Business, Innovation, and Managing Life (November 15, 2023)

Okay. Hello, everyone. Welcome to another episode of Q and A about business, innovation and managing life. And I'm not in my usual habitat today, so all sorts of strange things might go wrong. All right, let's see a few questions that we have saved up here. Icky asks how you describe what you do. Can you contain it to a single sentence? You know, it reminds me. Sometime in the mid-1980s, I was staying in some hotel in Boston. I was visiting some company I was doing consulting for, and I get woken up by this phone call, and it's somebody from, I don't remember Time magazine, Newsweek magazine, someplace like that, who was like, oh, we're running the story and we're mentioning you in this story, and we want a few word description of who you are and what you do. And it's like, I was asleep. I'm going back to sleep. Bye. So that was my experience. That was the main takeaway from that experience is from then on, it's like, don't set it up so you can get called. And if you're staying in a hotel. But any case, the. That was my experience of. Can you summarize yourself in. In a sentence? So what I usually say when people say, what do you do? I usually say, I run a tech company and I do some basic science. Sometimes I leave out the word basic because people don't necessarily understand that in more detail. I usually say that I've been lucky enough to be able to sort of alternate between doing basic science and doing technology development a bunch of times in my life. And you do basic science, you figure out a bunch of things that are interesting for science and that are kind of seeds for making technology, then you build the technology, then that makes tools that let you do more basic science. And I've done that iteration about five times in my life. And I think one of the things that's always tricky for in a case like mine is when I start talking about science kinds of things, there's a thin set of people where it's kind of like. Because the science that I tend to do tends to be, I would say, big story science, so to speak. And that's not what typical academic science is. So it's very disorienting for the typical academic who say, well, what are you working on? Well, working on the fundamental theory of physics. That's not what academic scientists do. They work on some little tiny slice of. Some tiny slice of something like that. And so it's very, very strangely disorienting if you say that. If you're talking to academics and so on. And so it's always a tricky thing. And in any case. So I haven't got a great way of describing the what do you do in a sentence or two. I usually say we make software for R and D and education and things like that when people ask what our company does. And then I suppose I would say that I spent a large part of my life trying to make sort of a formalism for describing the world computationally that also gets implemented as software. So I don't do a very good job of explaining myself in a sentence or two. I think I'm a bit of a long form story. All right, let's see. Who you asks, what advice do you have for future programmers? Well, you know, it's a funny thing because if you look at the history of programming as an activity, it goes back about 70 years. And in the early parts of that time it was always pandering deeply to the machine. It was always kind of the humans do a lot of work to feed the machine a thing that's going to be easy for the machine. And I think the. It's always funny that history always repeats itself. Like today, if you're going to do some big neural net training that's going to be on lots of thousands of GPUs and all this kind of thing again, that's a thing where the computer time is expensive and you kind of have to have figured out for yourself pretty well what you're going to do so as not to waste that time. But that's how things were for anything you did with computers until basically the end of the 1970s, even beyond that for many people. And so sort of the first story of programming was do a lot of work as a human because you want to tell the computer as precisely as possible in its terms what it should do. Now I have to say there was really a transition. It's an important thing to understand. For me, it was in the well happened by 1978 or so that computers, my situation in the world and what computers existed, had sort of gotten to the point where I could run any program I wanted just for fun, so to speak, and where I didn't have to have a oh, I'm going to plan this and I'm going to do this run this batch job and it's going to cost all this money and so on. It was just like I'm sitting at a terminal, I'm communicating with a computer, I press go, I run whatever program I've written that for me Happened around, well, probably 1978, maybe a little bit earlier than that, 1977, perhaps. But the computers that I used, like by 1977 I was using computers that cost millions of dollars at the time. By 1979 I had access to a computer that cost only $100,000. And those dollars, which would probably be, oh, closer to half a million dollars today that I could do that with. And then the transition that happened in probably 1982 was that the computers on which one could do that cost only $10,000 in the money of that time. And so it was kind of. And there were also at that time also personal computers. Things like the original Apple II and things like this. Those were still. And then the IBM PC, but those were still a little bit too toy to do the kinds of things that I wanted to do on them. But in any case, the thing to understand is in the early days of programming, it's like panda to the computer. And that kind of idea has been an overhang in programming ever since. In those days it was like if you wanted to write serious code, you wrote in a summary language. And many of the sort of the operating systems, the big application programs, they're written in assembly language. Then at the time for sort of scientific computation, it was Fortran. Then 1979, I started building my first big computer system. It was written in C, which was at that time a very newfangled. Most people hadn't heard of it, type language. You know, at the time I was doing that, people said you can't use C, you have to use assembly language to write a serious program. And that even towards the end of the 80s, that was still the case. That sort of for the big, very serious programs, it's like it's got to be as low level as possible. It's got to be assembled language. And I think that kind of mentality of how to think about programming, it's like you are operating the machine. It's like you're operating a piece of machinery and you pull this lever and you pull that lever and that's what the machine does. That has been the story of the design of most programming languages. And there's been sort of the additional effort to deal with the fact that people write large programs and you've got to organize those programs in some way. And that sort of object oriented programming and various kinds of package management systems and all this kind of thing are for dealing with large programs. But the thing that my kind of thing I've been sort of pushing for the last 40 years or so is the idea that one should have as high level a language as possible, a language which includes which it gets as close as possible to being like the way people think about things. And then it's up to the implementers of the language to make the computer do what one thinks it should do, so to speak. So, you know, that's been the story of Wolfram language and our whole sort of computational language stack. And the real thing that is the limiting factor in using our language is now not kind of the implementation of these algorithms, getting all the relevant knowledge, getting it hooked up to the appropriate systems, getting it deployed in the appropriate server configurations and so on that stuff. I like to think we've done a pretty good job of automating. The big limiting factor is how do you think about the thing you want to do in a computational way? How do you take the problem you have and kind of break it down into something computational? And I think that's an interesting question, sort of how to structure thinking about that. And I think it's really two stories. It's like you've got a thing and you want to say, well, what is the computable way to describe that thing? So what is the thing? I don't know, it's. I'm describing something about, I don't know, the meetings that I'm doing in a particular day, my calendar, what say, a good clean way to represent that in terms of time intervals, in terms of kind of different sort of meetings that have different people that are entities that are connected to something, et cetera, et cetera, et cetera, on locations and so on, and what's the kind of structure that I should use to describe that? And to what extent can I root that structure in, for example, the constructs of Wolfram language, which would be things like time intervals where we already know how to compute lots of things about time intervals. How far apart are these time intervals? Do they overlap those kinds of things? Or even something as simple as a time in a particular time zone, a location, a geolocation, a something. I'm not sure there's the best example, but something where you have entities that correspond to people and so on, that's something which wouldn't be built in for the particular people you might be doing meetings with. Probably those people wouldn't be built in entities in the system. You have to specify those entities and then what would you. Okay, so sort of the first thing is. And that isn't the best example perhaps, but the first thing is to kind of take the things that you're thinking about and put them in some computable structure that is rooted in the kinds of things that we have a computational formalism for, which in Wolfram language is a huge number of things, whether it's things about positions on the earth or whether it's things about music you might be listening to, or things about, I don't know, historical events or who knows what. So that's sort of the first step is how do you represent the things you're thinking about in computable form? And that thing might be sort of the internal structure of a company, or it might be something about the geometrical configuration of this thing that you're, you know, these bookshelves you're building or some such other thing. How do you represent those computationally? First step, second step. What do you want to do with that? What functions do you want to compute? So, for example, let's say it's a calendar or something you might want to compute. Is it the case that my calendar. I'll always have enough of a break for lunch between this and that? Or is it the case if I go back and look at my previous calendar, what's the average number of hours per day that I'm working on something? I don't know what. For me, that's most of the time. But in any case. Or let's say you're doing things with your bookshelf construction. It's kind of like, okay, I've got this block of books. How do I take these five blocks of books and fit them in bookshelves so that they are. A block of books is maximally together on a bookshelf and not split between bookshelves. That's something where. That's an interesting example because that's one where, well, you could say, well, how would I do that? You could start thinking, how would I implement that in C or something? And it's like, well, that's a real pain to figure that out because it's a combinatorial optimization problem. In morpheme language, there's just a function that does combinatorial optimization. You have to describe your setup in a way that kind of corresponds to how that function sort of expects to get its input. And if we've done a good job in designing that function, that will be an extremely natural thing to get from, for example, the geometry and so on the problem. So that's kind of the figure out the structure, figure out the functions. That's kind of a way to organize your thinking. And now the question of Implementing those things is sort of a separate question and is something which, if we've done a good job in language design and language implementation, should be fairly easy. But the real non trivial thing is figuring out how to set up what you're talking about computationally. And that's mostly not what's taught in kind of computer science school and so on. It's not about that. It's usually about the innards of programs and how they work and setting them up and so on, and how algorithms work and so on. And that's really, you know, we've for sort of the standard algorithms, then they're all implemented. You don't need to know how they work inside. It is of some cultural interest to know how some of them work. And also as you think about inventing a new one, sometimes it's worth knowing how the existing ones work to invent the new one. But. So I think you're asking about future programmers. I think that the key thing is understanding how to think about the world computationally. And I have to say, I think our computational language is the shining example of, of something which takes you right to the place where you're really thinking about how to think about things computationally and you're not in the middle of the implementation details. Now, once you've got a way to describe what you want to do computationally, then you can go use one of our chat notebooks, ask an LLM to help write some code in Wolfram language. It'll be a short piece of code typically, and you can hopefully read it. It's a language that's intended to be read as well as written when you know, if you do the same kind of thing with a low level language, you'll get a big slab of code and people are like, wow, it wrote this big slab of code. To which my response is it shouldn't have had to write that big slab of code. It should have been the case that you have a language that's high enough level that that piece of functionality is just available as a function in the language, not something that has to be regurgitated by an LLM, so to speak. So I would say that the number one thing is learn to kind of think about the world computationally. And as I say, I think our language is sort of the best case for doing that. And by the way, if it ends up that later on there's something which needs to be implemented right down to the metal in C or whatever else, not clear that that's a. These days a compiler in Our language, if you annotate your code to the point where the same level that you would, or a little bit less level than you would with type specifications and so on in C, we can compile just fine to LLVM and make something which is just as efficient, if not more efficient than sort of the down to the metal C. Which again is the story of what happened with assembly language, transitioning to things like the C like languages and so on. Back some number of decades ago, it was no longer useful to write directly in assembly language because writing in a language like C, the compiler could produce assembly language which was as good or better than what you could produce. But any case. So I think the number one thing is how to think about things computationally and learning that if you need to learn the down to the metal version of how to do things with some low level language, for some reason that turns out, in my observation, to be very straightforward for people who've learned Wolfram language, for example. Now the next question is, well, what are the ideas you need to learn? Well, you know, the role of language designer me in this particular case ultimately is to make the language as principled as possible so that there are the smallest possible number of separate principles that you have to learn to use the language. But there are some principles, there are some ideas that are not obvious. Once you understand them, you're like, aha, now I can use these going forward and be very efficient in the way that I think about things and the way that I implement things. But you need to learn those ideas. It's like saying, well, I'm doing math, I don't need to learn calculus. Because it's all going to be obvious. Well, it's not, it wasn't historically and it's not to any given person. There are things to learn. And so it is in learning computational language, learning kind of ideas of symbolic programming and so on. And there are some of these ideas to learn. I think it's also very useful to know lots of examples of kind of how the world gets made computable and how things get computed in the world. I'm thinking pretty actively about how to explain these things because I'm trying to write a book course like thing I haven't started yet, but hopefully we'll have a chance to soon that tries to kind of explain how to think about the world computationally. As a practical matter for programmers, as I say, the number one skill is learn to think about the world computationally. And that's the skill that's going to be in incredibly high demand. The can I show up 9 to 5 and churn out a certain number of lines of code? Not so important. If our technology was sort of more broadly used in kind of the pure software engineering world, that would be gone. Even now the LLMs are biting away at that. That will be gone. Just like assembly language programming is not really a thing anymore. Direct programming of large slabs of code in languages will be a thing that is gone. And there'll be a small set of people who do it but. And who do it, who are needed to do it. But that will not be the mass of the few million programmers or something. Most of them will not be doing that. Whether those people will, they will not have the same kind of jobs that they have today. The jobs and it won't necessarily be quite the same people who are well optimized to kind of think about things computationally. I mean, there are people who are really good at kind of, you know, knowing about a lot of different things and being able to kind of conceptualize a big picture. And that's kind of the important skill for kind of understanding the world computationally. It's less important to be the kind of the person who is very, I mean, systematic design of what you write is important. But I would say that some kinds of sort of. I don't know, how do I describe it? I mean there's sort of the. Just, you know, I'm going to implement 17 dialog boxes today and I don't mind that it's kind of dull and monotonous to do that. That kind of work will probably go away. And so if that's, you know, if people have this idea that, you know, programming is kind of an easy way to make a good living where you kind of don't have to think very much, you just sort of turn the crank, that's going to get automated away. I think the programming that is much more human centric and is much more about. So what do you want the program to do? How do you think about things computationally? To me, that's something that's much more. For many people, it's much more fulfilling to do that kind of work than to do the kind of manual labor work of kind of grinding out large amounts of code. You know, I think that's the side of programming that is absolutely going to grow, take off, be important, et cetera. The sort of, the 9 to 5 grinding I don't think is going to survive in its current form. And you know, increasingly that's going to be like why would you pay somebody to do that? You can just define that in computational language and press go and then you have what you need. It's. I mean, we see this a whole bunch because, you know, with, with Wolfram language, it's really common for people to be able to implement, you know, a thing in a couple of hours. And then people would say, but it's going to take a month to program this. Well, it doesn't need to take a month. People still have the idea it isn't real unless it took that month. But that idea is going to go away. I don't know how long it's going to take, but that idea is going to go away. Just like when I started out in sort of computing, it wasn't real unless it was written in assembly language. And that profoundly went away. Let's see. Okay, so. Well, let's see. Well, I'm not sure I'm going to have much useful to say about from Des. Any advice for someone content to just get by financially with zero interest in the usual understanding of career, probably no kids, just looking to focus on other things. You know, one of my beliefs about the world, maybe I'm just an optimist, is that if there's a thing you passionately like doing, there's probably a way to sort of fit that in to a life, so to speak. That is, it's. There are, you know, it might take some creativity to see how do I make a living doing that thing that I really like doing. But I claim that there's a very good chance that that's a thing that's possible. You know, when it comes to people and living in different places and so on. And you know, I, it always, I find it interesting when people say, oh, I really want to do this stuff and I don't really want to do boring things to make money and so on, but I need to live in Palo Alto, California. Okay, but it's really expensive to live there. And you know, maybe you have some whole elaborate network of people you know, there. Well, then that's one thing. But if it's like, you know, kind of, there's a. There's considerable. In the US is probably a factor of three difference in cost of living. And if you go to other countries, it's even much more than that between different places. And it's like, yes, if you're going to live in Palo Alto, you know, or in the Manhattan or something, then there's a certain level of income that you have to have from what you do. Just in order to kind of survive. Although the other thing I find interesting is that people have very different sort of tolerance for what constitutes a reasonable place to live and do things and so on. But you know, I suppose I would say that the thing that. What I've noticed is people, you know, people say, for example, I mean, there's sort of a well known phenomenon that people say, oh, if I make more money, I'm going to be happier. Eh, it's, you know, there's a. Up to some level, maybe that's true, but beyond that it's not. Let's see, I said I was not in my natural habitat. So that means that I'm subject to all kinds of potential disturbances. Well, maybe not. Oh yes, hold on one second. Just a second here. Okay, and we're back the. Okay. In any case, what was I saying? We were talking about, you know, the, the value of money, so to speak. And I think what endless studies have shown is that if you have a certain amount of money, it will make you happier. If you have much more, it doesn't necessarily make you much happier. It really depends on what kinds of things you want to do. I mean, for somebody like me, there's kind of like living a life that doesn't have distractions and where it's sort of. I'm not spending a bunch of my time kind of dealing with the infrastructure of my life, so to speak, that takes a certain amount of money. And that's highly. For me, that's important to be able to do the things that I want to do. For me, also being able to. And this is something which is not a thing you buy with money, but is a thing where you have to have a structure, an organization to be able to achieve this is you want to do a project, it takes resources. Do you have those resources at your disposal? Which doesn't mean kind of money in your pocket. It means kind of. That you have the appropriate kind of sort of control of some set of resources to be able to direct them in the way that you want. And I think again, I suspect that scales at many levels of what and for some people it's, for example, doing the sort of infrastructural things of one's life. I don't miss the fact that I don't spend time doing sort of infrastructural things in my life. If I was kind of built a little bit differently, I might find it kind of relaxing to not have to be intensely focused on some highly intellectual thing that I'm trying to do for some of the time and to just be doing some sort of infrastructural kinds of things, that's not my particular way of being set up. So for me, it's kind of like if I can avoid, you know, doing grocery shopping myself, I'm going to avoid it. I'm going to be spending that time working on basic science instead. And. But I think in the, you know, the question here was this question about careers. That's a complicated issue as well, because somehow there are, you know, careers. Sometimes it's like, well, I'm going to get myself set up in a situation where I can just keep turning the crank and I'll make a reasonable living doing that. And that can be quite a trap for people. There's another one which is kind of I'm going to put myself in a position so that I have a sort of platform to do things that allows me to do more of what I want to do. Whether it's I built a following on, you know, YouTube or something that allows me when I want to do something, or I've built a. I have, I'm, I'm kind of, you know, I've, I've built up the ability to, you know, I'm a, I'm a travel blogger or something and I built up enough of a following that I can, you know, get invited to go to, you know, hotels all over the world or something. You know, these are things where you kind of build a platform which is sort of a career like thing. You build a platform to do more of what you want to do. You've invested something or even in an area like, let's say you're even doing academic work and you sort of invest in getting to know some community. Or maybe it isn't academic work. Maybe it's something completely different. Maybe you're doing some kind of charity work or you're doing some kind of, you're collecting some weird kind of, I don't know, porcelain penguins or something. And you gradually get to know the community of people that collect porcelain penguins. And that's a thing, that's sort of a useful investment because then you get to trade them and things like this. And that's not career in the ordinary sense, but it's making an investment in things that you're doing. And that's the thing that I think can be important and satisfying. Now, different people work differently in this regard. There are people where they built, they've made a lot of investments and a lot of things, and those things sort of build up and up and up over the decades. I'm probably an example of somebody who's done that. There are other people I know who do a thing for a few years and then they do a different thing and it's almost like they just change. And on a particular day they're like, okay, I'm quitting that thing, I'm going to do something different. And people can be very happy doing that. I have to say, if in, you know, it depends on one's personality and so on. I mean, sometimes it's very refreshing to do kind of a new and different thing and to not have the baggage of something else. I mean, one of the features of quotes, careers and investments and so on is that investing in what one does is that, yes, one builds up a certain sort of, a certain amount of baggage. Given that one is in the porcelain penguin community or whatever. It's kind of like you're going to have to spend some amount of time interacting with those people to maintain that relationship. And if you kind of just go cold turkey, penguin whatever on the porcelain penguin people, then usually what happens in my experience is that if there's a community of people you've been dealing with and then you, for whatever reason, stop dealing with them. There are a few things that degrade fairly quickly, but other things that are surprisingly long term, I mean, where you come back 20 years later and people are like, oh, hi, I haven't seen you in 20 years. But you still have that sort of basic relationship there. And it's easy to kind of reinvigorate it, of course, as you go even longer and it gets to be sort of the complete length of a career, people retire, die, et cetera, et cetera, et cetera, and you sort of lose that investment, so to speak. And you know, for me, for example, in, for example, the physics world, I know a lot of people who are in their late careers now because I knew them when they were in their sort of earlyish careers and I was at the beginning of doing physics when I was, you know, late teenager, 20 years old, things like that. I would know people who were 35, whatever years old doing physics. And I know a bunch of those people, they were very good group of people back then, but they're now considerably older. And the younger folk in that area I don't know as well and sometimes don't really know at all. And that community is kind of turned over and it is something where you kind of know one group of people, but you don't and sort of how you meet the younger group of people, for example, in that case. I mean, I know. Actually, it's kind of funny. I just did an event for our Wolfram Institute. It was an in person event, and it was interesting to see that there was a very bimodal distribution of ages there of a bunch of quite old people and a bunch of quite young people who I've kind of interacted with through the things I've been doing last decade or two versus the things that I was doing, you know, several decades earlier than that. So I'm not sure how useful any of that is. But I think the, you know, there's this question when you talk about not investing in a career. There's the question of there's other kinds of sort of systematic investment in communities and things like that that can be very valuable, I think. You know, and there's also the question of, of how much do you want to sort of consistently build up expertise in one area versus I'm going to do this this week. That. That week. You know, generally when people get into a particular area, it takes five to 10 years before they're really truly sort of fluent and expert at doing that particular kind of thing. And there are people who, for whom sometimes you can be an expert at doing many different things. But nevertheless, if you're like, I'm going to be, I don't know, doing some kind of, let's say I'm a writer or something and I'm going to be writing about lots of different kinds of things, well, you sort of become an expert at being a writer who writes about lots of different kinds of things. I think that that's another kind of thing that I think tends to be quite satisfying for people is to get to the point where you're really fluent in a particular kind of activity. And then the trap, of course, is. But once you're fluent, you might spend the rest of your life doing it. And I think that can be something where, well, you might really enjoy it. And for me, and for many people I've seen, you have to change it up a bit. It's kind of like the core of what you do might be the same, but the kind of the, let's say you're a writer or something, the core of the fact you're writing might be the same, but you might want to change up what kinds of things you're writing about. I saw an interesting statistic actually just yesterday from a study which I don't know how great the data was for this, but looking at Lots of academic citations and so on. The question is, to what extent do typical academics in science, for instance, change the things they do versus kind of stay doing the things they did for their PhD thesis, for example? And this data seemed to indicate, I mean, the kind of folk statement people make is half of all scientists are still doing what they did for their PhD. This data was actually, in a sense, much sort of less optimistic even than that, because it seemed to indicate that the vast majority of scientists who stay in science and don't sort of exit quickly after like getting a PhD or something, that the vast majority of those people end up kind of continuing to do the kind of the same things they were doing for the course of their career. Thirty years or more, people. And this is something. And yet if you look at where science innovation comes from, almost all of it comes from people changing fields, people being put into different fields, et cetera, et cetera, et cetera. So it's kind of if what you find fun in science is innovation, then the story of sort of do a thing for your PhD thesis and then keep doing it is kind of a bad deal. But there isn't a good forcing function to get you to do something different because a lot of the forces are telling you keep doing the same thing. You're getting the grants to do those things. You've got this job that is about this particular thing. And by the way, academic work is one of the more diverse kinds of work where you can decide, you know, I'm a tenured professor of this, and you can just decide, I'm going to do something slightly different from what I'm doing before. It's a rather unmanaged field in general. And, you know, much not necessarily to everybody's benefit, because there are plenty of people who, in an unmanaged setting like that, will just do what I just described. For example, it's like, what are we going to do next? Well, we're going to do something similar to what we did before. And there's not a lot that's going to push and do something different. If they were in a more managed situation in a company or something like this, it might be that the company just says, well, we don't care about that anymore. You've got to do this different thing. Let's do that. I can. Just a second here. Okay, so that's a few thoughts about that. All right, let me see. There's a question here from Eli. Why don't I quit CEOing and commit full time to investigating whether nature is computable does running everyday things help or do you just find it fun? Oh, it dramatically helps. I mean, in the. Just sit and think for the rest of your life. It's kind of a terrible sentence, at least for most people. I think there are rather few people where that's going to work well for them. I think that for me, first of all, as a practical matter, even if my only goal was to do science, building the tools to do that is something highly worthwhile and highly necessary. You know, when I worked on. I spent a decade working on my big book, A New Kind of Science, and I had spent the previous five years before that working on building Mathematica and what's now Wolfram language, getting that started and bringing out the first version of that, first couple of versions of that. And the. I had always thought that, well, I, I sort of stopped doing science for five years to do that. The amount of time that five years saved me was probably pretty close to a lifetime. So in other words, doing that technology development for five years accelerated the other things I could do by 40 years or more. So it's a huge win, huge leverage to develop tools to do, even if your only goal is do basic science. I have to say that for me, it's a satisfying thing to make technology that can be widely used in the world and that other people get to have fun with and get to discover lots of significant things with and so on, and get to learn lots of things with and all that. I mean, that's a satisfying thing that makes it for whatever, you know, none of us can really perhaps explain from first principles why we have the motivations we do. But for me, the idea that people can make use of the things we're building is a. Is a worthwhile motivation, so to speak. And I think the. So that's, that's, you know, that's that kind of thing. The fact is that the things I do instead of figuring out how to run a company, figuring out sort of the strategy for the company, I find that it's sort of invigorating. It's kind of like you take. It's kind of a. Both a, it's an enlivening and a, you know, you kind of get used to every day. I make lots of decisions about lots of things. And when I'm doing science, for example, I also. It's cheap for me to make decisions, so to speak. It isn't something where I'm like, oh my gosh, I have to make a decision. Let me worry about this for a Long time. It's like I'm used to making decisions. Let me just apply this to that. I think also that it's another thing that's important for me, at least in the CEOing world, is there's a whole group of people that I work with, and I find that invigorating. They're great people. And also they're things which even if I was only trying to do science, it's like, well, I need to know this particular thing, or I need help with this particular thing. Well, there's a big pool of people where I can get that kind of information, that kind of help from. So to me, it's a totally symbiotic kind of story between sort of CEOing and doing basic science. And I don't think having an extra few hours in the day to do basic science would be. I suspect it would be a net negative for me personally. I think that it's kind of the quality science time is really more worthwhile than just a few extra hours in the day. Now, having said that, there are things which I will avoid doing because it's kind of like, well, this would take me a week where I'm just not able to do kind of intellectual work and so on. I don't want to do that. And that's, you know, there are plenty of things which sort of come with the CEOing package for many people, which are like just big blocks of time, which I tend to avoid doing. And my version of CEOing is much more on the strategic side. And. And that's. That's kind of the way that I go with that. I think that this whole question about having a portfolio of things that you do in your life and finding that it's invigorating to have all those different things. I certainly find that. I mean, even, for example, doing these live streams, I find this fun. Hopefully other people find it interesting, but I find it fun. It helps me to understand things better, to try to explain them and so on. And for me, it's kind of like I have a live stream on my calendar, and it's kind of like that's a fun thing to do, and it makes me probably more energetic for the time I spend kind of yakking at you. All the amount of energy that that adds to the other things I do probably absolutely makes up for more than makes up for the time that I spend sort of doing this. And I find the same kind of thing with some sort of educational stuff with kids and things like this, that I find those things Invigorating. If I was doing them 100% of the time, I'm sure I would get kind of bored and cynical about them. But if I'm doing them for a few hours a week, it's kind of like, hey, this is fun. This is something that helps kind of liven things up for me. Let's see. Ollie asks, do you think there will come a major shift in business planning with AI? Oh, I don't know. I don't think so. I think that the strategy of what to do is a deeply human thing. AIs can think about all possible things to do, but computation gives one that opportunity. But what do we actually want to do? That's up to the humans to decide. I think in terms of the sort of the things which are expensive and the things that are cheap, and that does affect business strategy in general. There are things where there was a time when if you wanted to move information from one place to another, it was horses and carrier pigeons and it was very expensive. And now kind of moving information is basically free, very cheap. And there are things. So there are different kinds of business strategy that are enabled by certain kinds of things getting cheap. And the general sort of globalization of labor, for instance, is another thing which, you know, changes certain calcul calculations about what can be done and what can't be done. You know, what's, oh, we can't do that. It's going to be too expensive versus well, there's somewhere in the world where we could do that and it actually makes sense to do it. And I think with AI there are some things where it's like, well, we couldn't, you know, that wasn't a thing before. Now it will be. I mean, for example. For example, I don't know how well this is going to work. But a AI tutoring is one example. Personalized education that may become very much a thing. And it may become the case that will enable lots of kinds of things, not just purely in ordinary education, but even in educating new employees. Somebody is retraining for this or that thing. It may become much cheaper to do that, which may affect kind of how companies think about sort of human resources and their deployment and so on. So I think that those are things that can happen. I think it's more, you know, what's now cheap. I mean, certain kinds of. And certain kinds of activities are now kind of, we got this boring report to write. Let's just have an LLM write it. Or we've got this thing where we've got to Roughly figure out, you know, we got lots of different sort of essays that got written for something. Let's try and sort these in some way. Let's get an LLM to do it. It's not going to do it perfectly. They're going to be mistakes. But there will be mistakes with humans, too. And maybe it's worth it to say, yeah, we can read 10,000 essays and look at 10,000 applicants for this thing where before we could only look at 50 because that was all a human could do. And even though we're going to get it wrong some of the time with the 10,000, it's still the case that it's a win to be able to look at a broader range of applicants, let's say, than if we had to do it all with humans. Let's see. Well, there's a question here from Eli about how much control do you maintain over Wolfram Institute? I assume he's asking about, do you find loosening your grip on management of sort of research allows for a higher chance of success and discovery? That is an interesting question that I'm wondering about myself. This whole question of, let's say you've got creative people and they're trying to discover things and so on, what's the best way to give them a chance to do that? Door number one, give them as much freedom as possible. Just say, go do whatever you want, and you can make a living just sort of doing whatever you want. Door number two, it's highly managed. Somebody like me is saying, here are the strategic objectives. It's kind of like a company where there are products which aren't products in the usual sense. They're intellectual products. There's a very well defined sort of strategy and leadership. Which one of these is more likely to succeed? I think it is. Well, I have to say I feel much more comfortable with number two than number one. Back in the mid-1980s, I started an institute center for complex systems research that was kind of a door number one type option, where it's like, bring in a bunch of talented people and just say, do what you feel like. They did some interesting things. I didn't think they had nearly the strategic oomph they could have had. Whereas I think the good thing about good, hopefully leadership, is it allows the whole to be greater than the sum of the parts. It allows one to have a sort of strategic direction where there's really all those people who are contributing. Talented things can really be pushing in the same direction, can build up to something that is bigger than any individual, one of them could have done. So I tend to think that that's the win. And actually, I think, in a sense, perhaps it's my responsibility in the case of Wolfram Institute, to provide that kind of leadership. I don't yet completely know how to do it because when I'm working on projects myself, it's kind of like, well, help me do this project. And that makes sense. But my bandwidth for projects is limited, and sometimes I feel like I'm pulled into spending more time on a project than I strategically want to spend just because I'm working with other people who are really into this project. And it's like, from my point of view, it's like, well, I'm going to give you the strategy. I'm going to produce a version of this project. You go write your own papers and embellish it and extend it in all the ways you want. And that's, I think, a possible model. But I don't think that the. I think one has to have some kind of. I think it helps in getting great research and even innovative research done to have some overarching strategy. I think also that I would say that for somebody like me, I've been in this business a long time. I'm a quite significant idea generator. There's a bandwidth limitation on the execution of ideas. But, boy, if you give me the resources, I can come up with a lot of ideas. And I'm not to say that all my ideas will always be the best ideas, but I have a pretty good idea source. And I think relative to a person, for example, early in their career, trying to say, well, what should I study? There will be some. Some exceptional cases where people are just great at coming up with what to study and just they're off and running. But there are many more people who can be extremely valuable contributors in basic research than the set of people who will come up with their own ideas and be off and running with them. I mean, there's another category of people who come up with their own ideas and cannot execute them. That's a different category. And I don't know how to plug those people into the setup either. I mean, I, I, you know, I see these things. People are sending them to me all the time. You know, I have an idea, you know, et cetera, et cetera, et cetera. It's like, well, you know, I have a lot of ideas, too. I'm not, you know, I believe more in my ideas, typically, than ideas that people sort of throw at me, partly because I'm more Engaged with my ideas, I understand them better than an idea that was just thrown at me. And I don't know how to plug kind of sort of ideas that are thrown in to kind of, you know, they usually to get an idea executed, it is more than just saying I have an idea, here's the idea. Now to the people doing the execution of the idea, just okay, go execute this idea. It's the same thing in business and companies and so on. You know, I see this quite often. People say I've got this great idea, even at our company, I've got this great idea, somebody just has to implement it, okay? It's never going to happen. Because until you can take that idea and break it down into steps that somebody that doesn't require your imagination to fill in those steps, you've got to have unrolled your imagination to the point where you're explaining things so that other people have a chance to execute it. The people who are executing may have their own ideas and lots of cleverness and how the execution is done. But unless they understand more than just your general hand waving vision, they've got to understand things at a concrete level. And I think the same thing is true in the case of kind of basic research. For example, just say, here's an idea, let's throw out this idea. You got to go further than that to make it actionable, so to speak. And I think that the. But this question about how best. Let me give you some examples. So what are the schemas for getting things done in basic research? One schema for me is I'm going to write something about it. I'm going to do the research, explain the research that is a concrete thing that is going to be delivered. And I can kind of understand how to get sort of things to line up to make that happen. Another thing I'm going to. Or we're going to make some tooling software tooling that implements some particular kind of thing. That's another thing where we understand how everybody can be pointed in the same direction. In both of these cases there's a lot of uncertainty in what actually comes out. There's a lot of. That's the nature of research. We don't know. We don't have a sort of precise kind of timeline and flowchart worked out of. Here's what we're going to do to get to the end. Because we don't know where the end is because it's basic research. You don't know where it's going to end. And by the way. That's another reason why throwing the idea of the transom just doesn't work because you throw the idea over and it starts getting legs and starts being worked on. And by the way, the idea runs off in a different direction. And if the original inventor of the idea isn't somehow connected to that, they're not going to be able to deal with the fact that the idea ran off in a different direction. So I think it's. But in any case, I think the. My main. So that's another scheme, so to speak, is kind of we're turning it into a piece of tooling, a piece of software that's what defines the flow. Now, there are other things which sort of experimenting with. So for example, let me give you an example. Let's take philosophy related to a physics project. So for instance, the project going on right now of just saying, let's take all of kind of classical philosophy and let's see how it plugs together with a physics project. That's something where once we have kind of. Once we've done Descartes, let's say we've got a model for how to do Kant and Plato and you know, Democritus and Nagel or, you know, who knows what, or a whole range of different kinds of people are quine or something like that. And so that's a thing where there is value to the kind of the bulk, so to speak. Once you defined a direction, then it's kind of just like scale it up. It's kind of like, let's say you've got some software and you're going to make a version of it for Japanese. And then you say, well, given that we've done that, we can kind of know how we're going to do Korean and French and this and that and the other. There's kind of a roadmap for how to do it. And I think that's another possibility for the institute where it's kind of where there's value in bulk, so to speak. I think the notion of how to deal with the kind of sort of individual people with the I've got an idea type thing, I'm not sure. I mean, maybe there's a. Maybe it's. And perhaps it's my responsibility to sort of say let's, let's help develop that idea to the point where it is sort of actionable. I think, I think one of the things that, I mean, I guess if you take our summer school as an example, the. I would say, oh, probably 80%, 90% of the projects that get done, there are projects that I invented in some way or maybe some other mentor at the summer school invented. Maybe 5 or 10% are projects kind of people came with, I want to do this project. Often we push really hard to have people not do the project that they came with because it usually fails. Because the scoping of projects and the kind of designing a project so that it can actually be implemented, and particularly in the form factor of the summer school, three weeks, et cetera, et cetera, et cetera, there's a whole bunch of expertise in knowing how to do that. And people who just sort of come in from the outside on day one, they don't know how to do it. And they just, and they, they, you know, the project is too big, the project is not defined enough. The project requires some steps that aren't really doable. The project even at the end of it is too muddy. You don't really learn anything from it, et cetera, et cetera, et cetera. And I think there's a place where, you know, it's a question of where do you use the expertise that you have. And I think a really pivotal place to use expertise is in the design of what projects to do. And so I tend to think that that's kind of the right thing for our institute to sort of maximize the amount of basic science productivity that comes out of it and basic thinking productivity that comes out of it. So let's. Okay, there's a question from Nick here, Right? There's a question about remote working. There's a tension regarding remote working versus being in the office. Nick says, in my experience in a remote working team, junior, new starters take a few months before being efficient. It appears, say that you've mastered remote working with your teams. What do you think makes remote working a success? Every team at our company is a bit different. It is super useful if it's on the R and D side. For example, when people have come to our summer school, let's say, and there's a bit of sort of enculturation that happens just by being around people who work at the company and things like this. That's useful. I think some amount of sort of. There has to be some kind of mentoring type interaction with actual sort of people at the company to bring in somebody who isn't, you know, who's just sort of coming from the wild, so to speak. I'm not sure that I know. I think just the very fact that we're so used to remote working and have been doing it for 30 years, more than 30 years, makes the kind of culture of the company have certain expectations that make it easier for that to work. So I'm not sure what aspect of it, other than the fact that nobody thinks it's unusual, that really helps. And I think probably, I mean I honestly don't know what the front line, just starting person, what exactly our procedures are these days for that. I think though that there's quite a lot of one on one kind of mentoring education that goes on with new people starting. And I think in some groups they tend to have people starting in clumps. So they do that kind of education in multiple people at the same time. There's not a very good answer, but I would say that the thing that there are also groups at our company where they said it's never going to work remotely. We just have to have the people in the office. And then there was a pandemic and so people were sort of forced to be remote and by golly they realized, well actually we can make this work. And I think, I mean in our particular case our workforce is so remote and our kind of central. Well again there are exceptions to every rule. So there's some places around the world where we have teams that are actually very much more in person, but the majority are not. But if it's like well, why don't we get people to show up at our company headquarters for example? It's kind of like eh, it doesn't really make sense because it's like one person out of three in the team will be there and it just doesn't really make sense. So it's kind of like once there's an expectation of remoteness that really helps. Not a great answer I'm afraid. Let's see. This is a question from Slayer. Whenever you were or are learning new stuff as part of your independent research efforts, do you have a structured purpose like I'll learn subject X topic by topic or do you take a looser approach? How do you know how much time to dedicate to your various research interests? Okay, I am a project driven person. If somebody just says learn this for the sake of learning it, I'll fail. Just don't do that different from other people. I mean when I say that it's on a casual basis, I'm really happy to do that. I'll go to some museum about something and I'll be very happy kind of wandering around learning about that stuff. I'll run into a person and I'll be excited to learn from that person about some particular thing. But when it comes to me on my own, undriven, so to speak, doing things the way I drive myself is. It's always projects. It's always, I'm trying to achieve this particular thing. I'm trying to learn this. I'm trying to figure out this particular thing. Now let me learn what I need from all these different places to make it possible to do the thing that I'm trying to do. That is my approach to doing things. And I really don't have an easy time doing other stuff. The question of. Of how far afield, how much homework should you do? I mean, this has come up in some things I've worked on recently where, oh, I don't know, I'm working on things about quantum mechanics, for example. There's a vast, vast literature of all the different approaches and ways that people have of thinking about this and that thing. And I have to say, in that particular case, I've just decided I'm just drilling through my little hole, so to speak. I'm not going to make sure I know every piece. I mean, I have decent awareness, I think, of most of the tentacles, but, you know, really kind of going to the end of each tentacle. I'm not doing that. There's some fields that I've worked on. Oh, I don't know. I studied these things called combinators a couple of years ago where I just inventory every single paper that's ever been written about them. I even made a bibliography of those things. It's a small enough field. You can wrap your arms around the whole thing and then you can go on and do stuff. But there are other fields where the field is vast. There are hundreds of thousands of papers. There's no way you can read all that stuff. And you really have to just do your thing. And occasionally, when people say, well, isn't it relevant? Isn't this or that thing relevant? I would say that in my experience, even talking to real experts in all these fields, that it's one in five times that what they say is relevant. Really, really is relevant. Because usually the least stuff I'm doing, it's kind of original and alien and creative and weird enough that it's kind of going in a different direction than people have gone before. And occasionally there'll be technical things where, like, I can learn this technical fact from this place or whatever. But as a whole, it doesn't, you know, the cost to go do that homework completely is vastly higher than the kind of the potential risk of, oh, I missed that thing. I wasted a lot of time because people already figured that out. I mean, you know, whenever I have a technical thing that's well enough defined, I'll go try and figure out whether people have figured it out before. But when it's the big idea, maybe my own ideas are sufficiently alien that they're far away from other ones, usually. So. But in terms of kind of. Yeah, so I mean, my answer is I learn things for projects and I learn what I need to know, so to speak. Now, I also, for me, the history of fields is always important to understand the field. Just knowing random factoid. Random factoid doesn't cut it for me. And it's usually a mistake. And I usually want to know, how did that factoid get to be that way? What's the root of that tree? That takes me to something that I really do understand pretty well. And, you know, it isn't a floating factoid that just, you know, happens to be that way. It had some whole, you know, tree that came, that produced that factoid fruit, so to speak. And I kind of like to go back and understand that tree. Doesn't mean I understand every branch of the tree. It doesn't mean I understand every factoid hanging off the tree, so to speak. But I do understand going down to the root of the tree, you know, kind of what. How did it get there? That I find very useful. How do you know how much time to dedicate to your various research interests? Yeah, that's always an interesting question. I mean, I always go into every project as an optimist. I'm always like, this project is going to be easy, or at least reasonably easy. And either that or I say I'm just going to kick the tires of this project. Maybe it's going to go nowhere, but the thing that I'm going to do is going to be fairly easy. It often turns out not to be that easy. And once I've started one of these projects, I really should finish it. Sometimes I don't get it finished at first, and that's super frustrating. And I come back to it years later and I say, by golly, I should have finished it before. And then I go and finish it. I have one right now that I started maybe five years ago now, and various other people worked on pieces of it and so on, and I really should have just finished it back then. I didn't. I now perhaps will finish it now, the next next few weeks even. But, you know, I really. I hate sort of having these projects unfinished. And I I, you know, and I end up. They end up being things that bug me and eventually I come back years later. I mean, there's one project about the second order of thermodynamics, which I kind of started when I was 12 years old and I kind of finished it 50 years later. So that's my longest hanging. Talk about low hanging fruit. That's my longest span of. I don't know whether I can come up with a fruit analogy here, but that's the longest hanging fruit or something before it got picked. And so there's another question, which is, you're not making progress. What do you do? For me, it's like I'm digging straight ahead. I'm not making progress that way. Let me start going to the side and see if I can go around the obstruction. That's what I do in projects, both in science and in technology. It's very, very rare that I get to the point where it's like we're in a, a brick wall. We just don't know how to get around it. For me, there's always a, well, why don't we try this? Why don't we try tunneling under it that way? Why don't we do this? I'm very rarely at the point where I'm just stuck. I think if I was just stuck, I would probably do something different. I mean, I would probably just put it on hold. I have to say, I'm trying to think, when are some cases where I've just gotten stuck. It's usually like, I know what to do next. It may be a lot of work, I may not have time to do it, but it's not like I'm stuck and I have no idea what to do. I suppose for me, whenever I think I'm getting stuck, it's kind of like, well, let me implement what I've already got in a nice streamlined way in Wolfram language. And even just streamlining what I already have, that usually will give me an idea for what to do next. Because one of the things that goes wrong in research, in my experience, is you build up to something and you kind of just. It's sort of a rickety structure and you kind of just about make it to this thing and it's kind of rickety. And you don't have a streamlined way to get to that thing. You are just managing to climb up there in some very elaborate way to really make progress and go beyond that thing. You have to have something streamlined. You have to have a solid kind of structure on which you can build a platform to build on. And sometimes when I'm sort of stuck on the where to go next, a reasonable strategy is just go back and streamline what's already there, because that's useful anyway and because that often lets you kind of figure out where to go next. Let me see, PRAB is asking how is innovating outside the system different? Different from working within institutions? Is one better than the other for certain fields? You know, I just was seeing this thing yesterday about kind of the, the sort of the dynamics of, of the academic world based on tens of millions of citations and the whole sort of history and network of those things. And I have to say it made a terrible case for inside the System. Now it depends what you want to do. If you want to have a stable job and you want to know that you're going to keep getting paid and getting grants and so on for the whole of your career, then being inside the cocoon of the system is probably a pretty good deal. If the thing you really are excited about is kind of inventing new things, innovating, going where no one has gone before type thing, then that's not the best strategy. I mean, it may be the only strategy you can adopt and you have to kind of by subterfuge, do new things even though the system doesn't want you to do new things. But if you have a choice, then sort of the outside the system is the way to do new things. Now you know, there are upsides, there's downsides if you're outside the system. There is, well, there's no sort of support structure of just like you're going to get, continue to get paid as a tenured professor to do this, that and the other, or you're, or you know, there's a place to publish what you do. And you know, everybody publishes in the Journal of xyz. And so there's a well defined distribution channel for your ideas or there's a whole series of conferences and you know, everybody goes to those conferences and that's where you go to, to sort of talk about your ideas. If you're in a well established field inside the system, you have all of those mechanisms. That's the good news. The bad news is those mechanisms have a shtick. And that shtick will be an old shtick. It's not going to be a new shtick because you're just inventing the new shtick sort of by definition here and there isn't a structure for the new shtick. And there'll Be a lot of pressure to keep doing the old shtick. And in fact, in this data I was just seeing yesterday, it's dramatic. That effect is dramatic. I mean, as I was mentioning earlier, the people typically doing the same topic that they started their careers on, larger teams are much more conservative in what they do. This is the anecdotal thing. Apparently the data seems to support this. You know, large teams with lots of funding, they just keep turning the crank on what they've done before. It's the little speedboats that get to zip into different places. However, it takes a certain confidence and a certain kind of thick skinnedness and so on to do things that aren't what everybody thinks you should be doing, so to speak. And I think the general principle is most big innovation happens in new fields and new fields created between existing fields. In fields that don't exist yet, that's the time of highest innovation. I think one of my slightly cynical statements is by the time there's a prize for something, it's too late. That's not quite fair and true, but it's kind of because fields do change and so on. But by the time there's a well defined shtick for something, kind of the days of highest innovation are over. In 100 years they'll come back. That happens in fields quite often. It's kind of like there'll be 50, 100 years of fallow period and then there'll be enough ambient sort of technology sometimes or paradigmatic development that you'll be able to come back into that field, ride back and conquer that field again with something new. But it's, you know, I think that the, the tendency is an existing field, you're well supported but you're on track. And when it's, you know, when you're really outside the system, it depends how far outside the system you are. I mean, I think it is, it is when you have the opportunity to be sort of inside the system, like you have a job doing this but you have enough time to do something different. You know, that's a good deal. I mean in the case where, gosh, I mean, you know, one of the things that can happen is like in my own case, I was a fine card carrying successful academic before I went sort of outside the system type thing, that was a fairly, that was a win. You know, if I'd been sort of outside the system from day one, I think quite a lot of the things I've done would have been much more difficult to do. Now it's, and by the way, you know, the system, let's make no mistake, the system is a, has a big self defense operation going. I mean it's kind of like if you say to the system, well, I got something new and different, the system's going to push back. It's going to say no, no, we don't need anything new. You know, we're, we're just fine. You know, we don't want the new thing. We're happy just doing what we're doing. And from the point of view of individual people, that's a, you know, people don't want their careers disrupted. And it's quite understandable. And even in the data, apparently you kind of see that, that there's a kind of a cynical thing that sometimes said. I think it's due to Max Planck actually. He said science proceeds in a succession of funerals. By which he meant that people just keep doing the same thing their whole lives and until they die or disappear. For some reason it's hard to get new things to happen. And I'm afraid there is a certain tendency for that to be the case. And it isn't always true. I mean, sometimes people in mid to late career will kind of, there'll be a new thing that comes along. And by the way, that thing is the thing that if those people have been luckier, would have come along early in their career because it is the thing that is really the match for what they are really good at and want to be doing. And then they make a change. It's the exception rather than the rule that that happens. The much more common thing is people keep turning the same crank they've been turning before. Some of us are perhaps luckier in that the crank we've been turning, insofar as it's a crank, is a crank that's all about doing new things. And that means you get to do new things. I mean, in the case of somebody like me, I'm used to doing new things, although I suppose I have the overhang of existing things have a lot of kind of development and support and technology and so on around them. But that's, you know, that, that's a. Yeah. So I mean a lot of what happens is, is just you're keeping on doing the same things. But I think, you know, one of the issues about being quotes outside the system is there may be methods that are known. You know, it's like, well, how to explain yourself in an academic paper? Well, sometimes those aren't the best way to explain yourself. But even like, you know, I see People where they're like, well, I'm studying chemistry, but by the way, I'm also solving for the theory of everything. It's like, no, that's not a sensible thing to say. You know, it's like stick to the thing that, where you're making real progress and it just makes, it kind of, it ruins credibility to say that you're off solving everything else as well. Now having said that, I have to say I find it ironic that a lot of things that have come out of our physics project look like they touch other areas and we get to say sort of foundational things maybe about areas like biology and economics and so on. I will claim this is a somewhat different situation because as just a big tower of actual content in the physics project, point one and point two, I'm at least trying to do my homework in these other fields and really know them well enough to know where I'm injecting myself. And I think sometimes people, the thing to understand about different fields is they really have different ways of thinking about things. And people who just sort of say, well, I've got this thing from this field and I know this one well and I can immediately transplant it without change to another field. That usually doesn't end well. All right, I think we have one question we should wrap up here. Meltdown is asking what's my view on using thorium reactors for energy. Ask that on a different live stream. Let's see, is there anything else I can quickly address here? No, I think that that looks like everything I can can quickly address here. Well, thanks for, thanks for joining me and see you another time. Maybe I'll have a different non natural habitat location next time and see you another time. Bye for now.