Josh Wolfe: Our new world order is one where algorithms can wield as much influence as armies

Danny Crichton:

Josh, welcome to the program.

Josh Wolfe:

Good to be with you, Danny.

Danny Crichton:

All right, so we're talking about our Q2 2024 letter, which had the theme of V for Vendetta. The purpose of this kind of discussion was we're in the throes of the presidential election going on. Just a few weeks ago we had the debate. Extremely aggressive on all sides, lots of negativity, lots of nihilism, and you sort of took this theme of V for Vendetta, which is really about a singular anti-heroic nihilist who is fighting against a totalitarian government, is bombing buildings, is wearing the Guy Fawkes mask. There's sort of this lack of philosophy. There's no sort of point. He's not trying to build freedom. He's just fighting against a system in that way that I think is a through line between The Matrix and Fight Club and a bunch of these other movies. You sort of pivot from that and say, look, we can't have a world of this sort of nihilism, but we need to go towards something positive. There has to be a philosophy that you're working towards and you direct our attention to science.

Josh Wolfe:

Yeah. This is a through line of some of the conversations we've had in the past and a consistent theme at Lux, that it isn't this Pollyanna-ish, let's go towards the future and have this proto-utopian or techno-utopian pursuit. It's not the nihilism of the V for Vendetta, the lone individual who's trying to take down the institutions and just rally anarchy. It is about the preservation of a certain kind of institution. That institution is the institution of science. That does not mean buildings with columns and bureaucracy. It is the institution of lots of conjecture and hypotheses and criticism of those in the classic Karl Popper sense. It is the pursuit that many people believe of truth and progress, but really it's about the vainglorious nature of scientists in pursuit of their own ambitions and status and claiming Nobel Prizes and winning prize money and beating down foes and competitors.

They do form this collaborative community of comrades. They are looking around for what we call the cogent consensus that stands up to the scrutiny of criticism and competitive spirits of scientists. It is what progress comes from. It is a great institution to behold. When you look back at the annals of history, you pick your headline-winning crisis of the day, the Cuban Missile Crisis, assassination attempts of presidents, the first Iraq war, impeachments of presidents, scandals in the White House, market collapses. None of those things stopped science. None of those things stopped science, science continues as this great machine of progress forward fueled by these individuals. We'll talk a little bit about that. It used to be the sole individual in a lab and now it is so complex and institutionalized. That's a great virtuous thing because it's harder to destroy, but none of those things throughout history stopped science, but science stops lots of bad things.

Science stopped the Covid pandemic. Science stops an epidemic of obesity in the United States right now with the GLP-1s. Science stops the slow bureaucracy of limitations in compute as we develop ever better GPU. Science stops our limitations to extend to the heavens and launch rockets and satellites. Progress comes from science which continues endlessly and it's not guaranteed, but it is this innate desire of individuals to want to achieve great things and they want to achieve those great things, like I said, more in a cynical Hobbesian sense than a virtuous one. We all benefit from that. Yes, lots of things stop, but science doesn't.

Danny Crichton:

I think one of the things that go through this entire letter is this competition between scientists leads to superior science, which leads back to competitive advantage again. That's a theme that you sort of carry through this entire letter. Talk a little bit about that.

Josh Wolfe:

Yes, this idea that superior science begets competitive advantage. One of the things that we write in this letter is we have a new world order. That new order is one where algorithms can wield as much influence as armies, where vaccines vie with warheads for strategic importance, all of that is superior knowledge. The true measure of a nation's strength lies not in its ability to just manipulate its people or project weapons, but to harness the power of knowledge to further its ends. Whether that's for energy, whether that is for economic development, whether that is for housing, whether that is for healthcare, all of those things from forecasting the fury of monsoons to tracking those trajectories of missiles or sending them they're on their way, that is from scientific knowledge and scientific knowledge is power. Again, not a virtuous thing per se, but something that gives people advantages. It gives us individual advantages in the way that we deploy in new science. It gives companies advantages as they compete with other companies. Of course it gives countries advantages.

There is nothing more powerful than looking at Iron Dome on display, at the superiority of technology and science, the sophistication of optics and radar and calculations and computations to detect a projectile intended to hurt or kill people and produce human suffering and to prevent that because we have superior science. Yes, superior science begets competitive advantage and frankly, all of that depends upon a culture that celebrates that science.

Danny Crichton:

Well, it is interesting because you talked about forecasting the fury of monsoons, tracking the trajectory of missiles. You focus on Iron Dome. One of the things I learned recently is that the amount of turbulence we experience on planes has actually declined dramatically over the last 15 years because our computer modeling of weather has actually improved so dramatically. We're actually able to reroute planes up, down, left, right in space, in order to avoid little pockets of turbulence. Unbeknownst to me, but known to pilots and to the FAA, our flights are actually getting smoother because we're able to actually use artificial intelligence to predict those flight paths better than we've ever been able to do before.

Josh Wolfe:

Think about the complexity of turbulence, to be able to do fluid dynamic modeling in computations to see the unseen, I mean quite literally, pockets of air differential pressures, temperatures coming from hundreds or thousands of miles away. This is such a complex thing and we're able to identify it and then adjust this flying aluminum tube in the sky going 600 miles an hour, as Louis C.K. was like, "The entire time, you should just be gripping your seat and screaming, oh my God, this is amazing."

Danny Crichton:

Yes.

Josh Wolfe:

It truly is amazing. The miracle of flight is amazing. Just as you noted, little tweaks which derive from somebody saying, "This sucks," turbulence sucks, the unpredictability of it, the fact that people freak out for two minutes, that flight attendants or stewards and stewardesses are told to take their seats, that provokes fear and how do you reduce the fear and the psychological suffering? With superior science.

Danny Crichton:

I was recently just reading Antoine de Saint-Exupéry, the author of the Le Petit Prince.

Josh Wolfe:

When you read that, just so people understand, you read it in French.

Danny Crichton:

I do read it in French, I read both.

Josh Wolfe:

I look at the pictures. I look at the pictures of the Little Prince in the graphic, yes, but you read it in French.

Danny Crichton:

I do read in French. Wind, Sand and Stars, along with some of his other books, are really these testaments to flight. One of the amazing things, it's existentialism, but he is a pilot.

Josh Wolfe:

Yes.

Danny Crichton:

He gets into this plane and it is a rickety box of metal and he's flying over mountains. There is no radar, there are no comps. One of the things, I mean this is about the 1910s to twenties, thirties, forties, you have no way of communicating until you are with an eyesight and are able to use a lamp to basically do Morse code to the control tower and say, "I'm landing. I'm coming over here right now." For hours, you're by yourself. If you're in the middle of a storm, you can't see anything. You have no idea where the mountain is, which is why so many people fail to survive.

When I think about the miracle flight, it's not just what we are able to go into space, it's all these supplementary technologies we didn't even think about. The ability to predict weather, the ability to see where other planes are, the ability not to crash, to have systems that are able to take off in beautiful patterns. We just upgraded a lot of the control towers at FAA so that we're actually able to save several percentages of fuel on every flight today to help improve climate change. We're still iterating on this core technology. It's a hundred years in and we're still not fully done.

Josh Wolfe:

Well, the beauty of almost everything that we look at in modern society once technologies get adopted from some scientific principle better understood and capital and talent flowing in to solve that, is that we look at these things that we take as modern technologies and you almost always invariably look back and say, "I can't believe we used to do it that way." I think you're mentioning like airplanes, we still have not figured out ideal efficient loading patterns to get people on and off. It's one of these things like turbulence, it's a mild inconvenience and it's the kind of thing that probably in 10 or 15 years we will look back, if not sooner, and say, "We figured out an algorithm not only for loading people in the traditional configuration of this, but we've figured out a better model."

I'll give you one which is a little bit ridiculous, and it's inspired by the Star Tours ride at Disney World or Disneyland. When you are loading into that vehicle, you load into a bunch of seats and then those seats move into the flight simulator. You're sitting there already off the tarmac at the gate with your seats. Some people there with their coffees and their plugs, some people spread with their bags. If you had a preloading area, you could sit there, you could effectively load into your seat, and then boom, all the seats just swap out into another thing in sort of a modular transition. Somebody will figure that out, too capital intensive to retrofit our planes. If you were doing first principle thinking on this, you would identify as most people do, entrepreneurs and scientists are like, "That sucks. Why does it work that way? Let's improve it." That's how you get all progress.

Again, you go back through all time for the past a hundred to 50 years. I mean through all modernity, we constantly are iterating and improving and against all these crazy things in the world from wars to epidemics. You go to the 1940s, you had Norman Borlaug and the Green Revolution saving billions of lives. You had the decoding of the structure of DNA, a competitive process itself, with Watson and Crick and Rosalind Franklin laying the foundation for modern genetics. The polio vaccine in the fifties, saving millions and millions of people from paralysis and death.

The laser, which at first people thought, "Hey, this is going to be great for war." Little did we know that it would be used for Compact Disc players and the revolution of music and TV and whatnot. You had surgical tools that came from that. You had the Apollo 11 and the moon landing, which begot inspiration for so many and ultimately today's modern versions of SpaceX. GPS, the PC CRISPR gene editing, this never ends. When everybody is like, "Oh, productivity is down and America is in decline," just open up a science textbook and look at what we have today compared to what we had a hundred years ago. The marvel of progress is just a stunning one.

Danny Crichton:

I literally was writing two hours ago the newsletter about how productivity is down and stagnating.

Josh Wolfe:

You and I always disagree about this, and this is the one through line is that I am so skeptical about human nature. I think people suck. I just do.

Danny Crichton:

Yes.

Josh Wolfe:

They're vainglorious, they're ego driven, they're petty, they're spiteful, they're vengeful. We're fallible. We have urges that lead us against better moral judgment. All these things make people suck individually, and collectively they sort of filter out and we punish people and we put them in jail or we shame them and we degrade them. Science progresses, it doesn't matter. Sometimes as the old quote goes, "Science progressed one funeral at a time," because you need old ideas to die with the people that hold those ideas. Regardless, again, of what's going on in the world, I truly believe that the delay between a insight and a breakthrough and the impact on productivity, sometimes poorly measured in the statistics. The invention of the internet, and its promulgation and diffusion more rapidly than every prior technology before. ChatGPT and modern AI tools, again, more rapid diffusion than ever before.

GLP-1s my entire life, people talked about, "We're going to have a pill that will cure obesity." People thought this was crazy. It exists. People are going from being on seven different medicines to being on one or two. We don't know the longitudinal effects, but my gosh, these all comes from science. Science is about understanding and having better explanations for the way things work, including our own bodies, of which we're still learning so much. How does this work desire for cravings and food and cycles of different hormones that are triggering you to overeat or undereat. I just continue to be utterly inspired by the process of science and utterly repelled by human nature.

Danny Crichton:

Well, and you and I are both fans of Will and Ariel Durant's History of the World Book, and the thesis of that book is basically competition is the core unit of understanding human history. Societies compete, individuals compete, they may be vainglorious, they may be egotistical, but it's precisely those negative characteristics that create positivity. That competition trigger scientists who want to go discover, build, find new progress, etc. Now the question is you have been very concerned about culture.

Josh Wolfe:

Yes.

Danny Crichton:

The culture of science, the culture of progress in the last couple of years. You were just describing the 1950s and sixties, this amazing, really just panoply of time where new appliances, I mean you talk about the Green Revolution, which saved a billion lives, but that was also the era of the dishwasher. It was also the era of the laundry machine. Tools that suddenly made our lives dramatically simpler. Laundry went from a 10-hour chore to throwing a couple pieces of clothes into a machine, hitting a button, and coming back two hours later.

Josh Wolfe:

By the way, technologically ending a form of gender apartheid.

Danny Crichton:

Ending a form of gender apartheid. What has changed to the culture and what are you worried about?

Josh Wolfe:

Well, a late partner of ours, Larry Bach, who was an incredible scientist and entrepreneur and venture capitalist and started Illumina and Pharmacopeia, Neurocrine, Athena, done all these incredible companies and sadly passed away in his late fifties, but was a huge inspiration for us at Lux on creating new companies, particularly in biotech, particularly launching what we call Lux Labs, these new companies that, and it's never going to end doing those kinds of things. I'm sure we'll talk more about that in the future, but he had this great line that cultures get what they celebrate. If you celebrate celebrities, back then it was like Paris Hilton and Kim Kardashian, and maybe today it still is, then that's what you're going to get children to want to be. It's the difference when you look at TikTok that is infiltrated the minds of American youth and what they are fed and what they think and believe and spend their time, versus what China itself is doing in its version of TikTok. They are flooded with STEM, science and technology and engineering and math, and we're flooded with nonsense.

I think that there is a cultural imperative, a moral imperative for the leaders of our country, be they investors, scientists, great entrepreneurs. I'm critical of Elon on a lot of things, but I think he has a great inspiration for people far more than the Kardashians. It is just important because you can put the spotlight of societal esteem and you can shift it from what people believe are these prestigious professions. It used to be in the eighties and nineties, greed is good and people wanted to be investment bankers. Now seeing people want to go into aerospace and defense, that wasn't because they read a book, that's because they saw somebody like Lux family founder, Palmer Luckey and Palmer is interesting and intriguing and controversial and right. He's creating a lot of economic value in developing new weapons systems and really helping to make America great again in that sense of restoring the arsenal of American democracy. It's inspiring people to want to work on these hard problems.

That is exactly kind of the cultural ignition that you need to light an entire generation to say, "I want to go and do this and pursue it," as opposed to just, you ask many young people what do they want to be and they say, "Rich or famous," instead of, "I want to be a doctor or a scientist or an astronaut." Celebrating that I think is a moral imperative. It starts from K-12 with science fairs, the many prizes that are awarded from Kavli to Keck to Lasker to Turing to Nobel. Cultivating that culture that celebrates the vanguard of the visionaries, I think is absolutely critical.

Danny Crichton:

One of the things that you have as a through line through here is this balance between individuals and groups.

Josh Wolfe:

Yes.

Danny Crichton:

The tradition of science has always been the individual solo scientist, the Newton who sort of has an apple drop and figures out physics. Now labs are not one person, they are very large structures. How does culture play a role between those two dynamics?

Josh Wolfe:

I think that there's two main drivers of this. One is capital and one is complexity. On the capital side, big breakthroughs cost a lot of money. To be able to cross the frontiers of science, you have to cross the barriers of fiscal feasibility. You've got the CERN's 27 kilometer circular stage, you've got the square kilometer array for radio telescopes, the James Webb Telescope. All these things are not just like hundreds of thousands of dollars in a nice lab in Princeton, New Jersey. They are tens of billions of dollars. Only governments that effectively have a monopoly on that kind of capital and the ability to deploy taxpayer money into those things, which again, I think is a virtuous thing, can do that.

Now, that's the capital piece, that incremental discoveries at the far reaches of the universe cost a lot of money, whether that is going to see things out in the celestial heavens and explore the birth of the universe and get understanding from that, to the subatomic particles and understanding how life itself works. One piece of this is cost. Now the good thing on the cost, I believe, is that we spend about the same on R&D as a percentage of GDP at the government, as we do on defense. It's around 3%, up from 2.5% percent since the fall of the Berlin Wall, which is significant and we want to spend productively. You don't want it to be wastefully, but I think that that's really important.

Now, a second piece of that is that a greater percentage of that is actually coming from the private sector. That's also good because private sector, going back to this theme of competition, competes with each other. It is good to have Google and Microsoft and Meta and Amazon and Oracle all competing to build data centers and come up with more energy efficient approaches and foment demand for nuclear power to power those kinds of things. All of that is good. You see the private sector doing more and more basic research.

Some of that is correlated to the cost of capital. When interest rates are high, less of it gets done. When interest rates are low, lots of experiments get tried. One of the themes at Lux, which we have been taking advantage of over the past two years, and I think probably for the next two as well, are taking divisions out of big companies where they funded a lot in private R&D, doing really cutting-edge cool things that now are considered non-core, maybe fiscally irresponsible. They are creating the opportunity for us to come and knock on the door or sometimes break the window and say, "Come on out, we're going to start a new company around that."

We did that with Osmo, for digitizing smell out of Google. We've just did that out of Apple with a company that hasn't yet been announced, but it's their first real corporate spin-out, we're really excited about. We did that out of Apple years ago with some technology and people for 4D LiDAR in a company called Aeva. We did it out of Meta with Evolutionary Scale, which is the real frontier model for biology. That first piece of capital is really important.

The second piece of scientific collaboration competition, is it used to be a lone practitioner in a lab and they would sometimes labor in obscurity or sometimes have a great breakthrough. Now I always say chips on shoulders, put chips in pockets, and you want nothing more than a scientist who is doubted by his peers or her peers so that they can push through. Because science gets more and more complex, you have lots of interdisciplinary collaborations. That means a computer scientist is working with a biologist and maybe they're working with somebody in Tel Aviv or somebody in Taiwan or somebody in Mexico City. The average number of authors that you see on a scientific paper today are astounding. It went from one or two or three from the same institution to 20 different people at 12 different institutions.

I think what you will see next is probably an AI scientist that is actually involved in the production replication conjuring, and they will get credit in the materials and methods if not as a named author on some of these papers. Those are the two things I think that are shaping the culture of science. One is the availability of capital, really the monolithic monopoly hegemon of government spending and we should do more and more and more of it to have these big projects. The second is the more and more scientists that are becoming more and more authors on papers and the interdisciplinary collaboration, all of which underpinned by this culture of wanting to advance progress and knowledge.

Danny Crichton:

I think you also read Benjamín Labatut's The Maniac, and I think you read his earlier one as well, When We Cease to Understand the World. I mean he to me, is the only fiction writer at least that captures this sort of notion of, in The Maniac is John von Neumann, the inventor of the Von Neumann architecture for chips and inventor of game theory, but basically a long list of radiology for the Manhattan Project, which just is involved in all these different institutions.

When We Cease to Understand the World is really focused on quantum mechanics and that world. It is one of the few works that I've seen that gets at this solo practitioner in this transition from at least in The Maniac, you go from John von Neumann, this sort of brilliant Hungarian mathematician to deep mind and this idea of AI went from a single individual to hundreds of people trying to create an AlphaGo, which is sort of the last section of the book. I just think it's interesting that I think even the fiction writers sort of picking up on this theme of how that sort of individual hero that is usually the center of any western piece of work, is becoming a group effort.

Josh Wolfe:

In our letter most recently to our limited partners, we have this line where we say, "No lone mastermind however brilliant, has the requisite knowledge to understand the mind-boggling complexity of science generally, but modern defense systems, let's say particularly." It is really complex. I mean you go back to the iPencil of how difficult it is to even or sketching a bicycle in the tacit knowledge of that, these things are hard. A modern war fighting machine, a missile defense system, a F-35 joint strike fighter, a cutting edge autonomous winged man that Anduril is developing, or the subsea drones, these are really complicated.

You go back to the 1930s, a combat aircraft nearly a hundred years ago maybe had a few hundred components by the 1950s, 20 years later, one generation, that ballooned to tens of thousands of components. You fast-forward to today that F-35 that I just mentioned, there's about 300,000 components. You have this mirroring between hardware and software. The F-4 Phantom, going back 70 years, 1958 or so, that had about thousand lines of code. The F-22, about 15 years ago, was 1.7 million lines of code. Today the F-35 code base, whether it's bloated or necessary, is 5.6 million lines of code.

That is important for two reasons. One, it's really complex, requires a lot of cooperation, competition, people trying to one up each other and then team. At a national level, it's also really important because rival nations can't just simply reverse engineer or pilfer that technology. You can't just take that and suddenly copy it. Copying a modern state-of-the-art defense system isn't just a matter of industrial espionage, it literally demands this lattice of knowledge that's not just embedded in a single individual, not somebody that you can cajole with money or honey traps and get them to come over and reveal the secrets. It's across entire organizations, some of them private companies, some of them government agencies, some of them academic labs, but it's really complicated. In that complexity, there can be great advantage. It's something that we look for when we're funding companies. How do you find something that is so complex it itself creates a moat that others can't copy?

Danny Crichton:

Well, talking about competitive advantage between nations, obviously US, China, artificial intelligence has been on the front pages of every story and you love page A19, but this is definitely like A1 every single day in our world. You just pointed out with the F-4 all the way to the F-35, this massive expansion around coding, the complexity of modern software. Obviously that's been a nexus in the last 12 months for a lot of AI startups, including some in our own portfolio, of how do you have coding agents, how do you have science agents? Sakana just launched her AI scientist as an example, as a working paper and a working kind of prototype. What impact does AI have on the production of science, on the production of code? Is that going to help the STEM complexity and reduce it so that these individuals sort of have power and agency again? Or are we going to see just another layer of complexity added into these systems?

Josh Wolfe:

I think it's going to be both. There will be more complexity baked in, some cases you may have a result that is derived and people want to understand, "Well, how did you get to that," and you don't have a human that can actually explain it. In the same way that sometimes you have serendipity and an outcome and you're like, "Well, I'm not actually sure how that happened, but let's go interrogate and understand what the experiment was to see if we can reproduce it." You're going to have the same thing where a result might derive from reasoning inside of an algorithm of a large language model, that people don't actually understand, "Well, how did it get to that?" increasingly you have this sort of step-by-step thinking to be able to, even in the latest ChatGPT version, that is trying to do that to elucidate what are the methods and mechanisms and step-by-step methodologies to come up with that.

One part is yes, it will become more complex, but I am a firm believer that AI is going to be a huge productivity boon for science. Why we've talked about this in the past, the idea of a scientist on their laptop or their iPad at the beach drumming up an experiment, having that experiment run in a cloud lab operated by robots, with the results coming off and reverse prompting the scientists to say, "Here's the results, here's the parameters that you used. Here's the paper from 1994 or 1972. It might be a spurious correlation, but do you want to rerun the experiment changing these particular aspects and parameters of the experiment?" You just click yes and all of a sudden it runs it again 24/7 and it's safe and it's applicable and repeatable and digitized, and that is going to be a boon to science.

Very bullish on AI contouring hypothesis, able to independently and autonomously run aspects of experiments, not entire experiments, but aspects of experiments. I think that that will be a boon for knowledge creation. Then the question will be, well, who's going to capture that and patent it? Is it going to be institutions? Is it going to be countries? Is it going to become general purpose technologies and insights? Who's going to own that knowledge? Of course, that will be something that is debated in the same way that people look at the provision of credit on a scientific paper when there's 20 authors and there's 10 different institutions, who's getting the intellectual property.

We had a system beginning in the eighties with the Bayh-Dole Act that awarded government money that went to universities and universities then were able to commercialize and own that intellectual property and license it to a startup, that we as venture capitalists could come in and give them a license and a royalty and an equity stream. That is going to become increasingly more complex itself. Bullish on AI in increasing the productivity of science and scientists, and sort of realistic that it's going to be really complex and it's going to raise all kinds of new questions that we don't have answers for.

Danny Crichton:

One of those questions, of course, is just returns. You described the business model of this. There's all these new discoveries. Let's say that actually pans out, the intellectual property, where does it go? More importantly, with billions and billions flowing in from venture capital, from the private markets into AI startups a day, and as we're recording this, there was literally several mega rounds, I think literally billions invested today in the AI world. What is the prognosis there going into the future and how should investors approach it?

Josh Wolfe:

Two things can be true, which is all of this can benefit society and can create great products and great utility, and billions of dollars, tens of billions of dollars, hundreds of billions of value can be destroyed. Right now, the beneficiary of most of the largesse of venture capitalists like us, excited about entrepreneurs who are excited about these breakthroughs, are pouring capital in. That capital in turn is going to one main rent seeker, which is Nvidia. That will change over time. AMD, we've talked a little bit about this online, about the distinction between training and inference and how AMD is likely to increase market share there.

I think that the sheer amount of capital, one of the things that we said in our letter described how the sheer amount of capital is going to just destroy returns for the venture industry. The torrent of capital that's flooding in which demands aggregate returns, stretch the bounds of plausibility. You look at about 125 billion in VC on average over the past two years itself, and we'll get down to AI, that barely matches the performance of the Nasdaq over the past two years. There you have a benchmark of about 11%. You factor in fees and average holding period. The math says you need about a three x return cash on cash, about 375 billion on that 125 billion of annual investment.

If you look at on average, not an individual VC, but most VCs own collectively about 60% of a typical company, then you multiply that, that 375 billion has to turn into 625 billion of market cap created. 625 billion for a company to be created. There's a handful of companies that can realistically reach 20 to $30 billion or more in valuation. There are obviously thousands of companies right now that still bestow this unicorn status that think they will be 20 to 30 billion companies, but they won't. The vast majority of this will see the mean reverting gravity, non-defying results that they basically fail and lose a lot of money.

Within AI specifically, if you were to look at just Nvidia and extrapolate from there how much has been spent on AI systems as a proxy using their earnings, it's about $50 billion just on chips. If you round that up with the annualized spending that includes energy and data center costs, you're at a hundred billion dollars. For the industry just to double the value of that investment, you need 200 billion of enterprise value in the public markets. If you add up all the revenue from our companies and OpenAI with all of its, you're probably in the mid single billions of dollars, so four or five, 6 billion or so just in the past two years. Undoubtedly that is extremely impressive, but far short of the 200 billion that you will need, I mean 99% difference.

I think that like many things in technology, I can be bullish about the technology and bullish about the progression of science and bearish about human nature. The great thing about human nature is that we get overexcited about things. We fund things in excess. The excess produces bubbles, it lowers the cost of capital, it provides that capital to people who can do productive things with it. The vast majority of people will end up broke or have lost.

Danny Crichton:

All right, so if everyone's ending up dead and broke, I mean it's a lot of fun, a lot of activity. There's also the argument you could just live in a shack and do nothing. You end this whole letter for our Q2 2024 letter with this quote from Robert Wilson, who was the former director of Fermilab, I think this is coming out in the 1960s, at least some time during the Soviet Union era, and he's talking to Congress. He's being asked questions of like, "Why should we be funding open-ended undirected science when there's so many high priorities," whether it's defense systems, whether it's social security, Medicare, Medicaid, why should we be investing in all this science, particularly when the cost of capital is going all different directions, people are ending broke, there's all this excess. What was this answer?

Josh Wolfe:

Well, this is a timeless question that will be asked, deserves to be asked, "What good is this? Why are you spending on this? Why waste money? There are higher priorities." It is a timeless answer that he gives. Now against the backdrop then of competition with the Soviet Union, the USSR, he says, "Only from a long range point of view of the developing technology, does this matter. Otherwise, it has to do with, are we," and this is the virtuous part, and I absolutely love this because it is timeless and it speaks to the good in human nature, "are we good painters or good sculptors or great poets?"

I mean, all the things that we really venerate and honor in our country and are patriotic about. In that sense, this new knowledge has all to do with honor and country, but it has nothing to do directly with defending our country except to help make it worth defending. I just think that that is beautiful. Funding all of this crazy far out stuff, it may yield a competitive advantage directly, but more importantly, it is part of the great virtue of what makes our country worth defending in the first place.

Danny Crichton:

Well, Josh, on that note, thanks so much for joining us.

Josh Wolfe:

Thanks Danny.

‍