Business, Innovation, and Managing Life (January 17, 2024)

Hi, everyone. Welcome to another episode of Q and A about business, innovation and managing life. I'm still not back in my natural habitat, so all sorts of strange things could go wrong. Let's see, there's a question here from 1h. Do you think a PhD is worth it when you're later in your career? Interesting question. I mean, PhDs are kind of a credential you need for certain kinds of jobs, pretty much for jobs that involve minting more PhDs. That is being a professor, for example. And so if you want to be a professor, you need to get a Ph.D. most of the time. I mean, when there's a new emerging field that is more kind of practically oriented, then there are usually moments when there are people who are professors who don't have PhDs. But in well developed fields, it's kind of like that's the credential you need to be able to mint more PhDs. It's not true. For example, in business schools, that's often not true. I think they're terms like professor of practice and so on, which are for people who don't, you know, who aren't like in the chain of PhDs, so to speak, as far as other things are concerned. You know, I think there are moments when companies, for example, have sort of, well, we'll pay more to the PhDs than we do to the others. But I'm not sure that that's much of a thing these days. I mean, I would think that there are, you know, okay, there are other cases where I don't know, you're working for a think tank, you're working for a government lab, you're working for something where part of what's relevant is being able to say, yes, we have lots of PhDs. I think that's in terms of the value of the credential, so to speak. Now there's the question of what about the things you actually want to learn? Well, One feature of PhDs is that in the end they're focused on writing a thesis and the thesis is about some typically particular, often quite narrow topic. The good news is you're building a big tower, and it's interesting to learn how to build that big tower. The bad news is it's a specific tower. If you're interested in being kind of a polymath who knows a lot about a lot of different things, the PhD is not the thing aimed at delivering that. It's the thing that's aimed at sort of giving you the experience of doing a specific sort of build it tall kind of tower. Now it also has to be said that sort of the requirements for PhDness are quite different in different kinds of fields and the amounts of time involved are different. I mean, there are fields where it's kind of like three, four years. Well, it depends on the country, but you know, four years, maybe five years and it's like you're done. And there are other fields, particularly in some areas of the humanities, where 10 years is not an unusual amount of time and where it's kind of where people are often doing a PhD alongside doing lots of other kinds of things. And it's much more of a kind of, I would say, I don't know, experiential type thing. And many people who do Those kinds of PhDs produce a book that is their PhD and so on. It's a different model than the model in areas like, I don't know, math and physics where it's like write a couple of papers and that will be sort of substantial papers and that's going to turn into a Ph.D. and that's it. And it's at least for the theoretical side of those things, it's a shorter period of time. Now there's the question of sort of if you're in mid career, what's the point of getting a PhD if you're already doing a bunch of stuff, then you know what, well, as I said, there are sort of credentialing reasons like you want to be a professor and you need that. You might not need it to be a professor as I mentioned, but you know, and then there's the, you want to kind of, you know, you. Another reason is sort of a personal priority setting reason that is you've got all kinds of things going on in your life and if you say, oh, I'm just gonna, you know, jump out to something quite different, I'm gonna, you know, go and get a PhD and I'm gonna quit my job and I'm going to focus on doing the things that I'm doing to get a PhD. That's a kind of personal management type thing where you might say, well, I'm intending to learn this stuff, but you know, I'm always, every day I've got stuff to do with my job and I never get around to it. So it's kind of a thing where, you know, it's a, it's a way to sort of force the issue of yes, I'm really going to concentrate on doing this kind of thing in terms of can you learn the sort of material that's sort of PhD type material. I mean one of the things about sort of the PhD story is it sort of teaches you to be an academic. Now that's different from it teaches you to study field xyz. There are perfectly good ways to study those fields that don't fall into kind of the academic pattern of writing papers and doing this and that and the whole sort of academic dance. And I think that that is, you know, it depends on the field. There are some fields where sort of everything. Another thing to say is there are fields where there is a sort of necessary amount of stuff you have to learn to be able to make a contribution in that field. And the, you know, you need to take. It takes a lot of effort to learn that yourself. It's easier if you take classes. That's true in some areas like math for example where there are sort of. Where it's sort of a tall tower of things and theoretical physics and so on. But there's a tall tower of things to learn. It's absolutely possible to learn those without taking classes. Certainly in my life I never took any classes about those kinds of things but it requires a certain tenacity and in my own case it sort of required having projects that I wanted to do that required learning those kinds of things as sort of a driver for that. I mean I think different people are driven differently. For myself, I'm much more into I've got a project I want to do. Let me kind of learn what I need to do that project. Other people are much more look, I'm going to show up to this class every, every Thursday morning and then that's the time I concentrate on this and that's, that's kind of how I sort of manage my life. So anyway, that's a few comments on you know I think again if you kind of want the hiatus of your life of jump in and do the Ph.D. that's one thing. Even if you maybe want to do it alongside other things as a kind of a way to manage yourself to do it. I think that's a worthwhile thing it turn, you know, I'm not sure. Yes, as I've said, I mean there are very specific career paths where it's important but others I don't really quite see it. I mean I think the, you know, one of the theories is oh, you know, very pragmatic kind of thing. Oh I'm in the middle of, you know I'm, I mean the reason people go back to school to get e.g. mBAs in mid career not Something I'm a huge fan of in most situations is because they are on a particular career track and they sort of want to kick themselves into a higher career track, so to speak, by sort of by having that extra piece of theming. And I guess that sometimes works, I think, with PhDs a little bit less well under, you know, well defined than what happens with MDAs, MBAs. Let's see, Let's see. Gosh. Hoggy is asking, as far as managing life, what has been your biggest takeaway in your career or personal life that you would like to pass on to the younger generation? I don't know. I think figure out what you like to do and try to figure out how to do it is a good thing. People get locked into these kinds of, well, I did well in this at school and there's a career track for this and so on. And it's really all a bit of a drone own activity. But that's sort of what I've got to do, I think with some creativity and maybe some tenacity and maybe some luck occasionally. I think there's a sort of a great potential to find the things you actually like to do and then go and try and do them and not settle for, well, I just kind of drone through life at least. You know, this is always complicated because there are people for whom kind of the, the career and what they do for a living is the most important thing. There are other people where kind of the thing they like the best is, you know, kind of the meals they have or something. I mean, for somebody like me, I don't, you know, I'm not, I'm not. I don't really, you know, I eat the same thing for breakfast every day type thing. That's not a thing. That's sort of an important, notable feature of my life. The things that are notable features of my life tend to be sort of intellectual projects and things like that, but different people are different in that respect. And so it isn't necessarily the case that the thing you're trying to optimize for is, you know, get the best career you can. Although it's really nice if that aligns with something that you want to do. I think that's one thing. Another thing for myself at least, and I think for other people is sort of one can figure things out. It isn't necessary that one just sort of takes the accepted wisdom and goes with it, so to speak. It's kind of a, you know, thinking is something an individual person can do and can create things from it's not something where thinking is something which you are only given programmed kinds of instructions for how to do. It's not something where everything you do, and it tends to be the case from education that you're sort of given a formula for. Now you do this, this, this and this. And that's how things work. I think it's one of the things that certainly has been a feature of things I've ended up doing. And something that I like to encourage other people to do is, you know, the thing that we humans are able to do is this actual thinking, so to speak, that isn't necessarily just following the mechanics of what was suggested before. So that's. That's a thing that I think is really worthwhile. And people sometimes lose sight of it. They kind of like, you can't just think about that from scratch. You can't create that thing from scratch. The only thing you can do is be part of the sort of production line of the world in which you're in this particular slot. I think it's always a thing that I find interesting is when you get. I mean, we have our WolframAlpha product, for example, which happens to have, you know, my name as part of its name. And I have this interesting experience with some regularity of particularly younger people who run into me. They hear my name, they say, oh, wait a minute, there's this Wolf Mouthworth thing, you know, sort of. Are you connected to that? And I'm like, yeah, you know, that's kind of my thing. And they're like, wow. There's this moment of surprise where it's like, there's a person behind this thing that appears to be this software product and all that kind of thing. And it's kind of like everything out there is something somebody had to initiate, invent, et cetera, et cetera, et cetera. And usually it's the case that the things one sees there was an actual individual person who caused that thing to exist. Much more rarely was it some giant kind of mechanistic committee, so to speak, that caused something to exist. Sometimes that person will be a very visible person. Sometimes that person will be a much more behind the scenes, the project manager who made this move and that move and caused this thing to exist. Let's see. Lola asks, and I'm sure I've talked about this before, but how do you approach risk taking in business? And what advice would you give aspiring entrepreneurs about managing risk? I think it's true of many people who try and do sort of things that haven't been done before. And certainly it's true of myself that I'm not thinking about kind of the things that the projects I do, I'm just thinking about getting them done. I'm not thinking about them as, oh, this project is a big risk. Now I will say, having said that, I certainly do think about what could go wrong, but I don't think of it as, I always think of that as a kind of a way to not have those things go wrong, to understand what to do to avoid those things going wrong rather than as a oh my gosh, all those things might go wrong. Maybe I shouldn't do this project, it's too risky type thing. I mean, I think the projects that I end up choosing to do are ones where I'm confident that it can be done. Now some of those projects turn out to be super hard. And quite often as one is into doing that project, you have to kind of change the, change the milestones, change the directions a bit because the project you originally imagined, maybe it wasn't quite the right project for one reason or another and you have to kind of make a mid course correction to go in a slightly different direction. But I think that one of the things that's true probably about many entrepreneurs is that there is a certain kind of reality distortion bubble. I think that's a term that people used to use in connection with Steve Jobs and communicating that to other people. But there's a certain, look, I'm sure this is going to work. Even if from the outside people say that's very risky, that's very, very risky internally for the entrepreneur. It's like, I'm sure that's going to work. You know, I have it, I'm, you know, I'm in charge, I'm going to make this work, et cetera, et cetera, et cetera. And I think the people where it's like a lot of self doubt of, you know, can I make this work? Et cetera, et cetera, et cetera. That's a bit of a challenging thing for an entrepreneur particularly because when you're leading groups of, when you're leading other people, you kind of have to communicate, we can do this team, this is something we can do. Not, I mean, you know, one has to, it's a complicated thing to communicate. Look, we're trying to do something that for example, has never been done before. It's really hard. There may be some issues in doing it, but it's like, well, I don't know if we could possibly do this and I'M not really sure and it might be too big a risk and maybe we shouldn't do it and et cetera, et cetera, et cetera. That's really not likely to work in terms of leading a team because people, people who are working in a team, there has to be some notion that look, we're really going to do this, it's really going to work. It's really, you know, go team. It's really a thing that's happening if everybody is doubting everything all the time. I think in a sense, if there is doubt, it should be concentrated on the leader, inside the leader, so to speak, rather than it's sort of not the responsibility of the team members to kind of spend a lot of cycles looping on, on the doubts, so to speak, that the thing for them is, okay, somebody has decided we're going to do this, they're leading it. My role is to do my part of what I'm supposed to do as a member of this team rather than to be the one who's worrying. Is this the right idea to begin with, so to speak? Let's see. Hyvie asks, do you have New Year's resolutions or big plans for 2024? Oh boy, I have so many projects, I'm not sure. Some of them are projects related to our company and trying to communicate better the kinds of products we have, new products we're building, all these kinds of things. I have sort of a personal project which is trying to write a book, course, etc about computational thinking, which is something that I think is. I've sort of increasingly realized people don't really quite understand what it is. And people say well, what is it? And I say well, there's pieces of it which I've explained and it's like, well, why don't you just explain the whole thing? And so I've embarked on the project to do that. I don't know how long it's going to take. It's a great example of something which to me is like, I'm going to do this, it's going to be possible. I don't know how big a book it's going to be. I don't know how complicated it is. I'm just getting into it. Turns out I can see that it is intellectually much more complicated than I had originally imagined. I'm not daunted by oh my gosh, I can't do this. But it is definitely something where there are non trivial things to figure out about sort of this presentation of something that really hasn't been presented as a coherent thing to the world before. Let's see. Okay, Ssh says, I'm a software engineer in the field of quant plus data. Science is very appealing to me. What advice do you have? You know, at any given time there's always sort of a certain gold mine, a certain place where you can kind of the gold is near the surface and hasn't been explored yet. I think quantitative finance is one of these things where sort of the gold was first exposed sometime in the 1980s and there's still lots of gold there. But at any given time there's always sort of the new area where there's kind of sort of a set of skills first become important. For example, logistics and things around logistics or ad tech. These are things where some of the same kinds of sort of quantitative quantitative finance, like thinking and modeling in both of those cases were relevant and sort of people moved from sort of fintech and quant finance and so on into those areas. I think one of the things that's sort of an issue sometimes is when it comes down to the raw question of money, there's sort of how do people get paid? And you know, one of the things that say, well, perhaps sad, perhaps not fact about the world is a good way to make money is to be close to the money and what you do. And if you're part of sort of the flow of money from here to there, if you're a quant finance person who's sort of managing the flow of, you know, these people are investing this and they're going to do that and it's, it's, you know, if you're very close to the money, it's easier to make money from it. It's also sometimes there are, there are fields where people say you're going to get a cut of what comes in. And that might happen in quantitative finance with commissions of various kinds of. It might happen because you're a partner in some firm, it might happen. If you're doing a startup and you get some equity stake in the startup and perhaps through no fault of your own, the startup ends up being successful, which is something quite rare, but it certainly happens. And one knows about lots of sort of newsworthy cases where it does happen. Some of these other fields, there's much less of a. You get a cut of what comes in from it. And I think that's something which affects people's view of these fields. Now in terms of what is there to do in, let's say, quantitative finance meets data Science. I think a large part of the quantitative finance story is a story of who's got the best data. And a lot of effort goes into. Sometimes it's like the world history is going to repeat itself. What's happening now is just the same as what happened in 1922. Well, you've got to have the data from 1922 and your model may automatically find this collection of prices that are doing this and this and this matches what was back there. But you have to have a clean data from that. And I think that's a big part of sort of the effort of doing a bunch of quantitative finance. And that's very kind of data oriented. It's also the case that quantitative finance is also often about is there yet another new source of data? You know, those new pieces of satellite imagery that show this or that, those new things where you can kind of plug into some flow of information about this or that process going on in the world. These are things where often that sort of that new data that nobody's has access to before. You know, whether you put, whether you're using satellites to look at how many cars are parked in the parking lots of retail stores or whether you're using microp loans to figure out how people are stomping their feet in the trading pits, which I don't think exist in physical form anymore at a commodities exchange, those kinds of things. Each time you get this sort of new source of data, there are new opportunities to do quant finance around it. But I think in terms of the mechanics of the market, the thing to realize is the quantitative finance. The successful quantitative finance operations are very small. The number of people involved is quite, quite small. It's sort of high talent, lots of work being done, but not a lot of people. I mean hundreds of people, not tens of thousands, hundreds of thousands of people. So it's a rather, and I'm always amused that there are quantitative finance firms that seem to be like their recruiting is everywhere and yet the Firm only has 200 employees. And it's kind of like, you know, you're doing, you know, you're sort of trawling for, for what you, you know, for a situation where you might talk to a hundred thousand people or something and yet you're only hiring, you know, a tiny number of people every year. And I think that's. So it's a very sort of the high end of that business is an incredibly filtered business that involves lots of smart people. I'm, you know, I must say that for myself, it's Not a way that I would find it fun to use my skills, but it's certainly a thing. I think that it depends on what one's interested in. It's a kind of a thing where it's a great intellectual challenge and it also happens to make a bunch of money if you're successful at it. But I would say that it's worth realizing that. That there's a small number of slots for sort of the very high end of that kind of world. And I would say that in kind of, you know, how do you use data science and be more at the front lines of sort of business and so on? I suspect that there are other opportunities for that in industries other than quantitative finance. Wow. Junior is asking, I teach my first class tomorrow. What advice do you have? You know, it's a terrible fact that when I was first a professor, there's no. At least at that time, and this is now 40 years ago, there was no kit that anybody gave you saying this is how you're supposed to teach a class. And in my particular case, because of my particular life trajectory, I'd never taken a class that was sort of taught in a standard way in a standard American university and so on. So when I started doing my teaching a class, I had no idea what to do. And I think it had certain disastrous features, like I didn't really know what the expectations were about sort of what homework should consist of and what. And how difficult homework should be and how much thinking should be expected on the part of students, those kinds of things. So I would say I may not be the very best person to give advice about this. I think that the, you know, somehow one of the things that people I think ultimately appreciate is being organized when you're teaching classes. Not something, you know, for the things I'm doing. And like these live streams, these are totally the opposite end of the spectrum. These are completely unprepared, off the cuff and so on. But I think students appreciate sort of a structure and organization to classes and that. That. That's a thing. I mean, it depends what you're teaching a class about. But I think the. And there's also a lot of dynamics about, you know, when you have a bunch of students and depends on what age of students, so to speak, in a room and how many there are. You know, are you teaching a class to 100 students where it's a sort of big broadcast thing? Are you teaching a class to 10 students where you expect individual interaction with each of those students and so on? There's Lots of different, lots of different dynamics there. I mean, I suppose I have to say in, in, in some of the times I've done this and I haven't done sort of structured teaching for, oh, gosh, 36 years now. So, and, and when I did, I, I don't think it wasn't a great match for me. I mean, in, in the sense that I like to think, I hope I'm fairly good at explaining things to people and getting individual, getting people to understand things. I'm not sure that I'm the best at saying this is the curriculum. We're going to go through this set of structured steps. And it's kind of worth realizing that those are two different things. I mean, giving, sort of inspiring and engaging and informing is a little different than saying at the end of this class you will know how to do xyz. And so, and those are somewhat different skills. And I'm more on the first one myself than the second one. Let's see. Ed asks, you have an advice on pricing software licenses. They say, I build bespoke design software and I'm kind of winging the licenses. Gosh, you know, software licensing is one of these things where sort of conventions have changed over the years. Like for example, in modern times, everything is a subscription back. Let me tell you a story which perhaps is helpful, perhaps isn't. When we first released Mathematica in 1988, we had to figure out pricing. How did we do that? Well, in those days, there were actually two different models in the software industry. There was kind of the Microsoft model, which was a $500ish productivity software type model. And there was the Borland model which was $100 and less model. And at the time it wasn't obvious which of those models would win. Many people probably never heard of Borland. So you know, which model won? It was the kind of professional software at the $500 price point kind of story that won out and the industry might have developed differently. It didn't. You know, one of the issues with software is, for example, when we first released Mathematica, I knew that if we were forced to have our basic software license be $100 or less, we couldn't make the business work. So it relied on the fact that we were kind of in the Microsoft bucket rather than the ball and bucket of how things would evolve in the world now. At the time it was like, well, can we, you know, what really is the right price point at the time, different world. There were computers that had floating point units and there were computers that just emulated floating point in software, and their floating point operations were a lot slower. So I had the bright idea, let me make a standard version of Mathematica for the Macintosh or something for the Mac that was $495 and let me make an enhanced version that was $795. And the enhanced version would make use of the floating point unit and the standard version would not. And I thought people who are very price sensitive, they admire it by the standard version and the enhanced version will be people who want, you know, the highest performance, but they're not as price sensitive. Okay, first observation is basically everybody bought the enhanced version. That was fine. I was happy that, you know, it was a. It was. We had a better price point there. But what was sort of interesting was that I think the. Well, perhaps the point was that the market. There just wasn't a market there for people who were that price sensitive. That the early market was people who really wanted this thing and they wanted this thing and they wanted it to work as well as possible. Now, you know, also at that time, I mean, there were lots of, lots of things that we did wrong in early pricing. Some things we did right. I think the, I mean, you know, okay, so what are other considerations in pricing? I mean, there are considerations like who's going to buy it and how are they going to buy it. And for example, if every sales transaction involves a human on the phone having a kind of detailed discussion about the software, you basically have to have a pretty high price point, I don't know, $10,000, whatever it is, to really, or at least a number, few thousand dollars to justify that. On the other hand, if it's something that, where, if it's cheap enough, people just say, yep, I'm going to give you my credit card, I'll buy it. It's all good. Then, you know, that's a different level. Now, when you get to selling to companies, there are often levels where it's like, oh, this person of this type of person has sort of signature authority to buy things of this price point and so on. And if you kind of the worst thing to do is to be right above one of those thresholds and to have something where it's kind of like the people who want to buy it have to go to their supervisor's supervisor to be able to buy the thing. And that's a huge sort of slowdown in what you're doing. I think another thing which has changed over time is sort of the negotiability of software Prices. And I think in, you know, it's important to have. I think just a. Here's what it costs. You know, one of the things that people do wrong sometimes on websites is they say, you know, it's full of. We have this amazing thing. It's a solution for this and that and the other. And it's just like, tell me how much the freaking thing costs. I'll tell you a story perhaps from a different industry, the encyclopedia industry. So long ago, Encyclopedia Britannica, which was sort of the dominant encyclopedia for a couple of hundred years, actually had a very weird sales model which involved actual salespeople who would, like, go and visit your house and bring the encyclopedia with them and kind of say, how does it look on the shelf? Type thing. I'm not sure if they actually did that, but it was a very individual kind of selling of the thing. And some point, this is probably early 1990s, I wanted to buy an encyclopedia, and it's like, just tell me the freaking price. And it's like, no, no, no. Somebody has to come and visit and do this and that and the other. It's like, I don't care. I'm just going to give you, you know, credit card or check, whatever it is. I just want to buy one. You know, you have a customer here completely on the hook. I don't need any more sort of dancing, so to speak. They simply wouldn't do it. And eventually, I was at some book street fair one day and there was an encyclopedia booth. And so I walked up to the booth and I'm like, I want to buy one of these things. And the person was like, look, I got to do this song and dance. And I'm like, okay, fine, I've got five minutes. Do your song and dance. And that was enough for them. And then they would just take my credit card and I bought an encyclopedia. But in other words, somehow I sometimes think of that when I see some of these websites where it's like, let's tell you all about this thing and how wonderful it is and how much we solve your problems and so on. It's just like, just tell me the damn price. Prav is commenting on the chat here. That call us for pricing is infuriating. Yeah, I mean, sometimes there's, you know, when you're selling a thing that's complicated enough, there isn't a choice. But I think people, you know, I think the world has evolved to a point of sufficient transparency that, you know, people just want to know the price. And if they say, oh, it's $5,000. You know, I was thinking it was going to be $50. You're never going to sell to that customer. That's just too big a gap. And the person who says, look, oh, you said $5,000. I was really expecting it to be $4,000. You're still going to sell to that person. But if the gap is too big, you might as well have the person just say, look, I'm just not in this market. I'm not going to get something that's at that kind of price point. So a few thoughts there. At least there's a question here about from Paula. What do you think of subscription versus one time purchase for software? In the past, when we first started selling Mathematica, what's now Wolfram Language, it was all one time purchase. Over the last 36 years, things have evolved to be really everything is a subscription. Everything has tentacles that connect to the cloud. Everything is kind of an ongoing thing. And for us right now, if we said, okay, just buy this thing today and it's yours forever, so to speak, it's like, well, 20 years from now we might be processing requests to our cloud from that person who bought the thing 20 years earlier. And that's not really a viable possibility. So with all these kinds of sort of cloud tentacles, I think subscription is really the way one has to go. And I think that's the way the industry has gone. Is it my favorite thing? Perhaps not as a customer, but it is the way that for all sorts of reasons things have tended to go. I mean, for us, I think we sort of, you know, there were countries in the world where certain kinds of organizations could only do, you know, this is a purchase one time. I think the last of those is now gone. And so we don't end up doing that. I think it's worth remembering that there are things where the benefit you get is sort of all a one time benefit. Like for example, when people buy a paperback book, probably they're going to read it once and then they're going to kind of sequester it away and not really get much benefit from it. Or if you're reading there are things where you're going to some event, you get the thing of the event you don't have. There's no tail to that. And but I think with software it tends to be the case that if you're using it and you're going on using it, the benefit you get is a, is a benefit distributed over time. And I always think that for any kind of pricing. It's a good thing to match the benefit somebody gets with the thing you're charging for and the way you're charging for it. Because as soon as you're kind of like, well, you know, we're going to, you know, it's going to be free but we're going to make it up in charging advertisers to advertise to you, things like that. That creates a lot of very complicated kind of incentives that don't ne that I think are more complicated to deal with than just whoever's going to get benefit from it. That's the organization or person or whatever who should be paying for it. And I think. Yeah. Another comment from LC here is upgrades and maintenance and support. Yeah, I mean I think another issue is that most people, computers change and computers get upgraded and operating systems get upgraded and there's sort of a force of, yeah, you've got to upgrade things and the thing that you bought. I mean I have to say we try to build software that will just keep working forever but sometimes we just can't do that because the manufacturers of computers change their operating systems or other things to the point where the thing we built 10 years ago just doesn't work anymore. And I think the now when it comes to sort of get the latest features, that's a good and important point. We just released version 14 of Wolfram language and Mathematica and it's kind of like takes a lot of work to keep innovating and delivering new features and so on. And that is something again where you get a good match to the people who are actually using getting benefit. If it's a subscription type thing where people are like continuing to pay for something you're continuing to work on. When it comes to support, that's a different issue. I think that I like to think for our particular products and so on that we have give good support and people seem to give us feedback that we do, which is great. And I think that the, I don't know, in all cases, you know, it's, I always get a bit, you know, would you like to pay for support? Well, you know, good products in a sense, you know, there's always a trade off. You can make a better product that needs less support but you can say I'm going to have my product be shoddy and I'm going to make the money on support. This is a, this is a common trick in the open source world of, you know, have the thing that is sort of the free thing be sufficiently unusable and crummy. You kind of have to use sort of the enhanced version with support and things like that. And that's the way to make money from it. You know, I think that's a, you know, there's a certain, you know, you can move that, that by putting more effort into the product, you can make it. You put better documentation, works better, needs less support. And I think that the question of, well, what is the quality of support? That's something that, you know, I'm not sure. I don't really have a good sense of how that works across lots of different companies. I know, you know, we're pretty pleased with the way that support works for our company. I have to say, it's an interesting thing in terms of sort of how this stuff evolves, because back when we started 36 years ago, one of my concerns was that we will become the homework help center for the world, so to speak, where by homework, I don't necessarily mean students. I mean people who were, you know, industrial physicists or something, trying to figure something out. That we would become the place where everybody went to just get answers and that they would just say, oh, we just want product support. Just tell us how to work out this thing. And that was obviously not going to be supportable. We couldn't do that as a, you know, we were making products where we expected to make a tool that then could be used by people, where people will be buying it as a tool, not, oh, by the way, can you spend 10 hours figuring out the solution to my problem? As it turned out, people's expectation, and possibly this was because much of the support of software, particularly in those times, was really bad. And so the people's expectation for what you get by calling the support 800 number or something, which is what it was at the time, was sufficiently low that nobody thought you could call up and ask your physics problem as a support question. And when they did ask their, well, how do we use the product to do this? Or whatever, they got good answers. They were happy. It was all good. But that was a case where kind of that was a potential concern about how the market would react, that in that particular case didn't happen. And I think people have been. That's where you have to kind of rely on sort of the ambient conditions of people's expectations of what people expect support is going to consist of. And so on. The. Sandra is commenting. Imagine mathematica with inline YouTube ads. Oh, boy. Yeah, yeah. I mean, it's. Boy, I never even thought of that. Actually. It's There are many things one, one doesn't think of because they're such obviously bad ideas. Um, but it's. We, we've never even run on Wolfram Alpha, for example. We've never run kind of. We tried at the very beginning we, we did consider running just straight kind of Google Ad like ads. And actually particularly at the time, it's now nearly 15 years ago, the, the matching of what ad you should run was so bad that it was just an embarrassment. I mean it was kind of like really, do you want to advertise life insurance to what's probably a, you know, a young student or something? Or do you want to advertise? I don't know what it was. It was just a very, very poor match. And so we ended up doing is just running sort of informational ads about things of ours that partly we, you know, they add some color to the website and they're hopefully useful and interesting and maybe they help us sell some things. But we didn't really emphasize kind of an ad supported model there. Let's see. The question from Sam. What's the coolest thing your company has ever done? Boy, it's a long history now, 36 years. I would say the, the big coolest thing is to define this idea of computational language, which is kind of, I think, the future of how one formalizes the world computationally. That's kind of a. I wouldn't say that's really a thing in the sense that that is a backbone of what we've been doing, what I've been doing for the last 40 something years, what our company's been doing for the last, what is it, 37 years or something like that. It is interesting that from time to time we do things where it's like there's a moment where people say, oh, that's a really cool thing. And we've had a lot of those. I wouldn't say that there's any one that stands out more than any other. I mean there are ones that are big things like WolframAlpha, for example. There are ones that are little things like the blog post about what's the most difficult word in hangman or something which can get lots of visibility. I mean, the world's a weird place where what will have sort of a burst of viral activity in social media may not be the thing that is even the most remembered by people. It's just a thing which happened one day and it was cool and everybody enjoyed seeing it. And then everybody goes on with their life and forgets about it. So it's, I think the things that I'm proudest of that we've done are mostly these very large scale, kind of giant intellectual structures, like our whole computational language and the design of our computational language and things like that. Elsie asked how much computational thinking and modeling do you do on our business itself, both for decision making and planning? You know, it's always been something of a, of a frustrating paradox that, you know, I know lots of stuff about quantitative kinds of things, modeling kinds of things, sort of complexity and the science around it, all that kind of stuff. I think the number of times we sort of explicitly say that's the top of mind for what we're doing in planning things in our business is not large. But the thing that I often forget is that a lot of thinking that I do, for example, is deeply informed by the fact that I am completely used to doing sort of precise, analytical, formal things and somebody else thinking about the same questions would think about them quite differently. Do I write down the formalization or is that merely part of my way of thinking about things? It's more part of my way of thinking about things. Also for us, it's kind of like there are things where we say, well, let's try this out and let's see how people react and so on. And then it's very cheap for us to say let's get a bunch of raw data on what happens and then let's write a piece of orphan language code that grinds it in a particular way. And one of the features of our company is, is that many of the senior people in the company are good Wolfram language programmers. And so for them, if they're running some product, for example, it's kind of like, in fact, I was just having a conversation this morning about a new product that we're developing where there's a question of sort of how will people actually use it? And so on. And it's like, you know, the person who's running that is an experienced Wolfram language programmer and it's going to be easy to take that raw data and answer the questions we want to ask. It doesn't require. And that's something where we're deeply using our kind of technology, stack analytical skills and so on. Even though for us it seems easy, for someone else, it might seem like, oh my God, that's a whole complicated thing. I got to hire consultants to do that or whatever else. So it's, to me, I might, you know, if somebody says, are you running your company based on Scientific principles, I would probably say. Not that I notice, but when I think about kind of how I actually think about things, it's very science informed. Though not in as sort of explicit and formalized a way as one might think of sort of a science project, so to speak. Oh my. There's a question here from Lizzy. Do I consider myself a celebrity? What has been your coolest encounter and weirdest encounter? Boy, well, all kinds of strange questions get asked here. You know, I think I'm a. Some kind of micro celebrity which means that for example there are, you know, if I'm kind of just hanging out in the world, in most places in the world, nobody's going to recognize me. There are a few, few square mile areas in the world where it will take only a couple of encounters before somebody will recognize me. Actually this is a, as far as I'm concerned this, it's a very nicely controlled situation because if you don't go to those few square mile areas where it so happens that people pay attention to the kinds of things that I do, you know, you're a fine anonymous person, so to speak, which I think keeps life simpler, so to speak. I would say that people, oh, all kinds of funny things happen. Like you know, in a. I like to have, you know, my, my backpack or something has a spiky logo on it for logo of our company and our products and so on. And I, you know, I partly do that just because I think it looks cool and because well, occasionally, you know, if I'm like at a trade show it's helpful to have that because people, it kind of gives people the clue, oh, we might want to talk about this product type thing. But it is amusing that from time to time people will recognize the logo and say, you know, kind of, oh yeah, you know, that's, that's this thing from. That's this thing I, I, you know, use for, you know, this Wolf Malpha thing that I use or whatever and their connection to that's sort of an interesting way in so to speak. I think the other thing that I was mentioning earlier actually this dynamic, particularly with younger folk where this idea that actually there's a person behind a product that they use is both very sort of surprising and very educational and perhaps even a little bit inspiring. As in humans do these things, so to speak. And that's kind of a, I find that sort of, it's kind of a, I don't know, it's some kind of good dynamic that you feel like people are learning something that's Useful to them just by the fact that they can see there's a human behind a thing that they use. Weird encounters. Well, yeah, I'll tell you a few. So, you know, funny thing is the name Wolfram is the, you know, the scientific symbol for tungsten. The Element tungsten, element 74, is W, which stands for Wolfram. And in quite a few languages around the world, Wolfram is the name for. Or some derivative of that word, is the name for the tungsten element. And I remember one time talking to. I've had these weird interactions. One in particular stands out from years ago, talking to a person who was, at that time, I think they were launching, maybe it was the Mongolian stock exchange. And the person knew my name. And they, you know, I think it was at some event, whether it was a name tag or something, and I simply couldn't communicate the fact that we were not in the tungsten mining business. That was just kind of. That somehow the name, you know. No, that's just my name. It has nothing to do with tungsten, so to speak. Oh, another thing that can happen, which is amusing, is with the Wolfram Alpha name is people where in some setting they sort of imagine one's last name is Alpha. That's a fun thing that can happen. I'll tell you one more thing about those kinds of funny interactions. So. My first name is Stephen, right? But turns out for Americans over, I think these days, about the age of 70, nobody will ever call me Stephen. It's always Steve. And I found this very strange over the years. It's kind of like it doesn't really matter to me. But it's just an odd phenomenon that even if you introduce yourself, hey, I'm Steven type thing, they'll always call you Steve. I think, actually, very recently, I kind of learned that part of the dynamic of that, which had to do with the fact that in a certain period of time, 50s, 60s and so on, to use a full name like Stephen was very pompous, which I don't think it would be particularly today. And so in a sense, it was a. That everybody was using a shortened name at that time. And maybe that's a sort of a holdover of, if one doesn't see my. I'd always thought that it was because in England where I grew up, it didn't happen that way. It was a very purely American phenomenon. And I always thought it was because there were so many kind of Steve was such a common name in the 1960s and 70s and so on. And there were so many sort of and that one somehow one, if one seemed dynamic, one seemed like a Steve, not a Stephen, I don't know. So anyway, silly points to make. But it's, you know, one thing about, insofar as one is sort of a micro celebrity, so to speak, it has both positive features and negative features. The positive feature is there's a good chance people will respond to email if I'm asking a reasonable question and so on. The bad thing is there are situations in which you would like to have sort of an ordinary conversation with someone and the person is like, oh my gosh, you're a micro celebrity. I've got to have a weird conversation type thing. And that's sometimes a challenging thing to deal with. Let's see. Christian asks, what do you think about organic education matters? Basically, can you use an AI to figure out sort of a fix for education? What do you want to understand versus testing for knowledge? So, you know, I mean, we're working on sort of AI tutoring products. We don't know how well they're going to work. One has to understand what is the purpose of education. You know, if the purpose of education is learning very specific skill where we humans can be trained up to do this or that thing, that's one kind of thing. And there it's kind of like, well, if you've got an AI that can tell you, look, you didn't understand this thing, learn just that, learn this piece and then you will have this skill and you will be able to ride the bicycle or whatever. That's not the type of skill one's learning from AIs right now. But if one is do the particular sort of mechanical skill, we humans can be trained to do those kinds of things. And that's a thing where sort of the goal is really the end point, so to speak, of get trained to do that. There's a quite different purpose of education. I mean people talk about this a lot of kind of the liberal arts type education. The learn to have a considered life, learn to be able to think those kinds of things. That's a rather different objective for education than this. Learn specific skills and so on. Now sometimes these get a bit conflated because in the way the world works, there's an awful lot of, you know, even if the, the thing you're going to do is something that involves broader thinking, it's like, do you have the credential to do that? And that becomes a thing like did you pass the calculus test? And you know you're going to be a doctor where you're never going to use calculus, but you're going to pass the, although perhaps you should in some cases, but different story, but you're going to pass the calculus test just to sort of show you can type thing. And I think that lots of education as it is in practice happening in the world today is for this kind of credentialing purpose, not for the sake of learn to lead the more considered life and so on. So I think in terms of that kind of goal based education, I think we'll see how it works out. My guess is AIs can help a lot. My guess is that that's a place where kind of the AI can learn enough about us to realize you should tell this person this, this is why they're confused. And you know, a good teacher might know the 25 reasons ways people get confused about X. And it's like, well, you've got confusion number 23, let me tell you how to solve that. I think the AI has a good chance to be able to do that and do that efficiently and pleasantly and in a way that's going to be infinitely patient with respect to, you know, you've got that confusion today, tomorrow you've got a different confusion. I'm just an AI, I'm hanging out here, I'm going to help you. I'm not going to get frustrated that. Can you really, do you really still get, are you really still confused about that point kind of thing? So I think that's going to be very helpful in terms of the kind of, you know, the kind of knowing yourself sort of unlocking of things through education. I'm not sure it's quite so helpful in that case. And in that case I have to say I think that doing unique things that are sort of oriented towards you, like projects you care about. I happen to think, and maybe it's just my personality and my approach to things that that's a really valuable way to kind of enhance oneself and be educated. It's something that is not quite so easy to run in the kind of industrial education thing that we, that we inevitably have just from the number of kids being educated and so on. And I think maybe there'll be ways, I don't know, that's, that's a sort of a higher level of sophistication. Maybe there'll be ways to do more project based education and so on through sort of AI technology. I will say one thing about sort of the AI tutoring effort is that if you are learning math or something and you are very Keen on butterflies, let's say, then it is certainly absolutely possible for the AI to state every problem in terms of butterflies. For example. Or if you're a. I have to say, well, if you're a person who's like, I want to think about money, but you're doing arithmetic and it's like 10% of this and that and the other and disembodied, you're like, I couldn't care less about 10% of this or that or the other. But if it's like, you know, you're buying a thing for a million dollars and somebody gives you a 10% discount, how much does it now cost? You're like, oh, I get it, I care, I'm engaged type thing. And I think that those are things that, that's, that's kind of low hanging fruit. I think for the AIs to have that veneer of engagement, how well that will work for people, I don't entirely know. I mean, it could be that the first three times you get the butterfly oriented problem that you think is really cool, and the fourth time it's like grown. Just tell me what the main point was. I don't want to hear about the butterflies anymore. I'm not sure. I think it's going to be a question of kind of the artistry of knitting the AI into this kind of curriculum type of thing. Okay, so there's a, there's a question from Fred here. You guys are asking me all kinds of stuff about myself. Can I ski? Answer? No. No, I, I cannot. I've, I've tried a couple of times to, to learn. Although I have to say the time I most remember trying to learn was in 1980, which is, I admit, rather a long time ago. It's, it's just one of those things I've, I've, I've never learned to do. And perhaps I've never lived in a place where that would be a notable thing to want to do. I mean, I think it's, there are all these things where sort of sometimes it's interesting to learn something for the sake of the process of learning it. And like years and years ago I learned to fly small airplanes and I haven't done it in years and years and years. But the process and the kind of, you know, oh, there's checklists and this is how you think about this and this is how you do that. That was sort of interesting. Even independent of the dynamics of, you know, press the right rudder to, you know, when you tip the nose up or whatever. Else it is, and I think so. Yeah. So, I mean, there are, there are all sorts of skills that I have to say I've never learned. And I tend to have mostly only learned things where I sort of have a pragmatic reason to learn them. I thought I did with flying planes, but it turned out not to be terribly useful. Unless you do it very, very seriously. You just don't have the level of sort of continuous skill necessary to do it as a practical thing. But I think the, you know, in other areas, I've not been a big one for learning these kind of skills for the sake of learning skills. I have to say that's probably been true of most of the things I've ever learned. I learn because I need to know them for some reason or another, rather than because I want to learn them for the sake of learning them. Let's see. Philippe asks, can you picture AI creating an alternative legal system? Oh boy, this is a complicated story. So, you know, contracts are an attempt to sort of make precise statements people make about what they want to have happen. In a sense, code programs are in a more precise way to say what you want to have happen. The only problem is that programs usually don't talk about the kinds of things that the legal system might talk about about things happening in the real world. They talk about the value of this variable, this array, this whatever it is in this computer. Now, the place where these come together is through computational language, through our Wolfram language efforts, where we try to have a language which can describe things in the world. And so that's a place where you can say, you know, if it rains today, send this Bitcoin to this person type thing. And that's a thing that's sort of all within the language to define. And the. So that kind of raises the idea of kind of computational contracts. And we've done a fair bit of work on these, of taking what would otherwise be sort of in legalese and making it be purely computationally described. I think that has. My feeling is there are many situations in which computational contracts that are automatically executed are a good thing for sort of the big ticket things, where things can go wrong. And there's kind of computational irreducibility about sort of things you can never foresee that's going to end up being a thing where humans and sort of human ethics and so on are going to play into what happens. But when it comes to sort of the small scale, kind of micro level of these things, I think computational contracts are certainly a thing. And there are certainly cases where we already see that happening. There are cases whether it's options, pricing options in financial markets, whether it's some kinds of service level agreements, these kinds of things where the contract is effectively a computational contract. But my guess is that that can be done much more broadly. And I think the picture that you end up having then is there's all these intersecting computational contracts. And it's kind of one of these things where when there are computational contracts, more lawyers will be writing code than writing kind of legalese. And there will be kind of, when you have computational contracts, there are all kinds of things you can do. Like you can readily simulate what's going to happen in the world in certain circumstances. Like I've sometimes made the claim, I don't know, I haven't really followed all the I's and dotted all the I's and crossed the t's on it. But I claim that, you know, if mortgages had been computational contract, the more contracts, the mortgage crisis might not have happened. Because, you know, a mortgage might be a two, I don't know how long they are, two, three page documents, something like this. It's a bunch of words and it's like, what's really going to happen if this happens in the market and you know, what are the levers that are going to be pulled in these contracts that are these mortgages? Well, if those things have been computational contracts, one would have been able to do a simulation of if this happens to interest rates and that happens to this, what's going to happen to all these contracts? And that's something one can do with things like options contracts because those are effectively they've been made into what amount to computational contracts. So one of the things, once you have computational contracts, you get to be able to simulate sort of what happens if the world, if this or that happens in the world. Here's how these contracts will play out. My guess is that in the sort of future of these things, there'll be much, many, many more contracts that will be mostly computational, mostly sort of executed at the layer of the AIs and not things that any human is ever going to notice. And they're just, oh, those are things are happening. This computer is communicating with that one. It's running this contract, it's agreeing to that. It's paying some penalty to this other computer because this happened and that happened. It's, you know, there's a, there's a lot of stuff going on that's just happening sort of at the AI level and not kind of exposed to the humans. So I think the, I think that's. That, that's kind of a, a picture there now in, in terms of. So, you know, I think that sort of infrastructure will exist. There are questions about how do the AIs relate to the legal system? At what point do AIs autonomously have rights? At what point can they be responsible for things? These are complicated questions. I've thought about them a fair amount, I have to say. I think there's a lot more to figure out that sort of a political philosophy story. And I think the, this question. Well, these questions have been sort of a pivotal question is at what point is the AI a thing unto itself? Where kind of if, if everybody thinks it's bad, it gets switched off or whatever else, but it isn't something where it's like. But the owner of the thing is so and so and is responsible for it. So lots of, lots of things happening there. I think the, this question of. I can imagine a situation where it's like, okay, both parties agree and AI will be. Sometimes you might have an arbitration clause in a contract, you might have an AI arbitration clause in a contract. And sometimes that might be totally horrifying, Sometimes an arbitration clause might be totally horrifying in what actually happens. You might say the arbitrator was crazy and you might say the AIs were crazy too, but you're signing up for something different. And probably the price point of the AI arbitrator might be different from the price point of the human one and so on. And that's a thing where when both parties agree to some contract and have it sort of adjudicated by AIs, that becomes sort of an interesting different layer, so to speak. Let's see. EI says, I'm wondering how you imagine your symbolic language might interface with AI. Well, yeah, there's a. Yes, this is a. I mean, you know, we built back, oh gosh, nearly a year ago now. We built a integration into ChatGPT of Wolfram technology Stack both Wolfram Alpha and Wolfram language. And yes, we are kind of if an AI something like an LLM that is oriented towards writing text, creating things, sort of trying to statistically figure out what to do next, so to speak. If it wants to sort of retrieve precise knowledge, compute something, I think we have the best, perhaps even the unique kind of technology Stack that makes that possible. And having the AI be able to call on us when it should is something that's worked fairly well. It's different. Particular LLMs are better and worse at tool calling than others. But, you know, we've developed, and some of the folks who worked on LLMs have developed, I think, pretty good apparatus for tool calling and calling when you need it. Now, the dynamics, I think, are one of the great things about Wolfram language is that it's something that can be read by humans as well as something that can be processed by computers. And it provides kind of this computational notation for the things you're talking about. And so I think the really good workflow is you talk to the LLM, it has this linguistic interface, it produces some kind of Wolfram box or something which says, here's the actual Wolfram language code that was generated by the LLM and here's what it did. And you look at that code and you can say, yeah, that's what I intended, or no, the LLM got it wrong and I need to change that and make it something different. But it's a way of taking what sort of. You kind of factor out the imprecision of what's happening with LLMs and their kind of hallucinations and all those kinds of things, and say there's this hard piece of actual computational representation of things that a human can read and that you can kind of confirm is doing the right thing. And by the way, you can take the contents of that sort of box and then you can use it. For example, you could build it into some standard program that then you build many other layers on top of and it becomes this kind of hard block that you can make use of. Whereas in something like an LLM, much like something like a search engine, you don't take the results of a search engine and directly build on them. You read some document, you conclude something from it, you retrieve something, but you're not expecting that those sort of AI ish search results are what you directly build on it. And same with LLMs. But what the LLM can surface for you is something computational and precise that you can then build on. And I think that's a really powerful direction, which is sort of just getting worked out. And I think the sort of. The LLM marketplace is still a bit confused and, you know, where's the value? And so on. I mean, I have to say it's always nice because, well, it has been nice last year because people keep on saying to me, you know, these LLMs are going to be tons of them. But the thing you have of this sort of computation knowledge system is unique, which it is, and it's something everybody's going to need and it's sort of, that's a great position for us to be in, and we're just sort of how the plumbing works and how all these things connect together is still a little bit chaotic, but I think that's something that's going to be an evolving thing. And I think. I didn't build a computational language for AIs to use. I built it for humans to use. I built it for. To have a clear representation of computational ideas. Turns out the AI's need that at least as much as the humans, perhaps more than the humans. And so it's been sort of arriving at the right place at the right time. And that's very, very nice. But the things one had built end up being useful for this purpose. And in a sense, maybe AIs will even be in some sense better users because, for example, humans, it's like, oh, I can't be bothered to read the documentation. Well, you just tell the AI, you know, when it does something wrong, you say, just read the documentation. It'll go off and read it and actually be able to ingest it and remember it and so on. And the AI is a good student, so to speak. The AI might be a better student than the human user. And so that, that's a, that's a good situation, particularly in a system like ours where our computational language is very big and very few people, probably even including myself, at this point, know every corner of it well. And, and yet the AIs can do a better job of doing that. So that's, that's a, that's a nice thing. Let's see. Few more questions here. It's comment from Max saying. Commenting about my mother saying she was a philosopher and anthropologist, which is true. Did she have any impact on me? Well, the fact that I exist is a consequence of her. So at that level, absolutely. I would say that the one thing that's sort of funny is that when I was a kid, my mother was a philosopher and philosophy professor in Oxford. And at the time, that was sort of, that was the, the world center of philosophy, I suppose. And so, you know, when I read books about philosophy today, as I, as I try to learn more about it, it's kind of weird because there are all these philosophers from the 1960s,'50s,'70s type time frame and all these names I know, and I don't know them because of the philosophy they did. I know them because, like, I knew their kids or I knew them because my mother knew them or whatever else. It's a very sort of, It's a strange thing that it's like, oh, that's what that person did. That was a very weird character. Now I know what they actually did for a living and it was interesting or whatever else. So when I was a kid I always used to say, oh, if there's one thing I'll never do when I'm grown up, it's philosophy. And my arguments were mostly things like, look, any subject where people are still arguing about the same questions they were arguing about 2000 years ago isn't making progress and can't be interesting. I don't think I was right about that. That was the 10 year old me being a bit too judgmental about things. I think that at the time it was always perhaps amusing that both with my mother and friends of hers and things like that, it would be like, I suppose I was completely used to having these discussions with people that were what I would now realize were quite philosophical thinking, sophisticated. But for me they were just kind of like the, the typical, you run into these people and they'll ask you, you know, some question and you know, you'll end up with. I, I think to some extent I was, I was viewed as being amusing for these philosophers to talk to because I think, you know, the, the 10 year old me or something could give them a reasonable run for their money with arguments that were a little different from the ones that they heard from the typical undergraduate or something. I think. I don't know, I'm always a little embarrassed remembering a few of those conversations. I remember one in particular with a philosopher of time where I was just, you know, you know, aggressively pursuing various things from physics, which I now realize this philosopher was actually a little bit right about kind of. Well, you can't really quite know that the way you thought you could know it. And so. But I only realized that decades later. But I think in, in this, you know, growing up in an environment around a bunch of philosophers certainly does, you know, the concept of thinking is something one can do is something that probably rubs off on one, so to speak. And the kind of thing where sort of it's not the fact that, you know, when you might read, you know, studies have shown that blah, blah, blah, blah, blah, and you know, to the philosopher mother or whatever or the, you know, the random philosopher one runs into, it's like, how could a study possibly have shown that? How could you possibly know that? So to speak. And that was probably a useful piece of early sort of meta education to realize that that was a thing. And also the kind of the Skepticism, I suppose, about delivered facts and so on is probably something that is interesting to be exposed to when one's a kid. I think I, you know, I have to say perhaps one of the things, because I had this I'm allergic to philosophy thing, I never really learned the kind of, oh, this is what Kant said, this is what Hegel said, this is what, whatever. I've had to learn that in much more recent times. And it's, it's, you know, it's sort of a pity I didn't learn that way back in the day. But actually the, the structure of education, particularly when I was growing up in England, was very specialized very quickly. I don't think I even would have had the opportunity to learn that even had I wanted to learn it through sort of the standard educational process. Well, let's see. You know, nightmare is commenting it will be cool to live in the future where you talk to your AI in your house and it does things for you. I mean, that's supposed to work now, right? With all kinds of, you know, Alexa and this and that and the other supposed to work now. And it does a little bit. I think home automation is one of these areas that sort of was supposed to be a big, big thing forever. And I remember, you know, we built a house what, 22 years ago. I think now that where we were like, you know, how do we, how do we future proof it for home automation? And the main conclusion was just run empty conduit in the walls because you don't even know what kind of weird Cat 5 Cat 6, whatever fiber optic cable you're going to want to run through there, just run empty conduit. And that turned out to be something of a win. Although everything went wireless and one didn't need that anyway for most things at least. But this idea, I mean all the systems that we got that were kind of home automation systems of that time are kind of, you know, sort of horrifyingly antiquated at this point. But I think also there was a, you know, the kinds of things that home automation can readily do just turned out to be not as interesting as people might have thought. I mean, you know, set all these lights to be in a particular configuration. Okay, fine. You know, set this, you know, notice when my, notice when this or that thing is happening. Notice when my, when my cat did this or that. You know, that's a little bit of that that's happened. I would say that the kind of everything is programmable sort of. It just turns out to be difficult to Wire in programmability to every light, every this, every that. It's just an awful lot of, oh, an electrician has to come and install that type of thing. And you know, construction is complicated and it, it. I don't think it's quite as easy to sort of retrofit kind of. Let's just layer automation on top of that. I mean, one's used to with computers. If you say, I want this thing that's a, you know, a nice plugin for my browser that does this or that, you just install the software and you're done. If you want to install that new lock on your door or something, somebody's got to take your door apart and put the, you know, put the lock in and wire it up and do all these kinds of things. And I think that has been sort of a slowdown to the home automation world. I think it's also the case that much more will eventually be programmable than is. I mean, there comes a point where you might as well make something programmable. I mean, there are plenty of things from, you know, I don't know, firewalls to, I don't know, ovens to whatever else where sort of you might as well, it's going to have a computer in it anyway, so you might as well make it programmable. And then the only question is, do you know how to program it? And you know, I think that. And there'll also come a time when displays are so easy that, you know, every wall can be covered with a display, you know, that's not quite there yet for that. But, you know, we're getting closer to that. And at that point it's like, well, every wall is covered with a display. But most of the time this quotes display is just going to look like wallpaper. That was something I could have just got out of a roll of paper and so on. But it's as cheap to have it be a display where in principle one day it flashes the giant warning sign or something. But for 999 days out of a thousand, it's just wallpaper, so to speak. It's a question of economies of scale, of building in programmability to everything. And I think we're not quite there where that makes sense. Maybe when that does make sense, sort of when everything is programmable and routinely. So then this sort of AI control of things in one's house becomes more of a routine thing and less of a sort of special case thing. I mean, I know, for example, in my house I tried to figure out the ux of things like light switches and pan light panels and so on. And yes, my wife is fond of pointing out lots of, lots of UX ideas like the double click, doing this and that. And it's like, well, it kind of works, but it doesn't quite work and it's kind of stupid. And I think that's a, it's, it's, you know, the things which are, you know, sometimes with these light switch things, it's like, well, you can double click it or you can just have a frigging light switch and you switch it off and it's a pretty good user interface to just click it's off. And you know, perhaps that's, you don't really. The fanciness doesn't really get you anything. All right, maybe, let's see. LC Comments when talking to chat gbt, I ask politely, as I would a smart human and say thank you way more than I should. You know, I had this policy, I've had this policy. It amused my children when they were younger that whenever I run into some sort of pseudo automated thing like you, you're paying for some, you know, a parking garage or you're, you're, you know, paying for something with a credit card and it says something to you, saying thank you. And I like say thank you to it. I kind of a. Kind of, I thought it was a funny habit. And one might as well kind of develop the habit of saying thank you to your AIs. I had an experience. I have a friend who just made a kind of interesting humanoid robot which you just saw a few days ago that's pretty good and has sort of powered by a rather impressive LLM and has rather impressive eye tracking and things like this. And the thing that was just shocking to me is how quickly you start sort of talking to it like you would talk to a human. And in fact, we had it for crazy reasons. We happened to have the thing in a restaurant and all kinds of people were coming up to it and talking to it and so on and asking about it. And it is interesting the extent to which people treat it politely. I mean, it's just an AI and you could just walk away from it, but people don't do that. They say, I got to go now, bye. And that's something we naturally do. I mean, we are built to be social creatures, most of us, at least, and to try and do these things that kind of sort of help lubricate the social experience. And so when we're confronted with something which is that sort of, which Even has those elements of humanity, we tend to quickly fall into this situation of being sort of empathetic with it. And I think that's a thing, you know, I don't know. What should one think about that? I think it's. I think we're used to doing that. And honestly, if it would be perhaps a bad thing if a lot of, you know, let's say you've got. Most of the time you're interacting with an AI, but just occasionally there's a human. You know, most of the time it's an AI that's delivering your food to you or whatever, but just occasionally it's a human. Well, if you get used to kind of being a jerk with the AI, then when the human shows up, you're going to do the same thing. You're going to kind of revert to sort of this kind of pre socialized experience if one gets to the point where one's being rude to the AIs just because they're AIs. So I think it's probably the case that sort of taking one's kind of common human politeness, so to speak, and extending it to AI is probably not a bad thing because otherwise you get to the point where sort of there's much less politeness in the world and that then becomes. That then becomes unpleasant for when there are actual humans in the loop. Okay, I should maybe. One last thing here. Okay. Actually, two maybe. Marie says, do you think there's harm in always learning? I think humans aren't built to be putting so much demand for energy on our brains. But it feels good to learn. They say, I don't know. I think we're set up to continue learning. I think that's one of our features as a species. And you know, we are. And I think it's something humans in modern society, maybe even true in the past too, but tend to forget. They tend to think there's this period of time when we're learning and maybe it's easier to learn when you're a kid than when you're adult. I'm not sure that that's true. I think the things I've learned as an adult, I would say if anything, they were a little bit easier to learn than when I was a kid. I'm much more used to the way my brain functions. I know a lot more stuff that I can use as an anchor for new things that I learn. And I kind of understand my motivation structure much better. And I know what's worth me learning and how I should Learn it and so on. But I think this idea that people have that you learn when you're a kid and then you get your last degree or whatever, and then that's what you kind of keep using for the rest of your life. I don't think for me at least, that's the most interestingly lived life. And I think that. And as the world changes, also one has a certain habit of being left behind if one just learnt up to some moment and then everything was frozen at that point. And so I think one of the things that, you know, keeping on learning is something that is really worthwhile, I think. And people sometimes just say, oh, no, no, I shouldn't put all the effort in to go on learning. But actually it's really worthwhile. And by the time you have a habit that you're going to continually learn, then it becomes easier. People who say, I finished learning when I was 22 years old, I'm done, no more learning. And then in their 40s, they suddenly say, oh, my gosh, the industry I was in just disappeared. I got to learn a new thing. Oh my gosh, oh my gosh, this is terrible. But if you just have the habit of keeping on learning it, I think it puts you in a better position. All right, last question here. I think from crucial. Would you want yourself. Myself to be automated? How far would you want yourself automated? Good and pertinent question. I mean, I think, you know, I'm a person who's recorded lots of stuff and personal analytics and so on for the last, I don't know, 30 years or more. And I'm a person who's done lots of these live streams and so on, sort of recording things that I say about stuff. Person who's recorded all the things I've typed, all the keystrokes I've typed for a couple of decades. So there's a lot of information that I've left out there about sort of output from me. And, you know, somebody recently trained a bot of me that was. Seemed like it was pretty decent from. I think I've left maybe 50 million words of publicly available kind of material out there, which is quite a lot. I mean, I suppose there's probably maybe, what, a hundred billion words or something that are sort of generally out there. But 50 million is not a totally shabby number relative to that sort of total of the 8 billion folks of the Earth producing things. So there's, you know, so there's a lot of material from me out there from which you can train an automated me now. What do you do with an automated me? I don't know if, if there isn't, you know, if I'm not around, maybe the automated me is, is something fun to have. But you know, realistically in there's a question, what can one actually do with it? So here's an example. I mean, I've been trying to think about this, this, you know, I get all sorts of email that I get sent about all sorts of topics. And the, the thing that, you know, I've, I've wondered about is should I send back an email, should I get my bot to respond to all those emails with the statement, you know, this is Steven's bot responding to you? Here's a few things based on your email that you might find useful. If this is helpful and answers your question, then we're done. If it isn't helpful, then tell us it was not helpful and a human will get to you in due time. But it may take a while. And I think that may be a good solution because it's like you want to get an answer quickly. The bot gave you an answer. If you want to get the human engaged answer, it might take two months or three months or a year. But you know, there's a different dynamic of what's needed there. So, you know, that's a use case for it. I think the, you know, start writing my next blog post, have my bot write my next blog post, wouldn't be much fun for me. And I don't think it's going to happen. I don't think that's the best use case for it. Is it the case that I would find is an interesting question, actually. Actually here's a use case for this which is, you know, I, I make great use of the kind of all of the archives that I've stored. I use them many times every day. Things like, you know, when did this happen? What happened? What did I remember about this? What, what? You know, I, I use these archives a lot. I search them just with straight textual search and use it many times a day. But here's a question. If I'm writing something and a bot could tell me, wait a minute, you've written about this before, this general idea before. That will be useful to me even at the very least in saying, like, make a link to this thing that I wrote before. Because sometimes there's a, I could just do a keyword search, but a lot of times I'm writing about ideas and this idea is similar to that idea, even though the words are a bit different. Is something that I don't necessarily can't necessarily pick up. And I think that that's, you know, that's a thing where I can see a bot of myself being a useful thing for me as sort of another layer of enhanced memory beyond the purely textual kind of enhanced memory. And actually I should hook that up. I think I actually was. I suggested that a little while back in terms of being able to find the right links to put into posts I write and things like this. Anyway, we should wrap up here and proud comments. Just joining these live streams has taken my learning a long way. Thank you. I'm glad to hear that. It helps me learn things as you guys ask all these questions and make me think about things that I that I haven't thought about at least with definiteness before. It's super helpful in terms of me figuring things out. So thank you all for that. Well, thank you for tuning in here and look forward to being with you again another time. So bye for now.