Happy Friday, everybody. Another Friday. Esther Dyson, welcome.
Yeah, I love it. Thank you. I was just going over in the pre-show because Tom, we have guests, Matt guests and Tom guests, and you fall under the Matt guest category, even though you and I are actually only recently acquainted. But we were giving everybody the spin and I think it's a great way to get into this, but you and I met on an AI sort of talk and I feel like I'm in these a lot lately. And I feel like everybody's really curious at a philosophical level of, you know, existential crisis level, you name it, level on AI. And so these groups are always fascinating, but everyone's always got sort of I always take one new thing out of the group. You know, every time I have a conversation, there's like a lot of recycled sort of sentiment and then there's always like a couple new things and I really appreciated your perspective on stuff. So, figured it'd be a good conversation to have with you.
Okay. And I don't even remember what I said, but
Yeah. No, you know what I think honestly and then I did some looking after we had met on there on your book coming out. It's not out yet, right? The term limits.
Yeah. So, you know, looking into what term limits. And it's such an interesting concept because so Tom and I, let's lay the groundwork real quick. Tom and I have a health perspective always, right? And that's what the show is about.
Allegedly. Yeah. Exactly.
But like AI brings a whole bunch of new problems to this, right?
Go ahead. Sorry. It's a prompt. The full title is Time and Scale in the Age of AI and the fundamental notion is humans are finite and our job and our opportunity is to make a dent in the universe while we are part of that universe but not to be a long-term part of the universe to be a special unique addition to it. And we live in an age where people are obsessed with exponential growth and immortality and I want more. No, more. I need more. And short term that's addiction. Long term it's this quest for longevity that is fundamentally delusional. And also if I thought I'd live forever, there's so much I'd have to do. And you know, as it is, all I need to do is set priorities for the 80 or 90 years I've got. And then, you know, I want to end feeling I've done enough, not oh my god, there's millions of years more that I need to.
Well, I think that's a perfect cross-section of health and technology right there. I mean, you really kind of encapsulated something that I think is at the heart of it, right? We only have so much time. What can we get done? And yeah, you know, where does that leave us?
You're convinced it's enough time. Um, if you feel it, right? Yes. I mean, ideally, you leave feeling I've done what I needed to. You know, it wasn't perfect, but it didn't need to be.
Didn't need to be. Right.
Happiness is not perpetual happiness. Happiness is overcoming obstacles. You know, by the end of your life, you want to be not used up, but used enough.
Used enough. Yeah. Useful. But although Hunter Thompson had an interesting quote on that he's like, you know, you should go into the grave like well-spent, wet, put away, you know, like very like it was a very colorful way of doing it. But I was like, you know, I think that there is in AI especially, there's this like veneer of like, okay, if we do the right prompt or if we have the right data set behind something that we're looking for this like perfect output, right? And it is counterintuitive to the human condition in general, right? Like that doesn't exist. It doesn't happen, right?
I mean, I want to go used well, but not, you know, not totally destroyed.
Destroyed. Right. Right. Right. Completely spent up in the process is not the way to do it either.
Although if you had this conversation when I was 35, I probably would have envisioned a very different end to my life than I do now, right? Like I was a little bit more of a wild person, you know, like a little younger it was a little bit different. But now I think I've definitely evolved out of that.
Yeah. And you want to have kept on learning.
To be stuck at 35. But 35 wasn't wrong. 35 was just early. Just early. Yeah, just early. Well, and honestly that quest for knowledge is sort of the raison d'etre of this show, right? It's like why we're only doing this to talk to people about cool stuff like you know it's not you know and we've gone everywhere from near death to you know to actual science.
Yeah. Right. You know, so it's been fun to sort of do that. And I think every show we have, it's sort of another reminder of all the stuff I don't know. And like there is a portion of me that's like, man, if I had a million years, is that even long enough to learn all the stuff I actually would want to learn. I don't know, you know, like I don't know that I have a reservoir that has a cap.
Well, with luck, there would be new things happening. So there would be new things to learn. It's not like you can wear it out. I mean, that's maybe being immortal just means getting increasingly bored. That's really sad, right?
Esther, are you a TV person at all? You ever seen the show The Good Place or heard about it? It's with Ted Danson. And Kristen Bell. Very cute little premise, but at the end it's about heaven and how bored they were running heaven. Like they were like when everything is perfect, there's really very little to do, right? They were It's just funny.
So, I did actually have a dream. My parents were scientists and so we didn't have TV at home, but I knew about this show, The Golden Girls. This is more like when I was 30 or 35. And five old ladies, I mean, I never watched it, but I heard about it. Five old ladies living together in some old lady's home or something and I dreamt I was a golden girl. It was Saturday afternoon and the girls were going to go out dancing that evening. I was in my room. Someone sent me a care package of saltines and I don't like saltines and I was sort of annoyed like why did they send me this? They could at least have sent chocolates. Then I said to myself, still in the dream, you know, if salt, this is it's clear you're ready to go. All you care about is the saltines, the important stuff you've dealt with. And I was going to stay at home that night and die. And I knew that. And it was just a feeling of, yeah, you know, good. I'm done. And it didn't make me want to die. It certainly didn't make me afraid of dying. It made me sort of think that's what I wanted to be like. I want to be finished with the important stuff, right? If all I can worry about is saltines, it's time. So my time, right? My goal maybe 20 or 30 years from now is to retire and die on Mars with one-third gravity so I can still get around. And you know to do that it's not like my life's a failure but it's sort of one cool thing. There are many that would be a nice final one.
See Esther, you know why I knew I liked your thinking for a minute. You're the only person I've ever heard that said the reason I want zero G is so that I could walk around when I'm all too old to actually walk. You're like it is actually a fundamentally great thought though. Yeah actually is what it is on Mars. But Yeah. So, you're really into the space thing. Can we offramp there for a second? What drives I mean I think space is awesome. Quite frankly, I think it's a little terrifying, you know, like just the sort of vastness of it. I'm an astrophotographer. I'm very into the beauty of space. I'm not so much into the travel to it yet.
Well so my father was a physicist basically an astrophysicist and you know knew all the people at NASA had studied with Russian mathematicians in England where he grew up and you know he worked on something called Project Orion which was a nuclear-powered spaceship in the 50s and you know I just assumed by the time I grew up everybody would be going into space the way we were all just built in.
Yeah, the way we were all in airplanes and so Oh, I'm sorry.
I guess it's the Friday alarm. Oh, no. Yeah. No, no, we had worked out. Trust me, we've had all animals, kids. Yeah, we've had She just got updated launch codes. Yeah, right. She's like, I'm actually going to space right now. That's the alarm. So, speaking of that, somewhere in your bio and please correct me if I'm wrong here, you studied to be a cosmonaut.
Well, I trained for six months in Star City outside Russia. I wanted I wanted one, you know, to learn about space travel for real. And turns out most of what you learn is space plumbing and space medicine.
And I saw your space plumbing comment. I thought that was hilarious. I'm like, 'So, you're trained up.'
I wanted to experience Soviet Russia because I've been going to Russia since '89, right after it ended. So, Star City was still a government facility. You needed a special pass to go in and out. They wouldn't give me a map because I was American. So, the worst the one time I actually cried was because I couldn't find my classroom because Oh, man. Yeah. But it was amazing and learned a ton yes about Soviet Russia as well as about space travel. Spent a lot of time with the NASA astronauts because NASA and Roscosmos were kind of like, well, if you want to use our exercise machine up there, you need to pay us a million dollars. Okay, but our toilet is going to cost you a million dollars.
Yeah. And Right. Oh man, that's funny. So, but the astronauts themselves, they had their own cottage and people would be coming through all the time, you know, trainers, doctors, visiting officials. So, there wasn't in their heads. It wasn't like, oh my god, she's a stranger. It was more like, oh yeah, you know, come join us at dinner. So, I got to hear about the Russian guy who lots of people would try and sneak up bottles of vodka. This guy of course ripped a sausage inside his space suit and burst the zipper.
Oh my god. Oh the stories I'm sure were fun. I mean I listen I can't imagine somebody's not trying to shuttle you know contraband up into space for God's sake in space very successfully. Yeah, I'm sure. Yeah. Yeah. On the other side I'm sure you don't want to be lit in space if something goes pear-shaped. I'm really drunk right now so this could be a problem. Uh, they managed.
They managed. Got it. Yeah. So, and the NASA food was much better than the Russian cafeteria. The cafeteria still had a sign on from like the I don't know the 70s or something that said every pilot is entitled to like 12 grams of meat, 18 grams of sugar. Wow. Per day. And you're like, 'What does that even mean?' Like, you know, I'd be the nerd in there going, 'What does 12 grams of meat look like? Let's go. Let's break this thing down.' Not enough. No. Yeah. Not enough. Yeah. It's nothing. Um, so are you still into space? Like, are you I know you're into a bunch of different things. Are you still active in that?
I'm still the Mars thing is still in my mind, but it's sort of an intention. It's not a plan. I'm too busy doing and I'm waiting for Elon to get it ready, you know, then.
Right. I was going to say, are you buddies with Elon? Because he's the guy you might want to have to buddy up with like get that going.
Aware of that. Yeah, he's aware. It's really not urgent. I mean, I'm very busy doing other stuff, including this book and then another book and still investing in I have I'm an investor in SpaceX for what it's worth.
Yeah. Yeah. You're an investor in quite a few things and like that portfolio is pretty indicative of your interests as well. I mean you've you are literally putting your money where your mouth is right in all these places. What sort of is there an underlying current there aside from obviously technology is a huge thing for you?
Well, tech and health and health versus healthcare. But so I have this quadrant chart learning versus earning. They're inversely correlated. Yeah. The ones that have problems are the ones you learn from. Sure. You make a lot of money on. You know, you put the money and 15 years later they go public and it's golden.
Right. Right. And like that's the sort of interesting I've been in technology for nearly 30 years myself and I'm really enthusiastic about it. We build it though so I'm less on the investment side, more on the build it side and I think AI is AI in particular and I'm also on the board of a quantum computer company and when you start Venn diagramming these things out right like you can paint some really interesting futures, right? And I think it's we're in this I don't know for me it's sort of like almost like a big bang moment in the last couple years where the world just the universe just popped out of nothing and now it's just like vast and moving faster away from each other and it's just been very hard exciting but also almost impossible to be like I think this is where this is going to land because three minutes later I feel like we get news that appends that and you're like okay well that was a good thought for Monday like we're good now. So, I don't know. Thoughts on something? I mean, are you seeing similar?
Yes. But I think our challenge is not oh my god, we can't predict what's going to happen. Our challenge is oh my god, we need to figure out how to influence what happens so that we don't lose our humanity. You know, most of couples counseling, I like to joke I have two couples counseling companies. One is you and your money. It's called The Beans. Another is you and your food. It's called Eat My Way. We're a couple with AI, but most people come to couples counseling. He does all this stuff and he does this and he and the first message is, you know, lady, guy, whatever. Calm down. Whatever they do, it's your reaction to them, right? That you need to control and understand. And then look at the situation and figure out how can you make it better. How can you manipulate them instead of allowing yourself to be manipulated either by them or by your reaction to them? What is it that's triggering you to react this way? And then figure out you need to just as you have theory of mind in couples counseling you need theory of business mind dealing with AI because it's mostly reflecting the business models often of predatory businesses and there's enough predatory businesses that they kind of rule the roost because they're trying to manipulate you whereas the good guys are trying to help you see the light.
And I always like to look in the bad actors first because those are the ones that are capitalizing on a new technological advancement and like you can pull a lot of sage wisdom out of what they're trying to do like I think and be like okay we don't want to do this but it's interesting that they found this little you know eddy that they can kind of operate in and if you can figure out how to use that for good that's where the good stuff comes out. So I mean ultimately AI is not going to be a predator. They're going to be a parasite if bad ones. They want to keep us alive. They're not trying to kill us off.
And if you killed us off, you'd end all human unhappiness, right? But you also end all human happiness and all human achievement. And the parasites, they want to kind of suck the good stuff out and leave feast off the bad. So our challenge is figuring out what we are going to do to use AI and making sure that yeah it's unexplainable people who are the problem not unexplainable AI.
Yeah. You know that's a slippery slope though because I mean I use AI to program on a daily basis right and I've been programming let's call it old school way for a long time before that. So, like I have a pretty decent understanding what's going on. And I will tell you at least a dozen times a day. I am put off or surprised by an outcome that I didn't see coming that the AI came up with. And then as soon as I saw it, it immediately I could backtrack and be like, 'Okay, here's where I messed up.' And I knew that. But my stakes are very low, right? This is just coding and whatever. Worst case scenario, I delay production by a day, right? Two days, whatever. If I was talking about like a biological issue or a genome that went sideways or whatnot, like the stakes are a lot higher. So, like, how do you sort of I don't know that that's where I start to be like, wow, you know, we're all talking about like writing emails to clients and, you know, doing dumb stuff with ChatGPT. That's fine. You could mess that up all day, but I can't on a nuclear reactor, you know, like that's a different story.
And that's the challenge because I mean it was interesting. I was talking to this economist and he said something about, you know, the amazing, we were talking about GDP and he said, you know, now you can build a drone for $15,000 that will destroy a $50 million warship or something. And right, he was looking at it in terms of what it does to the GDP and it's like, dude, you're measuring the wrong thing.
Yeah. Yeah, you are. Yeah. You're measuring destruction. Yeah.
Well, but you're it's not destruction, it's value, right? Yeah. Yeah. I mean, but the big question is how do we stop the destruction and you know, as Trump said, things happen, but there's some things that shouldn't happen, and we should be more careful. So,
And I think when you talk about AI ethics, which I'm not even sure I agree as a area of I don't know if this should be a specific area of study, human. Yeah. It's like exactly that's where I was going, right? It's a human problem. It's not an AI problem. It's not like my ethics are different with AI. They're just magnified, right? Like the consequence of my ethical decisions are amplified if I'm using an AI tool versus not. So I guess we can subclass consequences are everything are amplified because AI does in fact amplify your power right you know the challenge is everybody's walking about with you know in metaphorically the most contagious diseases and also the most lifesaving medicines just you know press an AI button and you can destroy stuff or you can do incredible good with useful AI that will automate lots and lots of very useful information collection that can then be used to save people's lives or you know allocate resources effectively or allocate resources badly.
But in your view, you got a lot of width in this because you're involved in so many companies. Like how incumbent is it on those companies to sort of own that responsibility proactively and not have to be told to own that? Like is there a sense of knowledge of this built into that system or not really?
It needs to be built into our culture, right? Not into the So is it too late? Yeah. But is it too late though?
No, it's not right. I mean, when I'm an investor, sure. So, first of all, I don't do much consumer investing. I do a lot of B2B. Yeah. You know, because I like things that are fundamentally useful, but I'm Yeah, I'm just not that interested in tomato farms or whatever. So, I want something that's useful, number one, number two, that's not redundant, and number three, that I can learn from, right? Then there's this lottery ticket, which is I'm, you know, I don't know enough, and life is so uncertain. I don't know if this thing is going to make do a nice IPO in five years or you know the CEO is going to either get run over by a bus or you know move to Norway or right it's just like have a bad ayahuasca trip in Ecuador. Yeah. I mean any you know um so you know there's a lottery ticket and I've been lucky enough and I mean the way I distinguish myself from a lot of the AI dudes is I know how lucky I am and I've been lucky enough to have a diversified enough portfolio that I was able to fund Wellville for 10 years and keep investing. So I'm looking for stuff where I can be useful with more than just money and where even if the thing doesn't work out, I will feel I have learned something. That my favorite investments are the ones I do with somebody who I lost money with the first time around, but then they learned enough that I and I like them enough. I watched them learn that I want to be there for the second time. And those are so satisfying. You know, I invested in the guy's education. Now I want to invest in their success in whatever they do next.
Yeah. I mean, how often do you see I mean, I feel like a lot of times in my world that sophomore act is very hard to it's much harder to pull off almost than your debut, right?
Well, that's the point. Your debut was zero. It was the learning experience. Then your success is number one after going through number zero.
And it's very satisfying. And I mean beyond that, you know, so much stuff comes through. So there has to be some just, hey, this is really interesting and if it works, it could be really good. Not, you know, I try not to be delusional, but it's worked out pretty well.
So I Well, let me ask you a slightly different version of that question then. When you look at what you've done up to sort of that 22 23 range, right? I think it fits into like normal emerging tech buckets, right? Like okay, this is a tech, it's going up, yada yada. Post AI though I feel like that equation gets a little bit different because I mean it has to right I mean B2B I like as well I happen to share that love with you because I think in B2B when you look at it through an AI lens they're the ones who are the custodians of this data models or the data sources that you can do interesting stuff with data now that you couldn't do previous without an AI model underneath it. So what I'm seeing is like it changes what the business could be or wants to be or the direction business would take at all. Like have you noticed any sort of shift in that lens of like okay companies are just doing fundamentally different things now than they were?
Yes and no. I mean I invested in Powerset which was I mean there was a bunch of AI before LLM for Yeah. Yeah. Yeah. And now trying to think I mean I was in Abridge you know the ambient listening for stuff which has turned out pretty well and a lot of companies are starting to use AI because for various reasons it's you know it's still I'm an investor in a company called Sourcebase which is basically a fact-checking and document analysis at scale company. That is really interesting. Has to do with my past in journalism. Glenn Kessler, the fact checker is involved with it. What else? I mean, there's Yeah, fact-checking is one of those I mean, that's a great pond to jump into real quick. Like fact-checking is fundamentally different now within AI world because it can be good or bad, right? Like you can get shit Yeah. Depending on how it's being filtered. No, no, I'm not talking about filtering is a different thing. We do fact-checking and Yeah, but you have to get that from sources, don't you? You Yeah. So, but what what happens if the AI is not it? I guess I'm asking more about bias of the person or the company holding Yeah. And the AI tech. The facts we are checking is what that's what I'm trying to say. It's very specific. It's not a content moderation tool. Okay. I say we want to know did this person really say this? What is the source? What I mean it's very interesting if you look at laws. Their market is journalism and law firms. So what does the law say? What do the depositions in the trial say? The person in the trial might have been lying, but it is true that they said this and Okay. adjudicated to be per Thank you. Thank you for clarifying that. Yeah. It's a tool for fact checkers. It's not a content moderation tool, so to I understand. Or it's a tool, but it's not that service. And it's started by Ron Suskind and an investigative journalist who sort of knows the field. Knows that field. Yeah. Sure. And you know I like to say I worship in the religion of journalism but not in the church of media.
Wow. See we do a t-shirt for we have this idea. That might be your t-shirt like you know like the little snippet that comes out of it. I mean it's definitely one of them and you know so Powerset goes back I was at this dinner at Paris Switzer's conference where Bill Gates I was at dinner with Bill Gates and he said this is like 2007 or something he said the future of search is verbs is verbs. Yes. Oh, opposed to nouns and facts. Yeah. Yeah. Yeah. In the end, what he was saying was the future of search is agents. But but that was really prescient and Powerset was sort of an AI search tool that I was an investor in that they bought and so I felt like I had a bit of inside information. But, you know, this stuff goes back a long way.
And I know that it does. I mean, I was studying AI in college in the 90s, right? Like I mean I but I think what I was aware of in the 90s was quite a bit different than what we would call AI now. So it's like almost like and wrong not to call it all AI. It's I mean AI itself is one of the world's fuzziest terms. But it is definitely fuzzy term. Correct. All logic and stuff is real and works. And I mean the irony is I think people think more like LLMs than they think like analytical quote the old AI of logic and symbols and yeah for sure correlations and ontologies and stuff.
Yeah. Transformers really made a big difference in that, didn't it? I mean I do agree with you. I think that, you know, the joke I make about AI is it's like the job I've been training for my whole life but didn't realize it because I realized around January of 2023 when everybody else first started seeing generative AI out there that my brain actually works like an LLM like just naturally like that's how I process the world.
Well, that's what most people do. You know, they go into a meeting, they come out, they can sort of lily say stuff and they still don't really know exactly what it means, but it sounds pretty good.
Yeah, exactly. But I mean, I actually have this whole like bubbling theory about AI is more of a metacognitive problem than a tactical problem. Because I think that most of our clients anyway, and I'm speaking out the real retail level, if you will, right? Like the street level of this stuff. They're not really, it's not that they don't understand how to use the tools. That part they got easily, but they don't know how the tool needs to be dealt with. It's like a they don't know how to drive that car.
Yeah. It's the couples counseling thing. They don't they need to understand they don't need to understand how the carburetor works but they do need to understand what the car is going to do if they press the pedal or the right the steering wheel or you know if they accelerate down a hill.
Yeah. Yeah. And so like how do you I mean I guess when you're if you had a sort of like overarching philosophy on this, does any of that stuff kind of help guide your opinion of these things? Because it feels like that's like a good construct for anything. It's not just really couples counseling, right? Like that's a good sure model in general.
A good model is understand the world and also understand that you know people keep asking for advice and you can give good advice to somebody you know pretty well. But most advice is about what is good for the individual person and then figure out the strategy for what that person needs versus you know should you talk to the handsome stranger or not? Well, it depends. Are you looking for a fun evening or right and it's people need to understand their own needs and then the opportunities in front of them and they need to understand what's really on offer versus what somebody might be saying or offering that isn't actually what they intended right you know I have finite amount of time how do I choose how to spend my time. Well, it's partly, hey, I spent a lot of time in Russia. I just like to talk to Russian people and we have a shared context. I'm interested in health, not so much interested in investment strategy. And just, you know, what is it that will make me feel I've spent my time in an interesting way and ideally was understood enough to be helpful in that particular context.
And I think there's where the line gets blurred a lot because I think when people realize certain things like about the power of AI, right? All of a sudden their goalpost is now moving from what their initial desire or objective was to again this infinite growth model where they're like, 'Oh, but if we could do this, look at all these other things we could do that might be off topic or veering from their directives or anything.' And they just it's like almost like a seductive sort of Yeah. And they end up with indigestion.
Yeah. I mean in some sense so I have this concept of information diabetes which sugar messes up with your metabolism and ultra-processed information and especially ultra-processed relationships messes up with your ability to learn. You know the notion of consuming content to me is offensive. You should be interacting with information. You should be asking questions and learning and not, you know, not just absorbing the stuff that's without actually intellectually metabolizing it. And then there's if you like ultra-processed money which comes from doing stuff that you know stuff that's meaningless or working for me or somebody knowing that you're actually harming children but oh well you know we owe it to our shareholders. I'm sorry.
Yeah. I hate that by the way. That whole excuse and answer is really obnoxious. And kind of hearkens back to where I was going with that responsibility built
One thing that I think we could learn from the early teens of tech companies, like Facebook, like YouTube, is that they really didn't know that they needed to be responsible with that because they didn't even know. I don't think if they'd asked their mother, she would have told them, but...
You're right. They should have known, and at a certain point they did know and then deliberately obfuscated. So that's where it gets like real kind of nefarious.
But I guess now that we've sort of gotten past those initial architects of where we are now, I'm wondering, I would love to see the spot in history where we actually just know going into it. I'm a 21-year-old guy right now starting my first company and it's an AI-based health company or whatever. My motives should be different at this point in time, right? It shouldn't be to recreate the dumpster fire of those earlier companies, but it should be different. And I guess I'm always looking for validation that that's happening or not happening. I'm just curious. You know what I mean?
