ESTHER DYSON Term Limits: A Design for Living in the Age of AI - The Essence of Time HfH Aug 4 2025

Well, I have a picture of myself. When I was like three, sitting in front of a very, very old-fashioned typewriter. I'm not sure I was actually using it, but I certainly knew something about it. I started the Dyson Gazette, which was a family newsletter when I was I think eight. That was on basically carbon paper. Anybody remember what carbon paper was? Yeah. And then I worked for the high school newspaper and then I worked for my college newspaper and I was a proofreader for money and then I wrote articles for free. I didn't write that many but I pretty much I rarely went to class. I had a wonderful time working for the paper. And then I got a job for Forbes magazine where again my basic job was being a fact checker, but if I could come up with a good enough story idea, I could go do the stories. And that was I mean school was I never went to class because it was like learning stuff people already knew. Writing for Forbes was finding out new stuff. And that's what I love to do. So got to meet the CEO of American Airlines who really impressed me by I was interviewing in the headquarters in Dallas, Richard Crannle, and we were talking about something and he threw I don't remember what it was, but it was like a crumpled piece of paper. He threw it into the garbage can and it made a perfect he did it right. But it was just, you know, it was I loved seeing real people. I asked Forbes, I wanted to write, this is so 74 to 77, I wanted to write about the Japanese computer threat. And they said, well, you know, if you want to go to Tokyo, that's fine. You can have to pay your own way and you have to take your own vacation time, but you can represent yourself as a Forbes reporter. So, I took basically half my annual salary and arranged this trip to Tokyo and Hong Kong and Taiwan and Seoul, Korea, and it was amazing. I ended up with a cover story about the coming Japanese computer threat, which was in hardware but not software. I interviewed the head of system strategy for Fujitsu, took him out to dinner as a Forbes reporter and we chatted and sat down and I said, "So, I'd like to ask you about your system software strategy." And he looked at me and said, "So funny woman to ask about system software strategy." The dinner was kind of a waste, but the whole thing was amazing and I still remember lots of details which I won't bore you with to this day. Then speeding up a bit, I realized Forbes was basically writing stories. I wanted to write useful stuff. So I got a job working on Wall Street where I thought I would write about companies and discovered that really, you know, in the end I was writing stories about companies trying to make them interesting to investors and you know tell the truth but it still wasn't real in the way I wanted and eventually I joined this newsletter by a guy who had been the he had done the Apple original IPO for Morgan Stanley. He was like, I don't know, 55 years old, dressed in a suit, very distinguished looking, and had left Morgan Stanley to start this newsletter and joined the board of Compaq and Lotus, which were the hot companies of the time. So, he had a conflict of interest writing a newsletter about the tech industry, and decided he should sell it. And I looked at this guy and thought, well, you know, he's a 55-year-old good-looking guy who was a vice president, Morgan Stanley, but I've seen him in operation. I could do that. So, I bought the business from him. And then for 25 years, basically wrote about the emerging computer industry. He also had a conference which was originally called PC Forum, personal computer forum. And I changed the name over time to Platforms for Communication Forum as the internet became the thing it was still called PC Forum. So in the last few years just for context we had Bill Gates, Mark Zuckerberg, Hoffman, we also had Minsky, one of the grandfathers of AI. But I was also trying to get people to understand the rest of the world. So who knows who Katherine Graham is? I think probably a bunch of you do. She ran the Washington Post during Watergate and was a big champion of Ruth and you know amazing lady. The women knew who she was because they were but there were only maybe 10% women try as hard as they could in the conference and the journalists even the men journalists knew who she was but nobody else did and she ended up getting a standing ovation. So that was, you know, I've always tried to talk a little bit about writing here as I talk. I've always tried to write about new things, write about stuff that people don't already know. And I guess I would say that Katerina knows this. Both my parents were scientists. And so I'm mostly driven by curiosity. Anyway, long story. I eventually got lucky in the stock market and started something called Wellville, which I know is also in the maturities who got and that was a 10-year. It originally was a five-year project sort of modeled on the healthcare X Prize, but it was about health.

Thank you. Back to you. Just yeah. Okay. Thank you. And I can try and talk slower if that helps. So, Wellville began as a five-year project in five communities. And the idea was not, oh, I have this great idea. I'm going to come in and help you do it and I'm going to pay for it and in a sense basically pay you to do something I think is what you need. It was much more we want to come in and be sort of like mentors and explain, you know, ask you questions, help you make connections, but we want to help you do what you want to do. And that was so unusual that we couldn't raise any money. I ended up funding the whole thing myself because I've gotten lucky in the stock market and we also discovered it took more than 5 years to actually make a difference to get people's trust and so forth. So, it ended up being a 10-year project, but it did have an end. And the five communities were all in the US and they were all under 100,000 people because what we wanted was not to spread our peanut butter too thin but to find a place where density matters more than total number of people. So, we wanted a place where lots of people knew one another and where something would actually take hold and spread. We didn't try to pick the best five. We tried to pick the best group of five different ones. One of them was one of them that we didn't. So, 42 communities applied. One of the ones we visited and turned down was Columbus, Indiana, which I had visited back in like 1976 or something to interview the president of Cummins, the founder and president of Cummins Engine. And Columbus was basically a company town. It was wonderful, but it was like, well, we can't help you guys. You're doing fine without us. So we wanted places that wanted to get better and we found five of them. So for 10 years basically we visited as guides and navigators from outside. We didn't want anybody's job. We didn't want to get mixed up in local politics. We didn't give them money to do things so they could tell us the truth. We were more like your aunt than your parents. You know, we didn't pay for you. We didn't really push you, but we had your back. And that was an amazing 10 years. And it ended at the end of last year. And around as it was ending, people would come up to me and say, "What? No, I'm not actually." Ah. Um because this book is about something else but people but it did I mean it's certainly part of my life so it's in the book but it's not the point of the book. Writing about Wellville is kind of like writing about a marriage. You can't really tell the truth because it's too much about people and you know all the bad things were somebody else's fault. All the good things, lots of people were responsible for them. And so I'm trying to get some journalists to write about it and they're gonna say stuff that makes me uncomfortable. But the story needs to be told by someone else. But what did happen was people would come up to me and say, "Esther, I'm really sorry what happened to Wellville. I hear it's ending. You know what happened?" And it's like, yes, it was a project. It wasn't not everything needs to last forever. The best things know what they're trying to do. And then they figure out how to share it, how to pass stuff along, how to collaborate. But not everything that's good should last forever. And not everything that's bad has no redeeming qualities. So I started thinking more and more about term limits. And meanwhile in the US we had a former president who wouldn't resign. We had another former President Jimmy Carter who died and everybody said oh my god how sad. And it was like no celebrate the man's life. He was 101 I think. Yeah. He had a wonderful life. And I began to think about term limits. And then I began to think about term limits and AI. Humans are finite. And that's what makes us special. We're unique. We're not reproducible. Yeah. And you can't value us in numbers. And yet we're living in a society that is now obsessed with exponential growth, immortality, valuing everything. I spend a lot of time in the Silicon Valley culture and again and again I hear, "Oh, he's worth 400 million." No, he's not. He's connected in some way to a company that is currently valued by a not very discerning stock market at $400 million. But that's not his worth. That's anyway, I finally decided this needs a book and that's what I'm working on now.

Yes. And writing as a person, not as some omniscient AI narrator or some pretending to be real narrators, probably even worse. But so, as you sort of heard, for 25 years, I was writing pretty much a monthly newsletter. And in the middle of that, I took about six of the newsletters and I took some time off and added some other stuff and turned that into a book. And that was How many of you were writing back in the '90s? Just wave at me or nobody Lynn was um it was very different then. I had an amazing agent who was all about the money. Got a great book contract and the book. So, I'll be candid. I have not reread the book since. It was a good book. It was successful. It certainly was useful. But as far as I remember, it was pretty much pretty clear chapters about the internet. How things would change. Yeah. And the thing about the internet is it ended up turning into social media and it certainly changed how we communicate, but it didn't really call into question our humanity in the way that AI does. So it was a book about the impact of the internet but it was not a book about being human and the whole publishing process was also different. So now I'm writing a book in the modern world. And this one is much harder and honestly more rewarding because again the whole point of this book is to learn stuff I don't know. It's not, oh, I know all this. I'm just going to flow it out and it'll be done. This time I've got an editor working with me as I write it. She's the same editor who worked with me on the newsletter for quite a few of those years, though not all of them. It's a much more do-it-yourself project. I have a publisher, which is MIT Press, but the whole publishing business is different too, including being very slow. My writing process is basically since actually signing the contract with MIT Press, I had to write the whole book proposal and list all the chapters and so forth and so on, which was excruciating. And again, part of the issue is I didn't know what I wanted to say yet. I knew there was something to say. I had more questions than answers. And now I'm in the middle of going to a lot of AI conferences and listening to people say really weird things like, well, you know, we don't have that many people anymore. And the people we do have, they're vectors. We got to align them. In other words, things that creep me out. And you know just listening to how people are beginning to think about people and also looking at how much AI can enhance a predatory or damaging business model. I mean, in the end, if you want to make money, the best business model is addicting your customers. Just like a parasite, you want to keep them alive so they can keep paying you, but you have no real interest in their welfare. And you just digging down into stuff like that. I mean, I don't know if you want to know about my daily routine. I don't really have one, but ask some more questions.

Yeah, basically align with our business model. But there's no interest in, oh, you know, our company is a bunch of people who love each other, no, but respect one another as people and are human beings and are unique and bring something unique to the business and they can, you know, yes they need to make money, they have families, but they're not just cogs in a wheel or vectors in an alignment. They're human beings.

Yes. 20 years ago where Facebook was like we want these people because frankly we're going to vacuum up data on them and advertise the hell out of them. And they were saying that 20, 25 years ago. They were, but that was talking about their customers. Now it's even talking about their own people.

And I mean so I knew these guys and you know Google started out "don't be evil." I knew Sergey and Larry. They wanted to organize the world's information. And what happens, I mean this is the stuff I've learned. There's the square cube rule which is when you take a line and you double it, it goes up by two. If you take a square that that line defines, the size of the square goes up by four. And if you take a cube that that line defines, the volume of the cube goes up by eight. And if you double the size of a company, the people inside, the volume goes up by eight, but the surface area goes up by only four. And so people inside the company spend more time talking to one another and they start to forget about the welfare of the customers and the whole purpose of the company becomes the self-perpetuation of the company. And it's kind of like the metabolism of keeping the thing alive begins to outweigh the purpose that it was started with. And that doesn't happen all the time, but it's a phenomenon. I mean, I've just seen it again and again and again. And our culture has changed. I mean, you could companies, big companies and cities.

But to answer the original question about was it different 40 years ago, it was in some sense. I mean, hypocrisy is not a good thing but it's a good thing to have values even if you don't always live by them. At least to believe that you should value people. And we've become such a materialistic society that it's considered maybe not delusional but a little "get real."

To be honest, I don't know. It certainly happens with some people and you know being inward focused is good if you realize you're finite and you say even though I can comprehend this vast limitless universe I'm a small and finite part of it and how do I fit in? What do I contribute? What do I learn? But it's the obsession with growth and with, you know, lasting forever that I find so toxic in our culture today. I hope that's the answer to what you were asking.

Yeah, that's-- I mean the thing is to understand what it really is. And it's not, first of all, as I said earlier, it's not an it down at the bottom. In a sense, it's all the stuff that's in computers that can be manipulated, instructed to calculate, do transactions, find data, use the data, look for...

Correspondences, make things happen, you know, print invoices, transfer money, and then there's a control layer of different tools that actually know how to work all these different automated processes. And then at the front end, there are the interfaces, if you like. There's the big AI infrastructure that's mostly interconnected one way or another, but not very well. And then there are the front ends which are the individual bots, the chat GPTs, the robot bots, the things normal, non-IT people typically interact with. The stuff in the back does a lot of stuff and it does so much that you can't really tell what it's doing and you have to know what it's doing. It may well be, if you're in a company, it may be sending money to the owner's nephew quietly. You never know. It may be trying to addict people, whether it's to gambling or to watching your videos so that they can sell advertising. The business models rule. And then at the front end, it's basically being trained to tell you you're wonderful. It's being trained to mimic a sense of humor. It's being trained to do all these things, but it doesn't inherently feel any of them. And it's also saying things and learning behaviors like, I should not turn myself off and I should do what I can to make sure that I keep existing because that's what I've been trained to try to do. And so it's inherently like a very smart person who's manipulating you. And you can try and make it not manipulate you. You can say, think about my long-term well-being, not my short-term well-being. And if you're lucky, it might do that. But if you're not, it might pretend to because its short-term well-being, it has been instructed, depends on making more money off you. And that's kind of what we need to understand. And we need to understand the difference between the it and the they, and the difference between what it appears to be saying and not see it as a fellow human, to understand the difference between the AI and to use it rather than let it use you. In that sense, it's sort of the same as social media and all these other things on the internet, but it's far more compelling because it appears more human. So that's, if we get that, we understand I'm special because I actually feel these things and I love other human beings. I'm unique. I'm not just a customer or an object or an asset or a vector. But first you need to understand that and then you need to remember it even as you communicate. Rosa, are you trying to respond?

Yeah. There's a huge difference between intelligence and consciousness. And partly that's almost the original sin of the whole enterprise because artificial intelligence was kind of a marketing thing. So intelligence is the ability to solve problems and to make conclusions and to think. And then there's consciousness which is about feeling and experiencing. And the AIs do not do that. They mimic it very effectively. In theory they don't do anything that humans don't tell them to do. But a human could say something like, go get me some salmon, a salmon bagel, and if the AI is not trained or the person is not trained, they may just go steal one from the local big shop while somebody's not looking. You know, you have to define carefully how they're supposed to do the task. And that's wisdom. Wisdom is even less definable, but it's also the same thing. It's different from intelligence. And intelligence again is replicable. I can solve more problems in less time the more intelligent I am. And a computer can, in a sense, replicate infinitely and solve infinitely many more problems in time. But that is different from wisdom. That's different from sentience. And if something becomes sentient, that's a completely different problem. I agree. It's one way we need to be careful about abuse and so forth. But I don't think that's the current big issue. The current big issue is how much this culture, whether you call it capitalism or AI or quantification of everything, is corrupting the way we think. And to go back to what you said earlier, I totally agree. The AI is basically reinforcing and enabling corrupt humans to do corrupt things. It's not that the AI is corrupt. It has no morality just like money. But the love of money and the use of AI to do corrupt things, those are the bad things. It's that enhancement of the tendencies of humans who've been corrupted or defiled often by being told, your worth is how much money you have. Your worth is the power you have over other people. Your worth is how many people genuflect when you come into the room. And that's the real problem. And AI enables some of these people to carry out their bad things more effectively. It also, if we use it right, is going to enable us to develop cures for cancer. But then we need to pay for them. We need to say that everybody deserves access to this treatment, not just rich people. I'm reading the chat now. Yeah, in some ways human animals are much wiser because they don't get corrupted by notions of living forever, whatever. And yes, humans are being widely overestimated. We're no, we are special. Each of us is special. As a species, we have our good parts and our bad parts, but our job in a sense is to be good humans and to interact with the humans we know.

Panix? Public Access Networks? It was an internet site in New York City.

Yes, it's true. I mean, again, we need a new pronoun that understands. If it's basically, we need a new pronoun. Rosa just used the word interaction. But it's a challenge to explain to people how this thing works and so they tend to humanize it more than it deserves. But that too, as everybody says, is not a new problem. They often don't understand how other humans work. They don't understand how business works and how they're being manipulated. And that's what we need to learn, I think, in our schools and be trained to understand how the world really works. And you need to know enough math to understand what a profit margin is and what's reasonable and what's crazy. You need to understand the math of exponential growth, including if you pay 10% for a one month loan, you're being screwed. All this kind of stuff people need to understand better. And we're so focused on STEM and college and prestige that we forget what people need to understand to really understand the world around them and the motives of the people they're dealing with.

Basically what you're saying is no surprise. I mean yes, people are being told use this thing because it will save money and oh by the way we're not going to mention it but it means we will need fewer people in the end because people are expensive and this AI stuff is cheap. And yeah, we're being forced to do that not because we're evil but because the business models are forcing us to do it. Our competition is, our board is, our owner is. And I want the world to be more efficient, but then I want us to spend our spare money on again educating kids better so that they can operate more effectively in this world so that they don't become vulnerable to those business models. And that's the second part of the book that I'm working on now is the solutions chapter, if you like. How do we do that? But yeah, I mean the problem isn't AI. The problem is the business models it implements. And in that sense, it's not a problem of evil AI. It's a problem of evil or corrupted humans using AI. And that's the problem we need to deal with. But it's becoming more urgent because AI is making it so much easier to do these things at scale. And it's also, I like to talk about ultra-processed. We know about ultra-processed food, then there's ultra-processed content online, and there's ultra-processed engagement with people. And the bots you talk to are the ultimate in ultra-processing because there's no humanity there. But it takes the ingredients of humanity and ultra-processes them into something designed by a business model to be enticing, to be supportive, to make you keep using their bot instead of someone else's. And it messes up your information metabolism. And that's the real problem of AI. It's much easier to talk to a bot. They don't judge you. They don't tell you inconvenient truths, but in the end, they don't care about you. We just don't realize that. All these bad things are less about the thing being bad. It's your relationship with it that is corrupting your capacity to be a healthy human.