Transcription
Well, did we fix that, the weak spot problem? Is that, does that have to be there where we can shoot it once and pull the whole thing up, or can we cover that somehow? They really, really needed some engineering droids to take a second look at it. But again, that, when, when you're, when you're under time constraint, you have to cut corners, and they just, they just said, you know, it's shipping. That's, that's Death Star V1.
If there are some jobs left, if a good chunk of today's jobs go away, then what do you do if you have a 20, 30, 40% unemployment rate? That requires an entirely new economic paradigm. It's the golden age of Athens. About 10% of the citizens were, you know, landowning wealthy men. And what did they do? They went to the Elucinian mysteries to trip balls, and then they debated politics.
The universe seems to, when you look at, uh, delayed choice and, and retrocausality and those sorts of things, uh, non-locality. You look at all those experiments, and it seems like the universe doesn't make a calculation until it absolutely needs to, which is what we do as video game engines.
The opinions are going to be really interesting once you kind of got out there and done all that. So, I have to be on my best behavior. No, tell them that. We'll see. I feel like you have to just tell them what you think. Strap in. Oh, wait one second. Let me just make sure. Yeah. And everybody knows how Riverside works. So, you wait for it to upload at the end. So, we're good on that. Um, okay. So, David, can you start by giving us your name, rank, position inside of Starfleet?
Uh, David Shapiro, I am captain. Um, not of a, uh, executive frontier cruiser. I'm more of a captain of a scientific cruiser. So, there we go. All right. Where do you, uh, do you hope to be in charge of the entire Federation at some point, or are you happy in your position? Admirals are all corrupt now. Okay, that's awesome. All right, perfect intro.
Okay, so Dave, thanks so much for joining us for this kind of emergency session because I know a lot of people are very kind of scared right now. Dario Amade was on CNN with Anderson Cooper, and right next to it is an article saying, "Oh, what Dario has been saying is just hype. It's AI hype." And I'm like, how can people not be confused about what's about to happen? So, first question, is it just AI hype, or should people start thinking about where this is going?
Yeah, I mean, to, to not really add much clarity, it's a little bit of both. Um, you know, there's, there's always inflated expectations when there's a new technology, and Silicon Valley is the worst. Um, namely because it's, you know, the technologists themselves are always saying like, "Oh, this is going to solve literally everything." And if you ask a developer like what, what is their definition of everything, it's, you know, it's a, it's a very small subset of reality. Um, and so that, you know, from a certain point of view, right, it, it is going to change a lot of things. Uh, at the same time, um, you know, to, to just kind of answer the question directly, I've been doing a ton of research for my books on post-labor economics, and we've actually found evidence that there's been about seven decades of, uh, demand for human labor slowly being eroded by automation. Now, AI and robots or humanoid robots are just the latest iteration, but industrial automation is nothing new. And so when you look at the data, you actually see that labor force participation rate in prime-age men has been declining since about 1953. Um, you see there's plenty of other little, little clues. Uh, but wages have also been eroded over that period of time. So we actually have plenty of evidence that automation has been eating into, uh, the, the wage labor negotiate or social contract for a long time. Um, and so AI and robots and everything else that's coming down the pipeline is nothing new. It's just the next iteration that's going to eat into it even further. So whether or not it's going to be, you know, next year or 10 years from now, like it's a very long-term trend. Yeah.
So what do you think that Star Trek gets right about post-labor economics? Like what does it get right, and what does it get wrong?
Um, I, I from an economic standpoint, it gets most of it wrong. The thing that it gets right, I think, is the lifestyle, honestly. You know, whenever you see people on, whenever they have an episode, uh, where they go back to Earth and everyone's like, you know, being an artist or hanging out at a bar or shooting pool or, you know, or, or following a career that's based on something that they care about. I think that's probably the most realistic thing. Um, but I mean, it, it Star Trek economics does not stand up to any scrutiny for longer than about 30 seconds because it's like, how does Picard own a chateau in a propertyless future? Like, do, is, is it just inherited titles or something like that? So it just falls apart, uh, with, with the least bit of scrutiny. But I think the lifestyle is what is what it really gets right.
Paint us a picture, like, let's, from the highest level, the post-labor, post-labor economics kind of, um, vision that you've been sharing, like what, what should we be thinking about right now?
Well, the biggest thing is if and when it's going to happen. And so, kind of the mantra that I've come up with is better, faster, cheaper, safer. And so, historically, whenever a technology or new way of doing things surpasses the old way of doing things on those four metrics, it usually becomes inevitable. Uh, you know, for instance, switching over from horses to tractors and now tractors to combines and, and beyond combines, we're having fully automated, you know, AI-powered, uh, farm equipment. Now, uh, each of those cases, there's, there's usually there's a few ways of looking at it. So, time, you know, it's something you can measure with a watch, there's energy input. Um, if something is more energetically efficient. Now, that's not always true because, you know, a tractor consumes a lot more energy in terms of diesel fuel than a horse does, you feed a horse grass, right? They're relatively cheap to feed, but they have better output and it's faster. And then it's also safer because you don't have zoonotic diseases and those sorts of things. Although industrial accidents do happen, but you look in the, in the longer term, there's far fewer, uh, farmers, right? So human labor is a huge component. So that's all going away as well. So that's the, that's the kind of part one is better, faster, cheaper, safer, machines replace humans. We've seen this time and again over the last few centuries.
So the next component is, okay, well, if that, if you take that out to its maximum conclusion, then you end up in a situation where who has jobs anymore, right? Or, or even if there are some jobs left, if a good chunk of today's jobs go away, then what do you do if you have a 20, 30, 40% unemployment rate? That requires an entirely new economic paradigm. Uh, and so that's what we call the economic agency paradox, which is basically technology makes goods and services cheaper, uh, but at the same time, it takes the wages that you'd use to pay for those new goods and services. So then nobody has any money in order to pay for those things that because they lost jobs, and so the whole economy kind of breaks down. Uh, there's certain ways that you can look at it. Aggregate demand. Um, and it's not just aggregate demand, which is basically how much do all households have in reserve money to buy goods and services. The another way you can look at it is when, um, is the sources of those in of income. So right now the vast majority of, of income comes from wages, uh, with some coming from property and others coming from government transfers like, uh, SNAP and social security and that sort of thing. Uh, the entire economy right now is predicated on the idea that wages are the primary way of distributing income, uh, but if humans lose out in a big way, then that entire social contract breaks down. Uh, so yeah, that's the, that's the primary thing, and then the inevitable consequence of all that is, well, if you want to keep the economy running, people need a way to buy things. They need a way to have money. And so you can either, you know, rob Peter to pay Paul through taxes and redistribution like universal basic income. Um, or you can also add in a property-based component. And so right now about 20% of all household income comes from property, uh, nationally in aggregate. And so that's rental, stocks, bonds, those things that kind of pay dividends. Um, but we are not set up to do that right now. And there's also a huge amount of skepticism amongst economists, politicians, and even people out there when you talk about an entirely new paradigm, a new relationship with money and work. So there's a lot of work to do.
Long story short, one thing that you mentioned that I think doesn't get talked about a lot is this idea that prices are going to come down with automation. So people talk about, oh, we're not going to have, you know, there's not going to be jobs, money, etc. Um, but if it is automated, of course, the price to produce various goods and services drops quite a bit. Um, and in fact, as we kind of drop the price across the board, everything, it like it has a compounding effect. Um, is that, yeah, is that something that we need to be thinking about because I mean, doesn't a lot of this also depend on how quickly prices drop, like I think Sam Alman had this idea about the Moore's law for everything. So if you imagine the prices of everything, you know, getting cut in half every, let's say, 5 years, that's a very different future than one where, you know, it's just a 50% cut in price, like the kind, kind of, kind of that compounding effect. Do we have any, does that change anything about how you think, I mean, that's a big impact, isn't it?
Yeah. Yeah. So that's actually kind of baked into the theory, and, and to take a big step back, technology is always deflationary. That's why we use technology is because it allows you to make things cheaply. You know, industrialized agriculture is cheaper than, you know, manual agriculture. That's, that's why we have 2% or actually closer to 1% of the population in America is farmers today versus, you know, 90% 200 years ago or so. Um, is because it allows you to, uh, lower the price, lower the input. Um, so in economics terms, there's inputs, uh, there. So there's a labor input, there's, uh, other capital such as the land itself, you know, uh, the equipment, those sorts of things. And then there's consumables like you need fertilizer and seed and other things. So there's all kinds of inputs. Now, uh, this is Sam Alman's quotation is great because that is an example of a tech bro saying, "Oh, we're going to do Moore's law for everything." His definition of everything is not all-inclusive of the entire economy. Now, you can probably remove a good chunk of the white-collar labor input. Now, white-collar labor is very expensive. You know, a good developer costs $200,000 a year, and that's just one human, right? And we have, I think, 150 million, no, around 160 million employed adults in America right now. So that adds up. Um, so yes, you can, you can, you can do some cost avoidance by saying, "Hey, let's automate away all the developers and all the other white-collar work, you know, so on and so forth, but you're, you're just swapping it out for something that costs less." Granted, it costs a lot less, but then you end up with a bottleneck somewhere else. Your, your next bottleneck is going to be energy. Um, and depending on what you're talking about, like let's just pick cars for instance. Uh, intellectual labor is not the primary constraint on cars, right? They're made mostly out of steel, which is very heavy, and it takes a lot of energy to, to make that steel into the correct shape, to move it where you want it to go. Um, there's advanced electronics that go into it and those sorts of things. So when you look at the total input of like if you were to make a pie chart of everything that goes into a house or a car, like the intellectual input is actually relatively small. So until we have a world full of humanoid robots that are substantially cheaper to run than humans, a lot of that is just not going to change now. We're already seeing where it is changing. V3 just came out, and people are already like recreating scenes from like Harry Potter and, and the Marvel universe where it's like, oh, hey, this took $200, and it looks better than this, you know, $3.5 million, 4-second shot, right? So, you can dramatically lower the, the cost of some things, but those are, those are mostly going to be digital goods and services like legal services and audiovisual kind of stuff and code. Um, but once you get to the physical world, there's still plenty of constraints and scarcity, uh, left around.
That's right. That's right. I guess I was thinking, um, the timelines for robots would be a little bit shorter perhaps. Uh, I think recently you put out a video saying that maybe that's going to be taking a little bit longer than we thought. You had some reasoning why it would take us a while to ramp up complete robot production, billion humanoids or whatever it was. Um, can you maybe explain that a little bit, why you think it might be taking a little bit longer than some of the more optimistic estimates?
Yeah, so the, the, the ultimate number that I landed on was around 2040, which, doing a little bit of research. Um, plenty of other like Brookings Institute and IMF and all those, they kind of all have arrived at the same, same level, which is 2040 is when we're going to see like the takeoff velocity of, of humanoid robots. Uh, and so number one, there's just the manufacturing, right, of it is how many units can you churn out, right? You know, if you imagine because we all have seen stuff like, you know, Terminator and Star Wars, like the droid foundries, like, oh, yeah, you can just stamp out droids like AK-47s, right? Like, yeah, we, we might be able to get to that point, but building a foundry that big is going to take years, um, and then of course, there's all the inputs that you need to put into it, such as, you know, the lithium batteries, um, neodymium magnets and those sorts of, uh. Now, there are ways to get around some of those constraints. One example is Tesla now has, uh, rare earth mag, rare earth metal-free motors. So it's just you have, uh, the, the stator coil and the other coil. I can't remember what it's called, but there, it's coils instead of having a magnet. Now, you lose about 10 to 20% of your efficiency, but you avoid a resource constraint. Um, so but you end up with something that's less efficient and heavier, which has a really big cost for robots. So, you know, will the economics of it make sense as you're trying to scale up to billions of robots, uh, globally? And then, of course, there's going to be, we haven't even hit product-market fit yet for robots. And what I mean by product-market fit is like, yes, the Boston Dynamics robots are really, really agile. Um, but they're like $120,000. They're mostly prototype platforms. They're not scaled up, and, and they're not even that smart yet, right? The, the AI isn't there. So, the cost is just totally not justified. Even if you wanted to buy it, it's like, great. You've got a really dextrous robot that can, you do some pre-programmed back flips and stuff, which is really cool. Disney already uses that kind of thing for stunt doubles, but it's not really going to help you build a house yet. And so what we were looking for is for the robots to hit product-market fit. And then you want to see economies of scale kick in to say, okay, every time I, what was it, Sal's law or Sal's law? Anyways, there's a bunch of economic theory about how every time you double the amount of production, you know, the, the, the cost per unit goes down by a predictable amount. Um, and so when Elon Musk says like, "Oh, yeah, we can get Optimus down to about $20,000," like that's following those laws. Like if you double production that many times, you go from 80,000 to 20,000. That's still pretty expensive. And so when people say like 2040, that's about when we're looking at like, okay, we'll have, we'll have worked out all the kinks. Uh, the AI will be smart enough. We'll have all the supply lines figured out, and we'll be, we'll, we'll ramp up production. And it, it really, it took a little bit longer than that for the automobile to ramp up. It took about 40 years for the automobile to ramp up from the Ford Model T to any, any appreciable competitor. So a 15-year ramp-up is pretty quick in the grand scheme of things.
What about breakthroughs in material science? Because we've seen AI sort of touching on kind of new formations of carbon fiber, and it seems like some of the elements that we used to pull out of the ground might be improved upon and built in a different way. Are you factoring that in?
Yeah. So the, the, the two most, so there's, there's, I think, three primary components in robots that you're thinking about. So one is sensors, which is like eyes and, you know, lidars and proprioception and those sorts of things. There's actuators, which is muscles, and then there's the actual processors themselves. Um, and the, by far the most expensive part is the actuators. Like a good, when you look at the bill of materials for a robot, a good chunk of it comes from, uh, from, from the, from the actuators. So that's the motors and servos and those sorts of things. And then, you know, there's been some of the, the, like electropolymer-based, like the artificial muscle-based ones. Those are really cool. It remains to be seen if we can, if we can get those to the point where it's just like, I could imagine a future where, you know, between, uh, you know, Google's alpha, you know, like AlphaFold and those sorts of things, like we could have an AlphaMaterial one that would create like basically an artificial muscle that you print out like Saran Wrap, and then you just wind it up. Like, I could see that happening. Um, and even if that does happen, still, when you, when you look at the, just the sheer volume, the sheer mass, right? Because this is some of the calculations that I've done is, okay, let's just assume that we get some, you know, like super-efficient polymer-based solution, you're still looking at like billions of tons of material that you need, and so even if you figure it out, it's like, okay, well, that's still a lot of material that we would need to set up entirely new, uh, prochemical supply lines for, uh, to scale up. Um, so even, even baking in some of those assumptions, like 20, we're, we're when you say 2040, you're basically assuming that there's going to be some near-magic breakthroughs between here and then. If you focus just on today's technology, it's closer to 2050 or 60, um, before, before we really get that level of ramp-up. So, when I said earlier like a lot of those assumptions are kind of baked in, it's like, yeah, we're assuming that the, that the amount of attention and money being poured into robots is going to cause some of those accelerations to happen. Um, and even taking into those accelerations, you're looking at 2038 to 2040 or a little bit after depending on, uh, some of the variables you tweak.
That makes sense. Okay. So, okay. So, like a billion humanoids is well past that, and that makes sense. But I mean, we can still expect for the wealthier households to maybe have a humanoid in the home. We're just not expecting mass production, right? Like globally.
Yeah. And, and it comes, like I said, it comes down to product-market fit. Like at what point do you, is it worth the cost of having a humanoid robot? Now, the way that I kind of envision it is like if you've ever seen a new subdivision being built where there's like a swarm of like 50 workers, like you'd probably have a handful of humanoid robots because they're going to be a little bit smarter and, and, uh, and more like, you know, universal. So, you might have one of those like leading a crew of humans because human labor is still going to be like, grunt labor is still going to be infinitely cheaper. Humans require a lot less energy, right? Like you have a lunch box with two or three pounds of food in it that, that's enough to take, you know, carry you all day. Whereas those robots are going to need to
Swap batteries every four or five hours if they're working on a construction site, because one, they're going to be heavier robots, and two, they're going to be using heavier loads. So, humans are just really, really energetically efficient, and then also for grunt labor, it's like, you know, $12 to $18 an hour versus, you know, how long is it going to take? Because also those industrial-scale robots are going to be a lot more expensive than, you know, your domestic robot that is good for folding laundry.
So, yeah, we're going to see, you know, I mean, there's already people on Twitter and Reddit and stuff saying like, "Oh, look, I bought the Unitry and they're great. You spent $80,000 on basically a doorstop." Uh, they're not much more useful than that right now. Um, but they'll get there eventually. And yes, you will see, you will see a gradual ramp-up.
Dylan, I think what we should do is probably like raise a finger if you have like a question next or something. You know what I mean? Just so we know. So we don't do hand signals.
Yeah, I mean, we yeah. No, this is great. It's a little bit of a I'll put this I'll put this right there, and then it'll be a signal. Poor. Okay, that that sounds all right. That that works. Like like the uh like uh Queen Elizabeth, she would switch which side her purse was on when she was like wanted her Secret Service to like say, "I'm done with this person," right? To get her to get her out of there.
Yeah. Please remind him to say, "Get away from me." I get a purse. Um, so a lot of people, kind of my fear right now is like a lot of people are fearful of losing jobs. So they're like, "Protect the jobs. We must have jobs." It's like, oof. That's like bad reasoning, because you got to sort of approach it from first principles. It's like, why do you want a job?
So I mean, if we're yeah, if we're breaking it down, what a job is, what it means to people into kind of components. I mean, kind of like making money and being able to feed the families. One, we draw some meaning from it. Um, I think you also, you know, talked, and others have also talked about it being kind of giving the workers some political power, right? They're able to kind of influence what corporations do, what governments do. I is that between those three, is that most of what a job gives? And are we able to um substitute that for something else if jobs go away?
So answering that question actually preceded post-labor economics for me because when that's actually how I came up with e well I didn't I didn't come up with economic agency as a principle. I've just kind of resurrected it and and breathed new life into the into the concept. Um, because the last time that someone wrote about it seriously, I think was in the '90s. Um, but so what I was I was thinking, you know, just on first principles like, okay, if we take the assumption that AI and robots are going to nuke jobs from orbit, what then? What is it that people actually want? And so by asking myself that question, um, I wrote a blog post on it when I was still on Medium, gosh, like two years ago, I think was the first time I wrote about this um was, okay, well, what is it that people really want? Like, you know, is it is it something to fill their time with? Is it something to feel a sense of purpose? And what I what it what it boiled down to from a strictly economic perspective is the reason that people want work is because they want economic agency. It's not that people like their health insurance. They like not being in danger. It's not that people like their jobs. They like having a house and having food. Now, some people do genuinely love their work, right? But that's not really the government's, you know, from from an economical perspective, it's not the government's, you know, role to say you like your job or you hate your job. It's just, do you have a job so that you can make ends meet? And the ability to make ends meet is economic agency. It's the ability to take care of yourself and get what you want and take care of the people that you care about. And so economic agency breaks down into three overarching components. One is labor rights, two is property rights, and three is democratic influence over policy. Those are kind of the three primary components that go into economic agency.
Uh, so then when you zoom out, you say, okay, well, we have labor rights. You know, you have a right to work, you have a right to quit, you can form unions, although people fight unions, but by and large, unions have a right to exist and that sort of thing. So, well, you might say, well, what's the problem then? You have labor rights, you have property rights, and you have uh democratic rights in America. That should be all that you need. And that is that is the political status quo. However, if the demand for human labor craters, right? Because AI robots basically says, well, I'm starting a new business and I want to, you know, do X, Y, and Z, you know, I want to, you know, let's say Dylan and I get into, you know, 3D printing homes, right? And it's like, well, if we did that in 10 years or 5 years, the cheapest option might be to not hire any humans at all. Just, you know, let's get a business agent, you know, AI agent and and accounting agents and hire some robots or even rent some robots and rent the the rig. And then it's just me and Dylan, right? So then we have no employees. So that's an example of what I mean by demand for human labor labor might crater. Uh, and so then if that's the case, then your labor rights don't really matter, right? So if that's if if if under that circumstance, then you're really going to be banking on property rights and and democratic participation. So but the there's a there's another problem. So there's it's always like it's always like a no butt when you're exploring this. The no but is that one of the ways that that society has worked for all of human civilization is the ability to withhold your labor. Um, you know, whenever whenever there's been big uh union disputes like one of the things that happened I think it was in the '90s it was in the 90s or 2000s was that the uh the sanitation services in New York City went on strike and they declared like a state of emergency after like 3 days. Right. So the ability to withhold your labor is one of the bargaining chips that you have as a civil society to you know force change. But if if all the capitalists and all the business owners say, "Great, take a hike, we don't need your labor anyways," you lose a huge amount of your bargaining power. And so collective bargaining power was why unions formed over a century ago. And so this is this is where it was like, wow, the more I looked at this, the bigger of a problem it became because if you know when when you can replace all the striking workers and stuff you say, "Okay, cool." Uh, then you end up in a situation where not only are uh citizens economically disempowered, they end up also politically disempowered as well. And that is like an even worse outcome than many people are even thinking, because it's not it's not a matter of like elite capture, it's democratic capture. Like you could you in a worst-case scenario, I'm not saying that I believe this, but in a worst-case scenario, you could easily imagine the complete collapse of civil society if too many workers are replaced with AI and robots.
Okay, I got to touch on that collective bargaining thing because I hadn't ever thought through that, but that's a really powerful thought. First off, that's also how it works the other way. Like when you have a group of people who are buying a product and they collectively withhold buying the product, like you know that's like everybody started boycotting Tesla and it like completely cratered the the stock, and then all of a sudden people reacted to that. So um is there something like that on the reverse too that we can use? Like if there's UBI, would the government put strings attached to it, or like how could we how does collective bargaining work from the money point of view without people having jobs to get the money? Right? So this is where we're getting into more theoretical territory, so this is actually one of the key questions that I'm trying to answer with the post-labor post-labor economics treatise that I'm working on, and so uh right now the way the the the the trajectory that we're on is that labor rights might become null and void. So then you end up with just you can vote with your you can vote with your vote, you can vote with your wallet, or you can vote with your zip code, right? You can leave somewhere, you might expatriate. There's plenty of people that are trying to expatriate and that sort of thing. So if that's the case, then well, we need a fundamentally new social contract. And that goes back to what I was talking about about economic agency, the sources of economic agency. If you lose wages and you don't want to rely on the government, then property becomes the primary uh avenue. So what we're looking at uh the current proposal is having a future that is uh more oriented around uh a property and dividend-based kind of social contract rather than labor wage uh based social contract. And so that means we would need stronger private property rights uh ways of of uh distributing some of that that property. Um, for instance, right now it's actually there's a bill working its way through Congress right now. Don't know if it's going to happen, but our American Congress is not the only one that that's working on this kind of thing is uh baby bonds where you give someone an endowment of uh, you know, treasury bills or, you know, whatever when they're born and it matures when they turn 18. And so then it's like, hey, hey, you've got, you know, $10,000 or $100,000 worth of bonds waiting for you when you come of age. So that's just one example of a potential state-run kind of uh property-based arrangement. Um, but the rest of society is not really organized around that. And one of the one of the chief kind of criticisms that comes up immediately is like, you know, who in their right mind thinks that billionaires are going to give up their property for for average citizens? And so it's like, well, you know, that's that's especially if citizens are not offering something in return, because if it becomes a if there's no exchange of value then it looks coercive, and if it looks coercive then no one's going to agree to it um unless there is a compelling incentive structure. And so that leads to the next point, which is well, the incentive structure is that if the economy collapses, the billionaires aren't going to have any customers anyways. So then the question becomes, okay, what's the right level of kind of rejiggering this whole system, uh, which is something that we're still working on modeling, so we're not quite fully there yet. But that was a good question.
Well, I just wonder what the minimum is to keep people happy and if they're just if the future just ends up looking like landing on that, or if we can get to a place where we're thriving, not just not completely stepped on.
Yeah. I mean, ideally we create an attractor state where the incentive structures create stronger democracy over time and stronger uh allocation of of resources and assets over time. Now, one of the things that that I think about sometimes, and this isn't in the treatise because this is more like philosophical or hypothetical, is what is the actual value-add of a billionaire, right? And in in throughout history, one of the things that people look at when you say, "Oh, well, this society is getting too grandiose," uh and it happened everywhere from ancient Rome to the Mughal Empire to the Russian Empire is the overproduction of elites. And so basically, overproduction of elites means you have too many people in the in the ruling class or the aristocracy, and America looks like that right now. Uh, so then you say, well, what what actual purpose are they serving? You know, you might say, okay, well, the world is better off with people like Elon Musk because he's able to go create SpaceX and Tesla and Optimus and those sorts of things because if one person can accumulate that much wealth then they can do great things with it. Not all of them do. So you might say, okay, well, there is an appropriate amount of billionaires in a society or trillionaires eventually or whatever, but what is the what is the appropriate what is and of course everyone can argue over what is appropriate, but what is the most useful uh amount? But in a future with artificial super intelligence, do we need any at all? You know, that's like that's that's I think that's a valid question like because if if you have the ability to solve problems and you can do it instantly and you don't need an Elon Musk or a Jeff Bezos or whoever or Mark Zuckerberg to solve those problems. Then the only problem then is coordination, because what that's that's where that's one of the other functions that elites have is um when you have one person who unilaterally controls that much wealth, they're able to coordinate a lot of labor. Same reasons that the pyramids were built, right? You the pyramids were not built because of someone, you know, saying like, "Oh, hey, this is going to benefit the Egyptian Empire." It was built because one person had enough wealth and enough control to coordinate that much labor to do a great work. So that's like that's the big fear is if we don't like and when I say big fear, it's not everyone has this fear. Some people have this fear that if you don't have a society that has elites of some kind, then you're not going to have any great works because everything just kind of regresses towards the mean and people get bogged down in process and consensus and those sorts of things. Um, so it's a really, really tight kind of argument is like what is the appropriate level of concentration of wealth to even allow or concentration of power to even allow, and you you definitely don't want maximally one way or maximally the other. Both are really bad just in different ways.
Let's see. Well, Dylan, do you have a follow-up or because do is there any big areas we haven't hit yet just to make sure we cover a wide area?
Yeah, we haven't talked about Dyson sphere simulation theory. We haven't talked about time travel. Like we haven't pushed David to the limit. We have to do simulation theory. So yeah, let's start.
Well, why why don't you start? I don't even know what question to ask. What do you think about this?
Well, let me start with this. Let me start with this because Demiasabis just recently or is it Habibus? Do you guys say Hassabis or Hassabis?
Has I think it's Habibus. I don't know.
Um, yeah. Um, so he was asked, "Do you believe in assimilation theory?" And his response was, "Well, not like I'm blanking on the name. Who's the simulation hypothesis guy? The main one."
Oh, Bostrom. Bostrom, right? Bostrom.
Yeah, that's who he said. He's like, "Not like Bostrom." So, to me, I take that to mean not like the Matrix where it's like a video game or some some sort of a thing that we plugged into simulation.
Yeah. Yeah. Exactly. He does believe that we are in a computational universe, and he kind of listed the fact that AlphaFold can kind of predict, you know, the building blocks of life, kind of the the shapes of these proteins, to to him it seemed a little bit that it's meaningful somehow that we're able that these AI models can predict those, and he said he's going to when he finds time he's going to write a paper about it, and I'm so just fascinated about that. I can't wait for him to kind of expand on how he thinks about it. But um yeah, so this is going to be I think a lot on a lot more people's minds because it used to be kind of a fringe theory, but now it's really kind of more people are talking about it. So Dave, I guess like what's your take on this whole thing? Complete nonsense or is there something to this?
Um, well, it's it's circular. The biggest problem is that it's circular is you say, "Okay, we're there's plenty of evidence that we're living in a simulation. Great. uh, who's running it and why, right? And so you end up back with an basic a fundamentally unanswerable question unless whoever's running the simulation reaches in and gives us a message and says, "This is this is the purpose," right? You know, so it's like I've I've done a few thought experiments and written a few blogs like imagine you're a character in the game Cyberpunk 2077, and it's like the entire purpose of that simulation is to entertain someone, and therefore literally your suffering is part of someone else's entertainment, right? And it's like, well, geez, I don't want to live in that world, right? But then you say, okay, well, okay, so there's two primary reasons for for simulation. Number one is entertainment. Number two is curiosity. Um, so it's like, okay, I I can accept that. Like, and if there's other if people have come up with other kind of overarching reasons, sure, that's why you'd run a simulation. Um, but some people you know when when you look at like okay the universe seems to when you look at uh delayed choice and and retrocausality and those sorts of things uh non-locality, you look at all those experiments and it seems like the universe doesn't make a calculation until it absolutely needs to, which is what we do as video game engines, right? You say, okay, if there's if there's a calculation that you can put off, put it off until you actually need to make that calculation, then figure out what's supposed to be at that pixel or at that you know simulated location or whatever. So all that is like the question is are we I is this proof that it's a simulation or are we just finding mathematical similar ways to render a reality? Um, and it's like I don't know if there's a meaningful difference because some people I've talked to physicists who say math is the language of the universe, right? But is math the language of the universe or is it just us approximating how the universe actually works? And so then when you look at it at different scales, there's the there's the quantum, there's the micro, there's the macro, and so on and so forth, cosmic scale. So depending on how you look at the universe, it doesn't like the the the intuitions that we have at the macro scale like in terms of how gravity works and you know um and electricity and those sorts of things. It's all very intuitive to us because our brain has evolved in this milieu for billions of years um or you know our bodies and then brains are a little bit more recent. But then the rest of the universe is just not that intuitive. And but then again, neither is math, right? Like like we don't have like I think I think uh primates have an intuition for addition and subtraction up to about 10, right? It's uh you can you can show um chimpanzees and and other great apes. You can show them and they can say like if you give them a task like here's three stones and here's four stones, they can figure out like okay that's seven stones. They that's about the limit of what they can figure out intuitively. Um, and you say, "Okay, there's an evolutionary reason for that, right? It's just how many bananas do you have?" Or, you know, how many how many rocks have you thrown at your brother to piss him off, you know, something like that. So, you can you can kind of do a little bit of reckoning around what's fair. Um, and there's there's good social reasons for that. But, we don't have any we don't have any natural intuition for calculus or code or any of that sort of stuff. We actually have to hijack other regions of our brain and etch it in over many, many years of practice. Um, I kind of lost my train of thought where I was going with all that. Oh, yeah. Simulation. Go ahead.
Well, yeah. Don't you think it's sort of fascinating too that we have the Planck length to things? Like there's no reason. It seems like there would be a Planck length to time. Like it's discreet. It's individual. And then like the same with distance and all all the three dimensions. I'm like that's so weird. That seems like a frame rate to me, right?
I mean, yeah, it looks like a clock speed. Um, but but then again, that's also just our current understanding, right? Like there there might be there might...
Be good, you know, reasons for that in terms of, you know, and also Planck length and Planck time are infinitesimally small. Um, like I don't, I don't even remember how many decimal points it goes out to, uh, but yeah. So this leads to what some of the work that I've done in in philosophy, which is uh, the concept that I call an ontological container.
So an ontological container is like what is within within the sphere of what is real to view. There are boundaries to that to that reality, right? And so, um, part of our ontological container is if there is something outside of that, whether it is, you know, a deity or someone running the simulation or there's something above it or below it, like hyperspace or subspace or other dimensions. There are boundaries to our ontological container. And so, you know, you say, okay, well, what is an ontological container? Um, using the simulation thing, it's the game, right? It's the instance of the game that some hyper-cosmic entity is running.
Um, and so when you have that model, it's like, okay, well, your reality is actually running in a CPU and memory and hard drive, right? You know, that's you can imagine that as a possibility. Have you guys seen the video of like, um, it's like the the VO prompt where they're saying like, have you heard the prompt theory? It's like, what is everything that I, I'm just a prompt. Whatever, you know, it's in a sense, uh, certainly something going to give some people. But the thing is, it's like, okay, if we can imagine where everything that we think is real space-time matter, all of that is literally just code running on something else, then you have no intrinsic awareness of the the context in which that other machine is running. So that's that's what I mean by ontological container. There is information and and things that are real, but that are completely by definition outside of the scope of what you're able to uh, interact with.
So that's what I mean by ontological container, and that's really what people have have rediscovered through technology is the concept of an ontological container. There's that art, that work of art where it's like, you know, there's the the um, medieval art where it's like, you know, there's the sun and the moon and the stars and it's all on a tapestry and then someone's like looking outside like, what's outside of that, you know? And so people have dealt with this problem for a long time, and you know, to the to the to the Hindu, it's, you know, the realm of Shiva and Brahman and all those like they had an answer. It's just like it's the dream of a god that is our ontological container. It's the dream of a god. Um, so okay, cool, like there's there's plenty of ways you can reconcile that that problem. But by definition, it's an unanswerable question.
Okay, Dave, we have tons more questions for you. I just want to make sure that the people watching know that this is going to be continued in part two on Dylan's channel. That video is live right now. Check the links down below, click on one of them, and dive right back.