The “makes everyone stupider” line is the right worry aimed at the wrong target. AI does not lower intelligence. It removes the friction that built intelligence in the first place. Learning to write is learning to think, Hayes says, and he is correct, but the mechanism matters. You do not get the thinking from the finished essay. You get it from the struggle of producing it. Outsource the struggle and you keep the output while quietly losing the faculty. The danger is not a dumber person. It is a person who cannot tell the difference, because the results still look fine.
Hayes feels this himself. He draws a clean line between thinking and grunt work, then immediately admits the line might be fake. It is fake, or at least it moves. “What year was this guy born” seems like pure retrieval, until you notice that the act of hunting through your own files is how you stay intimate with your own material. Every offloaded task feels trivial in isolation. The cost is only visible in aggregate, years later, as a capacity that no longer answers when you call it. That is the actual endgame worth a podcast. Not whether the machine can think, but whether we will still bother to.
I mostly agree with this take. I do think there's always opportunity cost and time allocation problem. A good historical analogy is memory and writing, where I believe there's academic evidence for tradeoffs. The
Homeric tradition, for example , relied heavily on memory, whereas "offloading" that process to writing freed up time and effort for exploring. Yes, searching your own files has value, but it also has a real cost.
I would posit a slightly different angle on what's lost when a chatbot is in the driver's seat: these models provide a "consensus" view, whereas I have a unique human experience and view of the world. When I'm steering, I can ask a model for feedback and it will often identify a "consensus" error, oversight, or omission. It's a bit like asking a friend for feedback, where the friend is a distillation of the median human. But when models drive, the world becomes beige (or worse).
What bothers me is the intent of AI’s tech bros masters: How can we trust them to act humanely vs. financially? The Anthropic mess pits moral v. amoral; and those tech bros prefer profit first.
I totally agree. I don't see the danger in AI per se; I see it in the men (and I do mean men) who are behind the curtain and who seem to only care about money and power. I'm 74 years old and grew up with TV (black and white originally) and all the knowledgable adults telling us our brains were going to rot. I watched a lot of TV (still do but now it's streaming shows and movies) and earned a PhD in Accounting, specialty in federal taxation, minor in law , CPA (university professor and associate dean in a business school) so obviously, it didn't rot my brain. I grew up reading a lot of fiction (still do) and then added technical reading. I love listening to music and I grew up when adults were slamming rock n' roll as immoral. Then it was MTV, music videos and video games. I owned the first IBM PC (smart phones are so superior today), used the first copy machines, fax machines, microwave ovens, and countless more innovations. Unfortunately, every innovation leaves some workers behind and we really suck at taking care of workers. I don't see the problem as the innovation. I see the problem as being tied to the greed and arrogance of the men running the corporations and in government. I think now is the time to write laws and regulations to oversee the innovations, and to strengthen unions use the wealth of the US to support workers instead of building ballrooms and arches and golden statutes.
We can't. Musk squawks about UBI but you know he'll fight tooth and nail to ever, ever have to pay more in taxes to fund it. Like all industries, he wants government to cover the damage he creates while allowing him to walk away with billions.
I agree, and I also think Hayes' comments on the economics are spot-on (for a detailes critique, see Ed Zitron's deep-dives). The US AI industry doesn't really have a competitive moat, and the Chinese open-source models have largely kept pace with their progress. Now the Chinese models are being deployed at a fraction of the cost on US and EU hardware -- i.e. privacy and security are solvable problems for Chinese models (ref https://opencode.ai/go).
Long story short, the US tech bros are front and center now, but it's far from clear that they'll end up on top. Indeed, recent reports suggest that integration of AI into real economy is proceeding much more rapidly in China than the US.
Really interesting. I don’t use AI hardly at all, and am in the end stage of my career (I’m almost 70), but if I was younger I think I’d be using it the way Hayes describes: basically data mining. In my case I’d be asking something like: “given these patient variables, give me the 3 most evidence-based ways to start building this patient’s care plan.” That would potentially save me and the patient a lot of time and aggravation. I sure wouldn’t want it doing my critical thinking for me, though!
Chris Hayes accidentally nailed the real AI fear: not killer robots, but intellectual atrophy. Humanity invented a machine that can summarize books, write essays, answer questions, generate art, and think faster than most people online already pretend to. There’s a real risk we automate curiosity itself.
I'm thinking about "enshitification". Facebook began with kind of a "noble" goal. And then...? Amazon could have been. ..but? Once oligarchs and corporate entities grasp/grap a concept for their own, any social value is out the window.🤑🥺
AI’s use as the sole purpose to eliminate the need to pay working people - which I contend is its current business use - is the *only* way corporations know how to use technology. Any other application feels like another way *to spend* money, keeping them perpetually uninterested. No one anywhere has shown interest in this fundamental problem. So far, AI’s application in business has been minimally successful & maximally frustrating for the workers who end up doing (read fixing) much of AI’s intended work anyway - and for consumers who are just looking to interact with a person. The potential security issues are monumental. #fail.
The military applications - that we know about - seem best suited in calculations and drone strikes. Most successfully carried out by skilled human operators.
Young people hate AI because they see it another way companies kill the environment.
Just like nuclear’s promise was stifled (to this day) by the Cold War, then Chernobyl (3MI) and then politics, young people equate AI with a million Chernobyls. But instead of politicians curbing AI’s use due to its danger to people and the environment, the corporate capture of government *is* cramming the tech into every aspect of life.
Congress is about 20yrs behind regulating this technology, because Congress is profiting from it. Until we as a nation muse about how to repair self-government, we’re just treating the symptoms of the problem, poorly.
It’s like asking the government for boards to build a boat and - b/c they are captured by the poster board companies - they bring you poster boards. They are in fact “boards.” But If they have no idea what poster board is, their solution doesn’t meet the reality of the situation. It’s not helping you build your boat, *and* it’s a waste of poster board.
We’re close to having Reps whose only interaction with technology *has ever been* AI.
I think the AI debate is getting framed completely wrong.
The real danger probably isn’t that AI becomes smarter than humans overnight. It’s that humans slowly stop developing the skills that made us intelligent in the first place. Writing. Research. Patience. Critical thinking. Memory. Attention span.
If calculators changed math education, AI could fundamentally change cognition itself.
At what point does convenience start making society mentally weaker?
The “makes everyone stupider” line is the right worry aimed at the wrong target. AI does not lower intelligence. It removes the friction that built intelligence in the first place. Learning to write is learning to think, Hayes says, and he is correct, but the mechanism matters. You do not get the thinking from the finished essay. You get it from the struggle of producing it. Outsource the struggle and you keep the output while quietly losing the faculty. The danger is not a dumber person. It is a person who cannot tell the difference, because the results still look fine.
Hayes feels this himself. He draws a clean line between thinking and grunt work, then immediately admits the line might be fake. It is fake, or at least it moves. “What year was this guy born” seems like pure retrieval, until you notice that the act of hunting through your own files is how you stay intimate with your own material. Every offloaded task feels trivial in isolation. The cost is only visible in aggregate, years later, as a capacity that no longer answers when you call it. That is the actual endgame worth a podcast. Not whether the machine can think, but whether we will still bother to.
Johan 🐌
I mostly agree with this take. I do think there's always opportunity cost and time allocation problem. A good historical analogy is memory and writing, where I believe there's academic evidence for tradeoffs. The
Homeric tradition, for example , relied heavily on memory, whereas "offloading" that process to writing freed up time and effort for exploring. Yes, searching your own files has value, but it also has a real cost.
I would posit a slightly different angle on what's lost when a chatbot is in the driver's seat: these models provide a "consensus" view, whereas I have a unique human experience and view of the world. When I'm steering, I can ask a model for feedback and it will often identify a "consensus" error, oversight, or omission. It's a bit like asking a friend for feedback, where the friend is a distillation of the median human. But when models drive, the world becomes beige (or worse).
Never thought of it this way
What bothers me is the intent of AI’s tech bros masters: How can we trust them to act humanely vs. financially? The Anthropic mess pits moral v. amoral; and those tech bros prefer profit first.
I totally agree. I don't see the danger in AI per se; I see it in the men (and I do mean men) who are behind the curtain and who seem to only care about money and power. I'm 74 years old and grew up with TV (black and white originally) and all the knowledgable adults telling us our brains were going to rot. I watched a lot of TV (still do but now it's streaming shows and movies) and earned a PhD in Accounting, specialty in federal taxation, minor in law , CPA (university professor and associate dean in a business school) so obviously, it didn't rot my brain. I grew up reading a lot of fiction (still do) and then added technical reading. I love listening to music and I grew up when adults were slamming rock n' roll as immoral. Then it was MTV, music videos and video games. I owned the first IBM PC (smart phones are so superior today), used the first copy machines, fax machines, microwave ovens, and countless more innovations. Unfortunately, every innovation leaves some workers behind and we really suck at taking care of workers. I don't see the problem as the innovation. I see the problem as being tied to the greed and arrogance of the men running the corporations and in government. I think now is the time to write laws and regulations to oversee the innovations, and to strengthen unions use the wealth of the US to support workers instead of building ballrooms and arches and golden statutes.
We need people to be educated, and to raise their consciousness. Maybe the latter mostly. Okay throw in some sense of morality.
Exactly! Trust Musk? Trust the Ellisons? Sure! And that bridge you got - the one in Brooklyn? - still available?
We can't. Musk squawks about UBI but you know he'll fight tooth and nail to ever, ever have to pay more in taxes to fund it. Like all industries, he wants government to cover the damage he creates while allowing him to walk away with billions.
Agreed. Even worse, though, his fellow South African Peter Thiel … who’s already got his talons in Vance.
I agree, and I also think Hayes' comments on the economics are spot-on (for a detailes critique, see Ed Zitron's deep-dives). The US AI industry doesn't really have a competitive moat, and the Chinese open-source models have largely kept pace with their progress. Now the Chinese models are being deployed at a fraction of the cost on US and EU hardware -- i.e. privacy and security are solvable problems for Chinese models (ref https://opencode.ai/go).
Long story short, the US tech bros are front and center now, but it's far from clear that they'll end up on top. Indeed, recent reports suggest that integration of AI into real economy is proceeding much more rapidly in China than the US.
Thanks for this thoughtful interview. Personally i am still frightened by the new ‘steam engine’….
Really interesting. I don’t use AI hardly at all, and am in the end stage of my career (I’m almost 70), but if I was younger I think I’d be using it the way Hayes describes: basically data mining. In my case I’d be asking something like: “given these patient variables, give me the 3 most evidence-based ways to start building this patient’s care plan.” That would potentially save me and the patient a lot of time and aggravation. I sure wouldn’t want it doing my critical thinking for me, though!
Chris Hayes accidentally nailed the real AI fear: not killer robots, but intellectual atrophy. Humanity invented a machine that can summarize books, write essays, answer questions, generate art, and think faster than most people online already pretend to. There’s a real risk we automate curiosity itself.
I'm thinking about "enshitification". Facebook began with kind of a "noble" goal. And then...? Amazon could have been. ..but? Once oligarchs and corporate entities grasp/grap a concept for their own, any social value is out the window.🤑🥺
AI’s use as the sole purpose to eliminate the need to pay working people - which I contend is its current business use - is the *only* way corporations know how to use technology. Any other application feels like another way *to spend* money, keeping them perpetually uninterested. No one anywhere has shown interest in this fundamental problem. So far, AI’s application in business has been minimally successful & maximally frustrating for the workers who end up doing (read fixing) much of AI’s intended work anyway - and for consumers who are just looking to interact with a person. The potential security issues are monumental. #fail.
The military applications - that we know about - seem best suited in calculations and drone strikes. Most successfully carried out by skilled human operators.
Young people hate AI because they see it another way companies kill the environment.
Just like nuclear’s promise was stifled (to this day) by the Cold War, then Chernobyl (3MI) and then politics, young people equate AI with a million Chernobyls. But instead of politicians curbing AI’s use due to its danger to people and the environment, the corporate capture of government *is* cramming the tech into every aspect of life.
Congress is about 20yrs behind regulating this technology, because Congress is profiting from it. Until we as a nation muse about how to repair self-government, we’re just treating the symptoms of the problem, poorly.
It’s like asking the government for boards to build a boat and - b/c they are captured by the poster board companies - they bring you poster boards. They are in fact “boards.” But If they have no idea what poster board is, their solution doesn’t meet the reality of the situation. It’s not helping you build your boat, *and* it’s a waste of poster board.
We’re close to having Reps whose only interaction with technology *has ever been* AI.
Think about that.
I think the AI debate is getting framed completely wrong.
The real danger probably isn’t that AI becomes smarter than humans overnight. It’s that humans slowly stop developing the skills that made us intelligent in the first place. Writing. Research. Patience. Critical thinking. Memory. Attention span.
If calculators changed math education, AI could fundamentally change cognition itself.
At what point does convenience start making society mentally weaker?
What a fascinating interview with Chris Hayes. Thank you. He is amazing and brilliant.
Fascinating and timely insight into this scary technology that’s already touched a raw nerve in the workplace, thanks for this.
Thx for this, love both Thor Benson and Chris Hayes, great topic for them to chew over.