The following is a rough transcript which has not been revised by The Jim Rutt Show or Pablos Holman. Please check with us before using any quotations from this transcript. Thank you.
Jim: Today’s guest is Pablos Holman. Pablos is a hacker, inventor, investor, and technology futurist. His experiences include doing cryptocurrency in the 1990s, AI for stock market trading, building spaceships at Blue Origin with Jeff Bezos, and 3D printers at MakerBot when they first came out. We liked it. In fact, we bought like five of them for our MakerSpace. Pablos helped start the Intellectual Ventures Lab for the very famous Nathan Myhrvold. I met Nathan once, long time ago. And currently, Pablos is a venture capitalist at Deep Future, hacking mad scientists, rogue inventors, crazy hackers, and maverick entrepreneurs who are implementing science fiction to solve big problems and help our species become better ancestors. That’s pretty fucking cool. I like that. Welcome.
Pablos: Thank you.
Jim: Welcome, Pablos.
Pablos: Thank you. I’m excited about this.
Jim: Yeah. This should be fun. I met Pablos at a Santa Fe Institute event, way back yonder. We had some couple of good conversations. Very interesting guy doing real interesting things. He made me even, when I got home, get out my pencil and paper and change my Bayesian prior about orbiting solar. That was right.
Pablos: So that was good.
Jim: Not sure it’s a done deal by any means, but it’s not impossible. Alright.
Pablos: There we go.
Jim: So today, we’re gonna talk about a book that Pablos published. What was it in early—
Pablos: Just last month.
Jim: Oh, even better. Even better. Like them when they’re fresh. And the book’s called Deep Future, and that’s pretty much true. So let’s start off with where you introduced the book, and that is with your concept of deep tech. Tell us what you mean by deep tech and contrast it to business as usual in Silicon Valley.
Pablos: Well, what I’ve been calling business as usual is shallow tech. So shallow tech is the things we think of as technology. That is this world we live in, which we think is full of technology, but it’s really just full of software. And so we have, for decades now, been getting more and more software in our lives. And that’s cool. Software is generally applicable, and it’s great if you can get an iPhone app to have Wii delivered to your dorm room by a drone or something. But what has happened is to the detriment of every other kind of technology, software just got us drunk, and we’ve been just doing that. And so we haven’t been advancing other kinds of technologies with the same fervor, and I think we’re paying the price. We’re missing out on all the upside that could come from improving all the other things that humans do. And so I think of deep tech as trying to take all the superpowers we got from our computers and what we’ve learned from Silicon Valley and how to make new things come to life, but aim it at bigger problems and older, nastier industries and things that could really be made a lot better.
Jim: Yeah. One number I recalled from the book was you guesstimated rules of thumb-ish that all the tech businesses in the world broadly construed was maybe two trillion a year in revenue, which was about the same size as the shipping industry.
Pablos: Yeah. Exactly. It’s crazy to think about.
Jim: On the other hand, I’ll bet the profit margins are far better at the tech businesses.
Pablos: They certainly are because the shipping industry is burning almost all their profit, literally.
Jim: Yeah. That was also quite interesting, but we’ll get to that in a minute. But, anyway, just to put things in context, I say this right ultimately regularly. If all that tech went away, say it was a solar flare and roasted all the chips, well, there’d be a fair big disruption for a little while, but humanity would figure it out. We’d get along just fine. We’d find other ways to do things. And if we needed computation, we could build analog relays.
Pablos: Well, Nathan used to say that humans are the cockroaches of mammals. And so you might kill off a bunch of us, but those last few are gonna be pretty hard to get.
Jim: Exactly. So but, anyway, let’s go on to your next idea, which is one that’s very near and dear to my heart, which is the idea of supermodels. One of the things we do get from all this computational capacity and, often written fairly close to the bare metal, is the ability to model almost anything. Let’s talk about that a little bit.
Pablos: Yeah. You know, in my experience working on AI and other kinds of all the kinds of things we could do with computers, I think it just gave me a different lens on what AI means than the hype cycle version, which is mostly chatbots and their derivatives. And I think it’s important to consider that because most of what has happened in AI is thanks to the glut of computation. Even the algorithms that turned out to be successful for LLMs are not all that sophisticated. And so you could—I mean, this isn’t exactly true, but you could almost imagine taking any other machine learning algorithm from the nineties and applying the staggering exoscale computation that OpenAI did to it, and you’d probably still get some pretty magical results. So I think there’s an important distinction that’s being lost in the way people talk about AI, and it’s leading them to think about AI with this lens that they use for social media and for search and for web surfing. And it’s so much more powerful and useful than that if you think about it as models of our world that we can use to run simulations and help us to make better decisions and help us understand the decisions we are making. That’s been my experience with AI going back for decades now, and I think that we’re at this exciting moment where the same scale of computation applied to chatbots can start to be applied to other kinds of computational models and make lots of other things better. So that’s what—and there are several examples in the book there, and we could talk about that if it’s interesting, but that’s how I think about it.
Jim: Yeah. I actually was involved with two startups—one as an investor and then chairman, and the other as an investor and then director—that used simulators in the loop to model analog circuits and to improve their design using something very close to genetic programming to explore the fitness landscape of designs. Purely compute bound. As compute got cheaper, our results got better. Of course, improved our algorithms too. And both companies were sold successfully to big EDA companies. So the crazy idea actually worked. Never since, I’ve sort of been looking for simulator in the loop businesses. Combine simulators with genetic programming, genetic algorithms, explore fitness landscapes, et cetera. And that’s something—and of course, one of the cool things about simulation, particularly of this agent sort or fine grain sort, is it’s embarrassingly parallel. I don’t really give a shit if the chips get faster as long as they get cheaper. Right? Well, if the cores get cheaper, you know, four gigahertz, that’s plenty.
Pablos: You can use last year’s GPUs.
Jim: Yeah. I don’t even need GPUs for most of that stuff. Of course, ASICs would help, but GPUs probably not, at least not much. Anyway, why don’t you give us an example of where you did something cool with modeling?
Pablos: Yeah. I think the one that I probably give the most detail about in the book is at the Intellectual Ventures lab, we were trying to take on solving these infectious diseases in the developing world. So, you know, millions of kids a year dying still of HIV and tuberculosis and malaria and these things that we don’t really even ever see anymore in the United States. And so we were trying to figure out how you could go about eradicating these diseases once and for all. And one of the things we put a lot of effort into was creating computational models of how the diseases spread. So in the malaria model, which is one of the ones we put the most effort into, it’s a model of the world just like SimCity. If you’ve ever played a video game, almost all video games are a model of some world even if it’s not ours. But SimCity is a great example because you get a model of a city that’s kind of like maybe one you live in. And what’s crazy is if you play a game like SimCity, you have a simulation where you can try putting stoplights up or deploying a Starbucks or try these different things and you can see if it makes your city improve or get shittier. And I won’t try to draw a comparison to real world cities today, but you can obviously see that we do not have a model like that for any of the cities we live in. Nobody’s doing SimCity in real life, and that is what I’m trying to suggest.
Jim: That’s not quite true. I mean, there are some cities—
Pablos: It’s not quite. I’m painting in a broad brush. There are certainly models being used to help optimize little things here and there, but not from a very large like, a macro scale city simulation. And I think there’s people who want to do that, and there’s researchers trying, but that’s not what’s being used in city council. And so I think one day we’ll get there, and I think we’re starting to get there with most businesses. It will be very, very normal for every business to have a computational model of their business that’s much more advanced than what they do now, which is run a bunch of spreadsheets and save them on Google drives.
So I think there’s a world coming where we have simulations. Anyway, in the epidemiological models, we have models built on climate and travel of humans and rainfall and all the things that would affect the spread of a disease like malaria. Then because we have a model, we can run simulations in the model, and it can help us discover what are the best ways to deploy finite resources. What’s the best way to attack malaria? What if we put bed nets here or spray DDT there? Could we figure out a way to contain that disease and eliminate it once and for all?
And it is possible to create scenarios like that in the models, and it can help you to plot an optimized eradication campaign. Now that work is ongoing, and malaria, we’re doing much better in part because we have these models. With other diseases like polio, we were able to use these models to help contain them and then eradicate them once and for all. There’s been no wild type polio anywhere in Africa for almost a decade, in part thanks to using these models to contain the disease.
Another example is Ebola. In the first Ebola outbreak, twelve thousand lives were lost. In the second Ebola outbreak, which was only a few years later, only twelve lives were lost. And that’s in part because using a model like this, we could simulate the spread of the disease. We could optimize the ring vaccination campaign. We could contain that disease before it spread. And there are three orders of magnitude improvement in a few years is fucking staggering. And that is what would have been possible with COVID if the world were trying. And so I’m trying to show people these computational models are superpowers for us. They’re capable of helping us improve how we do things in our real world, and we need to learn to wield these tools and aim them at things that matter. And, right now, they’re mostly getting aimed at selling you shit. So I think we’re gonna get predictably disappointing results, but we will learn to aim these tools at things that are bigger problems for humans, and it’s gonna make the world a lot better. I’m pretty excited about that. So that’s why that chapter is in there.
Jim: COVID is an interesting example. I mean, there was lots of epidemiological modeling that’s been going on for years. I know some of the people that do it. I’ve had them on my podcast. We talk about it quite a bit. And yet when COVID came, the decision makers were not appropriately connected to the modelers. There was a big impedance mismatch. There was no way to go from the modelers to some orange guy who was calling the shots.
It’s funny—I proposed, I never did publish it, though maybe it’s time to refresh it and publish it. Before COVID, I wrote, never published, an essay called “The Department for Wicked Risks.” Said that the US government should have a whole department extravagantly funded doing essentially complex systems modeling on everything that matters and have generalists at modeling who can model new situations when they arise—like, oh fuck, aliens half a light year away, what do we do? And I showed that we could afford to spend 20 billion a year and it’d be chump change. When they figured that COVID cost the United States a couple of trillion dollars—and we could have actually contained COVID if we had the proper impedance matching going from modelers up through undersecretaries to cabinet people to chief executive, but we didn’t. And we had the capacity.
Pablos: Yeah. And you would have needed some cooperation from China on the early detection. But it’s certainly possible. And I hope we’ve learned our lesson, and I sure hope we’re taking the lesson from that and trying to figure out how we’re ready for a response next time.
Jim: Because we were lucky this time. In my last conversation with Sam Scarpino, one of the top epidemiologists in the world, we both said, you know, you could think of COVID as a warning shot, something like a 0.5 percent fatality rate. There are definitely things in the same family even of viruses that could have 30 percent fatality rates. And MERS had 50 percent fatality rate. Now that probably wouldn’t at scale, but it might have 25. And if it was as contagious as late stage COVID-19, you know, might have whacked two, three billion people. That would—and so we gotta be ready. We have to be ready to make tough decisions. And we’re not. Anyway, why don’t we tell just a little bit about your youth as a hacker? But first, have to ask you a question. Did your mother really move to Alaska to capture a husband and wear the shortest miniskirt in Alaska?
Pablos: That is a true story. So that was 1969.
Jim: They were mighty short miniskirts in them days.
Pablos: There were, and they were acclimated to them because they came from back east, and they drove all the way across the US. They didn’t even know where Alaska was—they just asked at every gas station which way to Alaska. And they got there, and I don’t know, I think the nine-to-one men-to-women ratio might be a bit exaggerated, but they certainly were very successful at having her pick husbands, and they picked good ones. And I was born not much more than nine months later.
Jim: Plan worked. I like it when someone establishes a plan and executes it successfully. Tell me a little bit about what it was like to be a smart kid growing up in rural Alaska, not even in Anchorage, but you’re out in the fucking boondocks, aren’t you?
Pablos: Yeah. I started out in Anchorage, and then my formative years, junior high and high school, I was in a small town called Soldotna, which we thought of as being relatively cosmopolitan because it had road access. But there’s only one road into town. You drive on it for three hours to get to Anchorage, and there’s nothing in between. Nothing but nature. I’ve started to realize it’s hard to find a place even in America where you can drive for three hours without finding something, but that’s where I grew up. And I got one of the first computers you could have at home, one of the first Apple IIs ever made. In those days, nobody had ever seen a computer. So I was a ten-year-old kid armed with an Apple II and trying to convince people that it was cool. Nobody had ever seen a computer, but they were pretty sure it wasn’t cool. And so, here I am decades later still trying to convince people that computers are cool.
Jim: Very cool. And it sounds like you did something similar to what I did when I got my first Apple II in 1980. In those days, you could either write in kind of slow, boring Microsoft BASIC—AppleSoft BASIC, they called it—or you could write 6502 assembler. So I wrote my first few programs in 6502 assembler back in the days before there were even macro assemblers where I had vast pieces of graph paper, which I taped together, and I had to calculate your own jumps.
Pablos: The whole deal. Right? Had you used other computers before that?
Jim: Yeah. I had, when I was in college, I had done a limited amount, but at the other extreme, in PL/1 of all things, which is a very verbose high-level language that ran on IBM mainframes. I only took one computer science course in my life. People always ask me, “Well, how the hell did you end up as a CTO of a multibillion-dollar corporation?” I go, “I learn fast.” As it would turn out, learning about mainframes and all that shit in 1973 wouldn’t have been very relevant by 1993 anyway. So just being a fast cycle learner who was willing to invest in my own education was a far better competitive strategy than trying to exploit a very archaic academic credential.
Pablos: Yeah. There you go. Well, the Apple II, my experience is very interesting because it was probably about the same time, but I didn’t have anybody around me who knew more than me. And so I had to learn assembly by disassembly. I learned from just looking at the programs that were there and then poking at them, trying to tweak them, see if I could break them, see if I could make them do different things. So my early experience really was just hacking existing binaries. It didn’t make me a very—I mean, that’s an excruciating way to learn to program.
Jim: Yeah. I bought the actual instruction booklet, the chip manufacturer’s.
Pablos: Oh, that would have been nice.
Jim: Sold a little paperback book, which was sold in random computer stores in Kentucky where I lived at the time. And I bought one.
Pablos: Kentucky.
Jim: How about that? I think that’s how I actually figured out the essentials. But other than that, I just taught myself basically.
Pablos: Yeah. Yeah. Good.
Jim: Tell me about what you’d call the hacker mindset. How do you think that’s helped you over the years?
Jim: Tell me about what you’d call the hacker mindset. How do you think that’s helped you over the years?
Pablos: Well, I think it comes from that. You know, because my learning style had to be out of necessity, it had to be reverse engineering. I just had to figure out how things worked by taking them apart and then see if I could put them back together. And that was kind of true for the software and the hardware. You know, an Apple II is like a Jeep. You could just sort of take the screws out. And so I think that helped me a lot because it turned me into that kind of learner you’re describing—very motivated, self-motivated, attack the problem, learn as much as you can as fast as you can. And that’s what I think hackers are good at. They’re optimized for seeing everything as a puzzle to be solved, and learning a new thing is just another puzzle to solve. And figuring out what you’re not supposed to be able to do is just a way of making puzzles out of everyday things. And so hackers have, by their nature, I think, trained themselves to be really good at figuring out how to do new things. And that’s what I think we need in the world a lot more of. We need to invent new technologies. You can’t do that by reading the directions. You gotta figure out something that hasn’t been done before, and that is a very rare and special skill. That’s why I love hackers because they’re kind of optimized for that. Now a lot of them found a bottomless pit of puzzles in computer security. And so hackers are highly correlated with computer security or computer insecurity, but I’m trying to rescue them and get them out of computer security and get them aimed back at bigger problems. So I’m trying to hack everything but computers at this point. But I think that those are the brains we need to do it.
Jim: And you make a nice next step in your progression of your story here. One could argue that an inventor is kind of a more focused, more socially useful kind of hacker, basically. Talk to us about what you mean when you talk about inventors. Talk about how important and special they really are.
Pablos: A way to do that is to think about, like, how many artists can you name, like painters?
Jim: A hundred probably.
Pablos: A hundred probably. How about music artists?
Jim: Probably at least 500.
Pablos: Actors?
Jim: A hundred probably.
Pablos: Okay. Cool. Now try inventors.
Jim: I could probably name—
Pablos: That aren’t dead.
Jim: That aren’t dead. That’s the hard part. I know Lee Felfenstein pretty well. But—
Pablos: Okay. Cool. One.
Jim: Steve Wozniak. Two. Let’s see. Danny Hillis. Let’s see.
Pablos: Three. Good.
Jim: Bill Gates, I met once. I don’t know him, but there’s four.
Pablos: To know him at all, but that’s—but you’re up to four, and that’s like world-class inventor naming skill you got there.
Jim: Yep. I probably know twenty more if I thought hard about it, but no more than twenty.
Pablos: That’s right. Yeah. Good. No more than twenty, and you’re probably in the top one percent of people who could name inventors. And the point I’m trying to make there is that all these artists, we know them because we celebrate them for being these creative geniuses. And you could live without all of them. Every actor, every musician, every painter, like, novelist. And, I mean, as inspiring as they may be, you could live without them. Without inventors, you and I would not exist at all. And so to me, we’ve got this incredible and important creative class of people, but inventor is a less legitimate career choice than rock star. Right? Like, you can’t even compare them. And so I think we’ve really done ourselves some kind of societal disservice by not understanding, not recognizing that these inventors are actually our most important creative class, and we don’t even count them.
I wanna try to figure out how do we course correct on that. And so to me, because I am one of the few people in the world who has a business card that says inventor on it, I think I—and I got to work with a lot of inventors, and I got to help try to figure out how do we sort of scale up investing in inventors because it’s not a legitimate thing. We have lots of support for, say, scientific research, and we certainly always want and need more. But the job there is understand how the world works. That’s basic research. And then on the other end, you have entrepreneurs whose job is to go make a business, and we have extraordinary support for those people. Right? That’s all of venture capital and Silicon Valley and startups and everything there. But in the middle, the job of inventor is to take the output of basic research, you know, new scientific understanding and ask yourself, does this new algorithm, this new chip, new sensor, new scientific discovery, does it change anything humans have ever done? Can we do it faster, cheaper, better, more humane? And when you find the answer there, when you’ve matched that up and you’ve matched a new technology to a problem, that’s invention. And we do not have any structure, any support for that. That’s crazy hair and a DeLorean in a garage, and it’s super haphazard. So that’s the thing I think our mythology is off course here, and so I’m trying to figure out how do you straighten that out.
Jim: I’m gonna push back a little bit because I can think of a couple of domains where inventors are of the essence.
Pablos: Okay.
Jim: How about biotech?
Pablos: Name a biotech inventor.
Jim: I can’t name any, but I’m sure there must be a zillion of them. Because you take these ideas from places like MIT, my old alma mater, and there’s a whole—there’s 200 companies within half a mile. Right? And almost all of them are biotech, and they’re all actually inventing things, taking research and turning them into deployable either chemicals or physical processes. But you’re right. I can’t name a single one of them.
Pablos: Yeah. It’s okay. I mean, I actually am in Boston right now, and I can name some of them. But the point is, I think biotech has done a good job in its way of mapping. The reason that works at all is because we have more successfully mapped the commercial opportunity for the output of that research, and we’ve developed this extraordinary thing, actually, that track we’ve got for drug development. Because it is very expensive and resource intensive and failure prone, and we’ve managed to kind of make it all work anyway. That said, most of that support kicks in after the invention. And you could say that in biotech, maybe there’s more support for inventing while it’s still in an academic context. Because you’re able to do the kind of research to figure out, does this drug map to this particular thing? And so we do have a sort of system there. I don’t know why we would be satisfied with that. It is very error prone, very long and convoluted. It is so expensive that thousands of very promising drugs cannot be developed because they don’t map to a big enough market. There are a lot of problems there that independent of what I said about inventors that need to be improved, and it might be very difficult for you to name a second market where something is being considered functional.
Jim: Chemistry.
Pablos: Okay. So you could be a chemist at DuPont trying to figure out how to make better sticky notes, or that’s 3M, I guess.
Jim: Yeah. Yeah. Of course, 3M famously is an invention factory, but it’s also a microinvention factory. It literally was sticky notes. It’s better Scotch tape. It’s all that stuff. Anyway, let’s move on now to bigger picture issues. So you touched on something very important here, which you then address soon in the book after you talk about inventors, which is kind of the mismatch between our venture capital industry and things that take a long time. Actually, let’s start with a framing that you use throughout, which is the idea of a hundred years and then ten and twenty, and then try to take even the bottom part of that stack and connect it to venture capital. So let’s start at the top and work your way down.
Pablos: Well, one way that I’ve been thinking about technologies is I just look at them and anybody could do this. If you just ask yourself for any given new technology, would a hundred years from now, would we use this new technology, or would we stick with the way we’ve been doing it? And you could try that on, like, electric cars. A hundred years from now, are we still gonna be pouring dyno juice into our cars to run them, or would they be electric? And most people would probably come up with their own answer that it’s gonna be electric in a hundred years. And even before Tesla, you might have had the same conclusion.
What’s nice about a hundred years is anything could change in a hundred years. We’ll all be dead. We might elect someone competent. There might be a completely different funding and regulatory environment and robots to help. So you could easily imagine anything changing in a hundred years. So I try to ask myself that question about new technologies. A hundred years from now, will we be burning nasty bunker oil to move Happy Meal toys from China to Los Angeles in cargo ships? Or will we make those ships sail themselves with wind power and make them zero emissions? It’s easy to imagine a hundred years from now, we’ll totally be able to do that.
Jim: Or even better, the kid has a nanotech reactor in the kitchen, throws his old toys in, dials up the new toys, and they pop out five minutes later.
Pablos: The matter compiler. I love it. So I think there’s a way of just having kind of a first order answer to the question whether a new technology is worth working on or not. You can just do that exercise. And if the answer is yes, in a hundred years, we would totally use the new tech, then ask yourself, well, does it have to take a hundred years or could we do it in ten? And that ten-year window is very important because all of our societal and economic structures map to ten years. Venture funds are all ten-year funds. I get money. I invest it. I got ten years to get a big return and give it back to the investors in the fund. And that’s just how every venture fund works. Debt works on ten-year cycles. I mean, there’s just all these things that work in a ten-year window. And so if you can adapt your technology or project or company or whatever you’re doing to a ten-year window, then you can sort of co-opt the machinery that exists around you and make that work. And so the reason I’m writing about this in the book is that the conventional wisdom in Silicon Valley is that hardware is called hardware because it’s fucking hard. And it probably is harder than software. Software is really easy. Any app you can draw on a napkin, we can probably make.
Jim: Yeah. And nowadays with, you know, Kiro or something, you could make it before lunch. Right?
Pablos: Yeah. You literally draw it on a napkin from lunch, and before you’re done eating, it’s ready to go. So there’s those things happening. What I’m trying to show is that the reason software is so powerful and so cheap and so fast is largely because of rapid iteration. And rapid iteration, if you go back to what I was saying about simulations in a model, every shot on goal is a simulation. If I build an app and I ship it to users and they complain that it doesn’t work on their Android phone, I can update it and make an A/B test with a different version after lunch. So you steer your way towards success, and the cost of each experiment is so low that you can figure out what’s going to work very quickly and cheaply.
Well, once you start building rockets and Teslas and hypersonic jets, then it gets really expensive to iterate because you have to crash every one of those things and build another one, and it’s really slow and expensive. The whole point of this in the book is that we are now in a world, thanks to those supermodels, where we can simulate in software the rocket, the hypersonic jet, or the Tesla. We can design it in software, test it in software, crash it in software. We do that thousands and thousands of times, and then we build one.
When you watch SpaceX crash a rocket, that’s after crashing thousands of them in software, in simulation. NASA had to crash a lot of real live rockets, and that’s why it took them so long. It was so expensive to develop them and ended up with something really expensive. SpaceX can do that so fast in software that they only have to crash one now and then, usually on purpose, just to see if their models are accurate. We live in a different world where a lot of the efficiencies we got for software, we can now apply to everything else. And that is why deep tech is more possible than ever. That’s why we are at this inflection point where it makes sense to do these things. You can see the canary in the coal mine with something like SpaceX or Boom Aerospace or Tesla. It would be great to have a much longer list of examples, and that’s coming, but that’s where we’re at.
Jim: Sounds like, as we would say in finance, an arbitrage opportunity here.
Pablos: It’s an arbitrage opportunity. I am trying to co-opt the machinery of venture capital away from SaaS holes and aim it squarely at deep tech.
Jim: I mean, this is classic. Right? This is where something fundamental has changed, and the thing that’s fundamental has changed is a gross explosion of the amount of computation that’s available and the growth of expertise around modeling. You know, probably a lot of it driven by self-driving cars and aerospace and places like that where they could just throw huge amounts of dollars at it. But those tools and that thinking can be used in other domains. And hence, if you could reduce a fifty-year problem to a ten-year problem, might be an opportunity there. Right?
Pablos: Exactly right. And I think there’s a whole bunch of those.
Jim: Very cool. Why don’t you toss a few out?
Pablos: Okay. So if you go back to what we were describing earlier when I added up the revenues of the software industrial complex, counting Meta and Salesforce and Microsoft and everybody combined, it’s about $2 trillion a year in revenue. If you throw in hardware like NVIDIA and Intel, you get another couple trillion. Throw in the services of supporting all this stuff and you get maybe another trillion. So maybe max $5 trillion goes to the tech industry, which is in the headlines every single day. It’s our biggest industry, our richest industry, our most successful industry. It’s everything. Global GDP is well in excess of $100 trillion. So the tech industry is 5 percent of what humans rely on at most.
The other 95 percent—think about these things: energy, food, manufacturing, construction, even apparel, mining. These are industries every single human on Earth relies on. Silicon Valley has not touched them. So if you’re in Silicon Valley and you’re an investor, you’re fighting to get a piece of the 5 percent. I’m not. My TAM is $95 trillion. And if you look even in our portfolio already, we have companies in energy, in shipping, in manufacturing, in mining, and all these things, and they’re going after those trillion-dollar industries.
Shipping is $2 trillion. Durable goods manufacturing is $4 trillion. Energy is the biggest industry in the world, probably $15 trillion plus, and that’s even before AI. These are the industries that have been largely untouched by Silicon Valley. They have certainly adopted some software here and there to make things more efficient, and they have invented some new technologies now and then. Oil got fracking, and that’s helped them out a lot. But these industries largely are untouched by state-of-the-art technologies we could bring.
The example I mentioned earlier is we have a company making cargo ships that sail themselves across the ocean. It’s almost embarrassingly low-tech compared to everything else we do. You basically duct tape a Tesla to the front of the ship, and it’s self-driving. There’s nothing to hit out there, maybe like one documented pedestrian ever. And then the sailing part has been working for centuries. Meanwhile, the current shipping industry burns five out of six dollars. If you look on the ocean right now—there’s a website called marinetraffic.com—you can see every ship on the ocean. They’re colored by what they’re doing. Half of them are just moving oil and gas around, and they’re burning oil and gas to do it.
So, I think this is a pretty clear win. All we have to do is make a drone ship that drives itself across the ocean using wind power. We’ve already got the tech working. Silicon Valley’s big win is disrupting Yellow Cab. What about disrupting General Motors, General Electric, General Mills, or Maersk? The day our first ship sails, do we sell it to Maersk, or do we take the Uber playbook and build the next Maersk? That is what I think people are missing in Silicon Valley as far as opportunity. The big opportunities are not in AI. They’re not in SaaS. They’re in disrupting the biggest industries on Earth. We’re going to replace the shipping industry with a startup. That’s how it’s going to go down, and that’s just the beginning of the list.
Jim: What’s another example that you see as potentially ready for radical reinvention via the use of an inventor mindset and high-capacity modeling?
Jim: What’s another example that you see as potentially ready for radical reinvention via the use of an inventor mindset and high capacity modeling?
Pablos: So more than a year ago, Steven Wolfram emailed me and introduced me to this buddy of his that had invented a nuclear reactor. And I was sort of bummed out about nuclear reactors because we invented the most advanced reactor at the Intellectual Ventures Lab. It’s called the TerraPower reactor. And if you ever see Bill Gates talk about nuclear reactors, that’s the one. And that reactor is super cool. It runs on nuclear waste, but we’re not allowed to build it. The US government has outlawed advanced reactor technology for our entire lives. And so the only reactors you’re allowed to build are based on old technology. We call them pressurized water reactors.
So Wolfram introduced me to this guy, Rich Muller. Rich is a physicist at Berkeley, and Rich figured out how to change the design of a pressurized water reactor so it would fit in a borehole, and he buries it a mile deep in a borehole. So it’s literally under 10 billion tons of rock. It’s unquestionably safe. It’s a mile from anyone’s backyard. And so I started talking to Rich about this, and I got excited because I thought of it as kind of a regulatory hack. You know, this is a reactor that could get approved, and it’s totally safe and small and cheap. You could make it in a factory the way we make Toyotas, make thousands of them, and get economies of scale. This has never been done with nuclear reactors before, and that’s what needs to be done.
So, we invested in that company. Last week, the Department of Energy approved them to build the first reactor. They were on a track to get it approved and build a test core and then commercial reactor by 2029. The Department of Energy is pushing them to do it all by July. So this company called Deep Vision will deploy their first reactor at a commercial site by July, and this is incredible. So we are in a totally different world. The NRC has been overhauled by Congress. The president has signed a bunch of executive orders to push the deployment of nuclear reactors. This is the safest, the cleanest, the most scalable, and cheapest energy source that God gave us, and we have outlawed it. We conflated nuclear reactors with nuclear bombs when before we were born, outlawed the wrong one. If we’d done it the other way around, you never would have heard of global warming. That’s what’s possible. And so I’m so excited because this is the moment when the world gets to get on the correct fork in history. Got it wrong sixty years ago. Now we can course correct. And if you can solve energy, you solve a lot of other problems for free. And so nuclear reactors to me are the most exciting thing that we can do right now. It’s technology we have right now, and we could start growing supply around the world. And I think that is the most important thing to do.
Jim: Now there’s been a lot of startups that have tried to build modular, shippable, shipping container size nuclear products. I remember years ago, everybody was excited about the fluid bed technology in South Africa. A bunch of money flew into that. As you said, the one that Gates got in—I mean, there must have been fifteen of them I’ve heard of. Right? Probably more than that.
Pablos: There’s about a hundred right now.
Jim: Why have none of them shipped anything?
Pablos: Actually, that’s not true. Some of them have shipped, but all the ones that have shipped are in China. That’s interesting. China has a whole variety of nuclear reactors. They’re all doing well. They’re all shipping them on time and on budget, and they’re kicking ass with them. So, embarrassingly, all those designs came from the United States.
Jim: But probably a lot of them literally stolen. Right?
Pablos: Well, you know, there’s a semantic issue there too. I’m sure that we’re not qualified to weigh in on. But yeah. So we have seen it work in these, evil authoritarian dictatorships that are very well managed and run, but we don’t do it. So that’s what’s very exciting is that is about to change here, and we can finally provide some real leadership on how to do energy. The other reactors, look, there’s a lot of ideas. I don’t really want to disparage any particular idea because I think we need a thousand silver bullets here. So I hope everybody is successful.
But there are a lot of different ideas about how to build small modular reactors. The small part is kind of pointless in one sense because we need so much energy. We need a lot of them. The deep fission reactor works because you could just deploy a lot of them. If you need more power, you drill more holes and drop them down there, and that gives you containment for free. So that part is great. The nice thing about small and about modular is that you get the economies of scale back. You know, all of the reactors we have in America, we have 94 of them right now, I think. They’re all bespoke. None of them have any interchangeable parts. Everyone is like a Frank Gehry project that’s just completely unique and special. It’s one of one, and that is, as everyone listening knows, I’m sure, the most expensive way to make anything.
So if you could build a smaller reactor in a factory, then you could iterate on the design. You could improve it over time. You could improve parallelization for manufacturing. You could get economies of scale, and you can get the cost down. The deep fission reactor will, when it’s deployed in July, be the cheapest reactor ever made. It will put out probably 15 megawatts electric. And if you need 30, you do two of them. That’s no problem. And they have orders for lots of them already.
And then the other big classes of problems with reactors—like, you know, the literal specification for containment is that you have to have a Fort Knox of cement around your reactor, so much so that a 747 can crash into it without compromising the reactor core. Well, how much do you think that cement costs? That’s a lot of cement. And so, you know, even the small reactors suffer from this problem that they’d still need a containment vessel made of cement that’s that intensive. Well, that’s very expensive. So they try to offset the cost of that by saying, okay, we’re gonna put ten reactors in the containment vessel. So now they’re making ten reactors, but the day you deploy the first one, you still have to pay for all the containment.
So the economics are really difficult to navigate in the regulatory environment that we’ve had. Now the United States has gotten so much better about this. The NRC has been completely overhauled. They’re being very supportive and helpful. The Department of Energy is being very supportive and helpful. So some of these things might break free, and some of these SMR designs might become more practical. But, fortunately, with deep vision, we don’t have to worry about that because you get that containment for the cost of a hole. And we have a hole industry that’s really highly developed.
Jim: Yep. Yeah. Exactly. We are the world’s leaders in drilling holes in the ground. Right? And have been—
Pablos: That’s right.
Jim: Since 1870 or thereabouts. Alright. Now we’ve talked about hard tech. We’ve talked about money. We talked about venture capital, bad incentives, regulatory problems. Let’s change directions here, and you have a quite evocative chapter in the middle of the book about the moral imperative, and you start off by talking about a formative experience you had in India. Why don’t we start there?
Jim: Let’s change directions here. You have a quite evocative chapter in the middle of the book about the moral imperative, and you start off by talking about a formative experience you had in India. Why don’t we start there?
Pablos: Well, I think a lot of people have probably had some experience like this. If you travel around the world, one of the first things you notice is not everybody’s like you, which is very helpful, I think, for anybody to experience. My story was when I was in high school in the eighties, I went and spent a summer in Southeast India. And at the time, it was very much a third world country. And you might not use that term today, but in those days, we did, and that’s what it was. There was no running water, no electricity. People had to go fetch water from a well. You had to boil it over an open fire, so you’re literally burning some crap, creating all kinds of carcinogens, trying to kill off whatever bacteria is in the water so that it doesn’t kill you. This is not a very healthy situation. A lot of people suffered from very preventable health problems that we take for granted not having—sanitation and whatnot.
So it was a big eye opener for me as an entitled brat American teenager. But what I really learned from it, because I stayed there long enough to not totally be a tourist, I got to know people who were there, and I just felt like these people were pretty awesome. There’s nothing better about me, nothing better about the people where I came from, other than we just got lucky and were in a place where they didn’t have those same economic struggles. And so it changed the way I thought about people and about the world. I started to see that the world had really invested a lot in me. I was just an asshole teenager from the United States, but I had an Apple II, I had a pretty good education. I didn’t have to worry about dying of dysentery or malaria or any of these other things.
I felt like the world had kind of overinvested in me, not based on my own merit, just because I got lucky. And I wanted to figure out how to put that to good use. What could I do with that? Who can I become? What can I give back? I saw that I had a pretty unique skill set from getting started early with computers. Eventually, it sunk in—can I take these computers and use them to solve bigger and bigger problems? Can we use what I know and what I’ve learned to go after helping make life better all over the world?
People complain a lot these days about inequality, and I’m not even sure what people are referring to a lot of times. Not that inequality doesn’t exist, but why do they even care? I think the thing that matters is that baseline—how do you get people up over the hump who are living in extreme poverty? One of the ways I try to describe this in the book, going back to energy, is the average person on earth gets as much energy as one toaster. If you superglue the button down, run it twenty-four seven, a toaster consumes the same amount of energy as an average person on earth gets for all their heating, cooling, travel, air conditioning, water purification, and cooking combined.
It’s so radically different than our experience. We are Americans. We get eight bonus toasters. For you and me to survive and live our life the way we’re used to, an average American gets nine toasters of energy. I mean, probably—don’t know about you, Jim, but I’m definitely in the twenty, thirty, forty toaster range. In order to get those averages, three billion people live on less than one toaster, and less than one toaster is not enough. That is just a simple way to think about energy inequality, which I think is where it starts. We made eight billion people. We did not make what they need to thrive, to get them an acceptable living standard, what we would consider acceptable, which might not be nine toasters, but three or four or five would sure be good. I want to solve those problems, and I think that everybody should want to solve those problems because you would want to live in a world that’s more peaceful. Why are these wars happening all over the world in our lifetime? Those are not wars over religion. They are wars over access to energy.
Jim: Mhmm. Some are. I think more of them are testosterone dick waving, frankly.
Pablos: Yeah. Testosterone dick waving over who controls the most oil. I mean, honestly, that’s what—
Jim: How much oil is there in Ukraine? Not much, right?
Pablos: There’s rare earth metals, and we need them right now, don’t we, for our so-called electrification plans? Why do we give a shit what happens in Ukraine? Because we want those metals.
Jim: Yeah. Let me hop back to your toasters. In work I do, we’ve calculated that in, by 2080, we could, if we made the moral commitment to do so, provide about four kilowatts continuous, which would be what? About three toaster—
Pablos: Four toasters.
Jim: You’re using one kilowatt toaster per toaster. So four kilowatts for everybody on earth. Right? It was totally renewable by then without any problem. Wouldn’t even be hard. And four is actually a quite good number. To give you an example of two countries that are running—
Pablos: Europeans are living on four.
Jim: They must have seven or eight, six or seven.
Pablos: They used to be seven or eight. Now they’re down to four or five.
Jim: If you include their imputed energy in their imports. But two countries that are right around four everything included are Portugal and China, both of which are not bad countries. Right? It would not bust anybody’s chops to live like a Portuguese or to live like a Chinese, basically.
Pablos: It would be a dramatic improvement. And so I don’t know why it takes to 2080, probably because you’re trying to do it with solar panels.
Jim: Yeah. I do it assuming—
Pablos: And batteries that don’t exist.
Jim: Well, I assume a full portfolio. But anyway, if we can do it faster, great. If we can do it to a higher number, great. Right? One of things I object to is, and I think this is one of the reasons why climate remediation is unpopular with the masses, is a certain number of climate change people are really what I’d call aesthetics. They really do want to live in a mud hut. Right? And they want everybody else to live in a mud hut too. My view is if we can live in balance with mother nature, eight toasters or twelve toasters, that’s fine and probably somewhat better. Now I imagine it’s an asymptote, but there’s no reason not to as long as we can do it responsibly without damaging mother nature and actually giving mother nature a chance to recover because we have pushed her to the margins. Right?
Pablos: If we had built nuclear reactors since the sixties, we would be in that world right now. Probably. And then we are not on the right fork in history.
Jim: Yep. That is probably true.
Pablos: I hope that we can get there faster than 2080. It’d be great to—I mean, I would love to look at that analysis with you and figure out which tracks could accelerate that because that’s what we need to do.
Jim: I’ll reach out to you to do that. That had it done by a group at MIT who models learning curve and economics and projected future economics on alternative energy technologies. But you know, do keep in mind there’s a lot of inertia in the energy system. A lot of, you know, what they call it, landlocked capital, all kinds of perverse incentives, and the peculiarities of utility accounting and regulatory environments. So shit happens slowly and more slowly than you would think in these large core domains.
Pablos: Let me mention one thing. In your lifetime, have you ever seen Shell or Chevron advertised to get you to buy more gas?
Jim: Oh, yeah. When I was a kid, that was probably the number two or three advertiser was gasoline companies. Yeah. Oh, yeah. Constantly.
Pablos: To get you to buy gas from them versus their competitors.
Jim: Right. It’s like the equivalent of beer advertising. Right? It was Sinclair, Exxon, put a tiger in your tank. You know? Chevron, clean gas, cleans your engine. Yeah. These were used to be big time consumer marketing companies.
Pablos: But they don’t really have to advertise their product because we’ll buy as much gas as we can get. Demand is essentially off the charts for the biggest market in the world. They’re doing competitive advertising because they have gas stations. It’s retail. Yeah. That’s retail advertising.
Jim: Yeah. Just like beer. Right? Beer, you know, or cigarettes. Well, cigarettes are trying to get kids to buy ciggies, and probably beer ads are trying to get young teenagers to drink beer. But to your point and to your hundred year thing, yeah, I think we all agree that the idea of burning dino juice, hilariously, Sinclair gas, which was big where I live. Their logo was a dinosaur, a big green brontosaurus. How about that?
Pablos: Yeah. Wow. I forgot about that.
Jim: Look it up. Yeah. It’s true.
Pablos: It’s true. Have Sinclair, but, yeah, that’s proudly burning fossil fuels.
Jim: We’re killing—we’re burning dinosaurs. Even though—
Pablos: Were dead anyway.
Jim: And we’re not burning dinosaurs, actually, but the story is good enough. Now let’s talk about the next part of the moral imperative, which is population. One of the things I talk about in my Game B work, people don’t remember that—very few people know this when I ask them. In 1700, which I deemed to be the beginning of modernism with the simultaneous arrival of real science in the seventeenth century, something like democracy with the Glorious Revolution in England and the republics in The Netherlands and the predecessors of modern finance in the Bank of England and the Bank of Amsterdam. And a fourth one I would throw in there is the Westphalian nation state, which also arose in the seventeenth century. What do you think the world’s population was in 1700?
Pablos: Hundreds of millions, not billions.
Jim: I think that there is—I am a person like you, actually. You’ve well, you did lay this out in the beginning of your book. I believe you go either way. We could go into the toilet and have a major technological society collapse, or we could take off for a glorious future if we modulate the takeoff correctly. And it’s all in our hands. At the moment, not looking too promising with, we talked about this impedance mismatch between knowledge and execution, between values like, wouldn’t it be great if we just said everybody on Earth’s gonna get four toasters, and we’re gonna figure a trajectory to get there. You could never sell that in the United States ever. They’d lynch you. Right? There is a road to a glorious future for humanity. Whether we’ll take it or not, I think it’s an open question. And as to your point, it’s up to us to figure out that road.
Pablos: Well, look. I’m not trying to convince anybody what the right number of humans is, but I think we’ve gotta recognize that we made you and me and all of our friends and all the people on Earth—we made them, and we need to take responsibility for providing for them and make that possible. So I think right now the target is provide for eight billion people. We need a 10x global energy production to do that, which sounds crazy, but we’ve done it before. We’ve 10x global energy production in the last century. Now we did it by burning more coal and oil and gas, and we had to dig all that stuff up out of the ground, which was a lot of work. This time, hopefully, we learned our lesson. We don’t wanna 10x those things, and we can find cleaner ways to do it. But, you know, if it’s just deploying nuclear reactors, it’s a lot less work and a lot less fuel than last time.
So that’s what I think is the real goal here. And, you know, look, we’re having a lot of societal problems that are causing deep population because humans are having their own crises of meaning, but I don’t think it’s necessary. Like, you could find a lot of meaning right now just by helping take care of the people we got. So that’s kind of how I think about it. Population—we should be able to handle even more people if we learn to do a good job of it, if we learn to make more functional societies.
And you look around the world, different societies are good at different things. I think people are not doing a good job of using their shopping instinct and looking around the world saying, okay, who’s got the best solution for x, y, or z? Why don’t we have the trains that China has? Why don’t we have nice clean sports ball stadiums like Japan? Why don’t we have great tomatoes like Italy? Like, what are the things? We should be trying to learn from the best, and we’re not the best at everything. And so I think to make better future societies, we need to look around at all these experiments and cherry pick the best results and put them together. So I’m optimistic that generations yet to come will do that as long as we can leave them in a better place than what we found when we got here. So that’s how I think about it.
Pablos: Yeah. And one of the things I really take heart in is that if you look on the longer time horizon, humans actually do a pretty good job of improving and choosing the better route. You know, we suck in the short run. And I think the way I try to describe this in the book is I said, look, whoever invented the wheel was probably assassinated. And then for several generations, we’d probably kill off anybody with a wheel. And then at some point, the kids are like, fuck you, dad. Wheels are cool. And it’s been wheels ever since. And so you have to look at this kind of developmental stages or the life cycle that technologies go through. And we’re certainly, like, in junior high with social media and still very immature and poorly behaved. And we’re barely in preschool with AI. But other totally dangerous stuff like, I don’t know, like fire, we have graduate degrees and we’re handling it pretty well. I mean, a little regression recently, but by and large, we handle fire pretty well and anybody can get a hold of it. And that’s true for knives and hammers and all the other dangerous stuff. So you have to learn to wield these new technologies and apply them in a positive fashion.
Using the nuclear reactor example, that’s one where we’ve got this crazy mushroom cloud story in our head. We outlawed them for generations. We assassinated anybody trying to build nuclear reactors, and then my kid thinks Chernobyl is a TV show for old people. These kids are gonna build them with or without us because they see—so these technologies over generations, humans do learn to use them. Once we invent them, prove that they’re cost effective, prove that they’re good, we always adopt them in the long run. There’s almost no exception you can name of a technology that’s been through that, and we still don’t use it. So I think you can kind of take heart that even though you can look around you and say, okay, people really suck, and they suck at making decisions, they’re doing all the wrong stuff. The truth is in the long run, we do course correct, and we do the best things, and we do get better. And you know, you and I might not be around to witness it, but I think in the long run, these things will happen. What I’m trying to do is figure out how do we arm those generations with the best set of technologies, and can we prevent getting on the wrong track? Can we maybe get some of these things to life sooner so that they don’t have to take generations to recover?
Jim: Of course, the problem I would say is that we are the rate of change in so many things is accelerating rapidly. So the sheer in our society is increasing. You talk about the meaning crisis, I would call it the collective sense-making crisis. I’ve now moved that to the top list of what could cause a serious social collapse, that we can no longer collectively make intelligent decisions. Now the Chinese can still do it. They could show us how it can be done. You know, the Europeans, they have their strengths and weaknesses. US has some strengths and weaknesses too. But overall, the collective sense-making of humanity, particularly in the West, seems to be declining. So there’s a big question there. But let’s not solve all the problems of the day. Let’s focus in on one I know you’ve spent a lot of time thinking about, which is in addition to fission, what are some of the other renewable energies that we should be thinking about in this, you know, ten to a hundred year time frame?
Pablos: Yeah. So because I’m—you can tell I’m like pretty lit up about energy. Because I’m a lightning rod for these mad scientists, I attract a lot of perpetual motion devices and crazy concepts for energy production, and it would be insane to invest in any of them. But it might be genius to invest in all of them. So I’ve basically been backing a whole bunch of wild ideas in case any of them works. And I think some of these things could just turn out to be miraculous.
For example, one of the ones I didn’t want to like at first, but I started looking into it is that if you look at what’s wrong with a solar farm, it suffers from basically two big problems: clouds and nighttime. If you just look at how the relentless onslaught of night has been messing with our solar panels, we could actually just take the solar farm, throw it on a Starship, and launch it into space. A solar panel in space will get eight times as much energy twenty-four hours a day all year long, and you can beam that energy down to Earth using radio waves. Those radio waves can go right through clouds in the middle of a snowstorm.
This is a way to solve the biggest problem with solar, which is it is not baseload energy—it’s intermittent. If we put them in space, you can make energy all year long, turn it into baseload energy, and it can be cheaper than coal. It’s carbon-free energy that scales. You can beam it to cities around the world. You don’t need storage, and you don’t need transmission lines.
This sounds crazy. It sounds like science fiction, and it really was science fiction. But we have all the technologies to do this. People don’t realize—obviously, the solar panels are solved. They’re really good now. We have better ones coming all the time with new solar cells. But the other problem is beam forming and beam steering, and that’s not intuitive to people, but I had known a lot about it since we worked on those technologies at my lab before.
The real hard part of it is launch cost. Launch cost for putting an iPad in space would have cost you $40,000 on a space shuttle. SpaceX has that down to about $1,500 now. And their target for Starship, the big rocket, is $10—$10 a kilogram to put stuff in space. So I contend that by the end of this decade, you will store your old sports ball equipment in space instead of your garage. It’s getting to be that cheap. Because of that, we can start to build in space. We can build industries out there, and the most important one could be to solve energy.
We have a company called Vertis Solis that’s aggressively trying to commercialize this, and they have plans to deploy the first commercial array into space in like four years. That could end up being one of the most amazing ways of scaling up and solving energy. And so that’s one of them. I can go on, but you got the idea.
Jim: Yeah. And, again, this is what I’ve been interested in too. One I did a deep dive in twenty-one years ago is mass electrical storage. Utility grade and I did a six-month study, and I put it all in. I hired the world’s leading electrochemist, and I did build my models, got all my data. I put it in a box. I labeled “open in twenty years” because the technologies did not make sense, but the rate of change was such that in twenty years it might. Here we are twenty years later.
Pablos: So had you figured out that everything but launch cost was gonna be solvable?
Jim: Well, this wasn’t for space-based energy. This was for mass storage, utility grid battery.
Pablos: Oh, mass storage. Oh, energy storage. Oh, battery. Okay. What do you think now?
Jim: I’m not sure. I haven’t done a deep dive, but I have seen some signal like flow energy, for instance. I got a very good model in my head of what the cost was. The thing was amazing. Even back then, there were billions being invested by big companies on very foolish things that would never work like flywheels. I could, in five minutes, tell you why flywheels wouldn’t work. Idiot company called Energy Vault actually went public with the idea of raising concrete blocks into the—I could refute that in five minutes, but people invest in all kinds of stupid stuff.
Pablos: I think they’re still at it.
Jim: Yeah. I think they’re still at it.
Pablos: It’s actually one of the best right now.
Jim: But their stock price is down to like a dollar or something. Right? But the math doesn’t work. Sorry. Not gonna work. But flow energy could, right? So if their claims are true, it could work. In which case the intermittency problem then starts to go away. Because, for our listeners who aren’t in on all this, our current grid, particularly in the United States, can only take a certain amount of intermittent energy, 15, 20 percent.
Jim: But, their stock price is down to, like, a dollar or something. Right? But the math doesn’t work. Sorry. Not gonna work. But, flow energy could, right? So if their claims are true, it could work. In which case the intermittency problem then starts to go away. Because, for our listeners who aren’t in on all this, our current grid, particularly in the United States, can only take a certain amount of intermittent energy, 15, 20 percent.
Pablos: With storage, it’s one of these things where people are not being honest about the arithmetic, and I’m quite frustrated about it, because what happens is people get really excited about solar farms, and then they sort of wave their arms about batteries and just presume that batteries are coming. When I wrote the book, I looked up the biggest battery deployment in the world, which is also in China, was big enough to store energy from a solar farm and power the city of Spokane through the night. That might not be the best analogy since hardly anybody’s ever bothered to go to Spokane. But if you do go to Spokane, what you’ll find is that everything closes by 6 p.m. There’s nothing to do after 6 p.m. So I think that’s probably why you could power Spokane through the night. And that’s not to mention the fact that windmills are even worse. We don’t have any hope of a storage system to store energy across seasons. And seasonally matters for solar too. Like, in Germany, what are you gonna do? Store solar power through the winter? It makes no sense. I’m with you. I think the basic arithmetic needs to get done. People are being disingenuous about that. I’m kind of disappointed in how much attention has gone into battery storage as a way of trying to sell people on solar farms. I think solar is great, but you gotta put it in the right place. And that right place is—look, I’m from Alaska. It’s not gonna work there. So put it in space, and then we can make it work.
Jim: Interesting. Back when I was doing my deep dive in 2002/2004, the world’s largest utility-grade battery backup was in Fairbanks, Alaska.
Pablos: Yeah. Oh yeah, it was—
Jim: A boondoggle because it did make no economic sense, but it was a proof of principle. It used NiCad batteries, which are really great batteries. You can recycle them a zillion times, but they’re very expensive. And the whole city of Fairbanks had, like, four hours of backup on NiCad batteries that the government had paid for as a demonstration because Fairbanks was at the end of a single power line at the time, which was kinda bizarre.
Pablos: Fairbanks is a shithole. I’m sorry. Everybody there knows it. Alaska is a pretty rough environment. Fairbanks is in the middle of Alaska and really not a pleasant place to be, especially in the winter. And Alaskans—maybe less so now, I haven’t lived there in a long time—but when I was growing up, blackouts were a very normal problem. It happened all the time. So you can see why even though it didn’t make economic sense, why a city like Fairbanks would try to use battery backup. Probably, I’m sure by now they’ve jettisoned those batteries and they just have diesel or some other gas-powered generators. But it’s a tough problem providing scaled energy for places like that, and I think it’s very important. But yeah, we’re just gonna put nuclear reactors in Fairbanks and solve the problem once and for all.
Jim: Interesting. Now the other storage, like you said, the seasonal storage one—again, I penciled it, doesn’t quite work yet, but maybe a factor of two in electrolysis costs is, build massive solar farms on the Moroccan coast and hydrolyze water, and then use the Siemens process to convert hydrogen to liquid fuel. And then you can store it as long as you want, but at least within reason, fifteen years.
Pablos: Yeah. And I think there’s potential in some of these things. I’m not—again, I don’t wanna just kill them. We should try to develop them, but I think we need to be honest about the arithmetic and what makes sense. And that could vary widely depending on the deployment scenario. Like, I lived in Washington, Seattle for a long time. Washington State has the Columbia River Dam. That’s an amazing resource, and you should use it because you got it. But Arizona didn’t get one of those. So solar panels might make sense there at least to augment the baseload. But look, I’m not a total curmudgeon here. I just think that you have to be smart about the decisions you can make, and you gotta update your priors and take a look at what’s state of the art right now. Solar panels have made a lot of progress. We have gotten the cost down. We’ve scaled production, thanks to China and thanks to Germany for subsidizing it. But doesn’t matter how cheap your solar panel gets, it doesn’t work at midnight. And so I’m trying to help people see that our toolkit is much more advanced, much bigger, and can help us out here.
Jim: Yeah. As you say, you have to do the math, do the arithmetic. Right? Now, of course, you have to make some assumptions about the future, but I often laugh about Germany. I have a line: Germany is far to the north and one of the cloudiest countries on earth. I always wondered, did any of the politicians that signed off on this unbelievably massive investment in solar ever step outside?
Jim: Yeah. As you say, you have to do the math, do the arithmetic. Right? Now, of course, you have to make some assumptions about the future, but I often laugh about Germany. I have a line. You know, Germany is far to the north and one of the cloudiest countries on earth. I always wondered, did any of the politicians that signed off on this unbelievably massive investment in solar ever step outside?
Pablos: Yeah. I met one. I ran into someone from German parliament, and I had just been on stage at this event, and I threw Germany under the bus for shutting down their nuclear reactors and spending all their money on solar farms. Even so, she tried to argue with me and defended the choice, saying, well, it was important to subsidize the production of solar panels to get the cost down. And we do owe Germany a debt of gratitude for that. They did proactively waste a lot of money on solar panels and build that industry in China so that we could all get them for cheap, and I think that’s great.
But when you look at the actual power output from German solar farms, for half of the year, they’re getting two, three, or four percent of the energy that they paid for. At best, they’re getting eighteen, nineteen, twenty percent. That’s at noon in the summertime. They don’t talk about those numbers. They don’t talk about capacity factor. What they show you is deployed capacity, which is sixty, seventy gigawatts now, but they’re never getting anywhere close to that. The best they’re getting is twenty percent of that in the summertime. So it’s just not a solution. This is why we’re experiencing these wars with Russia now. I mean, it’s really a sickening thing.
And so I think we have to course correct on this. Actually, Germany seems to be doing it lately from what I’ve heard. They’re looking at reopening their nuclear reactors, which is great news. People don’t realize that when they shut down those reactors, the answer was just buy more gas from Russia. And then once COVID hit and they couldn’t get Nord Stream 2 up, they started buying up wood pellets from every forest in the world. All these protected forests that we thought we were saving to do carbon sequestration have been razed to send wood pellets to Europe so they can burn them. I mean, this is so regressive. It’s just sickening, and no one talks about it. But look—we don’t have to be on that fork in history. We can totally course correct, and that’s what we’re working on doing right now.
Jim: Alright. Well, one last thing before we move on from energy. If you’re thinking about good renewable energy, everyone always says, what about fusion?
Jim: Alright. Well, one last thing before we move on from energy. If you’re thinking about good renewable energy, everyone always says, what about fusion?
Pablos: Well, look. I, as much as anybody, want a magical fusion reactor. For my lifetime, it was one of these things that clearly was science fiction. People were researching it. They were making very small amount of progress. I think that we are in a different stage now primarily because of Commonwealth Fusion. They’re the ones who figured out they could use a new kind of superconductor to make a super magnet to do 1970s-style plasma fusion. So everything they’re doing is the same as it was in the seventies, but now they have a better magnet. And that’s really cool. I wish them luck. I’ve been over there. They’re doing amazing work.
There are some derivatives like THEO, which is doing an accelerator version of that, which is also cool using superconducting magnets. But these are possibly out of the window of science risk where they were for our lifetime, but they are in the window of extreme engineering risk. There’s a lot of very unsolved engineering problems with all of the plasma fusion reactors. And so I am hopeful, but I really don’t see those things mapping to a venture time horizon even though a lot of other venture investors have put money into them.
I think it’s a little dangerous because when they don’t get those returns on the ten-year horizon like I described, their investors are going to be pissed off, and they won’t have capital to deploy into things that actually work. So I think it’s important not to be disingenuous about how long it’s going to take these things to happen. But there’s a whole bunch of people working on fusion. We have backed one fusion company, but they’re doing what is called nanoconfinement fusion. This is not widely accepted, but we see it working in the lab. And so we think there’s a shorter, easier path to fusion reactor with nanoconfinement fusion.
We see other ideas in the lab that I think could make much simpler fusion reactors. One of them I wrote about in the book called sonofusion, where they use what was called bubble fusion in the past. And for a while, it got on a very storied track with some fraudulent activity and stuff. But the team at UCLA has figured out why it wasn’t working, and we think they can make it work now. And so I’m hoping that one makes some progress because if it works, it’ll be a very simple fusion reactor.
So there’s things like that coming, but I think in my mind, most of these things are not venture compatible yet. And there are other ones I won’t name that I think are fraudulent fusion reactor companies that are getting a lot of investment and support. Most of the fusion, with the exception of Commonwealth and maybe one or two others, most of them are running on some kind of secret sauce, so they won’t tell you how they do it or why they think it’s going to work. And I think they’re taking advantage of investors who can’t pick this stuff apart.
So I’m expecting a lot of unfortunate failures in the fusion space, and I’m sad about it because I think it pollutes the well for everyone. But look, it’s exciting on a longer time horizon. I think on a ten-year track, people are being a little too optimistic. And from a venture perspective, from a research perspective, I think it’s a fucking amazing time to work on fusion. I feel like a curmudgeon on your show right now. I’m really sorry, Jim. I know I’m ranting about a lot of things.
Jim: No. I think it’s all good. This is what I always say. You know, think sharply. Don’t believe bullshit. Right? Okay. Now, last big tech thing we’re going to do or not our big industrial thing to do, a step in a very different direction. We’ve been talking about hard physical things, you know, nuclear power plants, computers, etcetera. You have a whole chapter dedicated to the apparel industry.
Pablos: So apparel is another over $2 trillion industry. Every human on earth is a customer, and it is probably after oil the shittiest industry on earth as far as environmental damage, labor practices. I mean, we’re really talking about an industry that is probably responsible for at least 10 percent of global CO2 emissions, probably 20 percent of global freshwater pollution. And every year, 150 billion garments are manufactured. Thirty percent of those are never once worn by a human. So imagine if we drove a third of the cars we made straight from the factory to a landfill. That’s what’s going on in apparel. It’s almost as big. It’s just sickening.
And, you know, even in the best case scenario, over half of all the clothes produced end up in a landfill in less than a year. So the point is in apparel, the big problem is overproduction, and it’s because you are guessing usually six, seven, eight, nine months in advance what is gonna sell. What colors? What size? What design? The volumes. Because you have to order that stuff from a sweatshop in Asia, then it’s gonna end up on a boat coming back, go on the shelf, and you don’t find out what’s sold.
So you could flip all this on its head if you could produce on demand. And what I mean by that is you wait for somebody to click buy now, cut and print and cut and sew their clothes in an automated factory and ship them out the next day. And we have prototyped this. It works amazingly well. We have a company called CreateMe in the Bay Area. By, I think, early next year, they’ll be making the highest quality, lowest cost T-shirt in the world in California. And that’s probably the dumbest place in the world to make a T-shirt. And so this factory is fully automated. They’ve figured out how to make, have robots make the clothes. They’re a tenth of the labor and a tenth of the real estate of an equivalent factory in Asia. And so it’s a way to totally change that industry and now and stop making things that nobody wants. And I’m really excited about that. So that’s the kind of thing that’s coming.
Jim: Alrighty. Well, I want to thank Pablos Holman for talking a very interesting deep dive into all kinds of new things in his new book, Deep Future. And what’s the name of your VC firm in case we have some maniac mad scientists who are gonna send you their perpetual motion machine design? Where do they send?
Pablos: I can’t wait. The venture firm is also called Deep Future, and the podcast is also called Deep Future. And our website is deepfuture.tech.
Jim: Alrighty. Well, I want to thank Pablos for an exceedingly interesting conversation here this afternoon.
Pablos: Thanks so much, Jim. I love your show. I’m really thrilled that I got to do this with you.