History of Science and Technology Q&A (March 20, 2024)

Sam. Hi, everyone. Welcome to another Q and A about history of science and technology. I am not in my natural habitat today, so all sorts of things could go wrong, but we'll hope for the best. Let's see, all sorts of questions here. Doug asks, has any species evolved in a measurable way since humans have been observing and tracking evolution? Well, there's certainly some macroscopic examples of that. I mean, people always used to quote a certain kind of moth that as a result of pollution, things, trees had gotten darker, and so the moth that was darker was better camouflaged, and quite quickly, moths ended up being like that. But if you go beyond sort of the big critters, there's plenty of evolution that gets observed all the time. For example, the bacteria in our guts, our gut microbiome, is always evolving. And in any given person, it evolves quite rapidly. And depending on what you eat, it will evolve differently. Not particularly well understood yet, but that's something where you get to see evolution even within a single person. Same thing with tumors within a single person. This is one of the sort of bad stories of tumors, is that there will be evolution. You apply some chemotherapy agent. Oop. We have. I told you something might go wrong. This froze horribly. But so we seem to be back. Well, now you can probably see me again. Can you? Okay, well, we'll hope it doesn't freeze again. Okay. In any case, in. In things like tumors, you can get the situation that you apply some chemotherapy agent and the tumor cells will evolve to avoid that same thing. You see, actually, there are these nice experiments where you can see you have a plate and you have some bacteria growing on the plate. Oh, this is frozen again. Oh, bad, bad, bad. All right, let me try changing something here. Sorry, I think we have to stop video and. Okay, hold on one second here. Okay, let's see. I think we're back again. Let's see whether this fails again. Up. We're back again. Let's see where this fails again. We were just talking about evolution. You can see, and I was mentioning experiments where you have bacteria on some plate and you start putting down antibiotics and things. You can actually see sort of different colors when you've stained them and so on. And bacteria that successfully evolve, they make little fingers where they're kind of one particular sort of strand of bacteria will evolve to avoid the antibiotic, and then they'll keep going, and then there'll be more and more of those, and then you put down some other antibiotic, and then more and more of those and so on, and so you can really see evolution happening in a very direct way in cases like that. I mean, evolution is a bit faster for creatures with asexual reproduction, like bacteria and so on, than it is for creatures that have. Well, I mean, it probably helps also that the time to replicate the bacterium might be 20 minutes or an hour or something like that, as opposed to much longer periods for bigger critters. I think the. I've certainly heard of, of claims of lots of bigger critters who evolve over a very small number of generations to change their characteristics in response to some environmental change. I mean, it's worth realizing that when you have both genomic evolution and you have kind of evolution in the learning that happens in a species, so to speak. So in a sense, the genomics, you're locked in for a given generation, but what you learn is something that can change even as the fashion changes, so to speak, or the herd goes in a different direction. And that's something where you've got both the genetic evolution and the kind of evolution of the memes, so to speak, that are sort of separated from that sort of the cultural evolution is separated from the genetic evolution. And that can happen an even faster timescale of the creatures learn how to sort of get their food by opening this door or something, and then that gets passed on to the young creatures, independent of kind of the genetics, independent of natural selection operating. I mean, one can ask the question, when one looks at kind of the mechanisms of natural selection, people have kind of. Well, there's sort of a question, what's the fundamental theory of natural selection? I mean, the idea that originated with Darwin that, well, people had precursors of this idea that the species we have now are the ones that made it through kind of the gauntlet of having had the most kind of children, so to speak. And the forms that didn't make it are the ones that had been unfavorably selected by the characteristics of the environment or the other creatures in the ecosystem or whatever else. And that was something that I think Darwin really understood with clarity for the first time. I mean, Wallace was sort of on the track too, but that was something that was sort of a mechanism of you get what you get, because that was the thing, because there's this mechanism of natural selection where the form of the environment determines what you get. Now, at the time, there was plenty of artificial selection. People had been breeding all kinds of domesticated animals and pigeons and all sorts of other things. Darwin had all sorts of correspondence with pigeon fanciers and so on, who'd been breeding types of pigeons and so on. So it was very well understood that you could breed creatures to have particular characteristics of, you know, the fluffier cat or whatever else. I'm not actually completely sure that cat breeding was a big thing yet by then, but in any case. But the idea of artificial selection that breeders had been doing was well understood. The thing that Darwin really added was the idea of natural selection. That the environment of the world could determine what species would be successful and what would not. But sort of the question of, could you turn that into some kind of quantitative theory of what was going on? That was a different question. I mean, Darwin famously sort of thought that this process of natural selection was the thing that led to all the sort of amazing complexity of life on Earth. And he has the last sentence of Origin of Species. His book from 1859 says something like, as the Earth keeps circling around the sun according to the fixed law of gravity, so. So species have been and are being evolved. So his kind of concept was just like kind of this law of gravity. That was something that one sort of believed was a global feature of science. So this phenomenon of natural selection would lead to progressively more complex organisms. There's never been evidence that that's true. And the fact is there are pretty simple organisms that have been pretty successful throughout biological history. And it does not seem to be a necessary feature that sort of you have to get more complex organisms to be more successful. Now, it's an interesting question whether, in a situation where the organisms are competing with each other, whether that sort of forces one to have somehow a more complicated program for the organisms. That's still something, I think, to be determined. But Darwin at the time didn't really have a kind of a hard science, kind of mathematicized version of what natural selection meant. In fact, there's a nice picture, sort of a sketch that Darwin drew of kind of a tree representing the different sequence of organisms. His son, Francis Darwin, I think it was, that son was kind of more mathematically inclined. And that was kind of. And Charles Darwin worked a bit with him on these kinds of questions, but never really got sort of a mathematicized formula formalization of the idea of natural selection. And nor has anybody since, really. I mean, there are a lot of details about kind of using game theory and things like that to figure out. That was something that was very popular in the 1950s and 1960s. After the development of game theory in the 1940s, game theory had been developed for economic purposes and sort of strategic military purposes. Of kind of, there were these sort of payoff matrices of how well would you do if you played this versus that against your opponent? And that kind of idea turned out to be a good representation of sort of the competition between. Between members of a species and so on. And that led to all sorts of predictions about sex ratios and different kinds of species and so on. That was, I guess, had matured by the early 1980s because I interacted a bunch with a bunch of those kinds of people who were very proud of the fact that they'd successfully sort of mathematicized this aspect of biological evolution. I think this whole question of kind of to what extent can you sort of formalize it is an interesting one. Another thing that was sort of a different branch was the idea of genetic algorithms. The idea that you could have programs where you would get to the program you wanted by doing some kind of artificial selection where you say, this program isn't quite right, let me modify it. Even the idea of not only mutations in the genetics, the program being the genome, not only mutations in the program that are things like single mutations, point mutations, but also things like crossover of the kind that happens in sexual reproduction where you're taking a big chunk of genome and you're putting it. You're sort of copying it over to another. You're taking two different parents and you're sort of mixing together big chunks of genome from one parent or the other. And so a bunch was done on both. First, genetic algorithms. Genetic programs. Genetic algorithms. Trying to sort of see whether you could sort of evolve to a successful program by just applying constraints. And there were a bunch of successes, particularly in the 1980s and so on. 1990s successes in designing analog circuits. Kind of the big names in this area, both of whom I've known, John Holland and John Koza. John Holland, more on the kind of genetic programs. John Koza, more on the genetic algorithm side of things. And there were many kind of applications for sort of computer game kinds of purposes and so on. I think in the end, it's the sort of some successes in that area. One can think of kind of neural nets as being. And what's done with training of neural nets as being sort of a big version of something a bit like that. Except that in typical genetic programming, the programs are somehow discrete. Whereas in neural nets there's kind of this idea, at least in sort of the current neural nets that have evolved from the 1943 McCulloch Pitts neural nets, there's the idea that the weight are continuous numbers and you can kind of tweak them incrementally in sort of calculus like way. And that the thing that in genetic programs, genetic algorithms and so on, one of the issues is a program is something you write in terms of discrete, typically is something you write in terms of discrete constructs. And so what does it mean to make an incremental change? That's a tricky issue, I have to say. Actually, I've recently been doing some experiments on this myself because I'm sort of curious. What is the minimal model that you can make for biological evolution? And you can have very simple programs where you can say, can I reach this particular objective? Can I make the program do this funky thing? And how do I get there? And what you see, and it's very typical of what you see in machine learning these days, you can invent a kind of a loss function that says how far away is this program from achieving what you wanted? And then you can say, as you sort of evolve the thing, how much can you reduce that loss function? And so when you deal with neural nets, you have a lot of strategies that involve back propagation and calculus and derivatives, automatic differentiation, all this kind of thing that tries to kind of know exactly how you tweak the structure of the quotes program in the case of neural nets, the weights and so on to achieve the objective you want. But there's a much stupider thing you can do, which is what biological evolution seems to do, which is you just randomly make changes and you see what happens. And so you can ask the question, if you're in some state and you've got a program and it's not really doing what you want, but it's kind of close. It's, can you just try 100 random mutations and just see which one does the best of those? Or maybe the one you already had did the best. But let's say you have one of those other ones that does better, then you pick that one, you bet on that one and you keep going and you see whether you can evolve to actually be able to satisfy the constraint. Now, as stated like that, it's a very simple, simply stated problem. In reality, it's more complicated than that because in reality you don't just bet on one organism. There's not just one member of a species that's the sort of carrier of this is the creature. If this species is going to make it, this is the one member of that species that exists and is going to have progeny and things like this. Instead you have a whole range of members of the species. And at any given step, you're kind of betting on many organisms and you've seen what happens to each of those. And the one that you thought was, was last may come out first in the next generation. The kids may be smarter than the parents or whatever else, or better than the parents, even though the parents were kind of lame. The kids may be really good, and the parents who were great may have kids who are only as great as they were, rather than being able to jump ahead and so on. And so the definition of what's going on gets considerably more complicated when you start dealing with. It's not just the best of this generation is it's several different sort of members of the, of that generation and so on that you have to sort of carry forward. And I think one of the things that's interesting is you might just get stuck. You might get to a point where it's kind of like you've evolved to this point and you. There's just no way to get better. You're just, you're in this, in this sort of basin where in this mountainscape of, you know, can you get to the top of a mountain or the, or the, or the bottom of some, some bottom of the hill, so to speak. And there's just no way to get there because you're stuck in this kind of area with a bunch of walls around it. The thing that tends to happen, and I think the main sort of trick that biological evolution uses and that neural nets and machine learning use is by the time you have enough parameters, the chances are there's some direction to go in that gets you to the better place, so to speak. And I think it's the very, in a sense, it's the very complexity of biological organisms that makes it possible for evolution to work. If organisms were much simpler and you just had three parameters to tweak, it'd be much harder to actually escape the goofiness or something and get to the point where you have something better happening. So anyway, a few thoughts about observation of biological evolution. And, you know, it'll be an interesting thing to see whether one can more successfully, in a sense, formalize the process of evolution and the description of sort of what is really going on in evolution. My friend Greg Chaitin, originator of kind of algorithmic information theory, has kind of an approach he calls metabolism, where he's asking questions like if you have a program that, for example, runs for a long time and then halts and your sort of fitness function is run for the longest time and Then halt. One of the features of that is there's never a program that is the all time winner. There's never a point. You can always make a program that runs for longer than any program before and then halts. The program will be progressively more complicated, but you'll never sort of run out of, of possible programs that will always do better than the ones before. And in fact, as you start looking at the kind of the how long it can go before it halts, you're kind of ascending into numbers that are sort of absurdly large and even in some sense not computably large. So it's kind of a model for there's always a better organism. If that's the fitness function, you'll always find a better algorithm, a better organism. So it's kind of like there's no end point. There's no limit to biological evolution. I think one of the things that's sort of an issue in this kind of modeling is one thing is what's the environment of the world? What are the niches that organisms can exist in? Do you say there's a fixed set of niches and the organisms are going to optimize themselves relative to those niches? Or do you go even more complicated and say, given that there are these different organisms that are interacting with each other, how does that then work? And do you get a situation where sort of this collection of organisms is the only one that you can get to and you'll never get to a better place where there's another sort of branch organism that gets even better? I think the sort of, the intuition from this sort of metabolic approach and from thinking about sort of evolution computationally is there's never an end to evolution. There's always a way to make something better, so to speak, and with any definition of better. But that's something which is still for the future to really formalize what all of this means. And perhaps that formalization will be important for some of these cases of very rapid evolution, whether it's in the microbiome, in tumors, in antibiotic resistance, these kinds of things, or for that matter, in viruses, which also have very fast evolution. And I mean, just to be clear, I think the original question was about, have we observed evolution happen sort of on human timescales? Well, absolutely. In the case of the COVID virus, SARS 2 virus, there were endless trees that were drawn about. This is the evolution of all these different strains of the virus. And it was like, here was the actual person in which that new strain developed, and so on. And then you got to see the more and more complicated tree of what happened. Let's see, there's just one comment here, but this is more about general science than history of science. But I'll make talk about this anyway. Second asks, there's more bacteria than blood cells in. In us humans. Why? You know, it's always a funny thing. You talk to people who study kind of, I don't know, biology, physiology of mammals like us, and they say there's so many cells, there are trillions of cells in the body, and we study all these different things. Then you talk to people who study prokaryotes and bacteria and so on. They say, by the way, you should know there are more bacterial cells in a human than there are eukaryotic human cells in a human. And then you talk to people who study viruses, and they say, by the way, you should know there's more viruses in a human. Viral pieces of virus, little viruses. Then there are bacteria, which is again, bigger than the number of sort of human cells. So it's partly because the bacteria are smaller than human cells. Viruses are smaller than bacteria. I mean, human cells have a huge range of different sizes, from giant muscle cells to little tiny red blood cells and so on. Very great range. I mean, even though it's the same DNA, what's transcribed, what actual proteins get made is very different in different tissues and so on. The question, I think there are probably more bacteria than blood cells because there can be, because they're smaller. And it's probably as simple as that. Let's see, Jamie is asking, this is again about evolution. Have humans aided in the evolution of domestic animals? Well, gosh, I mean, you know, humans have done a lot of artificial evolution. I mean, the, you know, one case of artificial evolution is that one might consider controversial is things like gain of function, mutations for viruses and so on. That's a case where artificial selection was used to kind of pick out the virus that was the winning virus. In terms of domestic animals. Well, yes, people have created all kinds of different critters. You know, the dogs that have this characteristic or that characteristic, the cats with this characteristic, that characteristic, you know, the snakes with this kind of different pattern on them and so on. And, you know, can you make a snake? People haven't gotten to the point where they can make a snake that spells out somebody's name on the, on the, you know, the pigmentation of the snake. But. And that's sort of a challenge for theoretical biology, I suppose, to say, could you imagine doing the reverse engineering if I want the snake to spell out hello and to, you know, grow to have that pigmentation pattern. Can I figure out what I should put in the genome to make that happen? And we're not. We're not there yet, but in. I think the. That's part of the story there. Let's see other questions. The question from without here, have all of Grothendijk's works been understood yet? Okay, so Alexander Grothendijk was a mathematician, French mathematician, who was probably one of the most abstract successful mathematicians that there's been. And he invented kind of a way to really think more generally and more abstractly about lots of kind of areas of mathematics, bringing together these areas of mathematics. I mean, basically what happens in mathematics is you'll have something like algebra, you'll have something like geometry. Well, it was known since Descartes, since the 1600s, that algebra and geometry have some sort of common. You can represent geometrical things algebraically, some algebra you can represent geometrically and so on. But as we look at different areas of mathematics, algebra, geometry, topology, all these kinds of things, there's a question of sort of what are the common themes between these different areas? And I guess Grothendieck's big contribution was getting to a level of abstraction at which these very different areas of mathematics seem to have common themes. And I think that's sort of one branch, a somewhat different branch originating in the 19. Well, so Grothen Dieck was, I guess, most active in the 1960s. He had a slightly complicated history, and I think I never met him, but he, he was wound up at a place called ihes, sort of top French math research institute, and was sort of a big fixture there and sort of generated a lot of very, very magnificently abstract mathematics, which, fortunately for him, was. Was fairly well appreciated. It might not have been. It might have been like people just say, we don't understand what you're talking about. But no, it was quite well appreciated. But then I think he got rather disenchanted with things. He kind of went, as happens, from kind of thinking about pure mathematics. He started thinking about kind of the structure of the world and applying kind of mathematical, pure mathematical thinking to the way the world should be set up. And that doesn't usually end well because the world really doesn't seem to work like kind of the kind of abstraction and perfection of pure mathematics, so to speak. And so people got kind of annoyed with him. And he then went off and started hanging out in the sort of rural France, more or less, and Lived out many years, unclear quite what he was doing. But it's kind of an epic sort of tale of reclusiveness, so to speak. But in any case, during his sort of golden years of producing mathematics, he was producing extremely abstract mathematics in this particular direction. A different branch was the development of category theory, which started with folks like Saunders Maclean in the 1940s. Category theory is another kind of way of thinking about an abstract representation of mathematics, where you're just saying sort of, how do things structurally fit together? Rather than this is a particularly a polynomial or a triangle or something like this. But it's kind of what is the flow of. What is the kind of flow of these concepts connecting with each other? Now, it turns out that sort of higher category theory and Grothendijk's work kind of merged together at some point into a sort of a blaze of abstraction. I have to say for myself, I had thought, I'll never kind of need to know about Grothendieck's more abstract work, except it turns out that this thing that we kind of developed, that I call the ruliad, this kind of entangled limit of all possible computations, is also something that is essentially aligned with the most abstract thing that Grothendijk produced, which is what he called the infinity groupoid. And it's kind of a strange thing that this most abstract output of our efforts with our physics project and so on, and thinking about computation has sort of merged with the most abstract efforts of this kind of pure mathematics. And Grothen Dieck's main idea and conjecture is that in this infinity groupoid there is sort of an inevitable geometry to this thing, which is very similar to what happens with the ruliad and the inevitable emergence of physical space and so on from that. This is a long story to try and explain this in more detail, but that's kind of. That's kind of one direction there. Now, the question was, has all of Gothen Dieck's work been understood yet? I think the answer there is probably more or less yes, although I certainly don't know every corner of it, and that it was quite voluminous. But this kind of this real pinnacle of the infinity groupoid and so on, and sort of its merger with higher category theory and such like. I feel like that's something which, again, I had thought I would never personally care about, but it's something that has again merged with a lot of the abstract things that have emerged from our physics project and so on. Let's see, questions here about. Let's see. Question from O. Lemon about what was the earliest use of radioactivity for light, heat and so on? I mean, I think radioactivity. I mean, the. In terms of uses of radioactivity, the. Well, big use of not quite radioactivity was X rays, 1900. I think Rontgen discovered that what you got by crashing electrons into some target and getting something, in fact, X rays coming out that let you kind of like see through hands and so on. So that was one application, not precisely radioactivity. Then there were things like radium dials. This was a not very good idea. That radium radioactive element discovered by Marie Curie was something that has a pretty long half life. So it survives a long time, but it keeps on emitting radioactivity. And those radioactive particles, when they crash into other atoms, they excite those other atoms, and those other atoms can produce light. And. Well, particularly when you have some fluorescent dye and things like this in the picture. And that led to the idea of radium dials for watches and things which glow in the dark. And when I was a kid, for example, the 1960s, those were still around. And they were probably a profoundly bad idea because it's like radioactivity is not the. The same particles, radioactive particles that produce those flashes of light also will destroy the DNA in your cells and so on if they. If they interact with that. But I think the. Well, sort of the whole kind of story of the applications of radioactivity mean I think by the time there were alpha particles that were. That's one of the sort of channels of radioactivity. Those were used for many scientific purposes, like discovering the nucleus by scattering alpha particles from. And seeing that sort of most alpha particles, when they went through kind of a gold foil or something like this, just pretty much went straightforward. That was Ernest Rutherford's experiment. They pretty much just went straight through, but just some of them would hit something hard and they would bounce back. And that was sort of how the atomic nucleus was discovered. Many years later, in the 1960s, late 1960s, sort of a similar thing happened where people observed, actually this was beginning of the 1970s, that this was finally sort of clarified the. That when you bombard a single proton with a bunch of electrons, then again, the same kind of thing happens that usually the electrons just sort of more or less go straight through. But occasionally there's what was called deep inelastic scattering, large momentum transfers. The electron would bounce off something that seemed to be a hard thing inside. And that was what initially, around 1970, they were called partons by Dick Feynman and sort of they've been hypothesized quite separately by Marigold Mann and so on, that inside protons there might be these things called quarks. And it became by the mid-1970s, and this is something I was a bit involved in, it became clear that sort of the partons are in fact quarks. The hard bits of a proton are these seemingly point particles that are the constituents of protons, just as the nucleus is the constituent of atoms, so to speak. So it kind of was a case where history sort of repeated itself. But the real uses, I mean, there are many, many uses of radioactivity that have been discovered at different times. Whether it's for know, carbon dating, for using the fact that when in the atmosphere, there are most carbon atoms in the atmosphere are carbon 12, but some are carbon 14. And that happens because cosmic rays, mostly from the sun, hit carbon atoms and, or hit other kinds of atoms and end up with this slightly radioactive form of carbon. And if you're a tree, for example, and you're ingesting carbon dioxide every day, you, you ingest some of that carbon 14 from the atmosphere. If somebody cuts down the tree, you stop ingesting carbon 14. And the half life is what, 200 years, I think, I don't know, maybe 1400 years or something, I forget. But in any case, once the tree is cut down, the amount of carbon 14 in it will gradually decay off as opposed to being continually replenished because the tree is getting carbon 14 from the carbon 14 in the atmosphere. And so this was discovered as a way to date things like trees, wood, whatever, through radioactivity. There are other forms of dating that work with radioactivity. Various, I think potassium, argon, fission trap dating, I think that's dating by using again, the fact that there are some small amounts of radioactivity that get produced, for example, by cosmic rays, and that's one I think particularly useful for pottery, that before the thing gets kind of set, it's still ingesting these nuclei which are radioactive, and then afterwards they just decay same. Well, there are lots of kind of. Okay. Another sort of use of radioactivity had to do with understanding the heat of the earth. Why is the earth hot inside? Well, for a long time it sort of debated like I think Rutherford claimed it's hot inside because there are radioactive elements that were created at the time when the earth was created, and they haven't all decayed yet. For example, uranium has multi billion year lifetime and so it has uranium 235 has a shorter lifetime than 238, but it's sort of still decaying. And as it decays, it produces energy. It releases things with high energy that heat up the surroundings. And that's sort of one theory for why the Earth is hot inside. The other theory is just it was hot when it was formed and hasn't all cooled off yet. I think current belief is that both of those effects are there in roughly equal quantities. But that was a long time discussion of kind of why would the Earth be hot inside. And obviously, the big application of radioactivity was to nuclear power and nuclear weapons. And that kind of emerged in the 1930s as something that began to be understood. And the idea that one could have both spontaneous radioactivity and one could induce radioactivity by, for example, taking the particles like neutrons that have come out of the first radioactive decay and having them induce more radioactive decay. So that one got a chain reaction either slowly for a nuclear power station or very quickly for a nuclear weapon. Let's see, There's a question here. What current technological advancement will be most beneficial to ancient societies? Gosh. I mean, an awful lot of the technology we have today is built on top of a huge tower of capabilities. I mean, you don't get to make a Roman iPhone very easily. I would think that there are things like antibiotics that would probably have had a huge effect on kind of quality of life. And their supply chain is not so complicated. It's more kind of just knowing. Well, you know, penicillin was discovered by Fleming, as. I mean, it's one of these typical stories where it's just like, he left out this moldy bread and then noticed that the bacteria he was growing didn't like the pieces of mold that were there. And then sort of concluded that that meant you could use kind of mold as a. Things that came from mold, which turned out to be refined as penicillin as a way to sort of kill off the bacteria. I think the true story there is considerably more complicated. And as is often the case, the kind of one. Gosh, now I notice something. Story is really a story of kind of a long development of different kinds of thinking. And then you notice the amazing particular thing, and then you realize, yes, this is how things have to work. It's kind of interesting that even in current times, the kind of the idea of using sort of leveraging, the kind of the antibacterial properties of things like molds, funguses, things like this, this is very much a currently Interesting topic. And people are kind of doing things like going through the complete literature about every fungus that might have existed and then trying to use modern machine learning methods to figure out what might be relevant as a possible antibiotic source and so on. So this is a story that, even though it originated, I guess, in the 1930s or even 20s, is still a very active story today. So that would be one example, I think, in terms of other. I mean, so much of modern technology is built on this big stack of capabilities of doing precision engineering and this and that and the other. I mean, I think that if we look at the early stages of what led to modern technology, things like the invention of steam engines, the invention of various kinds of spinning and weaving technologies and so on, these are things that kind of launched the industrial revolution. And, well, also the discovery of which people. How do you get you can mine coal and do things with it and so on. But I suppose there's a question of, instead of recapitulating the actual history of human engineering development, if you could just go from today, jump ahead and then go back, what would you be most successful with? And I do think that when it comes to conceptual ideas, sort of the ideas of computation, the theory of computation, so to speak, the notion that you can set up these systems of rules and kind of deduce what happens from them and think about that abstractly, that's something that people were pretty close on in the 1600s, like Leibniz was pretty close on that in the late 1600s, didn't quite get there. And in the end, it took a whole chain of development and mathematical logic and so on before people were quite positioned right to get to people like Moses Schoenfinkel and Emil Post and Kurt Godel and Alan Turing and so on in the 1920s and 1930s. And that was. It was sort of. And then this abstract notion of computation emerged, which really was something that could have been discussed back in the 1600s, but wasn't because people didn't think of it. Building a computer, quite different story. I mean, building an electronic computer, just a huge stack of kind of engineering capabilities. I mean, I think when Charles Babbage, sort of as interpreted by Adan Lovelace, were thinking about the analytical engine back in the 1840s, that was going to be a giant. They called it the analytical engine because it was like a locomotive, like a steam engine, a giant mechanical device that could do computations, never got built. But that was something where, again, the concept of that as an abstract piece of computation, Ada Lovelace was sort of the closest to that idea. But one could have taken that abstract idea and kind of run with it independent of the actuality of making this giant mechanical engine to implement it. But so I think that would have been something that had that existed in the 1600s, the idea of computation calculus, for example, would not have been the only game in town in kind of formalizing science. And that would have probably been a very different history. And I think we would have had, well, many things that are just now beginning to be understood and developed, including lots of things I've worked on. A lot that come from this sort of computational paradigm would probably have arisen much earlier. But it's always this very difficult thing to take all the details of what's going on and abstract from it to something like the abstract idea of computation. In fact, I've just been thinking about some things recently where I'm quite certain that there is an abstraction that is kind of a little bit like kind of the abstraction of information theory or something. It's an abstraction that once you see it, it's going to be really obvious. But getting to seeing it is not easy. And that was kind of the story of computation. And the getting to seeing it was a multi hundred year adventure in that case and in the case I'm looking at, I'm sort of hoping it's going to be a months rather than centuries kind of proposition. Harry is commenting something I don't know much about. Getting to significant industrial scales was really hard for penicillin. I, I don't know, I mean I, I. That there's, it's a general feature of engineering that sort of, you know, first you get it to work and then you get it to work smoothly. You know, first you get kind of, you can make one of it and then gradually you improve the yield of what you're making and it becomes more viable. I mean, I know something I sort of really noticed was back in the 1980s, I had seen flat screen televisions. It's like, okay, they're going to be flat screen televisions everywhere. But there weren't. And the reason there weren't is because to make a flat screen television, every pixel has to work. And the failure rate of pixels was high enough that to make that one successful flat screen television was just, you took it, you had to scrap an awful lot of, you know, candidate ones that you'd manufacture that had some pixels broken. And it took took decade and a half before that got to the point where enough were working that you could really economically manufacture These kinds of things. And I think that's true of most kinds of engineering processes that, you know, when you start it, you just get something to work and then you progressively refine it to make it better and better. And I don't know the details of the making of something like penicillin, but I'm sure it worked in something of the same way. Let's see just a couple more things and then I should run along here. Let's see. Cody is asking what kind of technology present today do you see becoming nothing more than history? What are examples of previous technologies that have been discarded? Oh gosh. I mean, I have a whole collection of weird computer like things that people made at different times. One of the things about discarded technology is it has a terrible habit of coming back. You know, there were the computers that sort of had two pieces that unfolded. One was the display, one was the soft keyboard. And that seemed like it was a, like a goofy idea, but it's coming back or kind of the notion that kind of the, the idea of, of do you centralize computing or do you have it be kind of personalized edge computing type thing? And that keeps on oscillating back and forth. So you know, I don't think very, you know, these technologies that. Another great example is virtual reality. I mean I first played with one of those Systems around late 1980s and it was like, this is going to happen real soon. And you know, finally now we have sort of pretty good virtual reality type hardware. Not completely clear what the sort of use cases that are really going to be strong, that develop from it are, but it's something where it was kind of like you might have said, oh it's dead meat, never going to happen. But actually it's eventually going to happen. I mean there are things where you look at and you say that's really a bad idea. And maybe at the time it is, maybe with the way it's been built at that time, time, it's a bad idea. I mean, I, I, I'm, I'm, I'm reminded I have a cover of a Scientific American from this must be from about 1900 and has a picture of a guy with steam powered roller skates and basically says, you know, our, our steam powered roller skates are the way of the future. Well, obviously it didn't, you know, didn't work out that way right then. But you know, you see people with all sorts of different, you know, personal locomotion devices that do or don't work out. Another good example of this was The Segway kind of sort of gyroscope adventure of, you know, the wheeled object that you could just sort of lean forward on and it would go forward and it would keep its balance and so on. Kind of a derivative, I think, of various kinds of go upstairs wheelchairs and other technology like that. But that seemed like a really cool thing and, you know, for some case or another, and. But you. And you see them occasionally, but they didn't take off the way one might have expected. Now, will they take off in the future? Maybe, yes. I think it's hard to know. I think that the amount of technology that's just, this was a bad idea. Let's just not do it. Is perhaps not so large. It has a habit of coming back. I suppose a case where this does happen, where it's like various kinds of medical interventions. Oh, that particular piece of our physiology isn't important, let's take it out type thing. There have been a whole series of episodes of that where it's like, we don't know what it does. You know, it got infected, let's take it out type thing has been. Or it. It's doing bad things, let's take it out kind of thing has been. That's. That's a case where there are kind of theories that. Those are more biological theories, biomedical theories, where they, you know, they just turned out to be a bad idea. And usually the reason they turned out to be a bad idea is because the absence of knowing what something is for doesn't mean it isn't for anything, so to speak. It just means you don't know what it's for yet. And so that's one case where I think things that. And I could guess about some things now that people do or don't do, for example, for medical purposes, or where people will eventually say that was just a really bad idea. You know, throwing this drug into the situation was just a really bad idea because, you know, it was, you know, you thought you knew what it was doing, but in fact it was doing lots of things you didn't know it was doing and so on. So a few, a few comments there. I don't know what. In terms of kind of, well, things, I mean, there are so many of these, you know, is it a good idea, Is it a bad idea type things. As I say, the, the amazing thing is that these ideas come back. Most of these ideas come back. And, you know, whether it's the quad. Quadcopters which people looked at, I don't know, back in the 1940s and so on. Couldn't make them work at that time eventually. Batteries are good enough and control systems are good enough. Yes, you can make a little quadcopter drone and so on. All right, I should probably wrap up here, but thanks for joining me, and I will hope to see you on another live stream sometime soon. So bye for now.