#782: Legendary Inventor Danny Hillis (Plus Kevin Kelly) — Unorthodox Lessons from 400+ Patents, Solving the Impossible, Real Al vs. “AI”, Hiring Richard Feynman, Working with Steve Jobs, Creating Parallel Computing, and Much More

26:46Dana, you run an invention company right now. And is that something that you also apply to your clients as well as when they come in? Are you trying to tell them or help them make a story out of the inventions that you were working on? Do other people really care about it as much as you do? People care about it, but I don't necessarily talk to the clients about that. Because depending on their perspective, the thing they may care about is financial sales or the reliability of the machine or the rate of production. They have something that they think they care about, but very often behind it, there has to be some story for it to make sense of the people who are operating the machine or buying the product.

27:31And so I'm thinking about it that way. I'm not necessarily explaining it to the client that way. But yeah, I would say pretty much everything I do is influenced by that way of looking at things. And it causes you to look at the experience of it rather than the engineering of it. And actually, it's interesting. I mean, I was really lucky because I got to work with Steve Jobs when he was first making the Macintosh. And it was during that period he had been kind of kicked out of the campus of Apple and was in an apartment with a few pirates making the Macintosh. And at the time, this was before I had been to Disney and before I learned this.

28:11And it drove me crazy because I got called in because I knew how to make chips. And Steve wanted to make a custom chip for the Macintosh initially. And it wasn't going to happen. And I was the one that had to look at it and tell him it wasn't going to happen in the timescale he wanted, which is not a fun thing to do with Steve. And so I was like, well, you know, this is just reality. No matter how much you yell at me, it's not going to change. but you know you've got this simulator that Andy is making all his software work on so why don't you just sell the simulator and he basically blew up at me and I was missing the point of everything but what I realized looking back at it is Steve was like really wrong about a lot of technical things but what he was really right about was the story of how people would relate to the machine He had a vision about that that other people didn't have.

29:09And in some sense, it didn't matter that he was wrong about a bunch of technical things because the story was so correct. And the way that you related was so correct that all the technical things were fixable. But if you've been wrong about the story, no amount of technical excellence would have fixed it. And I think I only understood that in retrospect after I kind of saw people relating to the Mac. So, Danny, you were doing AI for a very long time. You made the first computers that were in parallel. You called it thinking machines. Our slogan was, we want to make a machine that will be proud of us.

29:46Right. Exactly. So, what is the story on AI that we're not getting right now? There's a lot of focus on all these LLMs and neural nets, which are very old, actually. What do you think the story is? and what's the story that we're not hearing? Well, I'll tell you a story I told a long time ago about AI, which I call the Songs of Eden, which is, in some sense, it was a story about where human intelligence came from. And it was a story about a bunch of monkeys that kind of grunted and repeated each other's grunts. They sung along with each other, didn't really mean anything, but they started noticing the mood of the other monkeys by the grunts they were making.

30:31And their brains began to develop to keenly notice the moods of the other monkeys because they're social animals. And so they started evolving the ability to distinguish sounds. But at the same time, there was another kind of thing that was evolving, which there was no name for it then, but we call it memes now, which is things that got repeated and that were very catchy tunes and things like that. And so there was this kind of co-evolution of these two things. One of them with the monkeys got sort of better and better at distinguishing between the grunts. And the ideas got better and better at helping the monkeys because that's how they got repeated.

31:15And so we're sort of a symbiosis of those two things, of the monkeys and the song. So this song, in some sense, evolved into human culture and human ideas. And we evolved into the monkeys that were able to hold those ideas and transfer those ideas. And so I kind of told the story that that was the way that human intelligence evolved and predicted that that might be the way that artificial intelligence evolved. That we would build machines that were sort of powerful enough, but then we'd kind of infect them with human culture. Now, the internet didn't exist then, but I wasn't quite sure where you were going to get the human culture or how you were going to do that.

31:55But I think that's sort of what's happened is what we've got is not so much artificial intelligence, but we've sort of got a substrate on which human intelligence can live that's not human. And the human intelligence is all of the things that was learned from all the data that we train on in some sense. And this is the early stages, so you might say it's just kind of imitating right now. But it's got so many examples, it's really good at imitating. And that's always sort of the first part of intelligence is imitation. I mean, child begins with imitation, and then they understand more and more.

32:29So I think we're in that imitation stage right now where we've built machines that are able to do a pretty good job of imitating, and they'll go beyond. They're just beginning to peek beyond the imitation stage, reason, things like that. But in the end, it's not really an artificial intelligence. It's human intelligence on an artificial substrate. That's a new phrasing and lens that I have not heard before. And we'll probably come back to AI, but I want to maybe ask a 30 ,000. And that's not the only possible form of AI. Oh, that's what we are. We will almost certainly come back to that. But I want to zoom out for a second.

33:07You said earlier at some point, I never really had a plan. but there are people who don't have a plan and have no direction and end up, I think as Mark Andreessen put it once, as a sort of a rabbit pivoting every 10 seconds going a different direction in the maze and not making any progress. Clearly you are not that rabbit. So it seems that there is some underlying scent trail or way in which you choose projects or what you will do next. How do you do that? What is your guiding sense of how you choose where to direct your attention? And you said at some point you want to do anything other than the computer stuff.

33:50So you shifted to the imagineering, right? And there were other lifestyle factors, but I'm just wondering, broadly speaking, how do you choose what you're going to do next? And then once you decide on that, I'm stealing from Kevin here, but how do you proceed once you decide that you want to get into a new field? So first of all, since I sort of love the process of invention, I have to say that I think it's a misunderstood process because what the inventor does is actually a very small piece of it. What society does is it creates these preconditions for invention. And once those preconditions are in place, then it's just a matter of sort of putting together the puzzle pieces and making it work.

34:36So I always love to see those moments where all the pieces are around and somebody just needs to. And usually they're not recognized because they're looked at by different people. They're in different disciplines and things like that. Could you give an example? A perfect example was parallel computers. It sort of now seems totally obvious, like how could anybody have not built parallel computers? But at the time, there were some pieces that weren't quite there yet until you got the ability to put multiple processors on a piece of silicon that required a certain level of complexity of the silicon production technology.

35:14So nobody had done that. So I made the first multi-core chips. And it was also, there were proofs that computers became less and less efficient the more processors that you added to them. There was something called Amdahl's Law that was how IBM basically pooh-poohed parallel computers, or people like Cray said you didn't need them. Danny, just for people listening, could you define what parallel computing is? Yeah, parallel computing is what you do in the cloud when you have lots and lots of computers that you put onto a problem. Or you do it on a single chip now that is a multi-core chip that has multiple processors on it.

35:54It's so obvious now. It doesn't seem like an idea. But just to be clear, the traditional way was you have a sequence and you would just do one thing at a time. That was the standard way. And this is, you're going to do things multiple at the same time, which is very complex because you have to do all kinds of things to coordinate, to converge. The complexity is incredibly more difficult when you're doing things in parallel. Yeah. And also there were all these kind of reasons why people thought it was impossible. hard to believe. And it took a while to understand why they were wrong. It hadn't been done, but I knew it was possible because I knew the human brain worked.

36:34So the human brain has these very slow components, much slower than transistors. So I was like, well, maybe they won't be general purpose computers, but if you're going to make AI, certainly that's the way to do it. So I had some confidence that doing the thing in the unexpected way was going to work. And the preconditions were there that I could design CMOS chips, make those work, build them. Compiler technology was at the right place. Television cameras were starting to produce digital things so that you could have eyes, you know, digital eyes on machines. So all the preconditions of converting audio to bits were there.

37:14So all the pieces were kind of coming together. And the only reason it wasn't being done was this sort of prejudice that it was impossible, which was sort of created for commercial reasons, I think. And so it was out there to be done, and I had a reason to believe that it would work. So that's a kind of example of sort of seeing the preconditions are all there. Now, that required an incredible amount of work on tens of thousands of great engineers to get to that point. So in some sense, all I had to do was take advantage of each of those pieces that were already there and put them together. Right now, it's very formal what we decide to work on because our partners that apply to mention, we put three tests on things.

37:59One of them is that, you know, one of the senior partners has to be really excited about it, which is usually because it has some big impact on the world or sometimes it's because it's just really cool technology, but usually it's because they see it has potential for big impact. And then the partners that are not the one that's the most excited about it, and often I'm the one that's excited about it. So the other partners get to look at it and say, does this make any financial sense? And it can make financial sense because we're guaranteed not to lose too much money on it, or it could make financial sense because there's a small chance of making a lot of money on it, or, you know, you have to have a portfolio of those things.

38:48But that sort of has to be evaluated by different people than the one that's most excited about it. Good idea. So there's a kind of practical aspect to it that I probably didn't do in the early days. I tended to do the things way too early before they made any financial sense. So now we have that bit of discipline added to it. But then the third thing that we do, and this is the hardest thing to do, and we call it the non-redundancy criterion. Because by then, you've got a project somebody's excited about, and you know it's going to make money. And why would you say no? Well, the answer is, you would say no if it's going to happen anyway.

39:27In other words, if somebody else is going to do it, why should you do it? You're wasting your time. It's like, there's some reason nobody's doing it. And, you know, in the case of the parallel computing thing, it was this crazy thing called Amdahl's Law, which seemed to prove that it was impossible. And so you have to say there's a unique reason why we're going to do this. We're doing something that won't get done otherwise or won't get done for a long time or won't get done right. So we only take projects like that. And that one is a tough self-discipline to enforce, but we do do it.

40:08Just a quick thanks to one of our sponsors, and we'll be right back to the show. This episode is brought to you by AG1, the daily foundational nutritional supplement that supports whole body health. I do get asked a lot what I would take if I could only take one supplement, and the true answer is invariably AG1. It simply covers a ton of bases. I usually drink it in the mornings and frequently take their travel packs with me on the road. So what is AG1? AG1 is a science-driven formulation of vitamins, probiotics, and whole food sourced nutrients. In a single scoop, AG1 gives you support for the brain, gut, and immune system.

40:43So take ownership of your health and try AG1 today. You will get a free one-year supply of vitamin D and five free AG1 travel packs with your first subscription purchase. So learn more, check it out, go to drinkag1.com slash Tim. That's drinkag1, the number one, drinkag1.com slash Tim. Last time, drinkag1.com slash Tim. Check it out.

41:13Quick question on the parallel computing example. So you mentioned, if I'm getting the pronunciation right? Amdahl's Law, which indicated it was impossible. You mentioned as a perhaps counter example, obviously in a different substrate, the human brain does it, but were there other pieces of evidence that led you to believe, given the constraints of the technology at the time, that it was possible? No, I think that was the one that really made me have faith in it. There's something wrong with Amdahl's Law. Because I actually, at the time, I couldn't tell you what the flaw was in the proof of Amdahl's law.

41:49It was pretty convincing. And now I can go back and tell you that the flaw was it assumed that you just kept doing the same size problem. But of course, if you have a bigger, faster computer, you do a bigger problem. And so you don't just use the same problem. That's the reason cloud computing works and these giant parallel machines work is because you use gigantic problems on them. And if you tried to use the little problem that you were running on a single computer, or they wouldn't be very efficient. But anyway, I didn't see that flaw at the time. I'll give you an example of something we're doing now that sort of fits that.

42:25Cybersecurity. Everybody agrees cybersecurity is a mess. Ransoms are going up. Nobody knows even how big it is because everybody hides the break-ins and so on. But everybody agrees it's getting worse rapidly. And the defense is losing against offense. And if you step back and really look at it, the reason that it's bad is because the internet was built on a sort of flawed foundation. The basic idea of IP, internet protocol, was that you'd look at a packet, and if it wanted to go someplace, you'd move it in that direction. And it was explicitly stated in the design principles that security was not the problem of the networks.

43:14because security was the problem of the thing that got the packet. And the packet can kind of claim to be from anywhere. So you get this flood of packets being delivered to you. You have no idea really where they came from, and you have to kind of guess which are the good ones and which are the bad ones. Unmarked packages from everywhere. Yeah, exactly. And so you can come up with very clever ways of guessing, but then as soon as you do that, somebody can come up with a very clever way of getting around your heuristic of guessing. and ultimately the attackers have the advantage if the packets are anonymous.

43:49And so clearly the right thing is to have the network have a policy of what it delivers. In some sense, we do that a little bit with firewalls. If you try to see, oh, this is a bad packet, I'll cut it off or something. But again, you sort of have to guess where it came from or what it's doing to do that. So I got together a bunch of people that had been involved in the early days of the internet and had built all kinds of things on top of it and had used it for very high security applications and things like that and said, how would we have designed internet protocol if we knew what we knew today?

44:28If we actually understood what cybersecurity was like, how people were really using computers, things like that. And, you know, that's a non-starter for anybody, any normal commercial company to ask that question, because obviously you're not going to replace internet protocol. But it was a great hypothetical that kind of captured a bunch of very smart people's imagination. And we got together and invented something called zero trust packet routing, where every packet carries a kind of a passport and a visa that proves it has permission to go where it's going. So the network itself kind of has a policy.

45:09It doesn't try to deliver everything to everything. It delivers things that are allowed to go to where they're allowed to go. And then it turned out after we built that, that actually we looked at it and said, you know, we could build this as kind of an overlay to start on the current internet. So people are starting to do that now. Oracle just announced a product that their cloud is going to start using this protocol. all. And so I think that that's going to cause a big shift in the internet eventually, because it sort of gets at the foundational problem that sort of no sane company would have looked at as a business opportunity.

45:51And it probably isn't a business opportunity, because it probably has to be open and, you know, a standard or something like that. But I think it's going to actually help the good guys and actually make the world a better place. What do you think, Danny, and I'm going to keep this pretty broad, but the future of cybersecurity potentially looks like? And you can choose the time frame, five years, 10 years, three years, whatever you want to paint. But I mean, there could be the dystopian sort of Cormac McCarthy version of what cybersecurity looks like. Then there's the utopian kind of island Aldous Huxley version.

46:25Then there's probably something in between. I think you will actually shift to this. And what this is, there's sort of two completely different layers of cybersecurity that have nothing to do with each other. So you'll have the kind of layer that we have right now that we depend on, which is the endpoints kind of protect themselves. So, you know, they force you to log in and identify yourself or exchange certificates. That will all still exist. But completely independently of that will be something like, it'll be zero trust packet routing or something like that, where the network itself is kind of aware of who's sending the messages, what permissions they have.

47:06And it's actually aware of the identity and sort of strongly authenticated identity of it. And that's a completely different system than we have now. So I think that two-layer system, actually, the defender has the advantage instead of right now, the attacker has the advantage. That's cool. So, Dan, I love your idea of the three criteria for deciding whether your company does things. I assume maybe that's also your personal one, too, where it's, am I excited? Is there some viable means to keep it going? And then thirdly, would it happen without me? That last one supposes a certain amount that you know something or you have some ability that other people don't have to do it.

47:49And so going back to you with the chips, like you're a young graduate student. Oh, I'm just going to design a chip. I'm going to make chips. That requires either a lot of knowledge about chip making. It's not every graduate student who says, I can make a chip. And how am I going to make? I mean, so how do you enter into this area of chip design that you don't have, but you're confident that you can make a chip? So tell me about how you get there. Maybe it just requires a lot of overconfidence. So, of course, it always turns out to be harder than you think. Right. But I guess I am gravitated toward learning new things.

48:24And I've also developed the ability to kind of search out the people who really know the thing. and hang out with them. So find the people that really understand it, hang out with it, learn it. And so it's not that I know things other people don't, but maybe I know a different combination of things that other people do know. And I'm kind of willing to learn the things I don't know and kind of have a technique of doing it, which is hanging out with people who are smarter than I am. Let me open that up a bit. So I feel like there are many different species of hanging out with people, right? So I could have as many group dinners with wine and banter with experts in AI as humanly possible.

49:15And who knows? Maybe I'd have a hangover and a few great ideas I thought were great, at least, jotted down in a notebook. Could you give a few examples of how you interact with people? Maybe, because the name was invoked earlier, you could start with Marvin Minsky, and maybe your first meeting, because maybe that'll lead us somewhere interesting. Okay, so Marvin Minsky is the person who named artificial intelligence. He and John McCarthy kind of founded the field. So when I went to MIT, I kind of knew that I wanted to do artificial intelligence, and I had read about Marvin Minsky. So I knew I wanted to work for Marvin Minsky.

49:57and I had to figure out how to do it. And the AI lab was sort of locked up in technology square. It was hard to even physically get into it. And you couldn't get into it unless you had a key and you couldn't get a key unless you had a job there. So I decided, okay, first thing, I got to get into the building. So it's like Ocean's 11. Yeah. Well, I did slip in a few times, but that wasn't going to work. It was pretty high security. DARPA was paying for all of the lab, and I got their proposals and the proposals to NSF. And I read their proposals to see what is I could possibly offer here. And I read their proposals.

50:40And you read those proposals because those were publicly available in some format? Because they were government funded? it well they were actually in the library that was in the lobby of the building which you could get okay here we go okay so i'm in the lobby so i can read the proposal marks nose and mustache cup of coffee don't mind me right yeah so i read them and and they came across there was one thing where they said you know we think it's actually important that young kids program computers and And we think even kids that can't read and write should program it. We don't know how to do that yet, but we think it's important.

51:18I was like, oh, they don't know how to do it yet. So I will invent a way for kids who can't read and write to program computers. So I went off and I invented this picture way where you manipulate blocks. And then that was enough. That proposal was enough to get me an interview with Seymour Papert. Who is Seymour? It was the first one that did sort of educational computing. And he had the gates to the kingdom in terms of getting you into the building? He had the gates, right. He was inside the kingdom. So what did you say to this person? Were you like, I was cruising in the library. I came across this.

51:54It seems important to your funding that you develop X. No, no, I didn't give him all that backstory. I just said, hey, here's a really cool way I've come up with for kids that don't know how to read and write to program computers. He's like, oh, I've been looking for that. Okay, got it. So it wasn't like a million things that were in these proposals. He would recognize. Yeah, he immediately recognized. That was something he wanted. What a coincidence. I knew who to go to. Right. It just sounds a little bit like logo. It was. He invented logo. He's the guy that invented logo. And what did you invent?

52:31I invented something called the slot machine, which is a way of programming logo with pictures. So you could arrange pictures. And actually, the squeak language is kind of the electronic version of what I invented. I invented physical things that you put together to make a logo program. What is a logo program? It was an early computer program language for kids. I got it. I got it. Programming language. You would say, move the square around in a circle or something. Very, very simple. Of course, people thought this was very impractical because we had to convince people, someday every school will have a computer.

53:05that was considered very implausible at the time. But that was our stretch idea there, that someday every school would have a computer. So I had a physical way that you could kind of program it by putting these, and so I got hired, and I got a key to the building. Okay, so now I'm in the building. Phase one complete. I'm building it. I go up to Marvin Minsky's office. Marvin's never there. But after a while, I make friends, and, you know, like, where's Marvin? And it's like, oh, he comes in at night, and he's working downstairs in the basement. And he's building something which is a personal computer.

53:42Whoa. And I was like, okay, that's great. But I had the key. You couldn't get into the basement without the key either. But I had the key. So sure enough, I go down there at night and there's Marvin Minsky with his graduate students around him. You know, in those days, they were wire wrapping machines and there were the diagrams lying around all over the place of the computer. And of course, I'm too shy to talk to Marvin, and I don't really have anything useful to say to Marvin. So I just sort of look around, and I look around at the diagrams of the computers, and I notice a mistake on one of them.

54:18I go up to Marvin Minsky, and I say, you know, I think there's an error here. And Marvin looks at me and says, oh, yeah, yeah, that seems wrong. Fix it. And it's like, well, do you mean fix it on the diagram? I mean, he's like, no, fix it on the diagram, fix it on the machine, you know. just fix it. And so, okay. So then, I look around and I find something else. I go to Marvin with him. I say, don't ask me every time. Just like, fix the problems. So, after a while, I just started working there and I think Marvin just sort of assumed I worked for him. And then eventually, you know, everybody, I mean, after I'm there for a few weeks and everybody else would get tired and go home in the morning and then Marvin, at some point, he was like, We're going, you need to ride someplace.

55:03I'm like, I need to go back by and where he's like, yeah, why don't you just like crash in my basement? So I kind of moved into Marvin's basement. I mentioned to Marvin that I didn't actually have a job and he gave me one. But I still sort of had my job at Logo working for Seymour. But, you know, that was how I got into the AI lab and started working for Marvin Minst. Hanging out with people. Yeah. So there's one other example that you're going to give Danny outside of Marvin. Learning by hanging around people. Yeah. Learning by hanging around people. or like the Hillis method of hanging around with smart people.

55:33Well, I was going to give the Feynman example. Oh, great. Yeah, let's do that. That was a fun one, too, because I had met Richard Feynman at a conference, and we had really hit it off. And for people listening, if they don't have any context, just a brief overview of Richard. So Richard Feynman was the Nobel Prize-winning physicist that invented Feynman diagrams and quantum electrodynamics and a lot of other basic techniques that everybody uses in physics. And one of the youngest people that was on the Manhattan Project. So totally brilliant, but also just a lot of fun. And we really hit it off, liked him a lot, and thought he was super smart.

56:15And so when I was starting thinking machines, I wanted him involved somehow. So I went to visit him at Caltech, and he invited me to stay at his house again. And I explained to him building this parallel computer and said, you know, do you think you have any students that we could hire or hire as interns or something like that that might be interested in working on this? And Hyman said, you know, no. He said, none of my students are crazy enough to work on something like that. I mean, that's nuts. And it's just a kooky idea. That's what he said. that's a kooky idea. And he said, wait, he says, actually, maybe there's this one guy I know that would work on it.

57:04You might hire him for a summer job, you know, but he doesn't really know much about computers, but he's a really hard worker. And I think he's pretty smart. And I was like, okay, well, that's good enough recommendation for me. What's his name? He said, Richard Feynman.

57:28So he came, he was actually showed up on the first day, which is the middle of summer, and he shows up. And I, of course, starting a company, you've got like worrying about closing financing and things like that. And I wasn't really thinking like, what's everybody going to actually do when we get all this set up on the first day? And he shows up the first day, he salutes, he says, Richard Feynman reporting for duty sir you know what would you like me to do and I'm like oh I hadn't really thought about this and so I think we're taking I say oh how would you do quantum electrodynamics on a parallel computer and he's like that's what you want on the first day it's like is that really what you need doing and I was like well actually the truth of the matter is we don't have any like pencils or paper nobody's gotten any supplies he's got great I'll be quartermaster.

58:17And so he goes out and he gets the supplies. That was his first job. But he kept on, you know, every summer he would come to Thinking Machines. And of course, we got more serious tasks. And he actually started the first quantum computing project at Thinking Machines. So we were, again, a bit ahead of our time on that. Probably way too ahead of our time. So, Danny, what I find interesting in your approach of hanging out with people is When you're going into a new field, you're not like reading the papers. You're going to talk to someone. Do you learn best by conversation and listening? Or do you learn by reading some fundamental papers?

58:57I read enough papers that I have questions. You know, because you're wasting the time of a Marvin Minsky or a Richard Feynman if you don't ask them something that makes them think. And so I would say most of my learning was from the people, not the papers. But I always do homework beforehand to sort of see where the interesting questions are. And in some sense, that's easier to do when you're coming into a field from the outside because the people inside the field have already kind of settled on a set of questions is the important question. But if you don't know much, it's sort of easier for you to see sort of the big holes that are missing.

59:37And sometimes your questions are dumb. And, you know, they explain to you why they're dumb questions. But sometimes they're like, yeah, that's actually a pretty interesting fundamental question. And, you know, if you can hit on one of those, that gets you into a conversation. But ultimately, I learned much more from people than from the papers. How did you, Danny, get into biotechnology or just the biological sciences? Once I set up these invention companies, people would start to come to me with problems that kind of as a last resort. The engineer of last resort. You want to solve a problem and nobody else can solve it.

1:00:19So that was the way I got into biology. It was a doctor named David Agus. It was an oncologist who was really frustrated with his abilities to diagnose and treat cancer. He came to me and said, you know, we've got a problem here. You know, cancer is all these different things. and the paradigm we have for treating things just isn't working for it. And I started talking with them about it, and that led to a big collaboration. One of the things that we realized was, in some sense, cancering isn't something you have like a disease. It's something that you do. like your body does. And your body is constantly doing it.

1:01:01And your body is probably cancering right now in three or four different ways. But usually it deals with it and stops it and occasionally gets out of control. So if you start thinking of it more like a verb, and then where's the action happening? Well, the action is happening at the levels of proteins being expressed and proteins interacting. So even if I knew all your genes, I don't know what your proteins are doing. I know maybe what possible proteins are, but proteins, after they get produced by the genes, they modify each other, and they also come in by food and the bacteria in your gut and everything like that.

1:01:41So what you really want to see is the proteins, what's happening in the proteins. And nobody had a way of looking at the proteins. So we started developing a way that you could take a drop of blood or eventually a cell and look at all, just measure all the proteins in it. and see how that changed with time. And we started doing it with mice and studying, you know, as they got cancer, we could see how the proteins changed and the cascades. And then you could look at ways of interfering with this process, which is different in every form of cancer. So once you start looking at it as kind of a runtime thing, rather than something that you have.

1:02:24What do you mean by runtime? There's two ways of looking at what's going on in a computer. I mean, I could stare at the code for a long time, but a better way of debugging the program is to try to run the program and look at what's actually happening. And in some sense, if you look at genetics, you're looking at the program. But if you had a way of looking at all the proteins, that's the equivalent of the debugger to see what's actually happening. and so you know it was a kind of became a different way of looking at cancer and the national cancer institute got interested in it gave us the money to actually you know make some real progress and then and so on so that's how i got in into that one so going back to that as you got into working with this doctor is your idea well you probably said i don't know that much about protein.

1:03:12So I'll start to hire people who will be the expert in this. My job would be to find the people who know the most and then start to work with them. Or are you trying to bring yourself up so you're now an expert on proteins as well? Well, so first of all, you know, he was one of the world's experts. So first step was just like learn from him. But then, you know, he knew people that were other interesting people to talk to and introduce me to them. And it was the same Same thing with Marvin. Marvin introduced me to other people. Or the same thing with Feynman. You know, once Feynman introduced me to his arch enemy, Murray Gelman, you know.

1:03:53So keep him close, Danny. Keep him close. So in that case, it was really David Agus was the doctor that brought me into it, was my mentor. And I think with all these people, they like explaining it to somebody who doesn't understand it because they get to sort of go back to the fundamentals. And then that's a process. If you've ever taught somebody something, you know how much you learn teaching somebody something. So that was in some sense what I had to bring to the party was I was the blank slate that didn't know anything that was asking the dumb questions. How did the doctor find you, Danny, at that point?

1:04:33How did he end up calling you or emailing you? It was funny. He kept calling me and I didn't, you know, because you get a lot of incoming calls. Yeah, unknown caller. You're like, no thanks. Mostly, I don't respond to them. And then finally, he was resourceful enough. One day he got, I think it was John Doerr, Al Gore, and Bill Bergman. He got like three different important people to call me up and say, talk to this guy. So I did. I mean, it was so far afield from things I knew about. Did he explain why he hunted you down in that way? I'm just imagining within the, let's just say, I think it's fair to describe medicine sometimes as a silo, just as there are many different silos, to reach that far afield to investigate some of the questions or to try to unpack some of these issues.

1:05:33At least I know a lot of doctors, MD, PhDs, researchers. Not a lot of them do that necessarily. No, he's a very unusual kind of a doctor to do that. Just like Dick Feynman was a very unusual kind of physicist and Marvin Minsky was a very unusual kind of computer scientist. That they all, first of all, They all share a kind of playfulness and curiosity, and they all share a kind of skepticism about the experts in their field. You know, they appreciate that they know a lot of things, but they also appreciate that they're missing a lot of things. And I think that that's probably rare in a field because really, you know, the best strategy for becoming important in a field is kind of go with the flow, work on the accepted important questions.

1:06:23don't question the things that nobody's paying attention to, and don't listen to people on the outside of the field and things like that. So yeah, these are all very unusual people to do that. And so I do have to find an unusual person that sort of is willing to put up with a dummy like me. Going back to your three criteria of you have to be excited by it, You've got to have some kind of financial basis and no one else is doing it. I bet that there are still three or four things a month that come into you that would fit those definitions. I would think that your opportunities are even within that space.

1:07:03You still have to make some choices about what you spend your limited time on. In addition to that, do you have like a fourth criteria that you're using? I'm kind of realizing, and I've never articulated this before, but there's always something that you kind of want to learn about. and so in that case it was clear that there was a lot happening in biology that i didn't know much about okay and so it was an excuse to learn about it so but how about today i mean i have it this month i'm sure you've got three opportunities something interesting maybe money no one else is doing it how did you decide what what new thing to do in the last month i I should say that the make money thing isn't exactly like you put it that way, because I've never really done things to optimize to make the great billion-dollar company or something like that.

1:07:58But you sort of have to have some financial model of how you're going to pay for all of this. I mean, it has to have some sustainable way of paying for itself. It doesn't have to make you rich. I actually like Disney's formulation of that, which is we don't make movies to make money. We make money to make movies. Yes. Yeah. I think that's a much better way of doing it. You have to have something that's sustainable. Otherwise, you're going around begging all the time. So your fourth one is I'm going to also learn something. This is a way for me to learn. Well, I'll tell you what, like right now, I've gotten very interested in agriculture.

1:08:35Part of it, I got interested in it because during COVID, I moved out to a farm in New Hampshire. And I started realizing, I mean, we just grew food in our own greenhouse. And I started realizing how much better this food was than what I could get, you know, shopping at Whole Foods. And started thinking about the whole supply chain and why was it food was so bad and expensive. And the more you look at it, the way we do food today kind of relies on finding some place where you can pay somebody an unfairly low wage to do something and bringing the food from there. And, you know, that's not really a sustainable future.

1:09:21And, you know, the land in which you can do that and just the social justice of doing that is not going to hold up. And people want more protein. People want better food. And it's incredibly energy inefficient. You're better off in California, but here you go to a grocery store. Most of the vegetables that you find in the grocery store are many weeks old. They've been shipped across thousands of miles in refrigerator trucks. Great cost energy. They're just about to spoil by the time they put them on the supermarket shelves. They've had all the flavor and everything bred out of them so that they can optimize their ability to withstand shipping long distances.

1:10:02And the rest of the world couldn't repeat this inefficient system that we've done. And yet the rest of the world wants to eat much better food, wants to eat more protein. Climate is changing. so denny when you're looking at a space like this you have the seed of an interest that is prompted by this time spent in new hampshire where i've also spent a bunch of time and then you start asking questions and the peripheral vision widens to include all of these different facets that you mentioned, right? So someone could get lost in that and just the sheer volume and complexity of all these different problems and challenges.

1:10:46How do you brainstorm questions and then choose which questions to pursue? I'm interested in, and maybe it's because of the technique I've developed to learning things is, are there ways to change the system rather than solving individual point problems within the system. You have a very systems view of the world. Yeah. Okay. Yeah. So agriculture is like the oldest technology. So it's amazing all the solutions people have come up with, like the point problem of, you know, how do you pull out a weed or pick a tomato or any one problem has been looked at in a lot of ways and lots of dimensions around it and so on.

1:11:30But it's surprising how few people think in terms of what's all the things that have to happen for food to get grown and, you know, it end up on the table. And a lot of it is stuff that you don't imagine, like predicting the weather, mining fertilizer, shipping things in refrigerator trucks. And, you know, it's not things that you would think of first when you're thinking of agriculture, but that's actually what a lot of the activity is. So people have point optimized out most of the specific solutions, done a pretty good job of that, but very few, if anybody, people have kind of tried to look at it as a system and how could you rearrange the system?

1:12:17Now by system, you don't mean, I assume, which is always a dangerous habit, but you're not talking about, say, some people might think of permaculture as a system, but you're extending the system to include many other aspects of food production, transport, supply chain. Yeah. Permaculture would be like a natural system. And so nature does build things in terms of systems, ecologies or systems, but typically we engineer things in terms of point solutions that get put together into systems. Kind of cobbled together. Yeah. And that's because that's probably, you know, that's the commercial opportunities.

1:12:57If you make a better point solution to something, you've got a market and you can build up expertise and a competitive advantage and so on. So there's a reason why people do that. And, you know, the system things are more complicated and more likely to fail, you know. And a lot of times I do look at it and decide this is too complicated. I can't do anything. But sometimes you'll look at it and say, wow, a lot of the easy things haven't been done. If you change this and you change that at the same time. Did you find that in agriculture? There was some low-hanging fruit, pun intended. Oh, yeah. In agriculture, it's very much like that.

1:13:34So clearly, for instance, things should be grown much closer to where they're eaten. I mean, they don't have to be grown like in vertical farms in a city, but they could be grown a few hours away out in the suburbs. and they'd be a whole lot better. But here I am in Boston. You know, you can't hire an agricultural worker in Boston. First of all, there are very few people who know how to do it and they would demand to be paid much more than, you know, you could afford to sell the tomato for. So you have to have a better way of using labor. You have to have a better way of building greenhouses so that they work in colder climates.

1:14:14You have to have different breeds of plants, different fruits and vegetables that are not optimized to be shipped 2 ,000 miles. So, you know, there are a lot of things you have to change. But if you change all of those things up at once, there's another equilibrium point. It's a nice equilibrium where there's another sweet spot of things working together in which many, many crops are grown much closer to where they're consumed. But you have to change a lot of things from the architecture of the greenhouses to the jobs of the workers to the microbiome of the soil. So you have to sort of be willing to take on all that, which means learning a lot of new things.

1:14:57Are you currently in exploratory learning phase? or once you have this grab bag of different issues that need resolving to produce the outcome of having multiple foods or maybe all of your food grown or harvested and sourced near Boston, let's just say, do you rank order those and then tackle one? Do you have teams or contractors and you try to parallel process at the risk of using that completely incorrectly? So one thing is you need to find kind of a visionary source of funding. The patron, right? You need your Medici, right? Medici or somebody who already has this idea and is trying to make it work and hasn't figured out how to make it work, which is what happened in this case.

1:15:45And actually, the doctor that I mentioned before was already working with a company that was starting to do this, but didn't really know how to make it work. And so they came to us for help, and we probably gave them more than they ever imagined that they wanted. And together, we made it into a much bigger project, and I think we're really going to make a real system. And if you solve a real problem, then that comes with actually an economic opportunity, which they'll be able to exploit. But it requires visionary funders who are willing to take risks and, you know, kind of like DARPA was for ALI initially or later other people were for ALI.

1:16:29I never would have been able to do the clock without Jeff Bezos kind of seeing the vision and saying, yeah, I'm willing to step forward and do this. Those are rare people. So I guess I've been lucky that I've run into a bunch of those rare visionary people that are kind of willing to take a bet on me. So you have many talents, Danny, so many talents. I'm wondering, which one do you feel is your superpower? I don't know. Maybe not being afraid to learn new stuff. In some sense, maybe it's a superpower we're all born with. So maybe I've kept a superpower that kids have, right? So kids, like they're not afraid to go in and see something new and strange and start playing with it.

1:17:16And then after a while, there's a lot of things in the world that that gets sort of beat out of you. You learn not to do that, and you get told not to do that in lots of different ways. And I guess I was lucky enough to be around people that didn't beat it out of me. Here's what one of your kids told me. They said your superpower was a mind shifter, someone who can easily shift into different mindsets and view things from multiple perspectives. And I think I would agree with that. I think that's your lateral thinking is, to me, one of your superpowers. That may have come from my childhood because my childhood was, my father was an epidemiologist.

1:17:57So we lived pretty much any place that was a hepatitis epidemic, which often came with a war and a famine too. So I lived in a lot of strange places around strange cultures. So you sort of had to mind shift into what are things like in the middle of the Congo or what are things like in Calcutta. So maybe that's how I sort of got that habit of being willing to shift my mind a bit. So maybe an angle into this, Danny, my understanding is that you homeschooled your three kids. Why did you do it and how did you approach it and how did that turn out? First of all, I can't take personal credit for homeschooling.

1:18:40I mean, my wife did a lot. And also, we hired a bunch of tutors. And we worked with a bunch of other homeschoolers. And so. So you guys jointly decided. I taught them some things. But it definitely takes a village. So when I was a kid, I did bounce around all these schools. And I remember I had some great teachers. I had some really bad teachers, too. And I remember sitting in school and thinking, I will never do this to my kids. And so I didn't. But when you had a great teacher, they would listen to where you were and stretch you a bit. One of my favorite teachers was actually a woman named Mrs.

1:19:23Wilner. She was a librarian. And I was really interested in collecting rocks wherever I went. I would always go in and ask for books on rocks. And she said, okay, well, here's some books on rocks, but here's a book on electricity, too. And I was like, whoa, this is, I never would have asked for a book on electricity, but she kind of led me there. And here's a science fiction book. It's like, what's science fiction? And that was like a juvenile science fiction book called The Wonderful Trip to the Mushroom Planet. But, you know, that was like, brought me into this whole other world. So great teachers are like that.

1:20:00They kind of see where you are and they stretch you to someplace you can get to. And that was wonderful. And you just have a lot more opportunity to do that in homeschooling than you do in a classroom. Were there any aspects of cognitive development, curiosity or otherwise that you cultivated through the homeschooling, understanding that it wasn't just you as a loan operator doing it, but were there things that you wouldn't really emphasize or touch in traditional schooling that you guys included or emphasized? We did, but it was interesting for me. Sometimes I would sit down to teach something that I thought was really simple.

1:20:41And then as I started teaching it, I realized, well, actually, this isn't so simple. There's this other thing underneath it and another thing underneath it. So I was actually not such a great teacher necessarily. You're like, wait, wait, wait. I know 2 plus 2 equals 4, but let's back up a minute. Yeah, exactly. You can back up. It sort of reminded me. I saw this happen in college. I had a math professor named John Calarota. And he was at the Blackboard once. And he's writing along. And he says, so you can see that it's obvious that this is true. And he stops. And he sits there for a long time.

1:21:17It felt like 10 minutes or something. We're all just like waiting. He's like, yes, it's obvious. And he goes on. But I think that's one thing that you realize about teaching is, you know, how much you don't know, like, you know, what that depends on. That was a wonderful thing. I think Dick Feynman was really inspirational in that. He really admitted when he didn't understand something. And he'd like, well, wait a minute, like, what's a one and a zero? You sort of have to back up and sort of explain that, and you never really thought about it before, but digital computers need to disambiguate.

1:21:57So they force everything to be the one or zero. If it's in between, you know, they force it one way or the other. That's what digital means, right? You don't put up with any in-betweens. You push it into a category of one or zero, and then you build it up from there. But you start thinking about that. Nobody ever really asked me that before when I was teaching them computers. And so, you know, Dick was always saying that he doesn't understand something unless he can derive it from first principles. So watching him do that in other fields, I realized, well, you know, I don't really understand an awful lot of the things I do either.

1:22:33And when you teach, you sort of realize the things you don't understand. Danny, I was just wondering, what do you try to optimize in your life these days? Well, I wish I had a lot more time ahead of me. So right now, time seems like the most precious thing to me. And you start realizing how much of it you squandered. It doesn't sound like you squandered very much time. I'm not seeing that. Where were you squandering? When did you squander anything? Danny, based on your bio. Objection, your honor. Yeah, right. I did a lot of things that didn't work out. Okay, so let's talk about some failures. What were some of the failures?

1:23:15Well, I mean, I think that thinking machines was my first big failure because if I had asked that financial sustainability question, really treated that as a problem of serious thought, like I was thinking about the machines, I would have done a much better job. That company didn't have to fail. It was awful when it did. It had like 500 people, almost none of whom had had another job. It was like I was hiring people straight out of MIT, and we were building the fastest computer, and everything was going great. And, you know, we just did a bunch of dumb things in how we started the business that, you know, we didn't pay enough attention to laws that were getting passed by our competitors in Congress that were making it illegal to export our products or making it hard for people to acquire our products.

1:24:07And, you know, we were just sort of blindsided by that. We did stupid stuff. We were growing up, up, up, and so it didn't occur to us that we might have something might cause a downturn and we might not have enough cash in the bank. And now going back, we just managed it badly. It's sad because, I mean, it was a terrible moment for me. I felt like I had let down all of these people, and I had let them down. So Danny, you look back on this life review and lessons learned and flash forward to today, you said focusing on, well, you look back and realize how much time you've squandered. What rules do you have for yourself or how do you think about not squandering the time you have left?

1:24:53Have you changed anything? Well, the non-redundancy is a piece of it. Right. If someone else can do it and will do it. Don't work on things that are going to happen anyway. I still think hanging out with extraordinary people is the right thing. But there's an interesting problem with that, which I've realized, which is I tended to hang out with a lot of people that were older than I was because of that, because they sort of established themselves as extraordinary. And, of course, that has been very sad to see so many of my friends die and so on. So I'm actually very curious. I'm sure that there's a whole generation of younger, extraordinary people that I haven't met yet.

1:25:35And that's something I'd love to do, is meet some of those unusual people that are thinking about things differently and learn from them. So that's kind of part of my agenda these days. Right, to find the young. So what do you think would most surprise your 20-year-old self about your life today? I guess one thing that surprised me is that it sort of has all worked out.

1:26:06You were not really expecting that? No, I think I was always kind of on the edge of failing in some sense. And, you know, many times I did, but it still worked out. And so I think I probably worried more than I needed to, because it always seemed like, oh, this is pretty dangerous, or I'm going to be penniless. There were times when I was penniless. I couldn't pay my mortgage. So I think I worried way too much. I worry less now. So Danny, this is going to be a fast left turn, but I've been staring at this the whole time we've been chatting, and I don't know why I can't use the term OCD, because I have been diagnosed with it, although I think it's actually a superpower in a bunch of respects.

1:26:52I've been staring at this prompt. What is the entanglement with capital E for like an hour and a half here? And just waiting for the right segue, but I don't know if I'm going to find the right segue. So what is the entanglement? One of the things that I've noticed about the world is it used to be that But nature and technology were very different things. Technology was something that we designed and we understood and we controlled. Nature was this mysterious, complicated thing that, you know, we didn't understand at all and pretty much had to take and work around and kind of riff with. But those two things are becoming entangled in sort of both directions.

1:27:38So things that used to be natural, like the atmosphere or our genes or our minds or my knee joint, are now technological artifacts, right? And the things that used to be technological and controlled and designed are actually kind of evolved and having. like the internet. Nobody can draw you a wiring diagram of the internet. Nobody can really tell you how chat GPT came to that conclusion. I mean, they could sort of make up a story about it, but they can't. They don't really understand it in the way that you used to understand a computer when it produced an answer because it wasn't really designed.

1:28:26It was kind of a combination of designed and evolved and learned. And so what's happening is that a lot of people's use of computers is now they kind of know the magic incantations that cause this library for you to do that but they don't really know like all the things that are going on underneath that make it work and so it's becoming more like nature nature we used to kind of know well here's you know the magic incantations we use for making beer we don't know really why this makes good beer this makes bad beer this makes champagne but we sort of know when we do this it does that and And that's kind of becoming our relationship with computers.

1:29:07So I think that what's happening is the distinction between the natural and the artificial is becoming entangled. And that idea may just kind of go away. So there sort of almost is no pure nature and there almost is no pure technology that we fully understand, at least not in the technology that we're using to have this conversation. for example there is nobody who understands every piece of it i want to put in a plug for my very first book out of control which was about that entanglement yeah i think you're one of the people that really got me thinking about that entanglement right and that book was probably a lot of what got me thinking about these ideas and artists of life got me thinking about it and the way i would say is there's one thing with two different faces and those basically we had two different faces to the single thing and we're recognizing that there's only one class which has two different perspectives on the same thing you're saying natural and synthetic or nature and engineered yeah there are basically different faces of the same thing going on in the long arc of the universe where danny might say they're being entangled i would say that they've always been entangled but we had kind of two separate views of them and now we have a better view of it.

1:30:32Well, I think there's also something very special about this instant in time. And by this instant in time, I mean plus or minus a century. It's a long now. Yeah. But I think when people look back at history, even really our lifetimes, I mean, over my lifetime, the population has more than doubled. The climate has changed. The computers have come out. Everything is really very, very different in a way that has never happened before. The population hasn't doubled in a single lifetime before, and I don't think it will again. Right, right. So I think we are at a special moment where our sort of technological powers have gotten enough to make these things that are more complicated than we can understand.

1:31:21And that's kind of a qualitative change. We weren't building stuff that was more complicated than we understood before. Or producing outputs that were completely unexpected. Danny, so a question about the entanglement, and this actually relates to a name you mentioned earlier, Jeff Bezos. So he's described AI, I don't think it was specific to LLMs, I mean it was broader than that, as a discovery and not an invention. It's something akin to electricity or fire. How do you think about AI? So I'm going to make a distinction between AI and what's called AI right now. Great. Please. So intelligence is a very complicated, multifactored thing like life.

1:32:10It's not just one thing. And in the beginnings of AI, we thought the things that were hard for us to do were the intelligence. Like we thought playing chess would be intelligent or solving calculus tests would be intelligent. Or translating languages. That came later. But the things that were hard for us, we thought that's intelligence. And that was the stuff that early AI concentrated on. And actually, it turns out that was really the easy part. The hard part was the stuff that we were so good at, we didn't even notice, like recognizing a face, jumping to a conclusion, having an intuition about something.

1:32:57Those things were way, way harder. So we thought producing speech would be hard. We didn't think listening would be hard because we just did that without apparent effort. But listening to speech turned out to be way harder than producing speech. In the early days, there was always a box called, you know, sometimes the, well, it was the neural network, the pattern recognizer that was sort of, was going to guess the obvious thing that was going to happen next and recognize the pattern. And we thought that was going to be the easy part because it was just going to be some neural networks that got trained.

1:33:36Now, it turns out that those neural networks had to be much, much bigger than we were guessing, way bigger than we were guessing. And you had to train them, or at least so far we've only known how to train them with way more data than we were imagining training them with, and so on. But sure enough, you know, that box has now gotten built. That's what we call AI right now is that little box in intelligence. And it's actually really good at kind of imitating human intelligence. And imitating is kind of a good first step. That's what my granddaughters do first. I have a granddaughter that can sit and talk to an electrician if she knows what electricity is just by using the right words and saying phrases that she's heard before and so on.

1:34:24And she can kind of fake it pretty well, but she has no idea what she's talking about. And that's mostly where AI is right at this moment. I mean, it'll be at a different place a year from now. And people understand that and putting it in different places. But it's just one little part of intelligence. It's a good start. But it's not going to do all of the things that we do that we consider intelligence until people come up with some other ideas. But people will come up with other ideas. The other big change is we've got a lot more smart people working on it than we ever had before. So those are the people that are going to come up with all those other ideas to make it work.

1:35:05So I do think AI is going to happen pretty fast just because we have so many smart people working on it. Is there anything you think people are, broadly speaking, overestimating and underestimating with respect to the development of AI? AI not in quotation marks. I think they're overestimating the capabilities of what we have now, but underestimating what we'll be able to accomplish over the long run. And it's interesting. I think that people get mixed up on timescales a lot. In general, I'm kind of a short-term pessimist and a long-term optimist. I think that probably applies to AI. as much as anything.

1:35:55I'll make an observation that I've been reading the early history of the discovery of electricity, way before Tesla and Edison. I mean, like, you know, Faraday and Davies and these guys. And what was remarkable was how the smartest people at the time, like Newton and others, were just so wrong. I mean, just so far off. The strangest ideas about what electricity was. And they really hadn't a clue. And it was just many, many years of going through. And actually, before they had the scientific journals, they had scientific demonstrations every week where they were demoing the latest discoveries in electricity for paying tickets.

1:36:35And each time a week would go along, they'd have another discovery. And what they were discovering, it was far more complex, far more unintuitive than what they thought. And I think that's exactly where we are with intelligence. We have no theory of intelligence. We have no idea what it is. we're just discovering some of the earliest primitives of what it might be. But I think we're as far from knowing what intelligence is as they were from understanding what electricity was in the 1700s. I think that's fair, except maybe it's actually quite possible we'll never understand what intelligence is because that was sort of part of the prediction of the Songs of Eden papers.

1:37:13It may be easier to actually make intelligence than to understand intelligence. Right. We used plants for thousands and thousands of years without understanding how they work. We used a natural world without understanding how actually they are made and are governed. So we can use things that we don't understand. So what's first is us being able to make things that we can use that don't understand. So a quick question on intelligence, though. Is it a useful term if we can't understand it? Or is it just so broad a label applied to so many things that it's kind of useless and should just be replaced by thin slicing and using more precise labels or concepts?

1:37:56I'm going to answer first because I think we're going to start to unbundle the concepts as we discover more things. I mean, my hypothesis is that we'll discover more about how our mind works through AI than 100 years of neurobiology has. And we'll come to understand intelligence is not a single dimension. I think it's a very high dimensional space in which there's lots of different primitives or elements. And part of what we're doing right now is we'll begin to discover some of those elements. And that intelligence is basically compounds. We have a compounded intelligence that's made up of lots of different kinds of cognition and stuff.

1:38:31And so I think we're on that path to not replace it as much as to unbundle it. And also, I would absolutely agree with what Kevin said, but I'd take it one step further. which is even if we unbundled human intelligence and did all of those things, there's still more to intelligence than that. But there's other kinds of intelligence that we can't even imagine. And actually, those are the ones I'm most interested in because I said I always like hanging out with people who are much smarter than I am. I would love hanging out with machines that are much smarter than people. Right. But smart in different ways.

1:39:14Or playing Million Color Connect Four with a mantis shrimp, you know? Yeah, exactly. The way I say it is that the space of all possible minds is huge, and that human intelligence, we're going to find out, is on the edge, like we're at the edge of the galaxy. We're not at the center. We're going to be an edge species of intelligence in the map of all possible minds. And the reason why we want AI is to arrive at these other places in the high dimensional space of thinking that we can't even imagine. That's the main thing. It's not to replace human thinking. That's boring. Nine months, we could have another human mind.

1:39:49But you want to have other kinds of thinking. That's the whole point. This is related to the transition thing. But I think humans as we know them today are kind of halfway between monkeys and what we're going to become. You know, we've still got a lot of monkey in this. We're not the far right in that diagram of the monkeys. No, no, no, no, definitely. We're in this transitional phase. You know, we still got a lot of monkey in this. And I'm really excited by that thing that we're going to become. So Danny, I have a question for you related to the short-term pessimist, long-term optimist. I'm sad to report that there are lots of people, my vintage, younger, just people kind of close to my, I'm 47.

1:40:33so close to my age or even quite a bit younger, mid-30s, who are on the fence with respect to having kids or have decided not to have kids because they look at climate change, they look at what they might fairly consider some of the unpredictability around AI and the fear around Skynet. And if we could go down this list of concerns they have that they cite as compelling evidence that they do not want to bring a life into this world because the future to them looks so bleak. How do you think about the long-term future? I mean, there is a value in optimism. There's like utilitarian function to optimism, but if we're able to put that aside and maybe we can't, like, how do you think about a hundred years from now, 200 years from now?

1:41:24So I understand that, but I also understand that like when I was a kid, We were taught to hide under our desk for when the atomic bomb was going to get dropped. And thinking even as a kid, this isn't going to work. I mean, I knew people who died of smallpox. That disease doesn't exist anymore. When I was a kid, most other kids were hungry and were malnourished, were likely to die of childhood diseases. That's not true anymore. So, you know, when I was a kid, you know, I had friends I now understand were gay friends and only understood later and understood what they were going through, but, you know, they couldn't say that to anybody.

1:42:12So there were so many things to be frightened about, and yet there were so many ways in which the world just got so much better. And even in my lifetime, it did. And it's true that, you know, it also, we created a lot of problems in that process, but we've always created a lot of problems. So, I guess, you know, if I just look at the sweep of history, there isn't any time when you'd say, oh, I would do better going back 100 years or at least, you know, not in history. You would not want to be alive 100 years ago compared to being alive now. Especially if you're going to be born in a random place in sex.

1:42:54But even, you know, I wouldn't want to be a king 100 years ago. Much better to be a peasant today than to be a king a couple of centuries ago in terms of, you know, your health, your food that you ate, how you spend your time, so on. So, your comfort, everything. So, I think that there is a general trend. It's possible that some catastrophic setback that could happen. But even if that happens, I kind of believe that humans are adaptable enough or nature is adaptable enough that it'll pick up and start up again. I suppose there's a scenario where it's without humans and something else. But I'm certainly optimistic, you know, the Earth is going to be fine.

1:43:39And I actually do believe that, you know, there are people that are going to see that 10 ,000-year clock, decide what to do with it when it comes to the end of this 10 ,000 years. but you know it won't be steady progress it never has been and so there's a bunch of things to worry about i see why people are worried but the bigger the picture you look at the more you realize i guess progress isn't a steady upwards thing it's kind of two steps forward one step back so let me ask you just a question about rank ordering sort of existential concerns because i I am very fortunate to effectively as a job, talk to the smartest, most interesting people I can find.

1:44:23And behind closed doors, generally not on the podcast, sometimes on the podcast, I have brilliant, brilliant friends, some of the smartest people I know, who are very preoccupied about climate change and basically view us as the frog in the heating pot of water that's going to eventually reach a boil. and it's not too late, but everyone needs to act now. And there just don't seem to be the incentives in place for that to really happen, frankly. Political will or competency as one piece of it. Then you have folks equally brilliant, in some cases you might even argue more brilliant, who say the preoccupation with climate change is completely ridiculous.

1:45:04It's just patently absurd that people would consider not having kids, citing that as a reason. And these are not people who are coming at it from a political perspective. They're just saying, if we actually look at trying to weigh the severity of certain risks, this isn't even top five. Where do you fall on that? So I don't think that people underestimate the problem. It's like a really big problem, and it's going to cause a lot of difficulties. But people do underestimate our ability to deal with problems. And so, yeah, it's going to be bad. And it's already starting to be bad for people. But people have dealt with a lot of bad stuff and come out of it and come out of it better and come out of it as improvement.

1:45:57So, I don't minimize the difficulties of climate change and the challenges. And it's going to be a mess. But I also know that there's a lot of super smart people. They're working on all kinds of things that are going to help with it in ways that are hard to imagine. And so some of those are likely to work. And, you know, it's kind of easier to imagine catastrophes than it is to imagine magic solutions. So it was easier for people to imagine that the population explosion was going to doom us. You know, that was a real easy idea. But But actually, Kevin was the first person to point out to me, I think, that actually our big problem in a century now is going to be the population shrink implosion.

1:46:43And people are starting to realize that already. And it was just hard for people to see. So we're kind of hardwired to pay attention to danger. And, you know, that's unpot. And also, there is an effect, which is that bad things happen fast and good things happen slow. Can you say more about that? So, yeah, if you read the newspaper, it's like… The bad things that happened closest to you or furthest away today. It's full of bad things. You know, the plane crashed, the war started, and, you know, there was no newspaper headline that said, the majority of kids aren't hungry anymore. Because that happened very slowly, and it's continuing to happen.

1:47:33And so I think that the world has actually been getting subtly better, but there was almost no headlines about the things that did that. They weren't the attention getters. also some of the better things are things that didn't happen most of the good things are things that didn't happen your kid did not die you did not get robbed on the way to work all those things and so there's no headline at all about the things that didn't happen but you know danny if you had to rank your worries what would you put at the top well i don't deny that ai is an existential risk for humans as we know them. Maybe what's good about humans could go on in AIs.

1:48:16I think that's a possibility. I actually think it's more likely that AIs will help us get out of this mess. What is this mess? For example, help us deal with climate change. Help us deal with the next epidemic. Help us avoid the nuclear war. Or even we're just talking about population. I think it would be kind of an amazing coincidence that at the very moment when we're headed towards a population implosion, that we have robots and AIs. That's another possibility. In some sense, it's much harder to imagine solutions to problems than it is to imagine problems. So this is kind of a trivial example, but when the technology for cell phones was being developed by Motorola, and I kind of knew about it, I went around and I told all my friends, you're going to have a phone in your pocket someday.

1:49:10It'll be just like Star Trek. And every single one of them, without exception, said, oh, I would never want that. And they gave all kind of reasonable reasons. They saw the problems. They were like, well, if people were like on the bus, everybody would be talking on the phone in a restaurant. People would be getting phone calls. You know, people would interrupt me in the middle of the night with the wrong number. You know, they could see all the problems very vividly. But they sort of couldn't see how much it would enable them. So they all predicted that they wouldn't want it. And there was also another part of that I was involved with.

1:49:48We were bringing the Internet to everybody. And there was the common response, almost invariably, every time I talked about it, was people were worried about the haves and the have-nots. What about all the people who don't have this technology? What are you doing about that? And my response was, I'm not doing anything because the benefits of this is so good that it's going to happen anyway. The thing you want to be worried about is what happens when everybody has it. There's going to be a lot more problems. When everybody has a cell phone in their pocket, that's going to be their problem. not because the people don't have it.

1:50:23So there is a sense in which there's an asymmetry where the things that break are easy to see and the things that work are hard. They're not equivalent. It takes a lot more energy to imagine something working than it is to imagine how it breaks. I'll give you a very specific example today. If you ask most people, you know, would you like a chip inside your brain that augmented your brain and helped pretty much everybody you talk to is going to say no. Wouldn't want that, you know. And they'll give you lots of very good reasons why they don't want it. And a lot of them are valid. But boy, I'm pretty sure that when that becomes possible, everybody's going to want one.

1:51:07You know, I think it'll be just like the cell phones. It'll just do so much for you that, yeah, you'll put up with the problems. You'll work around with the problems, but it'll be. you first is all i can see yeah so you can all watch danny glitching on video six months from now give you a beta tester so i want to come back to something and i'm gonna steal some cliff notes from kevin here but you mentioned talking in restaurants on cell phones and i'm very very sensitive to sounds and i see something in notes that kevin and i were sharing i don't know anything about this, but the name is descriptive enough that I feel like I kind of get the idea.

1:51:50Babel, the cone of silence. Yeah, I want it. What happened to it? What is this? It's like a Jetson's helmet that you plop on loud kids. What happened? No, it was actually a cool thing that somebody should do, but it was originally the problem of open offices and people overhearing each other's conversations. and so it turns out that the best thing to mask a conversation is somebody talking in exactly the same voice saying something different and so this was a little machine that people could put on their desk and and we tested it it worked which is that it sort of listened to you talking for a while and then it started talking kind of in your voice but saying it just making up sort of babble but in your voice and kind of your intonation and so on and sort of talked over you out to the people around you so the phenomenon was that the room got a little bit louder but mostly people didn't notice that mostly there was just kind of a little buzz but if you actually try to listen in on the conversation of the person on the desk next to you can't do it seemed like you could hear but you couldn't actually understand it.

1:53:09You got two violins playing. Yeah. Yeah. And then what happened with that was Herman Miller bought that technology for use in offices. It was actually a very sad thing. They set up a company to start it. Much to my surprise, the restaurants got very interested in it, which sort of bugs me because I hate all of the noise in restaurants. But, you know, things like, you know, the line at the pharmacy was interested in and so on. But it was very sad. because the CEO had a heart attack and nobody had the heart to keep going. So we'll never know if the technology would have worked. Kevin, what would you like to see Danny work on?

1:53:49If Danny was like, I'm out of my idea bag is empty. And he showed up and he said, Boss Kelly reporting for duty. What do you want me to do? I mean, another way to think about that is Danny is the inventor and he has a company that invents things. What would I like Danny to invent? Yeah. If I had to make a commission, I had a billion dollars. Oh, my gosh. Robot beard trimmer. Yeah, exactly. Something that would mean all his criteria. Well, no, no, no. No, just for you. This doesn't have to meet his criteria. No, no, no. But I mean, it's something he would accept. Well, he's got no ideas in this hypothetical situation.

1:54:31I have to think about this. Beggars can't be choosers. How about you, Tim? What would you like? You have an idea, right? Well, it's just, it's very front of mind for me. No pun intended in this case, but I have neurodegenerative disease on both sides of my family, Alzheimer's, Parkinson's, and more. It's quite the collection. And I've been interested and followed neuroscience. I was originally a neuroscience major way, way back in the day. and would love for you to take your blank canvas, no question is dumb, and apply it to neurodegenerative diseases. That's one that immediately leaps to mind.

1:55:12I know how to go about that. Somebody should do it. Which is, you know, it's the same thing with cancer, with the cancer thing. What we really need is we need a way to read out the proteins in your body dynamically, like we can read out your genes. And if we could really monitor that and read it out, you could find the processes that create disease before disease happens. So right now, when we treat disease, the cat's out of the bag. Your body does a great job of compensating for everything for a long time until it just can't handle it. But things have already gone a long way before you show any symptoms.

1:55:58Right. That's why so many Alzheimer's interventions fail. It's just too late stage. Yeah. So your body is great at masking things going wrong. If you could look at the proteins in the body, then you could see things are starting to go wrong before you're showing any symptoms. And you could see what was going wrong. and you could start treating it before the damage starts happening. Right now, we start treating things after there's already lots of damage, enough damage that your body can't hide it. And so we need to head off diseases rather than treating diseases. We need to treat when you're on the way to getting a disease, not when you have a disease.

1:56:40And the only way to do that is to have this debugger to understand what's going inside. And the only way to do that is to look at the proteins. And that's a technical problem. It's a very solvable problem. We got a long way to solving it, actually. And unfortunately, it was kind of all screwed up by the Theranos thing, which sort of gave a bad name to all that field and made it impossible to find. So that was sort of the tragedy of that is that, you know, sort of one fraudulent thing kind of gave a whole field a bad name. But it will come back. And it may come back soon enough to actually help you and members of your family.

1:57:19And there are people that are doing that. So look for people who are doing that. I'd love to meet people who are doing that. But I think that's the path. And procedurally, in terms of looking at proteins, would that take the form of, and I'm grasping for straws here, but something like a grail test currently. So for cancer screening, looking at DNA fragments. The first version of it would be a blood test. You know, it might be a finger prick that you would do regularly and just monitor it. But right now, it's just, again, we have so few, because of the way the medical system is set up, we have lots of blood samples, but we don't have the blood samples very well correlated with the medical records and so on.

1:58:03So it needs some big sort of population studies where you get a lot of regular blood samples, that you get good proteomic inventories, which is a technology that is not quite there yet because there's no commercial opportunity for it yet or limited commercial opportunities for it yet. But as soon as you start doing that and you start correlating what's happening with people or what was happening with their proteins before they got sick, as soon as you get that database, then I think we'll be able to head off a lot of diseases. before they happen. Systemic diseases, not infectious diseases. So Danny, I thought of two inventions I'd like to have from you.

1:58:46One sort of profound, the other one is sort of trivial. So the profound one was I recently had an MRI, which is an amazing piece of technology, but man, what a pain, what an unpleasant experience. And I just imagine like, well, in a hundred years from now, there has to be some way that they're going to have a machine that does this in a much more comfortable, easy, quick way. I got one I'm working on. Okay, there you go. All right. I just had two MRIs today, so I sympathize. Really? Yeah. There you go. My sympathization. Right, right. There has to be a better way, right? And I mean, I've had so many of these things, and every time I'm like, wow, this is a terrible experience.

1:59:29I am working on something. I won't do everything an MRI do, but I think it'll be more useful based on an ultrasound. Okay. And here's the funny thing. The great thing about an MRI is it produces this 3D image, and it can go to the doctor, the radiologist, and they can interpret it, or an AI can interpret it, because you've got an output that is disconnected from the process of measuring it. ultrasound these days is not like that ultrasound really the person that's doing the ultrasound has a lot more information than is captured in a picture or video they know how they're moving it around they know that they're pushing it past this they're shoving in this direction they're using this muscle to be a lens to magnify the thing behind it they have the intent of where they're moving it why they're moving it And so they can perceive a whole lot more.

2:00:30And then they have to take what they perceive and write it into a report with maybe a few numbers measured or something like that. But it's not nearly as satisfying. What gets to the physician is not as useful as an X-ray or an MRI or a CAT scan or something like that. Well, there's no reason that ultrasound has to be like that. if you had either a sensor on it or a robot moving it around, or you had the information of the pressures and the motion, and you had a model of tissue deformation and speed of sound through tissue and things like that, you could produce a three-dimensional image like from an MRI with ultrasound that would actually have information that you don't get from an MRI.

2:01:18And that just hasn't been done yet. I'd love to meet people who are doing that. If they don't do it, I might have to work on it myself. So the second trivial invention, Danny, that I'm going to assign to you is we all love a little microwave, which will instantly heat something up. I want the reverse. I want to put it in the machine and have it instantly ice cold. There's a way of doing that is laser cooling. You do it atom by atom. Okay. It can take a while to get that half chicken done, Kevin.

2:01:53That'd be a fun one. Yeah, that would be. No, I don't know how to do that. That would be a billion dollars for sure. Yep. Don't know how to do that. So Danny, if the sort of divine treasurer of the universe just bestowed upon you$20 billion, so one of your criteria can vanish, right? In terms of the sustainability, you're covered for the foreseeable future and beyond. Let's say then it came down to only what gets you the most excited. So it could be focused on that. and you were allowed to indulge, for every one or two serious projects that would have an impact, you had to do one trivial. Not trivial.

2:02:34I feel like that's underestimating how important something seemingly trivial could become later. Well, I'll tell you one that's already happened, but it was like that for me. It's traveling all over the world. Chris was super interested in maps and looking at maps of where I was and so on. And I always wanted to take a map and expand it and just go into it. I had that dream since I was a kid. Got it. It's like an infinite zoom. Yeah, like a pinch to zoom. Yeah, here we go. Okay, but there was no such thing at that time. But I really wanted that. I knew I wanted pinch to zoom. And so as I started building it, and I had worked with Steve Jobs, and I got kind of a prototype of it working, and I invited him over to look at it.

2:03:29And he said, ah, you know, people won't want, like, fingerprint smudges all over the screen.

2:03:38You wouldn't like my experience very much. I was like, but I kept on working on it. And eventually I made actually this touch table thing. And, you know, it was very expensive. It actually went into the Situation Room at the White House. So, like, that was a, during the Obama administration, Obama would show people, like, he had this map that he could pinch to zoom, right? Then, of course, Apple came out with the iPhone. You know, other people were working on it. And then, fortunately, when I did the table thing, I filed a patent. And then when the iPhone came out, of course, it did a very beautiful job of pinch-to-zoom and very refined version of it.

2:04:21And people started using it. And then the other phone companies started doing it. And Apple sued them. Apple filed a patent on pinch-to-zoom and sued them and actually, I think, won like a billion dollars from Samsung. But I had filed this patent. And Samsung went back and went to the patent office and said, wait a minute. You know, this Danny's patent like predates all of this. So the patent office, oh yeah, it does. And then invalidated Apple's patent. So everybody who had Androids or Samsung fund or whatever, you know, they could use Pinchless M2. So I think that's the invention that I'm kind of proudest of because even though I never got paid a dime for it, I see like little kids who had that same instinct that I had.

2:05:08Like I see them going to like a magazine and just like trying to zoom out the picture. And I know that nobody will ever remember that as ever having been invented because it's like kids are born with it. It's become so much of a part. So, you know, when you innately want something like that and you know you want it. So, Danny, how about some practical advice for people who are listening who may be inventor types about patents? Because I know you have a complicated relationship with patents. Here's the case where patents may have done some good. I know other times you're not so sure about the worth of patents.

2:05:45What would you suggest to people who are inventive? For instance, at Wired, we were involved with inventing the web. And Wired invented the click-through ad banner, right? You know, I mean, Brian Balehorf was the guy who coded that. And not one of us ever thought about patenting it. It just seemed obvious. It seemed like a really good thing. It was entirely patentable, but it just never even was in our vocabulary. And I'm not sure how much it would have been worth if we had. But Danny, what do you think about patents and people who are inventing? What would you suggest? So first of all, I think patents might be good for inventors, but I don't think they're very good for society.

2:06:31So if I had a choice, I would eliminate the patent system. Now, there might be particular things like pharmaceuticals and things like that where you could make the trade-off the other way. But I think in general around computers, I'm happy software patents are kind of getting rejected much more and so on. So I've always felt a sense of, I guess, ambivalence about patents. So why are you patenting if you don't believe? Well, so I patent because, remember, I'm often solving problems for other people. And so, I mean, they have paid for something to happen, and so they want to own something at the end.

2:07:14But I think for inventors, and I know inventors that have made lots of money on patents, but you have to sort of sue people. And so they end up wasting an awful lot of their life in courtrooms. And, you know, I hate it. I occasionally get dragged into court for something that I've patented that somebody else owns. You know, you have to get deposed. And it's pretty, you know, it's a big waste of time. It's a big waste of society's resources. And the whole idea of patent system was initially to help society. It was to get inventors to disclose their patents. But I think that things that are sort of self-disclosing, like pinch to zoom, when somebody sees it, you don't need to do any more disclosure about it, right?

2:08:02Or maybe you do about how you made it work or something, but they can take it apart and see how you made it work. So I would say that we ought to definitely narrow down the things that we allow to patent. And I think that to inventors, I typically say maybe file patents is trading fodder in this ridiculous game that's going to have to happen. But don't go off and sue people for violating your patent. And yeah, you might get rich that way, but it's not worth your time. It's not the way to spend your life. Are there any inventors could be past or present who really inspire you? If a intrepid inventor looked at you and they said, Danny, who are some people I should pay attention to or study in the world of inventing, broadly speaking?

2:08:57Anybody stand out to you? So the ones that I admire the most, and some of them have been my mentors, are people like Claude Shannon, who kind of look at something really complicated and messy and get a take on it that makes it simple and understandable in a way that gives everybody else power to do something with it. Who is Claude? So Claude Shannon invented the bit. Actually, another one of my mentors named it the bit, but he invented it. He invented information theory. So he invented a way of measuring information and encoding information. He worked for Bell Telephone, and they were interested in what was the theoretical limit to the amount of information you could put down a wire.

2:09:46Yeah, but even before that, like his master's thesis was the application of Boolean logic to switching circuits. You know, he just had this way of thinking about things that was so powerful that it gave everybody else a way of thinking about things and everybody else a way of solving problems that we just take for granted when we measure things in megabytes and stuff like that. Somebody I lived to know, somebody that was on my thesis committee, invented the bit, right? Or discovered it or whatever. Those are the kind of people that I admire the most because they give everybody else the power to imagine new things and do new things.

2:10:32And Newton did that in physics. Feynman did that in physics with Feynman diagrams. So those are the real wows of history. So speaking of people that you admire, I have a favorite question. What is a heresy that you have? And I define the heresy as something that you believe that the people you most admire don't believe this is a little strange when you're not gonna like us that's the whole point

2:11:02i don't believe in cause and effect oh wow okay okay explain what that is for people who don't know okay so we look at an equation like f equals ma and we say oh force causes masses to accelerate when you push on it, okay? That seems to be what F equal M-A says, just going back to Newton. But I think that's just a story. I think that we like to tell stories in which there are agents that cause change because we're social creatures and we like to sort of personify nature. But I could rewrite F equal M-A to be A equals F over M and say mass is caused by, you know, force acting on acceleration. And that creates mass just as easily as I could say that the force causes the acceleration.

2:11:58It's just a way we tell the story. And some stories are intuitive to us and sort of make sense and fit with our intuitions, and some stories don't. And so when we can tell a story about something that's kind of explanatory and helps us guess at what's going to happen next or things like that. It's a useful story. Then, you know, we believe it's true in some fundamental sense. And so, it's the way our brain works. We're wired to look for causes and effects. And that's like why we're kind of wired to believe in God, because if you have a chain of causes and effects, then there sort of has to be a first cause at the beginning, causing all the rest of it.

2:12:42But I think that's just kind of the way our brain works and the way we tell stories about reality. I don't think reality actually has causes and effects. Let me poke on that a little bit. So is it that cause and effect doesn't exist or is it that we simply over-apply cause and effect? And I was thinking back to the proteomics discussion and identifying changes in proteins over a sufficient data set such that you could have some predictive ability or ability to intervene earlier to hopefully mitigate or prevent disease states like Alzheimer's disease or otherwise. So does that mesh with what you are saying or does it not?

2:13:26I'm not saying that thinking in terms of causing and effects isn't a useful way of thinking. Just like I'm all for storytelling. Right, right. I believe in storytelling is a useful way. So I think, but we are, when we tell a story about a protein pathway causing something, we're making up a story. And we really look at what's happening in the physics. All those things work in the other direction, too. And the story isn't really what the physics is doing. It's a sort of simplified thread of things that we can understand in what it's doing. So it is useful to kind of abstract out these threads that we can tell stories about because that gives us a handle on it and helps us manipulate it.

2:14:13So I'm not saying that's not a helpful trick of thinking, but it's a trick. It's not really how the universe works, and we shouldn't fool ourselves into that, and we shouldn't get too enamored by it. And in fact, maybe when we get new kinds of AI, maybe they'll be able to think without using that trick. Right now, we can pretty much only think using that trick. Got it, yeah. And that's what digital is, okay? So computers are all about kind of playing out this fantasy of cause and effect. So by forcing everything to either be a zero or a one and nothing in between and making everything digital, We can kind of make things that almost work perfectly as if this and this caused that to happen.

2:15:00And so, in some sense, the computer is the ultimate fantasy of putting together causes and effects and piling causes and effects and engineering them into long chains that we write with programs and control them. and this comes so close to doing exactly what our fantasy is that it's hard to believe it's not true how does and this is way outside of my area of areas of expertise so who knows if i'm painting us into a corner here but how does quantum computing affect that presentation of computing and the forcing into one binary option or other effect that's exactly the right question to ask because if you really look at true quantum computers, it's much harder to sort of explain it in terms of causes and effects like we do as a digital computer.

2:15:56You operate on it and it causes this state to turn into that state as sort of a cause thing, but actually the cause also involves observation of the states and just looking at it changes it. And so one of the early things will be quantum key generation, where we'll sort of have a module and say, if we do this, we get a cryptographic key that has the right properties. Now, how that magic happens, very few people will have any intuition of how that happened. And the people who do have really deep intuition will realize that it's actually not causes and effects in the way that we're used to thinking about it.

2:16:41So I have a half-baked amateur hunch and prediction about quantum computing, which is I think in 100 years from now, that will realize that quantum does not want to do computation. It's actually not going to be used a lot for computation, but there'll be something else that we'll discover that is really incredibly useful for other than computation. Because I think computation does want to be much more cause and effect. Sometime in the 90s, I wrote a little book about how computers work. Patterns in the stone. Patterns in the stone, that's right. It was just, you know, kind of a high school student that was interested in computers couldn't understand.

2:17:21Now, it turns out mostly who likes reading the book are people who already understand everything in the book, but they like seeing it all explained. But it had a chapter on quantum computing, and, you know, this was written in the early 90s. And I got this funny call from the publisher and said, you know, there's this weird thing. your book is the only computer book we have from the last century that's continuing to sell. And boy, did that make me feel... That should be on the cover, I feel like. I don't know. It kind of made me feel pretty weird. Fortunately, they didn't say the last millennium, right?

2:18:04But they said, you know, would you like to revise it? I went back, and you know, there are a lot of things that's happened since then. Understatement. But it was interesting because most of the stuff I had in the book didn't change at all. In fact, some of it I would have talked about certain things more and certain things less and so on. But one of the things I talked about was quantum computing. Really, even in quantum computing, there wasn't much I would change. What I said about it is, if you want to look for where something could be a real game changer, it's quantum computing. and it's got all this potential and all these hints that it could work.

2:18:45And there's good theoretical reason to believe that it would be revolutionary, but nobody's actually gotten it to be useful yet. And that's pretty much still the state that it's in. I was surprised that in the end, I decided it was sort of more interesting as a historical document of how computing looked in the 90s, and I didn't change it. But most of it wouldn't have changed anyway. So let me, at the risk of this going sideways, I introduced a really slippery term. But we were discussing earlier the possibility, if my memory serves me, that AI and developing AI, different types of AIs, could help us get a better understanding of intelligence writ large, different types of intelligence.

2:19:32We might, as Kevin mentioned, discover we're on the edge of the galaxy or universe. Exactly. Possibility space, not in the center. Is it possible that through AI or quantum computing or other aspects of studying quantum phenomena, that we will get a better grasp of what consciousness is? Recognizing, again, that that is a term that begs definition. And I mean, there are a lot of people who take different stabs at it, but sort of what it is to be aware that we're aware perhaps would be one possible way of offering that, but also how that emerges from simpler constituent pieces that maybe at some requisite level of complexity suddenly have this emergent phenomenon, which is consciousness.

2:20:23Certainly that's possible. My guess, and this is really just a guess, is that consciousness is going to turn out to be way less important than we think in the sense that it's going to be a very small piece of intelligence, and it might just be a kind of a hack. For example, okay, so I have a complicated idea in my mind, and I turn it into a series of grunts and grunt at you and whistle and grunt, and somehow you listen to those grunts and you construct an idea in your mind. And so we sort of went through this translation process. And so But we have a lot of our brain is devoted to that compression process of turning the idea into grunts and turning the grunts into an idea.

2:21:12So given you've got all the hardware lying around, you've probably had the experience of misunderstanding somebody. But what you misunderstood is actually more interesting than what they said. Right. Sure. Or vice versa. Right. Because your brain took the thing that they said and expanded into a sensible idea. And maybe it was more sensible than the one that they had in the first place. Well, so you could do that within your own brain just by talking to yourself. And so probably given you got all this hardware lying around for compressing and decompressing ideas, a good thing to do with the idea is to compress it, tell it to yourself, and see if you misunderstand it in an interesting way.

2:22:00And maybe consciousness is just some hack like that. I've often thought that one of the main benefits of language was not so much that it enabled collaboration with other people, but that it gave us access to our own thoughts. Can you imagine Trying to think without language is just like, it almost doesn't seem possible. So language, I think, was a dual-purpose invention that mostly gave us the power of communicating with ourselves, basically. Yeah. I think consciousness may be that. I think consciousness may be our access to our own thoughts. And that may be useful, but it may not be the most critical thing in intelligence.

2:22:43Like maybe you could not have it and you'd still be very smart. And maybe I wouldn't even be able to tell the difference. I think in that space of possible minds, we could think like things are really, really intelligent to have very little consciousness. Things that have a lot of consciousness that can't communicate, things that communicate. But I think consciousness is another kind of elemental primitive. Yeah. And I think you're going to have multiple entities that have access to each other's thoughts. and that might be even richer, so super consciousness that might be better. So I think this might be another case of us looking at what's apparent to us when we think of our thinking and we're very impressed with the things that are sort of very visible to us, like we were very impressed with our ability to play chess.

2:23:35But ultimately, it might not be so important. The proverbial drunk guy looking for his keys under the streetlight at night. And they're like, wait, I thought you left that in the barn. He's like, yeah, this is where all the light is. So we were coming up. I could keep going for another three hours. We're coming up on three hours now, which has gone by very, very quickly. Kevin, do you have any closing as we start to land the plane? Questions for Danny? Comments or questions? Complaints? Old feuds you'd like to revive? I might want to go back to the question of what are you trying to optimize in your life?

2:24:10because you were saying you were trying to optimize your time. Are there any other things that you are, the general trajectory of your life, maybe in particularly recent years, where you feel this is what you were trying to optimize, maximize? Or another way of saying is like, again, when you're deciding what to do, how to spend your time, the little time that we have, what's something that you are trying to make more of? I guess I try to ask the question, will this make a difference over how much time? How long will that difference matter? If it makes a lot of difference after I'm dead, I'd rather do that.

2:24:53And I think a lot of people think I want to make a difference, but I think they weight it much too much to the near term. And so, for instance, I really admire Bill Gates as a philanthropist. He works really hard and he's super smart about it. But one thing that bothers me about some of the things that they do is because they try to measure everything, they try to do things that make a difference in the time that they can measure them. And I think that that is maybe not the right metric to be optimizing because it doesn't allow for the long tail of time, of impact, of things. It's like, you know, when Claude Shannon like invents the bet, the differences that that makes are just becoming apparent now after he's dead.

2:25:51You know, that's something that over time makes a huge difference. But if you tried to measure it during his lifetime, it would have been really hard to give it any credit. The long tail of impact. All right, I'll pick from my grab bag of favorite questions. One of them is pretty simple. It's a metaphor. But if you could put anything on a giant billboard to get a message, an image, question, anything to many, many people, hundreds of millions, billions of people, let's just assume they understand the language. Could be a quote. Could be a quote from someone else. could be a motto that you, or philosophy that you live by, could be anything at all.

2:26:34What might you put on that billboard? It could be an ad for your company. No ads. That's the one rule. Well, in a sense, I think I've answered that because I think the most successful example of that was Stuart Brand having a picture of the whole earth. I think he realized that when people saw that picture of the whole earth floating in space, they would think about everything differently, and they did. And so, to me, there's no picture like that of the future. You can't conjure up an image. You can conjure up an image of the past, of maybe the pyramids or something like that. But there is no iconic image of the future.

2:27:20And if you could imagine something to put on a billboard, that sort of made people see the future and believe that there was that future. That's what I do. And the 10 ,000-year clock is the best approximation I can do to that. But I think that that's what the world needs. I think it needs that picture that puts the context of everything today in the context of the development of humanity over tens of thousands of years. And I think that would make it a much more optimistic picture for everybody. And let me just add, because I don't think we have, this clock is real. It exists right now. It's inside a mountain in West Texas.

2:28:10It's inside a vertical tunnel with a spiral staircase carved into the rock, and it's hanging almost 500 feet. It's a mammoth, mammoth, monumental clock that is going to tick for 10 ,000 years. And my impression of having visited it is that it feels like the clock has always been inside the mountain. It feels ancient. The scale, the scope, the ambition, the tooling required, like this site identification, everything is beyond belief and i know we haven't spent a lot of time on it i will link to a few things for people who are listening also wasn't sure how much you could say about certain aspects of it so we spent a lot of time on other things but danny is there anything else that you would like to say about the 10 000 year clock no actually i think it's good that we spent time on other things.

2:29:08It will speak for itself. It's a story. And actually, my favorite thing about the 10 ,000-year clock is I run into people regularly who've heard about it but assume it's just a myth. Have you seen Bigfoot? No. Bigfoot? In West Texas? Yeah, well, you get all kinds of versions of it. It's in Nevada. I ran into somebody who said it was in China, you know. But that to me is very satisfying because stories are actually what really lasts. And really, you know, your question about the billboard is like that, of, you know, what's an idea you want to put in people's head that sort of stays there? And an idea has a lot more sticking power than any physical thing you could build.

2:29:58And so I love it that it has sort of become a story with a life of its own. And that to me is exciting. And the fact that there's this giant thing sitting in a shaft in the mountain. Yeah. It makes me think of Indiana Jones and the last crusade. And I can imagine the tagline, see you in 10 ,000 years.

2:30:22Yeah. If anybody believed it. Well, that's part of selling the story, right? And Danny, Danny, people can find Applied Invention at AppliedInvention.com. Is that the best website? Is there anything else? But there's nothing there. It'll just say, it'll have our address and zip code. All right. For those fans of zip codes, you can go to Applied Invention. But they can contact me with that way. All right, perfect. Great. And I do want to meet smart people. I told you that that's like what I'm seeking for is the brilliant people with different ways of looking at the world. So Danny, have you, have you spent any time with Derek Sivers before you guys met before Derek Sivers?

2:31:00Do you know this name? Oh, all right. Well, Kevin and I both know Derek. I feel like you guys would make for a fun meeting. Danny, Kevin, thank you for taking the time. This has been absolutely fantastic. I have tons and tons of notes. We didn't even get to the giant robot dinosaurs another time. And for people listening, you will be able to find links to everything that we've discussed at the show notes on tim.blog slash podcast as per usual. And until next time, as always, be just a bit kinder than is necessary, not only to others, but to yourself. And thanks for tuning in. Hey, guys, this is Tim again, just a few more things before you take off.

2:31:38Number one, this is five bullet Friday. Do you want to get a short email from me? Would you enjoy getting a short email from me every Friday that provides a little morsel of fun before the weekend. And Five Bullet Friday is a very short email where I share the coolest things I've found or that I've been pondering over the week. That could include favorite new albums that I've discovered. It could include gizmos and gadgets and all sorts of weird shit that I've somehow dug up in the world of the esoteric as I do. It could include favorite articles that I've read and that I've shared with my close friends, for instance.

2:32:16And it's very short. It's just a little tiny bite of goodness before you head off for the weekend. So if you want to receive that, check it out. Just go to 4hourworkweek.com. That's 4hourworkweek.com, all spelled out. And just drop in your email, and you will get the very next one. And if you sign up, I hope you enjoy it. About three weeks ago, I found myself between 10 and 12 ,000 feet going over the continental divide, carrying tons of weight, doing my best not to chew on my own lungs, and I needed all the help I could get. And in those circumstances, I relied on Momentus products every single day and every single night.

2:32:54Now, regular listeners probably know I've been taking Momentus products consistently and testing them, the entire spectrum of their products, for a long while now. But you may not know that I recently collaborated with them, one of the sponsors of this episode, to put together my top picks and I'm calling it my performance stack. I always aim for a strong body and sharp mind. Of course, you need both and neither is possible without quality sleep. So I didn't want anything speculative. I wanted things I could depend on and it is what I use personally. So I designed my performance stack to check all three boxes and here it is.

2:33:27Creapure creatine for muscular and cognitive support. The cognitive side is actually very interesting to me these days. Whey protein isolate for muscle mass and recovery and magnesium threonate for sleep, which is really the ideal form of magnesium as far as we know for sleep. I use all three daily and it's why I feel 100 % comfortable recommending it to you, my dear listeners. Momentus sources CreaPure creatine from Germany and their whey isolate is sourced from European dairy farmers held to incredibly strict standards. And I've chatted with the CEO about their supply chain, about how they manage all of these things.

2:34:02It's incredibly complex and they go way above any industry standards that I'm familiar with. And I am familiar with them. All Momentous products are NSF and informed sports certified, which is professional athlete and Olympic level testing. So here's the main point. What's on the label is exactly what you're getting. And this is not true for the vast majority of companies in this industry. So this is a differentiator. Try it out for yourself and let me know what you think. Visit livemomentous.com slash Tim and use Tim at checkout for 20 % off of my performance stack. One more time, that's livemomentous.com slash Tim.

2:34:39I'll spell it out. It's a long one. Livemomentous.com slash Tim. So livemomentous.com slash Tim for 20 % off. This episode is brought to you by 8sleep. I have been using 8sleep pod cover for years now. Why? Well, by simply adding it to your existing mattress on top like a fitted sheet, you can automatically cool down or warm up each side of your bed. Eight Sleep recently launched their newest generation of the pod, and I'm excited to test it out, Pod 4 Ultra. It cools, it heats, and now it elevates automatically. More on that in a second. First, Pod 4 Ultra can cool down each side of the bed as much as 20 degrees Fahrenheit below room temperature, keeping you and your partner cool, even in a heat wave.

2:35:22Or you can switch it up depending on which of you is heat sensitive. I am always more heat sensitive, pulling the sheets off, closing the windows, trying to crank the AC down. This solves all of that. Pod 4 Ultra also introduces an adjustable base that fits between your mattress and your bed frame and adds reading and sleeping positions for the best unwinding experience. And for those snore heavy nights, the Pod can detect your snoring and automatically lift your head by a few degrees to improve airflow and stop you or your partner from snoring. Plus, with the Pod 4 Ultra, you can leave your wearables on the nightstand.

2:35:53You won't need them because these types of metrics are integrated into the Pod4 Ultra itself. They have imperceptible sensors, which track your sleep time, sleep phases, and HRV. Their heart rate tracking, as just one example, is at 99 % accuracy. Conquer this winter season with the best in sleep tech and sleep at your perfect temperature. Many of my listeners in colder areas, sometimes that's me, enjoy warming up their bed after a freezing day. And if you have a partner, great, you can split the zones and you can sleep at your own ideal temperatures. It's easy. So go to 8sleep.com slash Tim and save between$400 and$600 on the Pod 4 Ultra.

2:36:318sleep currently ships within the US, Canada, the UK, select countries in the EU and Australia. This special offer is valid until December 14th.