History of Science and Technology Q&A (November 22, 2023)

Hello, everyone. Welcome to another episode of Q and A about history of science and technology. I am not in my natural habitat, so video and other things might be a bit odd today. But let's see. Well, all kinds of interesting questions here. Elsie asks, what are your thoughts about how R and D has been done and funded across history? So we're talking research and development? Well, that's an interesting question, I think. Well, let's go back to antiquity. What kind of research and development was done in antiquity? I think most of what was interesting then was military research and development. And from people like Alexander the Great to people who were making use of Archimedes as a consultant for the Sicilian, the Syracuse military and so on, a lot of kind of the, you know, it was a piece of a function of the military enterprise of the state. I think other kinds of. Well, I mean, certainly in many of these cases there were large municipal projects that were done building aqueducts, building. Building kinds of bridges and arches and so on. And I'm not sure whether that would be something which would be thought of as a sort of research and development, whether that was just more like the, well, let's construct this thing. And the construction of those things required figuring stuff out. I know. I mean, it's sort of a messy history, I suppose, because in this kind of idea that one would do sort of research and development disembodied from the actual projects and products that one was trying to get out, I think was a. Is sort of a thing of modern times. I think, back in those days, I mean, I kind of always wondered. The Antikythera device, this first example of a mechanical computer that came from sometime between 1st century BC and 1st century AD was found in a shipwreck around 1900 on an island off the coast of Greece. That thing, which is our unique example of a kind of computational mechanical device from antiquity, I've always thought, you know, how was that made? Was there Renantikythera factory somewhere? Would it be the case that if we could do excavations under the city of Syracuse in Sicily or underneath Athens or something like this, that we would somehow find kind of the Antikythera factory? I don't think it was probably a factory. I think it was more like a workshop where things were being made sort of individually by people. By the way, that whole process of doing sort of archaeology under existing cities is a challenging one. And I've kind of always assumed that one day there will be kind of autonomous sort of drilling devices that will just be able. You'll just Sort of put it in the ground and then it'll go off for six months and it'll slowly drill through and it'll find a bunch of stuff and it'll sort of reappear having put in its. In its cargo container all kinds of interesting finds from. From what was sort of discovered under the. In places where you couldn't sort of excavate from the top. I suppose also, when I think about that, the excavating from the top thing, if you were prepared to do much more work and if you had many more kind of autonomous digging devices and so on, and you'd sort of solve the problem of robotics, then maybe even digging from the top would be something that would be feasible, which it isn't right now because you can't take some existing bustling city and dig a giant hole in the middle of it to find archaeological objects. I mean, often archaeological objects get found when somebody, for whatever reason, digs a big hole in the middle of a bustling city. For example, they're going to build a new skyscraper or something. And I think this happened in London, for example. And, you know, they dig down and by golly, they find all kinds of interesting archaeological kinds of things. But any case, the, you know, how was the Antikythera device made? You know, it, and was it doing research and development? I, I'm assuming that most of if it was made, it was made by sort of an artisan or a group of artisans who then delivered it to customers. And it was a very simple thing where there wasn't kind of this. This complicated chain. Now, now, when I say deliver to customers, I mean there were certainly traders who would buy things and kind of travel around and then sell them to other people. So the supply chain wasn't necessarily person who makes it, sells it to people. But, you know, when it comes to folks like Archimedes, who certainly did basic research, I think he did that as, you know, he was kind of a person's. I don't know what his means of support was. I don't know whether that's even historically known. I mean, some people, like, you know, Plato, his family owned a bunch of vineyards, and so he had kind of a landowning business, so to speak, that was independent of his efforts with the Academy and so on, I think. Well, another place where some sort of basic researchy kinds of things were done, I think, is the Library of Alexandria, where, for example, Euclid worked. And the. I'm not quite sure what the funding source for the Library of Alexandria was, whether it was just a state function of kind of the Roman Empire, so to speak, or whether it was. I'm sort of thinking that it was. And I mean, somewhere there was a, a budget line item for Library of Alexandria. And you know, the Library of Alexandria was kind of like a university. And I think the other possibility, which I don't know the answer to, is whether people who went to study there would pay for it. I certainly assume that Plato's academy worked that way, that people who wanted to study there would pay kind of a fee, just like university fees today. And that will have been the way that sort of professors there got funded. I mean, certainly in many systems, kind of the idea of the university as middle person, where the university hires the professors and students pay money to the university, that's not always worked that way. I mean, even in, for example, continental Europe, particularly Germany in the 1800s, early 1900s and so on, this notion of a privat dissent, a person who had a license to teach, but the actual money they made from teaching came from individual students showing up to their lectures and paying them for being able to come to their lectures. They had to have a state government license to teach. But kind of like many sort of regulated professions today, you know, you have to have a state license to be a lawyer, a doctor, whatever else, but then it's up to you to collect the money from your individual clients. And so it was with professors, particularly in continental Europe back, you know, 100 years ago or so now. And I wouldn't be surprised that that system had existed. I think in some form or another. That system probably existed through the Middle Ages and probably back into antiquity. And it's kind of an extreme version of rate my professors, so to speak, that, you know, your living comes from. Does this individual student like the class you're teaching? A little bit more like the tutoring business, I suppose. But this was for, for larger scale classes. But it was done, you know, the professor collected the money. It wasn't like it went through a kind of a intermediate pipeline, so to speak. I think in, you know, if we look to later times, I think it's very common for the court, the king, the state to have a certain number of courtiers who were sort of science oriented or R and D oriented hanging around. Now whether those people were kind of, whether, whether they were kind of, I think many of them were kind of general purpose as in, well, you're going to do, you're going to, you know, help with our fortifications and you're going to help with our History research. And you're also going to occasionally entertain the sort of the monarchy with tales of science and so on. I mean, we certainly know many examples where that happened. You know, Descartes, for example, famously was hired by Queen Christiana of Sweden, I think late in his life. Unfortunately it turned out to be late in his life to be kind of a person to hang out there and teach her a bunch of stuff about philosophy. Unfortunately he caught a cold and died soon after that happened. But that was a case where it was kind of like bring in a philosopher. We want to have a philosopher around. And by the way, the Queen is interested in learning philosophy and I suppose a similar kind of, let's see, I mean we know that for example, you know, I think people like Kepler, for example, who tended to sort of travel around, he tried to support himself from various kinds of advising and you know, he, I know his, his book the Six Cornered Snowflake is dedicated to some particular king, prince, whatever, who was kind of his, his benefactor at the time. I think the. Now, now, Kepler, for example, so had sort of an entrepreneurial thing going. He was trying to figure out could he make a business out of figuring out, given a wine cask, how much wine was in the cask without sort of opening it and measuring it, just by virtue of sort of figuring out how deep things were. If you poked one hole in it and saw how that worked and then you looked at the outer solid revolution, so to speak, of the wine cast, could you do sort of pre calculus to figure out the volume of wine in it? And that was kind of a business he wanted to have, I think. I think also there were other people who were kind of using math for businesses at various times of. And by the way, in terms of something like, I'm jumping around a bit here, but like in Babylonian times where kind of the scribal schools were the places where people were getting taught sort of math kinds of things they're taught to do, writing taught to do, kind of figuring out of mathy kinds of things. Now was there some person who was like, I'm going to be the core R and D person who's going to figure out what you can teach? I don't think so. I think rather what happened was that individual people would have their methods for teaching this or that, and those would be relevant to merchants and merchants would pay to get educated, have their children educated about those kinds of things. I know if you go much later in time to like Fibonacci in 1200 AD in Pisa in Italy. His business was providing kind of books and consulting to merchants who wanted to know how to do calculations. And he was like, the Hindu Arabic number system is a much better way to do it. You should use this. And there are these different methods for calculating. So this was kind of a combination, I think, sort of I'll do the calculations and I'll show you how to do the calculations kind of thing. Again, individual, sort of an individual achievement is kind of more like somebody will figure out stuff to write a book, and then they'll make their living from selling that book. There were other systems. I mean, I think that the idea of sort of research and development for its own sake, particularly basic research and development, its own sake, was not something where it's more like individual people would do it. Now, there were systems like in England, for example, Isaac Newton was famously Lucasian professor of mathematics. I think he. In Cambridge, I think he was like the third one in the mid to late 1600s. And that was kind of a grant from a person. I think there. There were also Regius professors who were granted, you know, had. The king had provided the funding for that. And my impression is that much like universities today, job number one was supposed to be teaching. And research was something that you did perhaps to support your teaching or for your own entertainment, so to speak. I mean, Newton, for instance, kind of later in his life took this gig being controller of the mint, you know, the kind of chairman of the Federal Reserve type thing, because it was a step up in the hierarchy of those times relative to be a professor, relative to being a professor. And it was also more money. So, you know, I think that the research in support of teaching is probably a thing that's been going on for a long time. The occasional sort of research for the state is another thing. And people would get hired to do things. Like Gauss was hired to do a bunch of things related to land surveying, which is what kind of led him to invent differential geometry. I think there were also quite often things like state observatories. I mean, the Babylonians had those, and those existed much later. I'm not quite sure why people were funding kind of state observatories. Back in Babylonian times, it was probably all tied up with astrology and watched the heavens to predict what will happen on Earth. But certainly by the 1600s, that was not what was going on. Now, it also has to be said that there were plenty of people up through the 1800s for sure, who got to do basic science because they were independently wealthy in some way or Another and they were, you know, either it was, I think Napier who invented logarithms was a Scottish baron of some kind. The many people were kind of, you know, part of the aristocracy in England or in other countries. And you know, they, rather than just sort of hanging out and managing the land, they also sort of had a hobby of doing basic science that led to many kinds of things getting figured out. And sometimes there were sort of combinations like I think James Clerk Maxwell was a little bit of a I've got my fancy house in a very obscure part of Scotland and I'm also, you know, hired in London or in Cambridge University to be sort of a professory type. But I think the concept of kind of. Well there were different forms of, you know, usually governments would fund things for very pragmatic reasons like you know, the admiralty in England would be funding various kinds of navigational research, effectively timekeeping research, eventually weather, weather data keeping. But they were for very pragmatic reasons like we want to predict the weather, we want to give people reports on that, we want to let ships be able to navigate well, those types of things. The idea that sort of research should be done for its own sake is something that occasionally would happen as a result of just somebody saying well it's interesting to think philosophically about this, which is kind of like basic research, the modern machinery of kind of government funded basic research, sort of anonymous government funded basic research I would say, which was not like we're hiring Gottfried Leibniz to be our court researcher type thing or we're hiring Archimedes to be our kind of court scientist or whatever. I think the mechanism of kind of giant anonymous things where the government says we're putting out billions of dollars and it's going to go to the most deserving in some sense folk sort of anonymously. I think that's a phenomenon of the second half of the 20th century. I think that before that time it was much more there's a professor at a university and people are paying for that professor and so on rather than something where just there's a blob of money allocated and it's kind of flow to basic research. That whole mechanism got I think largely invented while the Manhattan Project was a big piece and then subsequently the real acceleration of that was Sputnik in 1957 of the US being like oh my gosh, kind of the Soviet Union has pulled ahead and we got to scramble and support science. And part of that calculation was getting a human resource within the country where there would be people who Would know about science and be enthusiastic about science. And that's something by the way, in the Soviet Union one has the impression that that was with a very sort of state orchestrated mechanism that was a thing that was kind of made to happen. You know, there were universities created that were teaching people about things that were thought to be useful in science. And I think even from the founding of the Soviet Union, you know, strangely enough, you know, Lenin had written some calculus book type things. You know, there were, there were people who were sort of, there was a kind of an intellectual interest there. Things got very funky because. Well, on the one hand, for example, even in logic there were sort of a. There was the Soviet style of logic which I think had a different treatment of the law excluded middle that not, not P is the same as P and so on. And there were things that maybe came from some interpretation of Hegel or who knows what that fed into the sort of the, the. The communist sort of story and the, the. The whole Marxist Leninist dialectical, I don't know anything type, type mechanism. But it was sort of strange that even something like logic became political. I mean it's kind of a sad and cautionary tale for modern times that even things that you think of as being kind of. Look, it's just sort of, just sort of pure thought. If you go far enough, even those become political. I mean the fact that logic could become political is remarkable. I think one of the things that actually did drive a lot of mathematical work in things like the Soviet Union is that fundamentally wasn't terribly political. And if you were going to be some kind of academic intellectual type thing and you didn't want to run afoul of the Communist Party so to speak, then working on math was a pretty good bet because that really wasn't the question of what the value of this particular integral was. Wasn't going to be something where the party was going to have this opinion versus that opinion. So I think that that did help propel things in that case. But again that was a very sort of centrally managed research enterprise and there were many sort of charming kind of Soviet style efforts. There's one I was just reminded of recently, this thing called TRIZ, which was this kind of Soviet style plan for inventing new things. It's kind of a methodology where you say you've got these different fields and you say well I've got one field where people are studying elasticity of this and another field where people are studying flow of that. And there's got to be something in the Middle that is kind of the thing that hasn't been studied. And if it was, it would be an invention. Kind of reminds me of some of the AI doing science type ideas that people have today that there's sort of a machine for inventing things. And it didn't work out terribly well in Soviet Union. I don't know of any spectacular sort of success stories from that. But it was kind of a part of the machinery of kind of the state planned kind of research effort. Now I should say that, you know, in the US this idea of for example, government labs that got started kind of, I think largely in World War II and with things like the Manhattan Project and then this idea of universities being a place where the government could give them money and then they could do basic research. I think that's a phenomenon of the 1950s and the Sputnik type period. And really in the 1960s that developed big time. And I know when I was growing up in England in the 1960s, the brain drain of people who like myself, although I was a little bit later than that, you know, moved from England to the US because the US was a sort of better climate for doing science, so to speak, largely because of ultimately because of government funding. And that was kind of a thing that, that the US was a sort of big winner in, of getting sort of the brains from lots of different countries around the world. I think that still continues. But you know, there are other kinds of initiatives. You know, there are place times when countries would say we're going to solve this. I mean, in the U.S. the Apollo program, space program in general was a very successful example of we're going to solve this and we're going to align kind of the government. I mean that was still, it was a. What was it? I don't know, some fraction of a percent, like half a percent or something of the GDP of the country was spent on things like the Apollo program. You know, let's, let's, it's a sort of a mission driven idea of, you know, let's create this kind of, let's, let's do this thing, this sort of climbing Mount Everest type thing where we don't know what its pragmatic kind of consequences in the immediate time are, but it's something that is, you know, a kind of a, a thing that brings the country together to do that. You know, I have to say over the decades all sorts of people told me, you know, somebody should organize a Manhattan Project of AI to kind of get to the point where we finally have kind of AI. And it was always like, well that's a nice idea. We don't really know what direction to go, so we can't really do that. Although it is interesting in the development of ChatGPT for example, that that was a. Pour a lot of money into it was an important part of making the thing happen. Although that was commercial money in that case rather than government money. And I think, you know, but there were other initiatives that were very government led. For example, one the, the Japanese fifth generation project, I think that was initiated in 1981 where kind of the idea was, and it's kind of charming to look at this stuff today because Japan had been a country that had come up from kind of sort of poor economic conditions after World War II and so on. And it really developed particularly on the manufacturing side and kind of the vibe was a little bit like China and Taiwan and so on today that the vibe was oh, they're just copying stuff that's done in the west so to speak. And so the government in Japan and you know, a lot of that is what happened. Although there was a lot of innovation manufacturing methods and quality control and so on and the, the U.S. for example, in like the car industry and things like that was very complacent and really kind of just was, was just sort of raced around, so to speak, by something where, you know, you could say, well, the core idea of cars didn't come from Japan, but certainly the, the kind of the, the management methodology to do with manufacturing at high quality and lower cost and so on was that kind of story. But anyway, in the beginning of the 1980s, the Japanese government, I suppose fed up with being told that they were copying everything, said we're going to jump ahead, we're going to do the fifth generation computer project. I never completely knew what all the previous generations were, but it was very famous because it was like, oh my gosh, Japan is going to do AI. Because AI was a thing back then. It was mostly things like expert systems and so on. And Japan is going to do AI and they're going to jump ahead of the rest of the world. And it was a sort of a big government project in Japan. Like many of these things, it really didn't work out. It didn't help that it was sort of the wrong time in sort of the history of the development of AI was a time when people were very much into using things like the prologue language, kind of a logic programming language where you state goals and then you try and have the language fill in how to get to those goals. Kind of like theorem proving type stuff. And that wasn't a good language design for all kinds of reasons, which I could explain, but that's sort of a different issue. But that was an effort to kind of get sort of the government, in that case of Japan to go and sort of jump in and make something happen. Didn't work out very well. I think that's always this whole. I mean, one of the things that's happened in the US in modern times is that it used to be the case in the 60s and so on, that sort of. And the 50s and before that, you know, big R and D project, it was a government project. Nowadays it's much more, at least nominally commercially driven. Although often when you really follow the money, there's a bunch of government money in there that's driving what's happening. So it. Anyway, that's a little bit of a story there. I mean, I think the challenge of understanding why should one fund basic R and D is a big one. And what tends to be the case. I've always claimed that sort of monopolies are what help basic R and D. If you've got a monopoly, then any progress that's made in that field is most likely to benefit you rather than anybody else. Insofar as the US is the biggest economy, that's something which sort of applies to the US as well. But I think the question of why should a non monopolistic player support basic R and D? Well, there are usually several reasons. I mean, it's often a human resource development issue. Something where you've got a country where, for example, let's say you've got a country where you're worried about cyber attacks. I think Estonia is an example of this where it's kind of like, my gosh, we've got to get a lot of people educated in STEM like subjects because that's going to be important to have kind of the workforce we need when we need that kind of technical kind of ability to fend off cyber attacks. I mean, of course in earlier times in the origination of public education, back in the late 1800s, the, you know, in places like Prussia and so on, that was kind of the big story was we need people who can be educated so they can join the army and, you know, fight for the country type thing. And I think that, you know, that type of human resource development has been seen as a, as a sort of important priority for countries, even if they're not the country that will economically benefit the most from basic R. And D I think it's, you know, there are just many, many examples where things didn't work out very well where people said we're going to become, you know, we're going to buy our way into being leaders in field xyz. It occasionally works, it usually does not. And this, you see the same thing with people saying look at the startup entrepreneurial ecosystem of the US and Silicon Valley and New York and all these other places. You know, we want to transplant that to place xyz. How are we going to do that? Well, it's a government story and it's funding and it's money and it's this and it's that and it's. And then we import some people and usually it doesn't work. Usually the kind of the stars don't align to make those things come to pass, although sometimes they do. And you know, if you look in the U.S. for example, sort of interesting that Silicon Valley where had many near death experiences where kind of all the major companies there had imploded and then another, you know, other ones have always ended up growing from the ashes and now there's probably enough diversity there to support it. But at different times it wasn't clear, for example the software industry would be substantially based in Silicon Valley. There was a time when Utah was a big place with WordPerfect and so on. Seattle obviously became a significant place with Microsoft and, and many other companies. But you know, in Austin, Texas is sort of coming up in the world today. But these things are. It's a little bit hard to know. You know, it seemed like Scandinavia was doing well for a while in producing software technology. It seemed like Germany, Berlin particularly sort of for a while was doing okay and then hasn't really been much heard from since. The UK for example, I would say had a very lackluster performance in that regard for reasons that are a bit hard to understand. And I mean ironically a lot of the people who are sort of leaders in those industries in the US came from those other countries. So it's not something in the water when they were growing up type thing that it's something different about the kind of whole ecosystem anyway, the. Let's see. Well, let's see, here's a one, might be a smaller one from Nitro saying I've come across Petri Nets and never had the chance to dive deeply into them. Can I give some historical concept about context about that concept and their applications and so on. So Petronets were developed by a guy called Carl Adam Petrie. I think Actually, originally, when he was in high school, I believe he was interested in kind of finding an idealized model. As I think is the 1940s. An idealized model for chemical reactions. Where it would be kind of like. Well, you've got a certain amount of chemical that can flow in here. And. And in order to have this reaction happen. You've got to have some number of units of this chemical. And some number of units of that chemical. And when you have those many units. You can kind of make another chemical. And that can flow through to another thing and so on. So it was kind of a chemical engineering chemistry story. Where things were made discrete. Where you'd have discrete lumps of amounts of chemical and so on. It's sort of an interesting idea that then got many legs. I mean, Petri himself had a kind of a fairly long life. Died in the least 1990s, maybe even later than that. I exchanged letters with him, for example. Actually, he was a quite enthusiastic user of Mathematica and Wolfram language. And was very interested in making sort of connections between Petri Nets and that. Which we never really managed to make at that time. But Petronats became kind of. They were sort of used for this idealization of chemistry subsequently. Various kinds of idealizations of parallel processes of various types. And people have used them as sort of ways to idealize systems. I would say that some piece of what came out of that was things like uml. There are these. Okay, so let's go back a little bit earlier in history. So flowcharts. The idea of representing. There's this, then there's this, then there's this. That was sort of a thing. That was born of the idea of production lines for manufacturing. Production lines for manufacturing. That's around, you know, turn of the 20th century. Then the analysis of those things by things like flowcharts. The kind of the disembodied analysis of a production line. As opposed to Henry Ford going and just saying, let's lay it out this way. The kind of more theoretical level analysis was, I think, largely due to Frank and Lillian Gilbreth. Who were working in the 1920s. And were kind of time and motion experts. So there was this whole tradition of time and motion study. Which I suppose is sort of fed into. In some very circuit. To sway modern user experience design. Because this was people going to factories and saying, look, you know, this person. If they have to walk further to get from this to this. To pick this thing and put it on this. If you rearrange the factory, you can avoid that. And time in motion was people standing around with stopwatches and clipboards, kind of watching what happened in factories and so on, and applied to other things as well. The Gilberts, charmingly, I think, had 12, maybe even 13 children. And one of them. One of the children wrote this. Maybe it was the mother, maybe it was L. Gilbert. I'm not sure, but maybe it was. One of the children wrote this book called Cheaper by the Dozen, which was the story of applying kind of time and motion ideas to a family with 12 children type thing. I think it subsequently became at least one, maybe two movies about sort of cheaper by the dozen. But those were the efficiency experts, the time and motion efficiency experts. And by the way, I think it was Lohan Gilbreth invented flowcharts as this way to represent those kinds of things. And that was kind of a tradition. No doubt that tradition was relevant to Petri when he was inventing Petri nets. But that tradition also sort of flowed through to block diagrams and these kinds of things and then to sort of modern methodologies for representing kind of the flow of stuff in some process. And that's led to, on the one hand, things like systems engineering tools like our Wolfram system modeler, based on this language called Modelica, that is sort of a block diagram, a way of representing the relationship between things. Like, you've got this cog that is kind of connected to this other cog. We've got this wire that connects this resistor to this transistor or something. These are things where you're describing the structure of a system. In that case, you're then creating a bunch of ordinary differential equations and solving them to represent the system. In the case of these other kinds of flowcharts and things like that, you're more just saying this is sort of causal connections. And that's sort of another tradition in modern times is graphical models and causal modeling of sort of this event causes this event. I mean, I would say in our physics project, we've taken kind of that kind of causal modeling to the extreme. I was talking to Judea Pearl, who was kind of the originator of the sort of causal models type movement in the last, I don't know, 30 years or so. And I was asking him, you know, what's the biggest causal network you've. You've seen? And I think he said, a few tens of types of events. You know, each event in those cases is a name thing. Like the, you know, this particular thing goes to this point on the production line, or this. This thing happens as I Was saying, well, you know, in our model of physics, we've got maybe 10 to the 400 anonymous events that make up the, the history of space time, so to speak. So we've kind of taken the idea of these kind of causal nets and almost flowcharts to the absolute extreme of that's what everything is made of in some sense. But they're very, very huge numbers of events that are involved. But anyway, so Petri nets, by the way, our causal graphs, there is a limit of those causal graphs which is Petri nets. Essentially what's happening is when you are saying in a Petri net, remember sort of originally came up in doing sort of idealized chemistry. In an idealized chemistry, you are saying, well, how much CO2 do I have, how much water do I have, how much, you know, whatever. And you're just asking, I don't care which CO2 molecule it is, I just care how many CO2 molecules I have. It's kind of like it's then sort of this random soup of CO2 molecules that are reacting with other things. Whereas in our kind of causal graphs, every atom of space has a story. Individual atoms of space, not in aggregate, how many atoms of space are there? It's every individual one has a causal story, so to speak. And so the way you get to Petri nets from causal graphs is you're essentially equivalencing all of the, let's say, atoms of space of a particular type. Just say, I don't care which one it is, I just care how many there are altogether. And that equivalencing gives you Petri nets from causal graphs. And so that's one piece of the whole story. Another place where they show up is in the efforts, particularly in the 1970s, 1980s, to understand parallel computation, to understand sort of formal models, apparel computation, what get called process calculi. There are a variety of these things that was communicating sequential processes was one. And these things are relevant for things like designing operating systems where you have multiple processes running, you can have multiple threads, you can have multiple CPUs. They're also relevant for designing kind of heterogeneous parallel computing systems. There are other things like PI calculus and so on. I've never been a big fan of these kinds of formalisms. They've always seemed rather arbitrary and complicated to me. Actually, just recently I wrote a piece about expression evaluation and fundamental physics that was kind of trying to understand from the point of view of our physics project and our kind of constructs of causal graphs and so on, how should one think about things like expression evaluation where you can have different parts of an expression that can be evaluated in parallel or that are required to be evaluated sequentially and so on. That was sort of an interesting. To me, that was a great simplification and clarification of sort of the history of some of these process calculi. Now by the way, one thing to say is that some of those process calculi and the work on process calcul evolved into sort of an algebraic approach to quantum computing, I think largely through sort of the introduction of category theory into kind of this process calculus type area in the. In trying to sort of formalize these things and trying to find a mathematical framework, category theory seemed to be the relevant one. I have to tell a story. This was probably how long ago? Six, seven, eight years ago. I was at the south by Southwest sort of conference trade show type thing where they had this trade show with a lot of wild exhibits of people advertising kind of put your ad in the upper atmosphere with rockets that would send out sort of different colored things that would re enter the earth's atmosphere and robot arms and, and all kinds of wild sort of personal medical things. Just a real collection of funky kinds of invention type technologies. And I'm wandering around this thing actually with one of my kids and we stop at this one booth and it's like, what do you do? Because you know, the booth next door had a robot arm and the other booth was a reentering the earth's atmosphere with add material type thing. And the person there says we do categorification of Petrinets. And I was like, is that what you tell everybody who walks up to this booth? Because for most people that's just utterly, utterly mystifyingly incomprehensible. I didn't really quite get an answer to that, but it was kind of fun that that was a, you know, along with the robot arms there was categorification of petro nets. And what that was about was an attempt in that particular case to use kind of the Petri net idea mixed with category theory to be able to do things like represent the kind of flow control of things like blockchains and cryptocurrencies in order that you could essentially prove theorems about what could happen. And you know, can you double spend this token things like this? So that's a bit of a story on, on the. On Petri nets and so on. And I would say that they petronet sort of reappear every few years as being a thing that's kind of an idealization of kind of the operation of systems where there are many interdependent parts. Same thing that you get causal graphs for. I think petronets in the bulk are probably less interesting. Maybe, I don't know. They're things that have appeared, as I say, in our physics project. They're part of kind of the, the set of tools that get used in thinking about causal connections between things. And there's a bunch of work we've been doing about sort of generalizing the way to think about causality, whether it's things cause things or things prevent things from happening. That's a concept that does exist in petronets as well. Petronets also have, well, interesting questions about what's decidable and what's not. Petronets generally are quite decidable. That is, you can say, will this petri net ever be able to get into this state? Is there any way that this kind of system of, you know, checkout for a, you know, a self, self checkout, you know, store or something can ever get tangled up? Let's make a petronet representation. Let's prove we can never get into the state where, you know, the, the thing is just offering people free cookies at every second or something. And typically you can make those proofs in petronets. They are decidable, which tells one that they're not really a general representation of computation. But they do tend to be very computationally complicated and they tend to be sort of double exponential algorithms for deciding some of these things, although they are decidable. Let's see a question here from Aaron. Do I have any reflections on the recent leadership situation at OpenAI and the history of decision making for large scientific research programs? Well, I know almost all the characters involved in the, in the recent OpenAI situation and it's one of those stories where I think sort of it's kind of like technical debt. You build a bunch of software and some things didn't get that you sort of put in there at the beginning. You didn't go back and fix them again. This was sort of organizational debt of a structure that really didn't make a lot of sense anymore, but and really was built for a quite different vision. I mean, I remember visiting OpenAI when it was a, a young company, very much just a research group type atmosphere, very different from what you get with $10 billion of investment money in it and things like this. And it's kind of like, well, you kind of have to rethink these things. Now it's sort of interesting to ask questions about what kinds of corporate organization make sense, what do not. I think that, well, people say that the invention of the joint stock company was one of the kind of great inventions of the west, so to speak, that the joint stock company, which was invented what in 1600s and so on, and this idea that you know, you could own, people could own stock in a company and it wasn't just the owners weren't necessarily the operators that that really enabled a huge amount of industrial development and so on. And I think that, and that's an invention that didn't happen for example in Asian countries for the most part. And so what you saw there was yes, there was a craftsperson, an artisan, you know, a guild member. But it was like kind of the person who was doing the work was the person who was making a living from it. Not this notion of investors and so on. I have to say in my own life I've probably not been leading the quite the story that was enabled by the kind of joint stock company because I've had a company without investors and very much more of a you make it, you sell it type situation. But that idea of the joint stock company and the idea that you could incorporate, you could make a thing who would legally act like a person. But it was an abstract thing. It was a company that had its own limited liability. It's like the person can go bankrupt but that doesn't mean that all their friends are bankrupt as well. And so for the company it was kind of like the idea is you can incorporate it so that it is treated in the law much like a person, so to speak and who can independently go bankrupt or not. And it doesn't kind of and it doesn't flow through immediately to the, you know, to the owners of the company, they're a sort of separate layer. So that that kind of idea is, you know, a long time idea. And you know, in every country, you know, in the US it's you know, comma inc. In England it will be comma Ltd limited company in, you know, Germany. What is it GmbH in can I do a much spa is I think Italy the and I don't know what most of these things stand for but in every country it was kind of a designator to say this thing you're doing business with is not really a person, it's this incorporated thing and it's designated that way. Well, that was kind of a, you know, that was a major innovation. Then another thing was public companies where instead of it being kind of named specific. I just invested in your company. There's a public market where you can buy and sell shares in companies. And that was the thing that became, I would say, well, by the late 1800s that was certainly big. But there had been previous examples going back to the 1600s. You know, famously, the South Sea Bubble was something a little bit like that. I was really charmed. You know, there was the whole mechanism in the South Sea bubble where what happened. Yes, what happened? I think the, the UK government, the Crown, the King at that time had a lot of debt. I think they'd been involved in some wars, something like that. Been involved in a lot of debt. Meanwhile, there was the South Sea Company and the South Sea Company was going to sort of was going to explore and exploit the kind of the value from the South Seas, including many unsavory kinds of things that, that were going to be done. But in any case, the South Sea Company was like a. You know, this is going to be amazing. This is going to be, you know, this is a new thing in history. It's never, you know, it's, it's, this is, this is finally new and people started investing in it and the shares of the South Sea Company climbed in value dramatically, really dramatically. And then in a thing that just has a resonance with things that have been done quite recently, the King had this great idea. Let me swap kind of interest in the South Sea Company for tea bills, for essentially treasury bonds, for bonds for debt from the country. And so it's like for you investors out there, you know, it's, it's like, you know, I've got a deal for you. Instead of you holding on to this government debt that maybe it won't even be repaid, it's all kind of crummy. Get on the rocket ship, you know, go exchange your government debt for South Sea shares which are going to the moon. And so a lot of people did, including Isaac Newton, actually. And needless to say, it wouldn't be called the South Sea Bubble unless it had popped like a bubble, which it did because there wasn't real value there. And. Well, and I think it was probably also mismanaged and had all kinds of other craziness to it. And eventually there are people who made a lot of money in that whole run up to the South Sea Bubble person, what was his first name? I don't remember. Last name was Guy. And that's the person who endowed Guy's hospital in London, for example, was somebody who managed to ascend the South Sea bubble and get out before the pop, so to speak. But I was really charmed, I have to say. What was it? I think it was the people issuing cryptocurrency and exchanging it for sovereign debt was a thing that in the heyday of cryptocurrency of a few years ago, that was the thing that people were seriously talking about. And I was like, that's really amusing because it's the South Sea bubble all over again and people don't learn from history. Well, that's perhaps why maybe that could be seen as sort of the public service of live streams like this is. Let's talk about history because maybe one can learn something from it and don't repeat the South Sea bubble, so to speak. But in terms of the organization of companies, the. There were other kinds of pseudo governmental companies a little bit more like the way I think many companies operate in China right now, which is kind of interwoven governmental sort of private entities. And I think the, like the East India Company, for example, was a British one of those. And maybe there was some Dutch. I'm not sure if I remember that history well. But you know, these, these companies would get very big and they would start having armies and stuff like this. And that, that was, that was a thing. Now, you know, we got to sort of the stage of public companies more and more kind of ornate mechanisms for kind of the flow of money. You know, venture capital, that was really a thing. I guess venture capital got started in the 1950s, which was again a kind of a professionalizing of what had been just hey, would you like to invest in my company? Was kind of more like there are people who are going to put money. There are people who are saying, I've got money, I need to invest in something. What can I invest it in? Let me invest in a venture capital fund which will have people who manage it who go actually make investments in startup companies or whatever. That became a thing. I would say it became big by the 1980s and that was sort of another piece of the mechanism of companies. Now, you know, there have been in the U.S. for example, there are different kinds of companies like the traditionally S Corps and C Corps, S Corps kind of small businesses where kind of the tax that gets paid, gets paid by the individuals, whereas C Corps are sort of bigger ones which can have ownership from things other than individuals, where the tax gets paid by the company type thing. You shouldn't get me started about tax and R and D and the disaster of the U.S. r&D tax credit and the absolute I would say irresponsible kind of inability of the US Government, even though I think everybody kind of agrees about how this should work, this kind of whole R and D tax mechanism. The tax credit mechanism expired. And so that caused endless trouble for people like me dealing with companies that spend a lot of money on R and D and how that gets handled for tax purposes. But let me not get started on that. But any case, the thing that sort of developed by the 1990s was sort of increasing elaboration of kind of the concept of a company or like in the US the concept of an llc, limited liability corporation as kind of a more flexible vehicle for setting up a company with shareholders and so on, or actually members, as they're called in that particular case, and these sort of different mechanisms doing these things. Then there have been kind of, you know, nonprofits have been a thing in the US for a long time. Don't know quite their history. The sort of key designation is 501C3. That's part of the. I mean, there are 501C lots of other things. But 501C3 is kind of the scientific, educational, kind of R and D oriented designation of a thing which in the US the tax law says if you give money to one of those things, you can deduct it from your taxes, even though it's a philanthropic thing. That's not true in other countries. It's a US Feature which has been very helpful to the buildup of lots of kinds of philanthropic efforts in the US but in any case, so that's a thing. And the governance of non profits is a quite separate world from for profit companies. And it's had its share of. I mean, it has a lot more transparency in the US than for profit companies in other countries. There's more transparency of for profit companies as well, but. And there's sometimes there's even more transparency of charities and non profits in other countries, but even more regulation about what they can do and so on. But in the case of. So then there have been things which are kind of combo versions of these kinds of things. Like for example, B Corps are a thing of maybe a decade ago. So one of the issues with a corporation is if you have stockholders in the corporation, then you, as a, as a manager of the company, have a. Or as a board, part of the board of the company, have a fiduciary responsibility to the people who own the company, so to speak, the shareholders of the company, to make the company do good things. Now, you know, that can be Very long term, you can have a calculation that, well, in the short term it will be bad. In the long term it'll be good. So long as you use reasonable business judgment, nobody can really complain. Sometimes they'll try. But really, in the end, the law kind of says that if you're using reasonable judgment, that's you can do what you want to do to operate the business. But one of the issues there is that that responsibility is really just to the owners of the business. It's not to, for example, customers. It's not to the world at large. There's no. It's not to the environment. It's not to any of these kinds of things. It's just like, are the decisions I'm making ultimately going to benefit the shareholders of the company? And that's, you know, it could be the case that it matters a lot to keep the customers happy because that is something that, you know, is obviously important for the commercial success of the company. It, it isn't sort of for the good of the world. It might actually be at cross purposes with the greater benefit of the company itself. So this idea came up of B Corps and where the kind of the charter of the company says that in addition to it being for the benefit of the shareholders, it also has other kind of other responsibilities to the world, so to speak, typically to be a public benefit corporation, so to speak. OpenAI did not set itself up as one of those, I believe, although it did have sort of a mission statement and it had this very complicated structure of a nonprofit that had a capped profit entity underneath it. The. I found it somewhat amusing, a quote from one of the board members of OpenAI who said in response to, look, you know, the things you're doing can destroy the company, said apparently, well, that would not be inconsistent with the mission statement of the company. For it not to exist will be compatible with the mission statement of the company. Probably. I haven't read the mission statement, but it probably has to do with, you know, we're going to make AI for the benefit of the world and we're going to make it be good, not bad type thing. Thing. And, you know, if we can't make it be good, well then better not to exist. So I think that's a, that's kind of a. A thing of that term. But it's an, it's an interesting problem if you're trying to make a company where you believe you in informing the company want to make something that is going to benefit the world as well as benefits of the narrow Sort of commercial interest of for example, particular stockholders. I mean, I know in my own sort of life and times I've been interested in doing things that I hope benefit the world. And you know, in that particular case, because, because we have a simple private company, it's kind of just on me pretty much to say I'm going to do these things which benefit the world and I don't really know for sure that they're going to benefit the company. I kind of think they might, but I, I don't know for sure. You know, making Wolfram Alpha free to the world, for example, it's not obvious that that would be a benefit. I don't know, it's hard to tell. You know, it would be great if people use Wolf Malpha and then, then immediately started buying products from us. That's, you know, that pathway is, is not one that one can so convincingly draw. I mean one. So you know, that decision to make Wolf now for free to the world or to make math world or to make a lot of things I've done in basic science and to make these things free to the world, it's like just give it away. Those are decisions which are things that I hope and believe will sort of benefit the world. And I think that's a good thing to do. I can't say whether those are things that, you know, if we were a company that had kind of public shareholders, for instance, I'm sure we would get many questions about, well, is this really benefiting us, the shareholders, or is it just benefiting the world which we shareholders don't necessarily care about? So I think that's, you know, it's a little bit of background there perhaps. Anyway, so just a story of those things. I could give you some other. I'm not going to give you the gossip channel of the story of OpenAI. I think it's one of these things where. It's, well, let's just hope it all has a happy ending. Nightmare is commenting. I'm planning to write a fantasy story about the Antikythera device. Cool. Sounds like a fun thing to do. Oh boy. Crab is asking about effective altruism and I'm not sure I can represent effective altruism terribly well. And it's a couple of Oxford based philosophers who've been, I think, primary pushers. One of them actually is also interested in, in kind of sort of theory of computation meeting fundamental physics. So I kind of occasionally run across that person in that context and I hadn't even made the connection that he was also a big effective altruism person, I think effectual altruism. It kind of sounds cool. And it's kind of like, I suppose a rather unfair summary would be lead your life by making a bunch of money and then giving it away to useful causes type thing. And don't feel bad about making a bunch of money if you're going to give it away to good causes. That's probably an unfair sort of micro characterization of the approach. But at any given time in history there are these approaches that are kind of like, they just sound good and particularly they sound good to the young. It's kind of like, well, like communism was one where certainly in England, I think in the US as well, in kind of the, let's say 1930s and so on, 1920s maybe. It's like this sounds cool, this is a good theoretical idea. The fact that it maybe hasn't worked out so well so far in the Soviet Union, people didn't necessarily know that and they were like, well, they just didn't get the implementation right. It's still a good idea. And I think that there are these things that kind of seem like they're good ideas because they seem kind of a modernizing, abstract, abstracting kind of concept. Another one that I've run across from time to time is this idea of holacracy as a way of organizing companies. You know, companies tend to be organized hierarchically. There's a CEO and they have people reporting to them and they have people reporting to them and so on. You kind of draw a tree. I have to say, in a company like ours, it's more of a graph than a tree. And it's not as regimented as it might be, I think to its benefit. But in any case, one of the sort of more extreme versions of this is a so called holacracy where the concept is there is no hierarchy. The company is run by a collection of intersecting circles of people. And there are some companies that have been tried to be run like this. Zappos, the shoe company was run like that. Consensus Ethereum blockchain company was run like that. I would say that the on the ground experience has been extremely poor for that kind of approach. You know, these are things where it sounds good, it sounds cool, it sounds like it solves a bunch of problems. But by the way, this is typically the story of kind of history and things that develop over long periods of time. Often the conventional way of doing things by sort of a natural selection process has ended up being fairly sensible. Now the world changes. So sometimes these things that did make sense in the past stop quite making sense. But usually kind of these radical, let's just do things very differently things, particularly when it comes to sort of human and social organization types of ideas, they often don't end very well. And obviously effective altruism, well, it's, you know, the, the FTX story was in supposedly very much a story of kind of sort of following the effective altruism path. And I, I know some of the people who are involved in that and I would say that there was a, you know, a, I mean, I would say a good faith belief in sort of the value of effective altruism. And then it seems like some, some of this, the OpenAI story has come from the affected altruism direction. And I don't know, perhaps the way to say it is a lot of these things, they have a grain of interesting idea, but if you take them to too much of an extreme, they blow up in your face. And perhaps that's what's happening here. And of course the thing that can go wrong then is yes, there's still a grain of something useful. And then because one version of it blew up in one's face and it's like, that's a bad idea. Let's get rid of that thing altogether. So I don't know how that will land. All right, let's see. This is a meta question. Can you talk about the history of podcasts? Gosh, can I? You know, there's sort of a thing enabled by audio on the Internet, I think, you know, blogs were an earlier thing. Weblogs, as they were originally called, that was really the, the weblog story was I think in large part a. I want to put something up on the web. How am I going to do it? Oh, oh, I'm going to have to learn HTML, I'm going to have to get a server, blah, blah, blah. And then there started to be these platforms that basically let you sort of edit things with an editor and just say press the button and boom, it's a live web page. Social media, things like Facebook and so on also served some of that need and earlier than that, MySpace and so on. But you know, this was a, this was a particular form factor of delivering information. Where it was kind of the idea was it was a, it was a writing based form. And again there were, oh, several companies which have now long ago merged into things that were merged into things. And, and really, for example, WordPress, you know, became a thing that was sort of a second generation of kind of blogging platform type software. But it was kind of like, I want to put something out on the web and have it be easy to put out. And I don't think, you know, in the early years. When was that? Gosh, that must have been beginning of the 2000s, right? I think that was later than the MySpace story. And so there's that, and then there's kind of the audio version of that, of podcasts, and that required kind of Internet audio to work well. And then of course, live streaming required kind of, well, Internet video to work well, to tell you a little bit of a story about that back. Well, actually, funny, it's a small world. I first met with like, Sam Altman and actually Emmett Shear, who was briefly CEO of OpenAI. Might even be still today, I'm not sure. I met them both because they were part of the first cohort of Y Combinator entrepreneurs. Y Combinator was this operation started by Paul Graham, who I'd known previously. It was originally done in Boston, and it was kind of like, hey, college kids, come to Boston for the summer and start companies. And among the companies that were started in that first group were Reddit and, well, let's see, what were there? A bunch of them. And. And there was kind of. Anyway, so there was a. Why am I bringing this up? Oh, yes, I'm bringing this up because there were these two guys, Emmett Shear and Justin can, who had a company called Kiko, which was a web calendar company, which I have to say. Well, one of my kids who happened, who was quite young at the time, who happened to be exposed to all this stuff, said, that's the one to invest in. You should invest in that one. Which I almost did. But some people at my company who were like, well, we know about calendar systems, we use them all the time, were not impressed by the distance that this one had got. But any case that was. Those two were worked on that eventually they sold that company on ebay for like $50,000 or something. They started another company called Twitch. And actually in between those two, they started a company called Justin tv, which was a kind of a reality TV show of Justin K's life as seen by a camera attached to a kind of a hat that he was wearing. And that was sort of a gimmick type thing. That must have been. When was that? That must been 2007, maybe on that order, 2006 perhaps. And so that was kind of a gimmick that started sort of a live streaming of your life type thing idea. Subsequently that turned into Twitch, which then started what initially seemed like the absolutely bizarre idea of live streaming. Things like people playing video games and also live streaming now things like this live stream. But you know, as a sort of footnote to history, one of the places where Justin TV was used was in the launch of WolframAlpha in May 2009. We were sort of setting up this whole launch. We had no idea how many people were going to access WolframAlpha on day one. And we knew we had servers that would deal with a certain volume of traffic. We had no idea whether that volume of traffic was correct and we had no idea whether the whole thing would just sort of fall over and die based on the amount of traffic we had. So we thought let's at least make it interesting for the world to see this process happen. Let's livestream the launch of WolframAlpha. So we built this actually that caused us to build this kind of nice mission control like place with very nice displays showing kind of traffic volumes and so on. We still use many of those displays today. They were kind of conveniently built and they're nice and ergonomic and so on, but they were originally built for sort of the made for TV version of the launch of Wolf Malpha, which we streamed on Justin tv. And I'm sure you can still find that live stream. It was for a while it seemed like it was going to be the world's most boring piece of television because all that was going to happen was at some moment, you know, we're going to press a button and the thing is going to be live. But actually, of course things went wrong. The most spectacular of which was that this was the actual sort of data center we were using was in Illinois. And that day, that afternoon, that early evening, there was a tornado that was approaching our location. And right at the time it was expected to arrive, right at the time we had sort of advertised the launch and actually the tornado went a different way. So that became a non problem. But it was for a while it looked like a quite dramatic problem. But anyway, that was sort of a piece of the early history of. Live streaming and so on. All right, we should probably wrap up for now here and next time I'll be back in my natural habitat and look forward to talking about some of the things people have asked about here today and more things in the future. So thanks very much and bye for now.