this doldrums of trying to mimic birds. We just built ornithopters which are basically planes that flap their wings

and they never worked and we had to invent a totally different way of flying to um to to actually leave the ground. There's a famous quote about thinking machines. The where a computer scientist

was asked, "Can machines think?" And they answered, "Can submarines swim?" Uh, submarines don't swim, right? But they move through the water. They do some of the things we do with thinking.

You know, I don't know if you're married, but I'm I'm married. >> I am. >> And I often can predict what my wife is going to say. Uh, and my wife's

autocomplete I can >> can often predict what my wife is going to say because it knows the corpus of her language, right? She sends a lot of text messages and so it knows something

about the statistical likely statistical distribution of the replies to certain messages. But when my wife says something that isn't what I predict, I can advance a pretty good theory of

what's going on to make her say something that is out of character. And the AI is at sea. And that is the difference between understanding and extrapolating. Right? Understanding

involves a theory. And the promise of big data, which AI is the latest generation of, you know, 15 years ago, we were calling this big data and now we're calling this AI was theoryfree

inference. Right? Uh uh um the idea behind big data was that you could observe by crunching population scale data that these molecules taken in this

combination to people who had these lifestyle factors and these pathologies produced a cure or uh produced some kind of um uh improvement in their in their state without ever having to know why.

Right? that this was this was theoryfree inference. And so long as it works, you know, it' be nice to know the theory, but so long as it works, you don't necessarily have to know the theory. And

indeed, there's a lot of medicines we use that are theory free in the sense that we've observed that they work. We've we've double blind tested them. We've checked for side effects and so

on. And we use them even though we don't understand at all or completely how they work. uh and AI is this idea that we can just effectively give up on having a

theoretical basis for the world and switch to this these theoryfree inferences. And the problem is that um it's very hard especially when you're talking about unusual circumstances to

unwind false positives and detect them when you don't have a theory because you don't really know how to test the theory. Right? A theory free inference is by definition not falsifiable, right?

You can't say if this weren't true, I would do uh then then you would expect X or Y to happen under these circumstances, right? Um because you don't have a theory. You don't have a

way to know why it's true or why it's not true. And so as a as a way of knowing, it's quite limited, which isn't to say it's useless. You know, there are lots of theory-free things that we have

discovered with AI that are interesting. And if scientists were being told, here is a tool. You have all the budget you've ever had before. You have all the freedom you've ever had before. And

here's a tool that does some new multivariant analysis that uh lets you do stuff that you might otherwise have had to hire a programmer for or a statistician or that you might have had

to spend several days designing stats uh uh custom stats crunching software for or algorithms for. But this is a thing that will do it for you. it will make some mistakes. So, you're going to have

to double check it. You have to be careful about how you use it. I bet we would have scientists doing all kinds of great things. But that's not how it's working, right? What you've got is the

Doge situation. Fire scientists, replace them with software, make the survivors do more with less, and expect the same outputs. And I think that is very frightening. You know, I'll give you an

example from my life. So I um have an extremely treatable form of cancer and uh this means that I'm spending a lot of time with radiologists and in fact about three weeks ago now a radiologist told

me that I am cancer-free which was very good news. Uh and um one of the things that we hear a lot about is that AI might be able to spot solid mass tumors that radiologists can't spot. And if

there was a sales call at my hospital, which is the Kaiser Clinic in in um Los Angeles, next to the Church of Scientology, because that's how we roll in LA. Uh and if there was like some AI

salesman who like, you know, deaked out the guys from Scientology trying to get them to come in and take a personality test and went up to the CEO's office and said, "Look, here's the deal. Right now,

you employ 10 radiologists. They do a 100 x-rays a day a piece. They cost you um uh $300,000 each. you're spending $3 million a year on radiology and processing uh a thousand X-rays a day. I

want you to spend an extra million dollars on software and that's going to sit on the shoulder of those AI those radiologists and once or twice a day it's going to tap them on the shoulder

and it's going to say take another look at that X-ray and they're going to do 98 x-rays a day instead of 100. So, you're probably in addition to the million bucks for the software going to have to

hire a part-time radiologist to pick up the slack, but you're going to save people's lives. I'd be very happy that as a cancer patient, I love the sound of that. That's not how you turn a trillion

in annual expenditure on $50 billion in revenue into a profit. The way that you turn $50 billion of revenue on on a trillion dollars expenditure into profit is you have the AI salesman deut the

Scientologist trying to give them the personality test go into the CEO's office and say right now you spend $3 million on uh 10 radiologists. Fire nine of them. Save $2.7 million a year. Split

that between you and Sam Alman make the remaining radiologists mark the AI's homework. They sign their name to the bottom of every radiology report. And when someone dies, you blame them.

They're the accountability sink, right? That's that's the thing to worry about. Now, radiologists right now have a ton of labor power. When I talk to radiologists about how they're using AI

right now, it sounds more like the first scenario than the second one. But the point of AI, the reason that investors think that you can sell AI to people like hospital administrators is that it

will erode labor power so you can fire nine other radiologists. Right? Like that's the that's the reason it's been capitalized the way it's been capitalized. That is the politics of

this artifact as it exists in the contingencies that produced it. And while we could produce a different set of contingencies and a different set of politics, that is the politics of this

version of AI that we're living with right now. >> That all makes sense. But to play a devil's advocate just for a second, what gets me rattled are some of these

whistleblowers. So ex employees who are now warning against the dangers of AI. A couple of days ago, I checked out an interview with Daniel Cockot, who was a researcher at OpenAI and apparently

refused to sign a non-disclosure agreement for $2 million, which would ensure that he keeps his his mouth shut about their sinister affairs or something. He then created AI 2027,

which I'm heard everybody has heard of. Um, it's detailed scenario of how things might go haywire with AI soon. I mean, listening to the guy talk, he just seemed genuinely frightened by what he

thinks is about to come, and that gets me worried. >> Yeah. Look, you know, I don't know this guy. I've seen I've seen AI 2027. I don't think much of it. And, you know,

I've also seen the rebuttals from people who don't work for AI companies but do work in AI, like, you know, research the research team at Princeton who maintain um the AI is a normal technology

website, who wrote a very detailed critique of of AI 2027. Uh, and and I gotta say like we have a a a history in Silicon Valley in tech of guys like going into the bathroom, shining a

flashlight under their chin, looking in the mirror, and going, "Ai," and then getting really scared, right? We had it 10 years ago with a bunch of people who used to work at Facebook

insisting that they had been trained to be evil dopamine hacking wizards and that the reason your grandpy was a QAnon is that they were so good at hacking his dopamine. And not that your grandpy was

a racist forever that you he kind of, you know, kept his mouth shut about it and that what Facebook did was like trap him in a room full of a bunch of other racists who laughed every time he'd said

something racist. So now he can't shut the [ __ ] up about it, right? Like that's like a much more parsimonious explanation. And it doesn't require any wizards. And you know, there is

something I think quite uh exciting about having been an evil h dopamine hacking wizard because now you can promise that you're a good dopamine hacking wizard. And the operative word

here is wizard, right? Like how cool is it to have been a wizard? And and I'm not saying these guys don't think they're wizards. I think they might believe they're

wizards. I just don't think they're wizards. Um, I mean, on another note, this thing is being crammed into every single product whether we it makes sense or

not. At the same time, apart from a couple of like overzealous LinkedIn boomers, I don't know a living soul that would be excited about it. Nobody's happy. Half the time I I want to take a

shower after I've used one of these chat bots. >> And yet, it seems we're not really putting up much of a fight against this really brutally forced adoption.

I don't know. I I think you know you see a lot of workers who are getting pissed off about it. You see a lot of creative workers are getting pissed off about the data anti-data center movement is

gigantic. Um you know that there there's there's a lot of people are very angry about it. I do want to say though I know lots of people who use AI who I think it was quite credible and who say that they

got great great uses out of it. People who use it in VFX uh not to like render final images but to like automate parts of the process. Uh the best programmer I know is a guy called Patrick Ball who

founded uh an NGO called the human rights data analysis group hrdag.org and HR DAG they um specialize in using fragmentaryary evidence from uh war crimes, civil wars, uh genocides and

they use rigorous statistical methods to build out a picture of what aspects are probably missing from the data by interpolating the data. And then they they testify at human rights tribunals,

truth and reconciliation. They worked on Malassovvich and they worked in uh Rios Monte and and East Timour, South Africa, the Colombian civil war. Patrick's, as I say, one of the best programmers I know,

brilliant statistician, brilliant programmer. He tells me he's using a whole bunch of AI agents to write the best code he's ever written. Right? But Patrick also has more discernment than

any programmer I've ever met. So, you know, if I was ever going to trust a programmer to know which parts of the job to automate and how to double check to make sure that when you automated it,

you hadn't accidentally produced a bunch of tech debt. Uh, Patrick's the guy I would trust, you know, and so, uh, again, this is one of those things where you have labor driven, uh, adoption. Um,

look, I use I use a chatbot running on my computer, a local open source model, uh, to check my blog post for typos in the morning because I publish my blog on all these different platforms. I publish

it uh, as threads on all the different social media platforms. I publish it on Tumblr and Medium and I put it on a newsletter and I put it on the web. And when I find a typo, my readers really

want me to know about it. They're very invested in my knowing about my typos. And so, I have to go fix it or I'll just get like a million emails. And once I published it to 10 places, fixing just

like a period that's gone astray is like 15 minutes work. And so if there's six typos, that's like an hour and a half that I lose, right? So I drop it into this local chatbot. I type find typos

and finds all kinds of typos, right? It's just like and and many of them aren't typos and I ignore those because I have discernment. I've been telling search engines that they were or not

search engines, spellch checkers that they were wrong for 30 years. Uh, I've published 30 books, more than 30 books with major publishers. I've been telling copy editors and proof readers that they

were wrong for 25 years, right? And I'm perfectly capable of telling a chatbot it's wrong, too. The difference, of course, is the chatbot never learns. Whereas the proofer, when you say,

"That's that's not a typo. That's intentional. Doesn't send you back the manuscript next time with those errors in it." It's not perfect by any means. But like I was never going to hire a

proofer to read my blog post in the morning in the 10 minutes after I write it before I publish it. It's just ridiculous. So yeah, I'm very happy to have it. It's fine. It is absolutely

fine. Uh it is a normal technology. My first word processor was a program printed in a magazine that I bought at the corner store and took home and typed into my Apple 2 Plus. I was 10 years

old. It was 1981. I've used a lot of word processors since. A lot of them have had new features. A lot of those new features I thought were stupid and when other

people use them, I think it made their writing worse. Uh, some of them are really useful and I use them all the time. I've never seen a feature for a word processor where I said, "That's so

cool. We should bet the world's economy on it. And by the way, we should probably also put all the writers in a wood chipper now that it exists." Right? And and like that's the problem, right?

is we are treating what cashes out to plugins for common software that are sometimes useful and sometimes not and when workers drive the automation can produce things that make your life

materially better and we're treating them as as genies and gods and demons as the end of the world or as the beginning of a new one. Uh and it's not that if if AI destroys the world it's going to be

because the bubble is burst. It's not going to be because uh the software was so cool that like it woke up and turned us into paper clips. >> Maybe this is a mad thought that's

completely divorced from reality, but is it possible we just get bored by it all? I mean, how many more AI generated videos of Tom Cruz fighting a large Pokemon on some Chinese roof can I watch

on X before I just want to retreat to a prehis cave forever? I mean, I can't be the only person that's already tired of these smartass chat bots. Come on. >> Yeah, and I got to run. So this is maybe

a good place to end the the conversation on. I do think that like a lot of that AI stuff is getting washed really quick. And you know this has always been one of my critiques about a certain kind of

technological doomsaying that even where you have a thing that that works really well at the start right someone's found a little cognitive blind spot. someone's found a thing that, you know, makes

people curious or catches their attention that unless there's some substance to it that it very quickly becomes uh something that you're a nerd to,

right? Uh you know, do you remember when Upworthy headlines were invented and it was like 11 facts about socks? You'll never believe number three. And I clicked through like half a dozen of

those headlines and then I realized that like number three just wasn't that interesting and never would be. and I never click through one of those headlines again. Now, that's not to say

that that form of headline in the hands of someone who's a good writer, who knows how to infuse it with actual communication, who's creative, who's smart, and who has something to say,

because you know the problem with those upworthy headlines is they were always on articles that were stupid, right? The best thing about them was the headline. But if you have something smart to say,

there are probably aspects of those techniques that will work to capture people's attention and hold it because you're writing about something real and substantive. And I believe that we will

have AI generated text and images and so on, not in vast quantities, not [ __ ] emails, routine business emails that have been inflated three bullet points into five paragraphs of, you know,

floorid [ __ ] But like we will have AI generated material in our lives that will be being used in the background whatever you know to to help you get stuff done uh and to you know make

things better. And some of it will be stupid and some of it will be fine. And most of the people who are making it now will have moved on to something else because they're they have nothing to

say. They're just they're just fiddling around with a new web toy. And and in that regard, AI is a normal technology because I've just described the trajectory of many technologies that

we've had in the last couple of decades. Uh that's that's how it went with Buzzfeed quizzes, right? It's that's how it's going to go with other web toys, too.

Corey, always a pleasure. Thank you so much. >> Nice to chat with you. >> Thank you so much for listening. Please subscribe on YouTube, follow and rate on

Spotify and Apple Podcast. You would help me a bunch with that. Also, feel free to support me on Patreon. This is a oneperson podcast. I do everything by my little old self. It's a lot of work,

especially with no team, no ads, and no real marketing. But you can help by hopping on Patreon and finding the Smart Cookies Podcast on there and giving me your money, of course. Super hard for

small creators to get through. So, your support would mean a lot. And then last, but certainly not least, I have to thank my existing patrons. Without you, there is no Smart Cookies podcast, which makes

you not only absolute legends, but also in effect actual producers of this show. So eternal thanks goes go to Lenox Diholdis Chtomania Charlie Balbeck Zalver Matia Valinger

Sluga Kaek Machkin Sim Vid Johnson beauty Idam manino Goran Gordon Lorenzo urichuka car and Tai Chi [sighs and gasps]

you guys are the best I'm deeply touched so thank

========================================================== Generated by youtranscripts.com - Free YouTube Transcript Tool