Business, Innovation, and Managing Life (January 31, 2024)

Sam. All right, let me see. Okay. Hello, everyone. Well, let me just see. My setup here is not quite working. Right. Hold on one second here. All right, well, let's see. Welcome to another episode of Q and A about business, innovation and managing life. Talking of which, I need to just restart one thing here. Oh, I see. It's very odd. Just a second. Hold on. Oh, that's not good. My source of questions is severely depleted because the system I use for that is not working. Um, let's see. All right, I have another. Another source for these. Great. Okay, let's see. All kinds of questions here. So a question from Jamie here. This one is easy. Have you had a chance to try Apple's new VR headset? Answer is, ask me again in a week. We'll see what I think of it then. You know, it's. It's a fascinating thing. The. The life and times of VR. I first used a VR system probably around 1989, and I had been kind of every so often I've kind of thought, this is going to solve some big problem for me. In particular studying kind of these models for fundamental physics where one's dealing with these hypergraphs and big hypergraphs and so on, and trying to understand, Trying to get an intuitive sense of what's going on in this giant sort of spider web with perhaps 10 to the 400 nodes in it. Obviously you can't see that many, but a giant kind of spider web network. I kind of remember thinking around 1991, when I was first really thinking about these kind of network models of space time. I was thinking, obviously, the way I'm going to visualize this is using virtual reality. Well, we're now 33 years later. It still didn't happen yet. And I have to say we've done some experiments on this, and it wasn't terribly useful to me. Now, perhaps that's because I've grown up in a world of sort of ordinary objects, and my visual system and cognitive processes are set up to deal with kind of objects as they usually are. And if I even give myself in virtual reality something that is deeply kind of not our usual experience of physics, not our usual human experience, it's really hard for me to parse that. And so I think that's one of the. So that's one use case that I always imagined. I have to say, another use case that I imagined in the mid-1990s was I like to kind of walk outside and do work while I'm doing that. And I've had all kinds of schemes for this, where it involves, you know, the popcorn selling, like things strapped to oneself. One has a computer one's typing on or most of the time it's like I'll just do meetings where I just have to look at a screen and I don't need to type anything because that's a lot easier. But I always thought I had this idea in the mid-1990s. I thought I'm going to use some, you know, AR kind of headset type thing and I'm going to have some kind of, I imagined at the time some kind of chord keyboard that I could just, you know, sort of have in my pocket more or less and just type on by, by, you know, deciding which fingers to press at the same time type thing. And that's what I sort of imagined would happen. Okay, we're now, you know, 30 years later, more or less, and that didn't happen yet. Maybe it will happen now. I have to say, I think the more likely use case for somebody like me is just lots of screens. I mean, as it is in my normal working mode, I just use two big screens. And it's always a challenge for me because if I'm kind of trying to do a video conference, I don't usually like video conferencing except for purposes where I'm trying to, to talk to other folk. I don't find it useful as a kind of everyday working mode. I find screen sharing super useful, but I don't find the kind of watching other people as we talk about things so useful. Sometimes useful at the beginning of a meeting to kind of switch on video and have everybody know, yes, it's actual humans there. And then when you actually get down to work and you're screen sharing and so on anyway, the humans are going to be little boxes on the side of the screen on. I always find it sort of more tiring to be in a meeting where there's, you know, I'm also having to pay attention to the little box on the side of the screen in addition to just thinking about the main topic of the meeting and maybe looking at something that is sort of being shared on the screen. But so it's, but I, I. So my usual configuration is just two big screens. I usually have, I have a tertiary screen which I use which I can kind of put up because I have my camera sticking in between the two screens so that it's more or less looking straight on at me. But if I'm using those two big screens, then when I'm using those screens, even though the camera is straight on at me. I'm looking one side and the other side to use those big screens. And so when I'm really doing some kind of video conference thing sitting at my usual desk, I usually use this third, third screen which is basically just, it's actually just a little. It used to be an iPad, but it's now kind of just a, a separate little sort of iPad without the, without the brains type thing display was surprising how long it took before one could buy those things. But be that as it may, you can. So I just sort of usually I have that flat on the table and if I'm really doing some video conferencing I'll put that up and it's right by my, my camera and that works well. But so my. What I imagine we'll see is that one of the things that will be the case with kind of this new generation of VR is the resolution is good enough that it's like I could imagine I've got five screens out there and I can just put the thing on and it's like even if I'm kind of somewhere where I absolutely couldn't have my big screen set up, I can put the headset on and I can be as if working on all these big screens. And that may be sort of as a display maybe the main use case. I think this whole question of can you visualize amazing things with it so far, last 35 years, not a big success in that regard. But we'll see. The future might be different. I think another case, I mean it'll be interesting to see what really evolves with video conferencing. When you have all this elaborate. It's on your face but it's tracking your eyes and it's being able to reconstruct somebody else's face. That may be just totally weird or it may be sort of a fantastic way to have a personal type a person, person to person type conversation. We'll, we'll have to see. I think the other, let's see. I mean the, the other type of. Well yeah, so this, this question of, of sort of as we got mobile for example, there were and the very idea of, you know, ubiquitous touchscreens and so on and the fact that you could sort of always be picking up your computer and pointing the camera that's in your quotes computer really a phone anywhere that led to kind of different user interface modalities than one had on, on a straight computer. And I think the same is likely to be true for, for well executed VR And I think the other thing that will be interesting is this whole question of, in the scene that you're looking at, see what human eyes cannot see. So, you know, can you sort of have information in the infrared, in the ultraviolet? Can you have more detailed spectral information which will give you information about what kind of material something is? That's the type of thing you don't just get by seeing red, green, blue. Yes, you can tell it's green, but you could tell in principle, it's got this distribution of frequencies of light which corresponds to a kind of a bad piece of fruit or a good piece of fruit or whatever else it is, which would all just look green to us, for example, without ordinary vision. But if you use hyperspectral imaging and you pick up more frequencies, you can get more information that way. The challenge is what? Well, we're just stuck with our visual system with just our red, green, blue color sensors, and how do you display that additional information in augmented reality? And that's something which I think we'll see a bunch of kind of probably innovative solutions. I've thought about this at times over the last couple of decades and have various ideas, and we'll see what can be actually implemented once one sees how well the hardware really works. Let's see, Questions here. Hive asks talked a little bit about AI and the future of science previously. What about the role of AI in future of business, innovation, managing life? Well, you know, I think the. Those are all separate things. I mean, in terms of. Of sort of what one can do with AI, I think it's sort of, in a sense, it's extending and repeating human judgments. Humans judge something a particular way. Okay, let's extrapolate from that. Similar things. How would we judge them? I think that it's also a question of. For us humans, we have a certain memory, human memory of, you know, we can remember what happened and so on, but. And we can get further by just doing textual searches. But potentially with kind of LLM type AI, we also get to do a more effective kind of type of, sort of semantic search for. Well, when did I last have a situation where I had an issue with an employee of this type of. I don't, you know, I won't know the name. It's not like I'm going to search for this particular name unless I remember the specific name, but it's like search for the idea type thing. And I think that's a thing that we can expect. There is reasonable success coming, even in the short term with LLMs to do those kinds of things now in a sense to have the raw material to deal with that, you have to have a good supply of remembered things. Somebody like me has been storing all their email, all their keystrokes, lots of other things for three decades or so. So I have a good supply of kind of raw material to make use of sort of enhanced memory on. And I've been doing that for years with pure text searching and perhaps one will be able to do that more effectively with some kind of sort of AI powered semantic search. That's been a thing that really hasn't worked in the past. I think it might finally start working and, and it's really a question of, you know, remind me of a case where something like this happened. One could imagine a situation where one's in, I don't know, some commercial situation with sales or something like this and one is, you know, comparing what is being said in the current conversation with sort of previous conversations that are kind of like this one and oh, let me remember, I should have said this type thing. So that's the kind of thing I can imagine being a sort of a good coming attraction. I think in terms of managing life, it's kind of. There's sort of very minor versions of AI which we've mostly got, which I have to say I don't really use much of. I mean there's very, very minor things like, well, they're not minor when they first came in, they're a big deal of speech to text, auto completion of texts, all this kind of thing and sort of a reminder for what you should put on your calendar, those kinds of things. I think this whole question about auto suggest for life is an interesting one. I don't know how well it will work. I mean it's kind of like you've got a GPS which tells you, given where I want to get to, it will tell you a good route to get there and it might be intelligent about traffic or whatever else. If one could define one's objective in some sort of life objective, one could imagine some sort of GPS that says do this, do this, do this to get there. I'm not sure that most things are sort of able to be broken down to the point where one can really do that. I certainly have imagined the kind of what can you delegate from us humans? It's like, should I eat this, should I not eat this? Well, you've eaten X number of calories today and you've got this and that and the other. I wouldn't eat that. That you're not really hungry, are you? Type thing. You know, I could imagine something like that being a thing. We're not quite there yet because of the technicalities of being able to do image to food information type thing we actually happen to have for better or worse through our WolframAlpha efforts, we now have I think the world's best database of nutrition information. But we still that last link of going to the thing you see on the plate to nutrition information, that's been a remaining challenge for machine vision and so on. But I think in. Well, let's see other kinds of things that about sort of. I suppose I can imagine all kinds of. I mean I, I've done a fair amount of personal analytics on myself knowing sort of how I sort of storing all this information about myself. I have to say one thing I haven't done in the post LLM world and you guys give me the idea of doing it, is ask the question. Like if I look at all the emails that I have, categorize them in a way that based on their content, an LLM should easily be able to do that. What fraction are this, this, this and this? You know, break it into topics in that way, is it going to be useful? I don't know. Is it the type of thing where, you know, I certainly have learned in the past looking at sort of the flow of email that the sort of optimal times of day, like for example, if I were to get up really early in the morning, I would have almost no email that was yet come in. And it's much better to wait a bunch of hours until there's a kind of collection of email and have this one time when I can deal with that collection, then wait, there's another time I can deal with another collection and so on. And because for me, for example, I find it much more efficient to say, okay, I'm going to grind through 100 emails in one big block, then, oh, there's another email that just came in. Oh, there's another email that just came in getting interrupted in the things that I'm doing. For me it's much more efficient to batch these things up. But there are certain times of day where it's going to make sense to do that and not in terms of sort of the innovation side of things and AI. I mean I think it's very easy for AI to be creative, just enumerate different possibilities and pick at random different kinds of things. Most of them won't turn out to be sort of human useful in a sense. What the, you know, the story of lots of generative AI is taking what would otherwise be sort of pure randomness, but sort of adapting it to the distribution of things that we humans find useful, whether that's text or images or whatever else. I think this question of so now invent something for me. My guess is that sort of 95% of that problem is defining what it is that you even might imagine you might want to invent. Rather than let me brainstorm and try and have a creative idea, I think it's quite easy and there are quite a lot of ways to sort of, let me just sort of, you know, throw a dice, throw a die and sort of try this or that or the other. I'm not sure you need a fancy AI to do that. It's a question of can you define what it is you're trying to go for? Then you have to try and figure out a sort of creative solution to those things. Now, you know, so I'm less clear on that. I think that these things about so sort of remembering what's happened before because after all, all of AI that we see today, LLMs, generative AI, image systems, all those kinds of things, why do they work? They work because we have a large corpus of information that is sort of what we humans typically want in those domains. And the AI is able to extrapolate from those things. And I think that's the place where we can say, well, you know, the things, if you've got the personalized corpus of information or the corporate corpus of information in a business setting, those are things where one can expect that one might be able to sort of have some sort of useful advice being given by the machine. Now know, I have to say the other thing that always happens is it's, it's often easy to come up with, yeah, we should do this. It's often really hard to actually execute that. And I'm not sure how much AI in the short term will help in the, let's go execute that. Because it's something which involves people and persuading people to do things and so on. And so far that's not really been a forte of kind of AI and people's perception of AI. Let's see. Yes, Brady is commenting, if everybody had Vision Pro, then we could, we could have a, a great kind of virtual lecture. That would be really fun. I, I'd love to see that. Brady also asks, have you ever determined your Myers Briggs personality type? The other considered hiring based on personality tests? I think those are very coarse measures, I have to say. I think I probably in that classification what an ENTJ or something. But that's totally obvious from my. You talk to me for two minutes and you could determine that it doesn't need a sort of a fancy test and a categorization in terms of sort of hiring based on tests for most kinds of roles. I think that's sort of a somewhat doomed enterprise. There are some kinds of roles, particularly ones where one's dealing with skills that people don't even know they have. Where yes, a test is useful, like for example, in the past linguistic curation. That is, you're asked the question sort of in the end it's going to be sort of right. Little algorithms that allow you to say all the different ways you could say this or that thing. LLMs help with that a bit now. Not they don't solve the whole problem, but that's something where some people are really good at that and others are not. And it's not a skill anybody ever sort of manifested they didn't know they had it. And so that's a case where just saying, hey, try this. Can you do that well? Can you not do that well is something that is useful because people don't even know. Even if you ask them, are you good at doing this? They say, I really have no idea. I think in the. It really depends on probably the type of organization one has and the extent to which kind of it's, you know, can you tell how somebody would do in this or that role? Do you have to kind of get them to do a trial version of that role to see what will happen? That could often be useful. But a straight test that is like a, you know, fill out the multiple choice form in most of the kinds of things we would, for example, hire for. That's not terribly useful. The, you know, make us a sample of a piece of code that does this or a report that does that. That's more likely to be useful. It's more customized to the kinds of things we want. So I mean, I would say and in terms of, of personality types, you know, if you asked me what personality type would I want for such and such a role, many different personality types could be successful in that role. It kind of depends on how they approach it. And certainly in our particular company, maybe we even specialize in having a wide diversity of personality types kind of keeps life interesting. And actually it often, although sometimes it means that there's a certain amount of sort of kindergarten kind of personality conflict brokering that has to happen. I would say that that is a good trade off relative to just a diversity of views and types of people and approaches to things that you get from having very different personalities in an organization. Let's see. Atore asks what's good advice for starting to work on a super novel thing where there is almost no literature about it? I would say the fundamental question is what's the fundamental question? You know, given a thing you're working on, it's kind of like what is the core question? And that's the place to kind of concentrate. And when there's no literature about something, well, you have to figure it out, you have to build it for yourself. That's a. And you sort of have to have the confidence that you can do that. And I think it's in this sort of greenfield building of things I see that some people sort of have the confidence to do that and others do not. It really helps to have done things that were a little bit like that before because people say, how can I just invent this whole field? How can I invent all the names for all these different kinds of things? How can I decide what's important, what's not? Well, you kind of just have to have the confidence to do that. But a good thing in terms of doing it in the right direction is making sure, you know, what is the point, so to speak, what's the thing you're trying to address? And I would say that often it's useful to say, well, what are the analogies here and there? Maybe not so close to what one's actually doing. Maybe there's a meta analogy that isn't about the subject matter, but but is about the sort of state of knowledge in that area? What do you do if there's no knowledge in that area? What do you do if the last time somebody looked at this was 500 years ago and they looked at it in a very different way? And so people gave up since then. I think sometimes when there are things where people tried doing something, perhaps with old paradigms, old methodologies, they didn't get very far. It's sometimes not terribly useful to see what happened there, because to see what happened, unless you are very strong willed, you end up getting pulled into, well, they did it this way. Let me look at that way. Seriously. Well, it was just a bad idea. It was done in a different context than we have today. And you're better off just starting from what exists today and trying to build from that. I would say for me, and I think I would say it's relevant for everybody, sort of casting your ideas into computational language. Write the Wolfram language code that represents that thing. That's the way that you have something definite. It's. For me, I try to start doing that very early in a project because that's the way that's my kind of medium for thinking about things. Yes, I can describe it in words. Sometimes I have to just use natural language words. But the thing that really makes it kind of start to get leverage is when I can use computational language and when I can get my computer to help me kind of try and figure things out. Let's see. Let's see. Derwin asks, at this stage in your career, has your vision for your legacy reified? Fully? And if so, what do you imagine it being? Eliminating computational irreducibility within the RD and maximizing its positive impact on the world, for example? Well, that's a charming suggestion there. You know, I have to say I'm. I tend to. I think about big projects and big kind of trends and big things I want to want to do. I have to say I'm more interested in sort of perhaps in the future than in the. In the past of the future, so to speak. That is, it's more what do I want to have the satisfaction of doing rather than what do I want to be able to look back on later on or have other people look back on later on and say, that was a cool thing. I think it is something where in terms of the things I've tried to do, the sort of big theme is all about computation. I've been lucky enough to live at the time in history when that sort of became. Was a new theme and a theme where there was a lot of very juicy fruit to be picked over the course of many decades. I think that for me, there's sort of the conceptualization of things in science. There's a conceptualization of things in technology. For me, those have been closely related and they've kind of mingled over the years and decades and so on. And I think it's a funny thing because there are different kinds of things that one makes that have different utility in the world. I mean, there are different amounts of sort of one's ego that's in it, so to speak. I mean, there's things one writes, for example, where people are going to read the individual words one wrote now. So one feels like that's the thing where one is sort of directly presenting something that people can make use of. On the other hand, it's something where they read the words and the words float by and then that's done, so to speak. There are tools one builds where one has the. Where one's sort of exposing people to the ideas one has in the form of tools where they will get much more engaged with those tools and really make. If you had a good idea, you might have, you know, a thousand person years spent kind of making use of that idea in a way that is sort of more engaged than just read the words and move on. On the other hand, it's something where there's sort of more. If it's read the words, it's those particular words and you define them. By the time you get to tools, it's more like there's a bit more interpretation that's going on a bit more. Well, you provided those wonderful tools but people used them poorly, et cetera. And then the sort of third level is ideas, where you can define ideas and you sort of the interpretation of the ideas. There's so much that happens in the interpretation of the ideas that there may be more, in a sense there's more that can come from them. But yet one sort of egotistical contribution is perhaps diluter. I think in terms of. There are certain kinds of things where I've said I think this can be done at some point. I think one can understand how this works. There are somewhere I've not been sure, like things about the fundamental theory of physics where I'm very, very optimistic that we are very much on track. And it's really a question of just putting more work in to make more progress there. There are things about. I had long been interested in kind of general AI and what will be achieved there. And I have to say the way that things have evolved is not particularly how I had expected it to evolve. And I always thought there was sort of, there would be sort of foundational things about sort of as I would describe, have described it in the late 1970s, kind of fuzzy pattern matching. That would be sort of the key idea for sort of AI. And in a sense that's exactly what's happened. But the details of how it's happened weren't things that I could foresee. I think there are other areas where it's kind of like molecular scale computing, sort of the generalization of chemistry, making chemistry sort of compute all the way down is something I'm sure will happen. I'm sure that's doable. How it will happen and what it takes to get there, I'm not sure. And it's something that I've sort of long been interested in. Hope to be able to do. I would say that one of the things that happens is there are some things where I know this is the right direction. For example, the whole idea of computational language representing the world computationally, computation as a paradigm for thinking about the world. This is inexorably the way things both sort of the front line of things is going today and will be increasingly the thing in the future that will be just as people said, oh, math is an important thing. People will say, well, logic is an important thing. People will say, this computational way of thinking about things is important. And I don't mean writing low level programs. I mean kind of this conceptualization of the world computationally, which is what we tried to capture in our computational language. And I think this idea of computational language that we've been sort of uniquely pursuing for the last 40 years, that is the inexorable future. It's sort of frustrating how long it takes for the future to broadly arrive, but that is the inexorable future. For me, it's just like it's a sort of a straight line. That's how it's going to be. Then there are other things where it's like, this was a surprise. For example, the things to do with the fundamental theory of physics. While I thought in general terms things would work this or that way, the details and the extent to which things sort of came together more quickly than I expected, that was all a big surprise. For example, now, understanding this multicomputational paradigm, the fact that that is a way of thinking about lots of different fields that have previously been hard to science, ify so to speak, I think that's important. I mean, back in the 1980s, when I was first thinking about simple programs and the consequences of simple programs, and came up with ideas like computational irreducibility, that to me seemed to unlock a bunch of kind of areas that had been sort of blocked, sort of blocked by complexity, so to speak, and then realizing complexity was a thing, and that was the thing one should be talking about rather than the ways in which one had been blocked with previous paradigms. And so now I think there are a bunch of areas where this kind of multicomputational idea and the sort of ideas that come out of our physics project, where we're sort of using the formalism of the physics project but not applied to physics, I think that's going to unlock a bunch of other fields. I think, you know, economics is one, I think foundations of biology is another. I think things in neuroscience, things in linguistics, and so on things in evolutionary biology and so on, I think. And that's sort of an exciting thing which I didn't see coming. You know, if it had been like, do I think that this is going to kind of unlock, you know, I certainly had ideas about economics in the past based on things like computational irreducibility. But I think this kind of multicomputational idea has the potential to really unlock kind of a different form of scientification in these fields that have frankly had a hard time being scientific in the normal sense of science. And perhaps what one will see happen is they are scientific in a slightly different sense of science. Just as things where you see computational irreducibility and complexity, the kind of scientific statements you can make are a bit different than the statements you could make before. That said, I predict this number and it's such and such. Let's see. Jonathan asks, how have your experiences in education influenced your approach to innovation? Well, my own educational track was fairly, fairly unusual. I mean, I went to fine schools, but. And I, you know, did well studying the kinds of things one's supposed to study probably mostly because I have a decently good memory and that made it somewhat easier to memorize Latin regular verbs or something like this. Most of what I've learned I've taught myself. So in a sense that's my own personal education. That's what it's all been about. And from a pretty young age of 12, 13, something like that, my, maybe even younger than that in some ways. It was always a project that I would define for myself and then I would figure out how to do it. So I, I kind of very early, in fact, probably by the time I was like 8 years old, I was in the. I want to define what I want to do, then I want to go do it. And later on it was, it got more sort of academic and intellectual. What the projects, I, I guess they were always sort of intellectual, but they got more sort of high falutin in terms of content. Here's this physics problem. I want to go try and solve it. And what I would do to educate myself was learn the things I needed to know to solve that problem. So I think for me it wasn't something induced by education. It was something probably internal to me that I was like, I want to figure out. I'm interested in this, I'm curious about this. Let me go chase that. I want to make this, Let me go do that. So it was kind of that sort of internally generated, kind of, you know, I want to do this thing and then I will learn the things needed to be able to do that thing. So I suppose in a sense the thing that perhaps I've been doing kind of all my life is sort of make up the problem, so to speak, rather than here's a problem, now learn how to solve it. It's more been I've had vast amount of experience and sort of make up the problem. It comes in very handy, for example, when we do our summer school, summer camp programs and things and we'll have 150 students or something like this and it's like come up with original projects for all of those people. The fact that I've spent my life basically sort of coming up with projects that I think are worth doing is super useful when you have a week to, to invent 150 projects or something. But so that's been a thing and it gets better, it gets easier, sort of the more projects you've invented, the easier it gets to invent a project which can be done, which is correctly scoped, all those kinds of things, and which, where the results will turn out to be interesting as opposed to, well, this was a project, you could do this. But well, when you do it, it either isn't going to work at all or, or it's going to work in some very uninteresting way. So I think the. Perhaps my. I suppose one of the meta experiences that I had in education to be contrasted with many of the sort of kids going through sort of elite education today, is I don't think I was nearly as busy as the kids today are. I think that I was able to spend many hours per day pursuing the things that I was interested in, not the things that I was supposed to do for schoolwork or extracurricular activities or whatever else. And for me, sort of at a meta educational level, just having the freedom to do things one wants to do seems worthwhile. Now not everybody, if they have free time, will do things that are sort of creative things that they want to do. I mean, and perhaps if they're better that they are doing these structured activities that achieve something than that they do things which probably achieve very little, but who's to know? But I think having the freedom to be able to, if you really want to do something, make sure you actually have the time to do it, rather than, oh, I should do seven other extracurricular activities, but that means I won't be able to pursue my own interests and anytime soon, my point of view on that at least. Memes comments Einstein is claimed to have said, if I had an hour to solve a problem, I'd spend 55 minutes thinking about the problem and 5 minutes thinking about the solution. Yeah, I think that's. I mean, and by the way, the thing that makes that even more true is what we've tried to do with computational language and so on, is make it so that once you've defined what you want and you can think about it clearly enough and computationally enough, it's like, okay, computer, go do your work and grind through and see what the actual answer is. And that's something even more. With the sort of computational superpower there, that's even more a thing that becomes possible to say. The thing you need to be thinking about is what you want to be thinking about, so to speak, rather than how you solve the problem that you've defined. Let's see. Charlie asks, do any of your businesses have plans to build blockchain solutions? How does blockchain technology fit into your vision of the future? We actually have a business unit called Wilson Blockchain Labs that's been doing blockchain stuff for a long time. We kind of got into it because in defining things like smart contracts, people have been interested in, well, how do I know whether that contract should be triggered? Well, it may depend on something in the world you're writing, weather insurance. You need to know, did it actually rain that day? And in more generality, got interested in kind of computational contracts. Forget the blockchain side of it. Forget anchoring it to a blockchain. Just here's a computational definition of what would otherwise be written in legalese. And we kind of got into this because we are the main producer of what can be used as sort of an oracle to say what actually happened in the world. That's what our Wolfram knowledge base, Wolfram Language and Wolfram Alpha technology stacks provide is kind of a. Well, what actually happened? What was the sort of public knowledge about this thing? Was there an earthquake here? Did. Did the price of this go above that? All those kinds of things? And so that kind of got us in originally into the sort of the world of blockchain kinds of things, because people wanted to use our technology for that. And so we built out a bunch of capabilities for that. Then we got involved in things like blockchain analytics and so on. And our technology is extremely widely used in quantitative finance. And as methods from that started to sort of spill over into the cryptocurrency world, so we were sort of pulled into doing those kinds of things. So we've been kind of involved in that. We did build out very early, many years ago now, an ability in Wolfram language to just put things on. Well, both we have connectors to many of the sort of the top blockchains, Bitcoin, Ethereum, Cardano, Tezos, I think, but a bunch of others where you can, directly from within Wolfram language, you can both look at the blockchain and do transactions on those blockchains and do analysis of what's happening on those blockchains. And that's a thing where, yes, you can kind of throw your thing onto these blockchains. We also have, partly as a sort of test and partly because it's useful to some people, we have our own little blockchain setup where you can, you can put things onto a ledger, so to speak. And, you know, it's not as widely distributed as a ledger as a bitcoin is or something like that, but it is, it's. It's, you know, I'm not sure how widely distributed ledgers have to be to feel pretty secure that that ledger is not going to go away. And for some purposes, it seemed like that would be a thing that people were interested in, whether it's for recording, you know, educational accomplishments or validating sources of goods and things like this. There was sort of a great interest in that for a while. I think people have realized that sort of the distance between the true blockchain solution and just like I've got a database and it's been backed up a few times is not so great. And, you know, it doesn't help that with sort of the most typical blockchain, the time to mine a block, whatever else, the time to confirm a transaction is quite long. And there are many purposes for which you really want short confirmation times. And that's something you can easily achieve if you're centralizing things not so easily achieved if you've got to get consensus among independent parties. Now, obviously there are approaches to get tentative consensus and then keep going and so on. In fact, it has been my guess that our physics project gives you a way to sort of have a distributed version of blockchain that has this sort of sliding scale between you want it now, and you want to be sure of it, so to speak. A little bit like happens in physics with the uncertainty principle. You want to know that energy. Well, if you want it to know it really quickly, there's some uncertainty in it. If you're prepared to wait a long time, you could home in on the value of the energy precisely. And we tried to invent kind of a practical approach to doing that instead of a distributed blockchain setup. I was never satisfied with that. We didn't really finish it. And what it segued into was realizing, gosh, to understand how this works you really have to have a theory of economics. And that really segued into thinking about what is the, you know, how is there a good global theory of economics? What is the, what is value, what is price? How does that relate to sort of individual transactions and the aggregation of those things. Those ended up being the story that was sort of the underpinning for making it a distributed blockchain based on ideas for multicomputation, our physics project and so on. It's a project I want to get back to, but as I say, it kind of backed into. And now, you know, a little sort of module of the project and now we have to understand a new science of economics type thing. So it's. But in terms of other types of things, I must say I thought, I think NFTs, for example, was something. We had some fun making some cellular automaton NFTs really kind of cool pictures, I thought, and kind of doing a live stream generation, an auction of those things. That was fun and I think some people had fun with that. I think it was sort of a. I think the idea of kind of NFTs and being able to have kind of this sort of validated unique kind of objects and so on. It's sort of as often happens, the kind of the hype cycle was such that sort of got oversold and crashed. That doesn't mean there's absolutely nothing there. But I'm afraid it got so. It sort of. The bubble was so big that it sort of had to burst completely. I have to say, I think as sort of AI, generative AI and so on becomes more and more popular, the desire to have, yes, we can validate this thing really is what, you know, this photograph was really taken where it said it was taken, et cetera, et cetera, et cetera. That will be more and more significant. However, I think most of the time people are going to trust sort of impartial centralized databases. I don't think they're going to need throw it on a blockchain and sort of distribute the trust like that. Probably because I think that the sort of one of the things that happens is when you're dealing with kind of a low value issue where yeah, you kind of like want to make sure it's right, but it's not that Important and, you know, there's only 20 bucks riding on whether it's right or not. Why, why bother to go to all the trouble to get it distributed among, you know, 5,000 copies of the blockchain or whatever else? Just not worth it. And similarly, on the other end of the spectrum, if this is, you know, a $50 million deal and something might go wrong, you really are not going to trust a kind of a DAO blockchain type solution where there are no humans in the loop. I mean, this is exactly what happened to the DAO situation with Ethereum is, you know, it's just code. And then somebody did something that people said, no, no, you shouldn't do that. That wasn't the intent. Well, the code is just recording what the code wants to do. And by the time the sort of transaction amounts are large enough, people are going to want kind of the human backstop, so to speak, of. Look, if everything goes down horribly, you know, it'll go into litigation and people will argue about it and humans will be involved and they'll hopefully do something sensible, so to speak. So I think there's a middle zone there of things where the sort of transaction size is large enough or there are enough different parties involved in the whole situation that you want something where there's sort of an a, a generic, oh, yeah, put it on this blockchain. You know, nobody controls that blockchain. Well, at least one thinks nobody controls that blockchain. It depends on what happens with these, with these cryptocurrencies particularly, and whether there's sort of such a dominant player in one of these cryptocurrencies that in fact it is effectively controlled in as centralized a way as one ever expected. I mean, I have to say, it's one of the things that the folks who've been working on this for us, one of the kind of jokes that seems to come up rather often is when you talk to people who are in the cryptocurrency business, for example, they will often rail about kind of the terribleness of central bankers and individuals making decisions about things. But in fact, if you look at what they are actually doing with their cryptocurrencies and so on, what they aspire to be is essentially a central banker. So in a sense, it's kind of like the thing that is the rhetoric originally of we hate central bankers, so fiat currency is a bad thing. Then the cryptocurrencies come in and by golly, actually they just want to be a different central banker. That's them not Those other guys type thing. Maybe that's an overly cynical point of view, but I think that's a thing one's observed more often than one might. So in terms of the significance of blockchain in the world at large, I think it's one of these things that is sort of a driver of, let's rethink how we're doing stuff. It happened with the cloud and software as a service technology and so on. People said, well, we got all these systems, maybe we should rethink how we're doing them a little bit. The same with containerization and virtual machines and so on versus installed software, those kinds of things. But these things, particularly when it gets very sort of hyped, it ends up being a, hey, well, let's take a look at what we've been doing for the last however many decades and see if we can do it better. And for blockchain, I think a bunch of sort of transaction processing systems, it was kind of like a, hey, here's a new way to do things. Let's, you know, maybe we should rethink this. And that was a useful driver in the world. And we're seeing the same thing again with certain aspects of sort of AI. There's some things that are being done where it's really new stuff that couldn't be done before, but there's a certain amount of, hey, wait a minute, maybe we can just sort of use this as a push to redo the things that we were already doing, but do it in a more streamlined way and rebuild the thing, which we haven't gotten around to doing for the last however many years. Let's see. Jamie asks, what's a fun business idea you've had but never executed? Oh, my gosh. Oh, I have all kinds. I actually tend to sort of restrain myself from having too many of these because in a sense it's frustrating because it's kind of like you can see a thing, but you can't reach out and grab it. You don't have the wherewithal to do it. I mean, there are things where I've thought, somebody should build this and I'm not going to be the one to do it. Whether it's kind of the, I'd always sort of thought virtual tourism of get somebody to go and go into the rainforest and wear all kinds of camera stuff and you're sitting there in, you know, with VR or something like this, and you're, you're kind of, you know, it's kind of a virtual, virtual visit to a to a place sort of under your control, but without you being there with all the, the vaccinations you need and all of the, you know, all of the mosquitoes that you have to, you know, bat off. It's, it's, you know, have somebody who's set that up once and for all and can go, go do the visit for many people type thing, one random, random idea which, which, you know, maybe, maybe will come to pass with the, with the latest round of VR. I would say that, that there's a whole bunch around personal analytics that I've considered and I just, for various reasons, I just don't think there's a business to be made out of them. Because while there are things that I personally want, there, there's not a, there's not kind of a way to. Other people don't have the sort of wherewithal to be able to make use of it. I mean, I would say in that area, one thing that I'm planning to do is build a good system for doing sort of archiving. When you go visit archives of, you know, famous person xyz, in my experience is the people who run those archives, if they know my sort of story and the kind of tech that we do, they'll say to me, gosh, if you ever get around to making a good system for doing this kind of archiving stuff, that will be exciting to us because we really need that. And with notebooks and a bunch of machine learning and a bunch of the kind of computational representation that we have in our computational language, one can make really good ways of kind of organizing archives and things like this. Now I have to say, in terms of what are things that one might have done, so to speak, one might do, there's just an unbelievable amount of stuff that becomes possible on the basis of the technology we built with Wolfram language and so on. And I would say that it's a deeply underexplored area. I mean, there's certainly plenty of people have made a living building technology on, on top of Wolf language and, and selling it in various forms, but there is so unbelievably much more that can be done. It's just, just a remarkable amount of stuff. And I think, you know, that that's a, in a sense, it's, it's rather, it's, it's sort of, it's, it's rather easy to reach kinds of things. And I'm, I'm a little frankly disappointed at the world that more of that hasn't happened. It seems super obvious and very much within reach for people. And yet the initiative just doesn't seem to be taken to get those kinds of things to happen. I think in terms of other things I've thought about doing. Gosh, let me see. That's a funny one, because I. I'm always kind of trying to. First thing to say about products is that most products that I see being successful, certainly ones I've been involved in, are ones where I kind of want that product myself. If it's like, oh, somebody else is going to want this, but I don't want that, it's much harder to internalize what you would need to make that be good. And so I suppose the things that I sort of imagine doing are often things where I'd kind of like that myself, so to speak. And that's. Yeah. So that. That sort of limits it, I think. Well, okay, I can give you. I'll give you an example from a very, very long time ago. This is. We had. This is probably 1980, for various reasons, somebody was like, come up with a bunch of patentable ideas. And I'm like, look, whatever. Let me see what some kind of random dopey ideas are. So the two that I remember, one was an alarm clock that senses your EEG to decide when in the sleep cycle to wake you up. Well, it turns out you didn't need EEG for that. I should have known you could do that with accelerometry and things. And essentially things like that came to pass. The other one was kind of at that time, 1980 or so, video games for pets. And this sort of. We reiterated on this about 10 years ago, almost actually did. This must have been a little bit longer ago, just after the iPad came out in 2010. The. I thought that would be a fun thing to do because being somebody who's interested in sort of alien minds, it's kind of like if a cat could play Minecraft, what would a cat build in Minecraft, so to speak? What would be the expression of, you know, what is a cat thinking about that it could express in some creative thing on, you know, on an iPad or whatever else. That project, for various reasons, never happened. I think it still could happen. I think it would be a fun project. I think the most interesting thing would be this whole question about, you know, can we get insight into sort of an alien intelligence or not? Maybe, maybe the cat would just never be interested. The person who was helping us from an animal behavior point of view at that time said, you know, cats aren't going to be Interested dogs don't have good enough eyesight, but cockatoos are the creatures you really should target. Obviously, there are fewer cockatoos than cats and dogs, but they're smart creatures and they're social creatures and they would have loved a. A. Well, what now would be sort of X for cockatoos or something, Twitter for cockatoos. But we didn't build that. So that's an example of a fun thing which I, you know, I don't know what the. What the cockatoo market is, but I do think that, you know, it's a place where something at least scientifically interesting could probably be done there. Let's see. Isla asks, do you think we'll get to the point of textbooks by AIs? Well, something we're working on very actively right now is AI tutoring. It's a hard problem, but it's a problem where I think that the latest round of kind of AI has a serious chance of being able to do something. The main issue is you're trying to learn this thing. The AI gets to know you well enough to know how to tell you in particular, this thing. One thing to read the textbook that everybody else is supposed to read. It's another thing to have that personalized for you, personalized for the things you're interested in, for the things you don't understand, for things you do understand, and so on. I mean, I fully expect it will be the case that will get to the point where my AI knows me well enough that when I'm trying to understand something, it can point out, this is the one thing you should know which you seem to be confused about, which will then unlock this whole direction and let you go forward from there. And I think that's sort of, in the end, the concept of the AI tutor and also being able. In the AI tutor, one of the things that's sort of a new thing is being able to look at the working. If you're doing a math problem or a chemistry problem, whatever it is, and you say, I'm doing these steps, and the AI can look at that and say, look, here's where you got confused, which is what a human kind of looking at it sort of after the fact, could do a. But I think that's something where kind of the. The AI should be able to do it in a more effective way. Now, you know, I don't think that the traditional, you know, when it comes to take an area that's well known and say, how would I write a textbook about it? Yeah, the AI can produce something which will be kind of the average of the average of sort of what a standard textbook in that area might look like when you are dealing with kind of newly conceptualized things, sort of new paradigms for thinking about things. That's a much more challenging, much less likely to be AI able type problem. I mean, for example, right now I'm just embarked on a big project that to basically write a book and course it's an introduction to computational thinking, sort of an attempt to be able to explain to people how should you think about the world computationally? Not how should you write low level programs, but how should you think about the world, how should you conceptualize things in the world computationally? And a large part of doing that is just knowing how parts of the world that we already understand can be conceptualized computationally. So this project I certainly thought about is this project sort of waste of time because the LLMs will just do it. And I decided, no, it's not. It's actually something the LLMs will, will happily consume perhaps and be able to help with, be able to help teach people based on sort of that framework. But you need to build the framework first. And by the way, you know, the exercises, many of the exercises will be in the first instance doable by LLMs. It'll be kind of like given this sort of fuzzy thing, sort of define that computationally and the LLM might have a way to do that. But the human then has to say, yes, this computational specification that the LLM came up with, yes, that was what I had in mind. No, that wasn't what I had in mind. And then you can kind of build from that as a human. So let's see. Comments Wolfman was one of the pioneers in using notebooks for scientific programming. Do you think programming languages would be more expressive? I would like to say that the historical record would be that we invented this kind of idea of notebooks in 1987. It finally got copied many, many, many years later by various, gosh, 25 years later. I always thought that that was a terrible thing in the following sense, that that was one of the easiest ideas we had, you know, 35, 36 years ago. And it took 25 years for other people to kind of cotton onto it and copy it. There are much bigger, more important ideas that I think I realized the timescale for other people really cottoning onto them is much longer than I had ever imagined. And that's sort of a sobering thing to realize, so to speak. But the thing of using Notebooks as kind of an environment to do programming. It's an environment for making computational essays, for expressing oneself computationally. It's a very powerful environment for that. That's a little different than cubicle type programming. You know, one of the things that's an issue, and this is one of the sort of big, I would say, things people haven't yet understood, is that what we've tried to do in building our computational language is automate as much as possible. That means that when you try and solve a problem, build a thing, we've tried to automate most of the machinery of doing that. We can't automate defining what it is you want to do. That's for you to do, but we've kind of forced you to be, that's the main thing you have to do. So if you have, oh, I've got two weeks to do this project, if you're using low level programming languages and so on, then what you're going to end up doing is spending most of those two weeks just doing the machinery, grinding through, writing pieces of code. The whole point of computational language, and what I've spent the last 40 something years trying to build, is automate as much of that as possible. So once you can conceptualize and express the idea that you have about what you want to do, then it's up to us building the computational language to make it possible to sort of automatically get that. So sort of the flow of work ends up being, I got to figure something out. It's going to take me half a day to figure out. It's hard to figure out. It's difficult. It's hard intellectual work. But then most of what comes after that is pretty automated as opposed to I can kind of feel like I'm getting real work done every day because I just ground through another, you know, 700 lines of code or 50 lines of code or whatever it is. And I think that the sort of, sort of the minimum, if you write Wolfram language code, you know, you can get a lot done in a couple of lines of code and you know, other languages, standard programming languages, that's really implausible. It's really going to be a big glob of code. Just because it isn't as automated, it doesn't have built into it all this knowledge of the world and all these capabilities that we've spent so much effort kind of building in, in a coherent way into orphan language. So I think the thing that then the kind of idea of I'm going to write code and sort of the idea that lets me sort of have thousands and thousands of lines of code that is much more a thing that is sort of the key thing to do. If you're using kind of low level programming languages, by the time you're using kind of computational language, I think it's much more sort of the interwoven sort of code as an expression of ideas, words as an expression of ideas. That's kind of much more of a notebook story. And you know, to take the thing, write it in a notebook, then when you deploy it as an API or something, you're just feeding in the code from that notebook and it sort of gets turned into something which can be delivered as an API or whatever else, but I think or as a packlet or package, whatever it is. So I kind of think that the style of programming, this sort of computational language style of programming, the style of programming where it's all about expressing ideas and where, where the actual language you're writing, the computational language you're writing is expressing ideas. Not just implementing code, but also expressing ideas. I absolutely think that's the future. It's a future which. It's sort of depressing how long it's taken to arrive and certainly still not here. I mean, I had thought back from nearly 40 years ago now, things like academic journals, sort of the thing that is should be a lot of the work product is what I would now call a computational essay. Something where you can read the words, see the pictures, and there's code behind that that you can read and understand as a form of understanding more about what's going on. And by the way, you can pick it up and just run it. And you know, the things that I write my whole writings.stephenwolfram.com, all the quite large number of things that I end up writing. One of my principles for myself is every picture in there that is computationally generated, you should be able to click it and there's computational language code right there and you can paste it into any Wolfram language system and you can run it. And I would say that that has been incredibly powerful. If you look at what people have been able to do based on things in our physics project and so on, they start from all those interesting computations and visualizations that we can already do. They start from that and can build on that rather than that I read the words, I re implemented it, I had to figure out whether it was right, those kinds of things. So it's kind of the ultimate way of having sort of incrementally buildable and verifiable science and other kinds of things is that that way of sort of having readable code, readable computational language alongside sort of natural language descriptions. And yes, I think that is absolutely the future. The future is computational language that you can read. Maybe it wasn't even language that maybe it wasn't even something you wrote. Maybe. And this is kind of the shocker for people who see themselves as, as low level programmers. Maybe, oh gosh, the AIs can write that low level code. I don't have to do it. It doesn't quite necessarily work as perfectly as that. But, but the, the thing that is, you know, that's code intended for just sort of filing away and implementing on the computer. The whole point of computational language is it's a way of expressing yourself computationally. That is a way where you say here's this thing and you can read the thing just like you can read a piece of mathematical notation. You can read the computational language and actually understand an idea from that. And I think that's kind of the future of how one thinks about quotes programming. I think it sort of segues into things where one's seeing computational contracts, seeing, well, what do you want to buy here? Well, here's it written out in computational language. This is what I want for my dinner or what I want for my whatever else. It's a thing represented in computational language. I then feed it to the execution layer, which might be the sort of the it might be. And the example for making food is not as fictional as one might think. Because for example, that is possible now with automated labs. You can take orphan language code and you can have that, you know, cloud lab, just go execute that code and it's actually, you know, putting, you know, squirting chemicals into test tubes and, and running analyses and things like this. So, so that becomes a perfectly realistic way you're expressing yourself in computational language and then actuating it in the real world. Let's. See. One more question and then comment and then probably should go on my way and go back to my day job. Let's see. Okay, there's a, there's a question here. From chat. Do you think philosophy and psychology or an applied version of those, for those less theoretical, should be taught throughout school years, just like English and math in addition to everything else. You know, one of the big challenges about education is kind of what fundamentally you trying to teach? What's the point? What's, you know, are you trying to get everybody to be aligned in some kind of common Cultural framework. Maybe. Maybe you see that as indoctrination. Maybe you see that as kind of keeping a coherent society. I'm not sure. Maybe you see that as preparation because in times past, you know, you wanted people to join the army and all be pointed in the same direction or whatever else it, you know, that's one purpose. Another purpose is get people to the point where they have the skills necessary to do some particular job. Another would be get them to the point where they can sort of think and where they can sort of reflect upon things for themselves, you know, have a kind of a considered life type thing. Sometimes those last things get very much trampled in the effort to sort of just define, you know, skills that you can. And partly just to go through the process of education and say, yes, you know, I got this particular certificate. And then the world knows that, you know, I'm ready to go on to the next thing, so to speak. I think this challenge of getting people to sort of think creatively and so on, which is both good for them as individuals and probably good for society. I mean, it really depends. You know, you can, you can have a society where you don't want people to think too much, you just want them to follow the direction that's been defined. I think in many of us would, would think we prefer the society where people are thinking and doing different things and have different points of view. But you know, that that's. I don't think there's a. You couldn't say that there's a theoretically right answer. I certainly prefer that. But I think this, you know, how do you teach people to think? You know, one of the things that as you teach sort of mechanical processes and you say you're going to learn the subject, if you learn steps 1, 2, 3, 4, 5, you can sort of present, you know, you can effectively execute in the subject. That's the exact opposite of if you're out there in the open field, so to speak, what are you going to do? Can you figure out what to do that's, you know, achieving the opposite objective? Yes, you can have people be able to do that circus act, so to speak. But as you, as you teach it by 1, 2, 3, 4, 5, you are almost going the opposite direction for what do you do if there's no circus there and you have to make up the act for yourself, so to speak. And I think so how do you best teach that? I have to say, I think different forms of thinking and computational thinking is a very nice one, which I think is still and that's kind of one of the points of this book and course that I'm trying to write is, okay, computational thinking is a framework for thinking. It's a framework for thinking that has the extra little feature, not so little big feature that computers come in. They bring you a superpower, so to speak. Once you can think computationally, your computer can kind of fly in and give you lots of leverage. More so than, for example, let's say, if you're doing, I don't know, well, math without a computer. Let's say it's kind of like, yes, it helps you to streamline things, to think about things with mathematical notation, with mathematical concepts. It helps you kind of clean up your thoughts, but it doesn't immediately give you. Unless you're using Mathematica to do the math or whatever. It doesn't immediately give you kind of the extra superpower. Computational thinking, sort of, by its definition, immediately gives you that extra superpower. I think it's a very powerful kind of philosophy, like activity. It's something where you're thinking about the world. It's, in a sense, it's like applied philosophy with this extra superpower that allows you, as soon as you've got, as soon as you figured something out, you give it to your computer and bam, you get to see some big result from it. I think that kind of getting people to think about things, it's always a challenge. I think philosophy has the nice feature that many questions in philosophy sort of have been debated throughout recorded history, at least since ancient Greek times. And they're things where any random kid is probably going to have opinions. You present some ethical dilemma. People are going to have opinions. They may not have thought through their opinions very well. And they may be like, they immediately say, you should do this. And then you can point out, well, if you do that, then this is going to happen. That's going to happen. That's probably not a consequence you foresaw, but being able to think through this, what was the thing that you didn't foresee? That's a useful. Can you think about what you didn't think of? That's in a sense, less the sort of mechanical. Let me go from the beginning to the end of learning how to do something, thinking of what you didn't think of, so to speak, from thinking about things from the outside is, I think, something that is a really useful skill, but it's something you don't really get from that sort of regimented way of learning things. I think similar to this question about do you Learn to answer questions or do you learn to invent new questions? Most of education just never talks about inventing new questions. I think in philosophy one's a little bit more exposed immediately to sort of the raw material of ideas and a little bit more of sort of this notion of yes, you could have another idea as well. I mean, if you're studying, let's say, physics, for example, there's a lot to study there where it's just like, well, that idea was had in the 1600s and now you're just going to learn it. And rather than. And then from that one idea a lot of stuff flows, a lot of mechanical stuff flows. And I think in philosophy it's a little bit higher density of here's a new idea, there's a new idea, there's a new idea. All these ideas are often contradictory with each other. That's not something you find in these other fields. And it's useful to see that these ideas are in kind of this foment rather than just they're all fixed. Because I think it gives much more the notion that you too could have an idea if it's like, let me just tell you the following 10 ideas which were established historically, and that's what's true. That's very different from there are these ideas and, you know, everybody can have an opinion about them. By the way, in computational thinking, it's the same type of thing. There's sort of a wide range of different kinds of things you can explore, most of which nobody has ever explored. It's you and your computer getting to explore them for the first time. So I think that that would be what I have to say about that. I, you know about psychology, I probably know less about it myself, and I have to say I'm a bit skeptical of. I mean, I think it's kind of like, well, come up with the mind hacks that let you understand other humans better type thing. Maybe there's some of that, I'm not sure. I think that's the kind of thing which could end very badly. I mean, particularly if it gets very structured. It's kind of like, this is what you do in this situation, this is how you react to your fellow humans type thing. It's almost like programming humans to be kind of like artificial AIs, so to speak, saying, you know, in this situation, do this, in that situation, do that. You know, I think there are, there are undoubtedly things there that do expand one's thinking about things, but I think there's a certain tendency to just say there's going to be a formula do this, this, this which I doubt will be all that useful. Maybe in the, in the short term, it's like, look, you know, you can it's like, can you you know, are there these hacks that if you're the only person who knows them, you can get lots of mileage out or probably there are I, I doubt I know these hacks. Maybe I've, maybe I've gradually learned them over time. But, you know, I'm not sure that intensely studying those is, is necessarily the best thing. All right. Well, I think we should wrap up for now today, so thanks for joining me. I think by next week I should be back in my natural habitat and perhaps some of the mechanics of doing this will be a little bit easier. So thanks for joining me. See you another time. Bye for now.