Future of Science and Technology Q&A (February 23, 2024)

Hello, everyone. Welcome to another episode of Q and A about future of science and technology. I am traveling again, so I am not in my natural habitat and my setup for live streaming is a bit jerry rigged. Okay, let me see. What do we have here? Gosh. Well, there's a question here from Jamie. What is the future of business? You know, I think a lot of these questions about, you know, how the future of things evolve, the way to start thinking about them is to look at how they have evolved in the past. And if you look at kind of, you know, what have been the big things that have changed business. You know, for example, there's the invention of companies with, you know, joint stock companies, that's what, 1600s or something. And the invention of things like. Things like the sort of the improvement of communications, the improvement of ways to transfer money from here to there and so on. It's only been a few decades that there haven't been really complicated arrangements that you need to make to have money paid from one country to another and so on. So that that kind of globalization of the flow of money, the fact that sort of the, the fundamental problems of logistics have been solved and you can make something in one place and have it shipped to another and have expect that that will just work. These are things that have changed over the last, I don't know, few, few decades to century or something like this. And then there now, of course, with respect to logistics, if people start saying we're going to do manufacturing in orbit or we're going to do manufacturing on the moon, then logistics will again be a little bit of a thing. But I think it will be sort of a factored issue of there's the logistics problem and then there's the what are you going to make, how are you going to make it, where are you going to sell it, and so on. And then there's the question of sort of the organization of, of, of companies and so on. Clearly the Industrial revolution had a big effect and sort of. Then the development of production lines and things like this had a big effect on kind of the, the way that, you know, it had been more or less artisans, where one person builds all the way through, so to speak, giving way to this kind of much more modularized, segmented kind of way of approaching things. I mean, it's, it's. It's often claimed that in economies in general, the, as economies develop kind of you get this more specialization of activities, but even in the terms of the building of an individual thing, you get more of that kind of Specialization, I would say that there's perhaps a trend in the opposite direction. As more things get automated, it becomes more possible to have kind of the one person band where one person is sort of in control and is initiating, you know, creating the software and doing this and doing that and doing the other. As more things are kind of able to be kind of outsourced in effect to other organizations. As more things get automated, it becomes I think more possible to imagine smaller companies that can do big things. I mean, certainly the effort with our company over the last 37 years or so has always been automate as much as possible so that we're in this position where, where we can do more with fewer people, so to speak. And my guess is that that's a thing one will see in some generality. Another question is sort of the arc of people's lives and how that will evolve and the extent to which people will, for example, be doing one thing consistently for a long time versus be changing from this to that over the course of years versus doing things where it's like everything is a gig type thing. You know, I think that, that there tend to be these swings. So for example, the possibility of the modern gig economy has happened because of the web, because of communications mechanisms, because of mobile phones, because of the aspects of kind of the things that one just couldn't have imagined if people didn't have some sort of sophisticated way to connect to the outside world on them at all times, so to speak. And I think that, that. So, you know, there's sort of the question of, there's also the question of how what people do relates to how they're educated. For example. I mean the, you know, education in a sense there's more to know in the world. So one would think there's sort of more education that needs to be done. There's always the sort of counterbalancing force is more things get abstracted so it's less like you just have to learn more facts. But it becomes all these facts get turned into one general principle that you can then learn so you're not having to learn all those facts. But still, as a general matter, if one looks at the last few decades, the typical sort of amount of education that people go through has somewhat increased. Again, I don't know whether that will continue because I think as there's more automation, as efforts like our effort to sort of make as much of the world as possible computable, as that advances, it becomes less necessary for people to themselves have various skills, know various things it becomes more possible to expect to sort of delegate those to automated systems. So I think in terms of education, I think that the value of generalism in education will dramatically increase in the future. I mean, the idea that what has been the case for the last quite a few decades of increasing specialization, I think is probably not that will reverse. But there's this question of what is the sort of arc of a typical life? How many years do you spend getting educated? I mean, there's some professions where you don't get to actually be in action until you're at least in your mid-30s, maybe late-30s, because there are just so many years of kind of education and fellowships and this and that and the other to get to the point where you can be sort of an actual person in action. And obviously lives are getting longer. So that isn't. But that's still, you know, that whole dynamic of how long, what's the sort of education period. Then what do people expect to do? Do people expect, for example, that, you know, how do they think about the jobs that they do versus the other things they do in life? What do they think of? Do they. Are they able to do things, you know, the things. Where do people mostly have kind of the. The, you know, what does the sort of work life balance look like? What does, you know, how do people. And that affects both. Both the things people can do to put productive stuff into the economy and the ways people are going to consume things. And I think another thing to point out is that if you look over the course of time at what jobs have people done? Back 150 years ago in the US for example, most people were doing agriculture that got largely automated. And then what tends to happen is that you see more and more fragmentation. There are more and more different kinds of jobs that people do. But I think it. And there's sort of a progressive evolution of what jobs make sense to do and then what kind of corporate systems make sense to deal with those jobs. So, for example, if the main thing you're doing is manufacturing stuff that has a certain rhythm to it because, you know, you've got to build the factory, you've got to spend a bunch of money building the factory. The factory is in a definite place, et cetera, et cetera, et cetera. If what you're doing is providing services of some kind, people can be doing that from wherever. Maybe they have to see people in person, but they can in general. It's a much more distributed thing. If you're doing something like your business is kind of I don't know, content creation or something. And maybe your business is aggregating content creation. That's a, again, a different, a kind of business that has different kinds of requirements. You know, I suppose it's another dynamic of businesses is how much, how much capital do you need up front to make the business possible. And that's something where one has seen kind of a progressive decrease in that in some areas where it's kind of funny because there are areas where there's suddenly needs to be a huge amount of capital like train your AI, spend tens of millions of dollars training your AI, or spend tens of million dollars crawling the web or something, or spend lots of money getting some distribution channel where you've got sort of people distributed all over the world. So I, I think this is a difficult question what the future of business is. I think, you know, the other, other thing that's another dynamic is when how closed are different kinds of distribution channels. You know, there have been times when if you want to make a such and such, if you want to make a drug that's a very complex distribution channel, it involves, you know, maybe as much as a billion dollars of regulatory compliance work to get the drug finally on the market. That's a. Whereas if you're, you know, putting up a website, there's much less that's, you know, anybody can put up a website. And there are sort of industries where things are more closed in terms of distribution channel. More open in terms of distribution channel. There's often a sort of all sorts of forces at work in terms of closing down distribution channels when there are a small number of big players who sort of want nobody else to be able to enter the market or cases where sort of, lots of, lots of scurrilous things happen when there are lots of players in the market and nobody's really paying attention to what's going on, those kinds of things. So there a bunch of dynamics. But I think that that also affects kind of what, what is possible in a business. Because if you have a sort of distribution channel where you have a certain amount of effort that needs to be gone to, to make it into the distribution channel, then it's, you know, that forces a certain investment to even start a business in that area. And I think one of the things that's happened in recent times is thanks to the web primarily and thanks to a bunch of e commerce things and so on, the minimum sort of viable business size has business that can have global reach has decreased. I mean, in the past, the size of business that can now sell things all over the world was a size of business that could only operate kind of the very local corner store type thing. So a question one could ask is, is the future of business more small companies or a small number of big companies? How much is what's going to happen in terms of kind of the fragmentation of that versus the consolidation of that? I think that these things like the sort of the easy to get into distribution channels tend to go towards lots of small operations. And when so the distribution channels close down, that's when you have sort of consolidation into a small number of big operations. It also is affected by sort of what the investment needs to be to get into each, each business and so on. I suppose in one might, you know, a lot of industries go through these phases where know, the industry starts up, there are a small number of players who kind of know how to do that thing. Then people realize, gosh, everybody can get into this business. And then lots of, you know, it becomes the, the next cool thing to make a startup about. And then lots and lots of people get into it. And then after a while, first of all, many of those entrants won't succeed. But then some of them, the, the ones who do succeed, there often are economies of scale that lead things to get consolidated until in the end, even though there was a period of time when there were lots of different entities, there end up being a small number of entities after a while. And maybe that's just a natural thing in the evolution of industries. That one goes from sort of the small ones who are starting it to the wide range of ones who are trying lots of different things down to the ones who can achieve economy of scale when the business is more developed. I'm not sure that that will always be the case because I think that as things get more automated, as distribution channels get simpler, that you could end up with a situation where there really isn't much economy of scale. Maybe in some industries which would sort of change the character of what kinds of organizations there can be there. Another question is within companies, for example, within, you know, how will they be structured? How should they be structured? And people try these experiments where it isn't kind of a hierarchical, you know, small number of people are in charge, larger number of people are doing things, they try experiments. Those experiments so far have been quite unsuccessful, I would say. I think that the, the thing that has, I tend to think when you have to, if you ever have to do new stuff, there has to be a person in charge who's making the decision about what new thing to do. If it's like if you've got the giant committee that's running things, it's really hard to have to decide for a big committee to say we're all going to decide to rush off in this direction. That just doesn't really happen. That's usually one person or a small number of people who say we're going to try and do this different new thing. I think that people. Another thing that's obviously a detailed dynamic is the globalization of lots of businesses of. Certainly our company has been a very geodistributed company for well more than 30 years now. And that's something that, particularly in these areas that don't involve sort of physical objects and so on, I think that's something one can expect to very much continue. But I don't know how that, in, in terms of things like, you know, how is a company organized? I, I'm not sure that that's affected that terribly much. I think that the thing that has, you know, there's this question again, it's, it's also affected by company size. If it turns out that in the future one can sort of get away with having a smaller company then and still do lots of stuff, then lots of. The kind of structure in the organization of companies doesn't need to be there. You know, it's always an interesting question, what's the minimum size of such and such a kind of thing like, you know, can you have a country that's really tiny because, you know, or is there a minimum size that you have to have because you've got to set up things to get your, you know, the, the mail delivered to you or the, or, you know, the Internet hooked up or whatever else where. So there's a minimum size to just deal with all the mechanics of, you know, signing the treaties, making the deals, etc. Etc. Etc. I think, you know, the, the minimum size of country people sort of experiment with a bit and it's probably gone down, but it. Probably it's gone down, but again, it's the same type of thing as right now for a company. For example, the minimum size of a company is, is trivially small, so to speak. And which, which again hasn't always been the case. I mean, it's, it's what else to say about, about business. I mean, I think people. There's always these questions about will. I mean, there's, there's. People are always saying things will be different because something, you know, it's all going to be AI's buying and selling things. And yes, I'm sure there'll be a bunch of. I just say generally what I want and then my AI goes and negotiates the price and all that kind of thing. Yes, that kind of thing will happen. I'm not sure how significant that will be in the overall structure of business, so to speak. I mean, I think the thing one might think about with respect to things like that is sort of the efficiency of markets. That is, you know, it's, it's like the phenomenon that happened over the last, I don't know, 20 years or more of, you know, commercial planes being full of people most of the time because sort of, there's been much more sophistication in pricing models and so on. And, you know, the, if they didn't sell a few seats, they'll, you know, sell them at the end at a lower price type thing, those kinds of things which. And that is something that was made possible by what I suppose in today's world will be called AI, but in, in, in the past was just algorithms and computers and so on. And I think this, this idea that sort of business can be made more efficient and you can kind of match up supply and demand more efficiently. That's surely a thing that has come to a bunch of industries. It will come to more industries. It's. That's something that will be, will be achieved. I don't know. You know, it. It's another, another question is how will people learn about stuff in the world? I mean, this is, this is one of the complicated things of. You know, there was a time when sort of everybody watched a few channels of evening news in some countries, just one channel of evening news. And then there were, you know, limited numbers of newspapers. And then there were sort of more of this and that. And then the web caused things to sort of explode in terms of the number of different places that one can get information. And then, of course, things became more complicated as there were, you know, a fairly small number of social media channels, but then lots of sort of algorithmic AI kind of selection of what different people got to see in those channels and so on. And I think it's this, this question of sort of how do people get to know stuff and how does that, how does the, how people get to know stuff enmesh with the way that business is done? And clearly that's been a thing. You know, advertising is a fairly big business. And this whole sort of dynamic of what do you make versus how do you tell people what you make and so on, I'm not sure. I'm not sure how that will evolve. I mean, maybe there will be sort of more effectiveness. If I could describe to my AI, if I could kind of prompt my AI with what I kind of want out there in the world, it would be, I mean, for example, one very straightforward thing which I don't think there's an easy way to do. It's like there are products. I can imagine that if somebody actually made it, I would buy it. But I don't think there's any mechanism for that fact to be in any way to be communicated or even for me to have an AI that watches. Oh, when somebody has a this or that kind of thing, I want one of those. You know, in the past there would be for sort of tech gadgets and so on, there'd be a limited number of kind of places that reviewed typical tech gadgets, but that kind of decayed because, you know, everything got much more fragmented. So, I mean, that's something which I can imagine where it's kind of like a feature of the market that probably is efficient now is certain aspects of pricing, at least in some domains. What's not efficient is things like things I imagine I want versus anybody knowing that I want that. Now it's not. Most of us don't actually spend a lot of time thinking about products that don't exist and imagining what we might. What if it existed we might want. And I suppose another thing that one could imagine happening is sort of a tighter feedback loop. I mean, over time there's been more feedback. But, you know, this is again another complicated issue because the more feedback there is about the things you use, the more you're kind of giving up your privacy to potentially benefit yourself or other people. I think a thing that could change some of that if the whole business of fully homomorphic encryption and so on actually works out and it seems like it's finally moving in the last year or two, then that might affect some of those things. So the idea there is normally when one does encryption, and by the way, encryption was one of those things that has enabled a lot of things in the world and it wasn't clear that was going to sort of consumerize in an efficient way and has. But, you know, encryption is like you can send a message which, you know, only the. Only the recipient can decode, or you can have a digital signature where you can guarantee that this really was. This was really created by the person who said they created it. But one of the issues is once you have the Key you can decrypt the message. The idea with fully homomorphic encryption is that you could for example, have a database that has all kinds of data. The database is encrypted. People could make queries of the database finding out aggregated facts like what was the total of all these numbers? But not being able to find any of the numbers on their own. And that's been a hard problem in kind of the mathematics of cryptography to figure out is there a way to, to sort of get to the aggregate, get to things where operations have been done without decoding every individual element, without somewhere having to have all those individual decoded elements? And that's a type of thing that I can imagine might affect some of this sort of mechanism for how one gives feedback about things and so on. Anyway, a few thoughts, much to say, I suppose, See. Oh yeah, Prabh asks what's the future of contracting and procurement and the question of smart contracts and so on. That's an interesting one. I mean I think that the idea of computational contracts, the idea where you write contracts in computational language, not legalese, that is surely coming. I mean it's come in a few industries. It's come in, it came a long time ago in stock options and things like that. It came, it hasn't, I think still come in things like mortgages. It has come a little bit in service level agreements for you know, providing Internet service and things like this. But it's something where I would, I, you know, I, in the future I have to believe that that contracts will typically be written in computational language, not in legalese. And that means that they can be automatically executed and that means they can be executed much faster. You can have much, much, much larger chains of contracts that affect things. It changes the work of lawyers who often have been writing what amounts to code in kind of legalese. Let me make sure there's no bug here. There's no way that this can sort of run wrong type thing. Well now it becomes just raw code that you're writing that represents what you want to have happen. And I think that kind of idea that you can sort of specify what you want in this very rigid in a sense way in many kinds of settings that will be a good thing. In some kinds of settings it's like, well, we don't quite know. We need some wiggle room. And particularly when extremely large amounts of money are involved, that's probably there's some wiggle room and some. If it all goes horribly wrong, we'll take it to court type thing. Is a thing that's probably needed, but I think that's something that I think one could see happen. And just as when you go to a webpage and an ad is going to come up, there's, I don't know what it is, 14 steps or something that are involved in kind of people bidding on whether they're going to show you the ad and who are you and all this kind of thing before you see that ad in a few hundred milliseconds. And I think we can expect that more things in the world will have this kind of sort of AI to AI communication that's happening, the sort of infrastructure of AIs that's happening outside of human, faster than humans think about things and sort of outside of the human domain. And there'll be this kind of, kind of whole, whole infrastructure of AIs as there already is in industries like Internet advertising, so on. I think it's, I don't know how important blockchain is in this whole thing. I think that a lot of these kinds of contracts, it's like they're small enough amounts of money. They're, they're, you know, a centralized database is just fine. Nobody's really going to be concerned about that. When the amounts of money get really large, people are going to want the sort of security of governments and legal systems and so on, and they're not going to believe in some sort of automated blockchain mechanism, at least not for quite a while. But I think there's sort of an intermediate range of values and cases where multiple entities in an industry, multiple different organizations in an industry, for example, who haven't necessarily ever agreed on standard for how to do things, they can all sort of throw everything into the blockchain. And that's a useful way to make use of that kind of thing. But yeah, let's see. Well, the question here from desperate. How do you see universities keeping up with, I think, mean, AI? It seems that universities will become a pre selection for corporations. What, what are your thoughts about this? Gosh, well, you know, I think, I think what people, first of all, the idea, that kind of education, that to operate in the modern world you need to go to high school, you need to go to four years of college, you need to do this, maybe you need to get a master's degree, et cetera, et cetera, et cetera. You know, I think as the world, the world's gotten more complicated in certain ways. But the world's also gotten simpler in certain ways because of automation. Because people, you know, Design user interfaces so that other people can find it easy to use things. It isn't the case. It's, it's. I don't think that the world is getting to the point where, I mean, you don't need to go to, you know, 15 years of school to learn how to work a typical smartphone. It's in fact, there's zero school for doing that because efforts been made to make user interfaces such that people just from their ambient experience sort of know what to do. Well, often the case, I mean, for example, if it came to somebody like me playing a modern advanced video game, I would have no idea what to do because I haven't kind of climbed that ladder of learning those kinds of things. But that's not something that kind of the education system is contributing to, so far as I know, unless you're an esports league in some school somewhere. But I think this idea that sort of everybody needs to go through all these years of education to operate in the modern world doesn't seem right to me. I mean, the modern world, as I said, has become more complicated, but it's also become more automated. And I think it is, it has become one of these things where there are lots of kind of commercial forces. Effectively, universities are, you know, they're a business and they're a business which has turned out to be a business that lots of people can be signed up for. I'm not sure how convincing that really is. I think that there are the, the role of universities and so on. My own view is if they can teach people broad, if people can learn broad things and can learn kind of how to think and can learn about lots of different areas and so on, that will be a good thing for their future lives. I think this notion that you go to university to learn to program is a bit silly. I mean, I know that most of the really good programmers I know never took a programming class in their lives. And this is, which is sort of an ironic thing, but it's something where you're, you're realizing that this kind of the trade school aspect of learning to program, I mean, I might say that if people like me do their job properly, that most of what you could sort of trade school learn like that is what we will have successfully automated. And the things that one needs to learn about sort of computational thinking are things that relate to thinking in general and are not things that are kind of learned in this, in this very linear way. But you know, my own view would be that sort of. It's not clear why so many people go to college now. Going to college, particularly in a place like the US is a whole package of stuff. It's not just you go to classes and learn a bunch of stuff. It's a whole social network. It's a whole kind of get away from your hometown type story. It's, it's all kinds of other things. It's become this kind of bundle of sort of become an adult type stuff that, and you know, there may be value to that quite independent of the, you know, may just be that the, the, oh, there are these classes that's just kind of an excuse to go do this thing for four years type, type thing. But I think in terms of, and I think it is an issue today, you know, what are universities for? Are they for sort of cultural development of people? And some people would perhaps unkindly say, you know, sort of advocacy, activism type things versus are they for kind of teaching facts? Are they for sort of passing on knowledge to future generations? What exactly are they for? I mean, my own view is that there's high value in having organizations that pass on knowledge to future generations and try and develop new knowledge. And that's a really worthwhile investment for the future. But that's not necessarily what universities actually are today. I think in terms of the sort of structure of what universities teach. I mean, it's a complicated dynamic because there's, there's, you know, there's a certain, you know, how do the professors change? You know, a professor goes and they get a PhD in this or that and you know, they'll, that's what they know about and that's what they naturally could teach for a long time. If new things come along, they can't naturally teach those things. This is a big issue with kind of teaching computational thinking, and it's going to be interesting to see how that plays out. I'm trying to make my own efforts, in addition to all the technology we've built, I'm trying to make my own efforts to actually define kind of what should you teach if you teach computational thinking? It's not completely obvious. If you're a professor of X, what X should you be a professor of to do the best possible job in, for example, teaching the computational thinking course that I'm trying to develop. It's not obvious, you know, being a computer science professor is not the obvious choice there, but it isn't really. Yet. There isn't really a category there yet. I mean, I've always thought that just as that's computer science cs, there's kind of a cx which is computation applied to everything computational X. And that's a little bit of a different kind of thing than what computer science has traditionally been. Now what happens at universities and it's often a complicated issue. Does a particular field evolve in a certain direction? Like in the 1950s for example, mathematics could have evolved to be computer science. It didn't. Mathematics stuck as what mathematics is, perhaps to its benefit. And computer science completely branched off. Similarly, something like molecular biology kind of branched off from what was sort of the typical biology department with, with you know, creatures in jars and things like this. And I think the, and there have been, well, there been some less obviously successful. It's sort of interesting to see the history of, of departments at universities. I mean applied mathematics has always been a complicated one because. And it's, and that's sort of a, I would say a non model for cx. You know, applied mathematics. It's kind of like, it's often confusing. What is it? Is it kind of studying partial differential equations? Is it operations research? Is it discrete mathematics? Is it? What is it? Because a lot of the time mathematics as it's applied to different fields gets absorbed into those fields. Not, it doesn't make so much sense to talk about. It's sort of the, it's kind of the, the, the, the, the thing that is reaching out, you know, it's the arm that's reaching out from mathematics to other things. But the arm hasn't actually touched anything yet. And that's a, that, that, that, that's something that, it's a little bit hard to define what that really is. And it has been differently defined in different universities and so on. Much it's much more diverse than you know, if you say what's in the math department, what classes do they teach and so on. There's a pretty fixed set of things that's to be expected in a math department. Applied math, much more diverse. And there have been other efforts. I know there are other kinds of departments that have had complicated histories like geography departments. That's one that, you know, in modern times there are endless things that are sort of geospatial, this, that and the other, but they tend to sort of absorb into their different fields. There was a time in the U.S. for example, when a lot of universities had geography departments in most places went away. I think in today's world it's a little bit like the computation computational X thing. There's sort of geography applied to all kinds of different areas. And there's a question of is there sort of a centralized thing? And another, another one that, well, has had ups and downs is linguistics, where there are many different languages you can learn. And then there's sort of a centralized science of linguistics. Maybe there have been sort of episodes of people thinking there's more and less of that. But you know, that led there to be a bunch of, starting in the 1960s, I think, a bunch of places that sort of had the big growing linguistics department and then maybe that got disbanded in favor of just, oh, we're teaching this language and that language and the other language. And then there's sort of, there's not so much of a centralized thing there. It's a bit similar, I think, to these other areas where there's kind of a core set of ideas perhaps, but then they get distributed to a bunch of different application areas. I mean, I think this whole question about how universities sort of make the jump to computational X is an interesting one. It's quite challenging structurally because it's like, who's going to teach that stuff? How does it really work at a university? Does it become a new department? Is that even the right idea? Does it become sort of a thing like often happens with mathematics, where the mathematics department teaches service courses that are taken by the people who are going to be pharmacy students or going to be physics students or whatever else. And sometimes those various departments say, like business schools often say, we don't want the math department teaching the math for business, we're just going to teach it ourselves. And, and sort of how that dynamic plays out for all these computational areas is an interesting one to see. And I think that is, I would say that the sort of computational X is much more the content challenge for universities. The, the question of, of. Of how things work with sort of the who should go to college and why is then a quite separate issue for universities and their evolution. And you know, I tend to think that, well, it's something where there is a lot of value in people doing what has traditionally been called sort of the liberal arts type education thing. And there will be increasing value to that. I mean, I found it interesting, I probably mentioned that on these live streams before when prompt engineering kind of became a thing about a year ago, that you say, who's going to be a great prompt engineer? Well, it's particularly people who are good at expository writing, it seemed not people who are good at doing the kind of, the obviously techy kind of programming kinds of tasks. Let's see, I probably have time for one or two more Questions here. M. Rudo asks, what is your opinion about terraforming the Moon? Sounds pretty hard to me. I mean I've heard about terraforming Mars, I've heard a little bit about terraforming Venus. I think terraforming the Moon seems pretty difficult. I don't think that much atmosphere stays in a thing one sixth the gravity of the Earth. So even if you tried to sort of, even if you imported or, you know, you did something to the ice caps of the moon to put lots of oxygen, I don't think it would stay. I, I should work that up. I don't think it, I don't think it will stay at all. I think that the, you know, in terms of do we extend our civilization to the Moon? I think it's kind of like have we extended our civilization to Antarctica? There are things that can be done in Antarctica that can't easily be done elsewhere. And you know, there are a small number of people with their neutrino detectors and things like this and telescopes and so on in Antarctica. But in general it's not a booming industrial place and there's a treatise that say it shouldn't be, but it's still, it's not clear. There's a lot of force to get that to happen about the Moon. You know, I think one of the big opportunities there is some that there's that people say, oh, let's put the really dirty heavy industry or something on the Moon and just ship the results back to Earth and we'll just let all that nasty pollution sort of float off into space. Don't know whether that really makes sense. I think it's probably easier, it's probably easier to recycle that stuff on the Earth than it is to take it to the Moon and let it belch out all kinds of, all kinds of stuff into space. By the way, I might comment, the intuitive machines. Moonlander, I gather, has this rather peculiar lunar library project which has had unfortunately two previous attempts to make it to the moon, both of which didn't quite work out. This one did work out and it has a bunch of our, and my writings on it, which is, which is I suppose, fun. Although it's in microscopic, you know, microscopically etched onto something and I'm not quite sure what it really means to have one's, you know, writings on the Moon. I'm not sure, you know, I'm not expecting that, you know, after our civilization is wiped out that another post to us and rulial space civilization comes along and says wait a minute. There's this thing on the moon. We can reconstruct all the things that these guys did, but it's fun anyway. Let's see. So, I mean, I think these questions about terraforming Mars, terraforming Venus, I keep on hearing different things. Every year or two, I will hear a completely different view on what's possible, what's not possible, and so on. I think it's an interesting ecological question, ecological ethics question. Should we terraform Mars? Should we terraform Venus? Clearly, if there was anything. If we think just as we sort of extend our ethics to other humans, we extend our ethics to animals, we extend our ethics to some things in the world. I mean, we don't have any qualms about saying, that species of virus that's been causing us all this trouble, let's wipe it out if we can. We don't. That's not a thing to which we say. Most people would say, you know, people would be pretty. Pretty unified, I think, on saying, hey, we should just do that. And similarly, it's, you know, when it comes to sort of, do you. Do you terraform Mars? And, well, what if there was something there already? What if there was something that is life like on Earth? Well, then we'd probably feel all squeamish about it. What if it's life that isn't like on Earth, where it's really a definitionally complicated thing to say, is it life? Is it not life? What do you even mean by that? I think then the sort of ethics of what we can or can't do with it are complicated. And, you know, by the time we just say it's just like a rock, there isn't a lot of ethics that people feel about what you do to rocks most of the time. And, you know, I think it's partly because we extend our internal feelings to that we have about ourselves, to things that seem like us, like other humans, but rocks don't seem much like us. So I think it will be an interesting kind of sort of dilemma of sort of what happens if we have that capability. Do we say, you know, well, if there was a form of life, let's say, that we do not yet recognize, but in 200 years we might recognize it. It's kind of like archaeological excavations, the ones that happened a couple of hundred years ago, people often blew things up that today we would very carefully preserve. They blew things up because they didn't really know that was an important thing, so to speak. And so similarly, we might say, well, gosh, you know, when it comes to Mars, maybe there's something very important about this aspect of something on the ice cap of Mars, if we blow it up and try and put water into the atmosphere or something to help terraform it, then we will have destroyed this very important piece of heritage of something. We don't know that yet, but it's rather hard to know because the thing that you might be destroying might be something whose purpose is only understood in the future. And then the question is, by the time you start thinking in that way, you pretty quickly get yourself all tied up in knots and it's like, oh my gosh, we can't do anything. So that also is not a realistic thing. I think one owes more to one's own time perhaps than one does to the unknown future. I mean the future has a way of sort of working itself out, whereas in one's own time one's living one's own time, so to speak. I mean it's kind of like people might have said at the time of Malthus, for example, they might have said, oh my gosh, the world's going to run out of food, we've got to stop. Let's, let's stop having children. Let's, you know, let's make sure the population of the world goes down, etc. Etc. Etc. That didn't turn out to be a thing because agriculture got much more efficient and that wasn't, you know, those kinds of predictions didn't, didn't happen. Let's see. Okay, one, one last question here. I'll, I'll try and tackle. I see quite a lot about more sciency kinds of things which I'm very interested to chat about. But I think it'd be another time the real asks, talking about companies and countries that have the forethought to have their own supercomputers do. How do I see the inevitable increase in the gap between ultra high tech nations and companies and the rest of the world? You know, that's been one of these things that's gone up and down like late 1970s. I used to use computers that cost quarter million dollars in those days, which be maybe a couple million dollars today. And I was proudly using these computers and being able to figure out all kinds of cool things with them. And these were expensive computers and other people didn't have access to those kinds of computers. But then it was sort of disappointing to me in some sense, satisfying in other senses that by sometime in the, I don't know, late 80s or something, early 90s, it's kind of like Everybody's got the same kind of computer. It's like, you know, we've all got laptops that have about the same power, same speed. You know, you buy different variants of how much memory, but it's all pretty similar. There's not this big gap between the kind of the fancy machines and the kind of the hoi polloi machines, so to speak. And I think that at different times in history there's been more or less of that gap. I mean, I remember at a time I was, this was mid-1980s, actually, I was doing things with some massively parallel supercomputer. And so on one of the dynamics, there was, There were things where the difficulty of setting up experiments and so on, on that machine was such that it actually made more sense for me to do more experiments on a. Just a simple machine that I could just have on my desk than to try and go to all the trouble to set things up for the fancy supercomputer. It was more efficient for me to just try and do, you know, do 10 experiments that would take more time on my little computer than to go and try and get that one experiment that would take, you know, would run 100 times faster, but it would take, you know, a thousand times more time for me to set it up type thing. And I think the, it's, it's an interesting question, the extent to which there will be sort of high value. I mean, one of the things that's remarkable in the world is that consumer electronics is so sort of widely distributed and universal, so to speak. It isn't the case. There are other forms of goods where there's huge variation between sort of the fancy version and the not so fancy version. In consumer electronics, that's really not the case. I mean, there's some small variation, but there's not a vast variation. I mean, it might be factors of 2 or 3, 5 maybe, but not factors of 100 and so on that it is in lots of other areas. So, you know, my, my expectation would be that in general sort of things will converge to the point where whatever you can do computationally, you can do fairly cheaply. Now we're in a moment right now where sort of the training, your AI happens to be expensive. I doubt it will stay that expensive. I think that will be, you know, there's a lot of, there's, there's no sort of law of physics that says it has to be that expensive. Actually. I just understood very recently something about why training is possible, which is one of the things. Big mystery from you Know when you, you have a neural net, for example, and you're trying to train it on to do some particular kind of thing and you sort of bash it, you know, over and over again with the training data. Why does that work? Why is it the case that you can eventually sort of mold it to go in that direction? Why doesn't it sort of get stuck and obstinately say more or less, you know, I can do this and that, but I can't do this other thing. I realized that this phenomenon of computational irreducibility that I seem to have been studying in different places over 40 years now is the reason that things don't get stuck. By the time you have a system that has sort of computational irreducibility, you've got enough randomness in the different things that can happen when you kind of change the weights in your neural nets that there's always a way out that you don't. Because of computational irreducibility, there's enough randomness in your local environment that you can always escape through some valley and you won't get sort of stuck by the walls of a, of. Of. Of something. So in any case, that, that. But I don't think there's a. To say, you know, I don't think that the train your AI thing needs to be kind of a, a thing that is forever that happens in the big places and the small people can't do it. Funny thing about the web, for example, the text content of the web is not that big. There hasn't been a big push for people to say, oh, I'm going to get a local copy of Common Crawl or something like this that has a version of a snapshot of the web. But that could have happened. People just weren't terribly enthusiastic about it. People were perfectly happy to have that kind of set on centralized service in most places at least we're perfectly happy about that. But this, this question about sort of what. So, you know, I don't think that there needs to be a big gap between the sort of haves and have nots as far as computational capabilities are concerned. As a practical matter, there is often quite a big gap, but it's more of a know how gap than it is a necessary gap in kind of the available hardware. I mean, yes, there are questions of it does cost a certain amount and there are places where that money is hard to come by or there are places where the geopolitics is such that you can't sell this to that and so on. But absent those Kinds of things are much, much more important effect is kind of know how and what people care about. I mean, I see this in Wolfram language, in Wolfram Alpha, its usage around the world, the timescales after which some innovation happens. How long does it take to sort of get to all these different places around the world? And it is interesting that there are definitely countries where there are somehow there's a sort of entrepreneurial tech spirit that even though the country is not economically that advantaged, people are jumping in and really finding out things quickly. And there are other countries which are economically much more advantaged where the, where it's actually quite slow, where it's kind of like we've got a thing, it's working, we're just going to keep doing that. And there isn't for whatever local cultural reason or something, tech cultural reason, there isn't an absorption of new ideas as quickly. And I think that's a much more dramatic effect. I mean those timescales can range from it's sort of the things are known about in a month, three months, six months, a year, something like that, to it can be 20 years or more before those sort of ideas sort of kind of take root in a particular country. And I think, you know, how does one. It's, it's interesting to see there's some countries that sort of have quite a, a tendency of kind of people from that country go off and they study in other places and they learn about other kinds of things and they come back. They're places where they grow up in that country and then they go study somewhere else and they never come back. And I think one sees a bunch of complicated dynamics of how things, how sort of information know, how not, you know, the information is out there. You could go to the website, you could go read about this, this or that thing. The question is sort of the, in a sense the motivation to go do that and kind of the way in which people decide that that's such and such a thing is important. I mean there are things where it, they're kind of, you know, it's interesting with, with modern AI stuff, it seems like that's reached pretty widely. I mean, I think, I don't know exactly how widely it's reached. I've seen some surveys where I've been surprised that some segments of even American, the American population sort of say, I've never heard of ChatGPT type thing. And, but, but you know, I think it's, it's always an interesting question how these things sort of spread, how the know how, how the how the desire to do something spreads. I mean I don't know. You know, when you take something like VR, you know, it's, it's if VR really finally takes off now, which it may or may not, you know, is it going to be the case that you're going to see sort of VR cafes all over the world or how's that work going to work or are people going to going to see people walking down the street in random places wearing VR headsets or is that going to be a thing you only see in San Francisco or something and it never seeps out of San Francisco or takes a decade to decades to seep out of San Francisco or la? Maybe LA is a more likely place for that to be common. Anyway, we should wrap up there. So thanks for tuning in. I think look forward to chatting with you all again another time. So bye for now.