Vaporware skepticism

Chris Gates:
Yeah, let's just start off with taking me through your argument for vaporware skepticism, and then we'll kind of go from there. Ready? 3, 2, 1.

Danny Crichton:
The countdown doesn't work because I think we have to do an intro, and there's really no intro.

Well, this week I really wanted to talk about a pattern that I've been seeing in the industry recently, which is that as software multiples have declined, and a lot of investors who raised a bunch of money, were investing in SaaS and FinTech in the last couple of years, have been looking for better opportunities. They've been walking over to the science and deep tech area of the venture world.

And so, over the last year or two, and there have been some dedicated firms, we're really not talking about them, there's just been a lot of generalists at other firms who only focus on software who've been, let's just say, meandering over into the science world.

I've noticed a couple patterns. One is VCs have been investing earlier in science-backed startups. So, instead of at that kind of commercialization stage, it's more typical. A lot of VCs seem to be investing at, I would call it the research stage, at the university academic lab stage.

So, I've seen this pattern in fusion and quantum computing and nanotechnology and lab-grown meets where we don't even know how most of this stuff works. We don't even have the right model. We don't even have test reactors that are functional. We've talked about fusion on the securities newsletter a couple of times, but we don't have a reactor today, a fusion reactor, that creates positive net energy. That there is actually more energy coming out of the reactor than that goes into it to run it.

And so, to me, it's nuts to think that there are billions of dollars of venture capital going into a sector where we don't even know if it works. We're going beyond technical risk to something far harder, which is long-term academic research that might be commercialized sometime in the distant future.

Chris Gates:
So, you're saying it's less about TAM, it's less about whether there's an addressable market, and it's just the inability to scale that technology generally. It's on a few of these cases like lab-grown meat, fusion, and your other example was crypto, that there's just not-

Danny Crichton:
I think what you're getting at is part of the challenge, which is a lot of folks who are used to software are used to evaluating products based on TAM, right? Total addressable market.

Let's look at the size of the market and say, "If they're winning in that market, then let's invest. Because it's a big market, and therefore if we can grow, we can do really well." And so, you always have this sort of market risk.

I think when you get into the science, yes, of course, a quantum computer that works is going to be a big deal. Yes, of course, if you have a fusion reactor that produces unlimited bountiful amounts of energy, it's going to be a big company. In most cases, the TAM is so effing obvious, they're like, "Why bother?" I mean, there's no point to the TAM conversation. If you solve cancer with a drug, as long as you have the IP, I mean, there's a little bit of it, but if you have the IP and you solve cancer, you are going to make money.

And so, to my mind, in the science world, it's almost never about TAM. The TAM is almost always obvious. The challenge is always, is what you're looking at realizable? Can you actually transition it either into a working prototype, which is actually the challenge we're seeing with a lot of these companies, or can you transition that prototype into a fully scaled-up product that is actually sellable and usable by everyday people?

And to me, you just can't bring that TAM mindset into this world because it's not the right diligence or analytical question to evaluate a startup.

Chris Gates:
I'm having trouble holding two things at the same time, and the two things are, "We believe before others see," this suspending of disbelief in VC, and at the same time holding the space for the dissent, the skepticism. How do you hold those two things at the same time? Of being a skeptic, but also needing to believe in something to fund it enough to be a possibility?

Danny Crichton:
Well, I think one of the challenges is, scientific advancements are rare. Particularly very fundamental advances, they're very rare. They don't happen that often. That's why you win Nobel Prizes. That's why they covered... Because how often does physics change? How often do we upend the way we think about biology that opens up new pathways for new companies and new therapies or therapeutics, whatever the case may be? It doesn't happen often.

And oftentimes, it's signaled well in advance. There's research that's been building up over a period of time that's explored. A lot of the companies we back literally are built off of... I'm thinking in particular like Icon, which won a Nobel Prize. I mean, it's not like last year they discovered this thing, and then six months later they plopped it out of the university and started to commercialize it.

No. I mean, this is decades of work that eventually cohered into a vision for what a company could be and do in this particular context. So, to me, there's a window of opportunity in each of these spaces where you can absolutely be too late, but you can absolutely be too early. And frankly, I think today, we're way too biased with a lot of VCs of being too early in a lot of these different sciences.

Chris Gates:
How does hype play into all this?

Danny Crichton:
I think there's an incredible amount of hype going on in all of these. And one of the core, I think, commonalities you see across all the hard deep tech sciences, including crypto, which I would consider a mathematical cryptography-based field... I mean, it's really hard math that's going in behind these algorithms and smart contracts and protocols... is that speculation and hype drive most of the investment activity.

And crypto obviously gets the most attention. It has a lot of reporters these days. But fusion has received billions of dollars of capital over the last couple of years, including one company, Commonwealth, which has raised $1.8 billion itself. And I don't have an opinion about every individual company, other than to say, a lot of smart people look at these businesses and say, "Look, we've tried to do this for decades. There's no working prototype in the world."

There's been some very, very early interesting results, something in which we've actually talked about in the securities newsletter over the last couple of months. But to put so much money at stake in a category in which no one has ever proven that the technology can even exist, seems really, really bad.

Quantum computing has had an enormous amount of investment over the last couple of years. There's no quantum computer that really does anything for anyone these days. I mean, you really have to step back. People are testing it out, they're playing with it. I just saw a news story this week that one of the companies had gone to 1,500 qubits, which is the kind of fundamental unit for a quantum computer. But even as an example, is the number of qubits the right metric of success to evaluate whether quantum computing actually functions? And there are folks who don't believe that that's the right metric. That you're trying to do this comparison between, I don't know, megahertz or gigahertz.

And in reality, it's actually the algorithms and the way it's designed that matters far more than the number of individual qubits.

And so, I think when you keep going into this marketing, and the narrative creation, and the speculation drives so much of the investment activity. And again, that's borrowed from software, where the hype and building people's excitement around some of these markets drives the investment activity.

It doesn't apply to science because science is up against the laws of physics. And so, I think that gets at, if you want to believe before others see, you also have to actually see. I think now people actually put on blinders and are like, "I'm just going to throw money at everything that sounds interesting, like fusion, nanotechnology. Let's throw some stuff at crypto." And there's no actual fundamental insight there.

So, I mean, look. If you're a blind monkey throwing money at anything, sure, that's great, but you're not seeing anything. That's a very critical distinction. You're not seeing something that others aren't seeing. You are just blind and throwing money around willy-nilly.

Chris Gates:
Okay, so let's try to see. You identified some dissent, some skepticism, in these three areas. Let's walk through some of them, especially the meat.

Danny Crichton:
Yeah. I mean, lab-grown meat has obviously received a bunch of investment. There's a number of companies in the space. It's very exciting. It has a huge amount of potential. And again, it fits into this kind of ESG climate bubble, which is live stock, and is very expensive from an environmental perspective.

And so, lab-grown meat has this potential being miraculous to solving part of the climate challenge. It won't solve all of it, but it could solve a large component of climate change and carbon emissions, and particularly methane emissions from cows.

One of the interesting pieces... So, we had an article in The Counter, which is a food-focused magazine, which I think is actually unfortunately gone defunct, but they published an amazing piece last September, a deep dive into the industry that was focused on a couple of scientists' work evaluating can you actually grow this stuff in the lab at scale?

And two conclusions. It's actually a really long article. I want to say it's like 15, 20 pages. It's a lot of stuff going on in there. But the two conclusions I really walked away from is, one, contamination is nearly impossible to handle. That a single viral particle on your gloves in one of these spaces... Because you're using centrifuges and moving very quickly to grow these meats in the lab, you're basically creating a growth medium to make the meat as fast as possible. And that growth medium is just as useful for a viral particle or a bacteria as it is for the actual meat itself.

You're growing meat, so it's a living meat, but the meat doesn't have an immune system. That blew my mind. I'm like, oh, right, right. You're not just refrigerating. You're growing a living organism that has zero immune system. So, just one tiny little virus, that normally anyone's immune system could deal with, is going to wipe out an entire batch. I've never thought about that before. Right?

Because what is the cow doing when you actually are growing livestock? It is a living, breathing animal, and that animal has all of the protections that an animal would normally have, namely an immune system, but that is not exclusive, and you don't have that in the lab. Now, one could argue that one of the solutions that will come up is building out basically a fake animal, an unconscious animal that has sort of the vascular properties.

That's the argument. You could build tissue. It wouldn't have to have a brain, it would not be conscious, it would not be a living animal. You could imagine coming up with a solution, which is, we're going to have an immune system that solves this problem. Now, that is a solution. And if you heard that, now you've heard the criticism, "Hey, if you have any single viral particle, you're going to have a huge problem." "Well, now we have an immune system. That's how we're responding to it." And you can go, okay, let's due diligence this new system to grow lab-grown meat.

But that's not where we are. It's essentially, it's a level of contamination that we basically can't handle. If you think that the baby formula crisis in America, or this recent one with Daily Harvest with their lentil product, which led to several folks having apparently organ failure and organ damage, I mean, this is traditional food, not grown in the lab, that doesn't have the scale up and growth media challenge. If you think lab-grown beef, this is basically an impossible problem to solve.

So that was A. B, there's this question of, well, even if you could do that, if you had the most secure facility, and there there's grades, and in the article they talk about a level-eight facility, and I don't know what that means. I know what it means in bio [inaudible 00:11:25], ironically, but I don't know what it means for food. But let's say it's a level-eight safety facility. What does that mean from an expense point of view?

And basically, the cost of operating such a clean room facility, that's cleaner than even how semiconductors get built, et cetera, is basically so uneconomical compared to traditional meat, that there's no way it would ever be cost competitive with traditional foodstuffs.

So, in some cases, they basically backed up the math using sort of the most bullish assumptions you could possibly make, and they came up with that a quarter-pounder of hamburger meat would be about $125. And that is in the best-case scenario. Right now, it's about $25,000. But if you were to really try to figure this stuff out, given the contamination number, people involved to run a lab like that, the fact that when any contamination shows up, you're to shut down the entire lab entirely and clean it from start to finish, and cleaning a lab when you're trying to find a single literal viral particle in the entire lab, it's not an easy kind of proposition.

It's basically going to be so expensive that it's basically not feasible from an economic perspective. And so, you combine these two together. And I remember reading this article, and this was right before I joined Lux, but I was like, "This is so convincing to me because there's so much evidence." And the response from the industry was like, "Well, everything is wrong and we're going to solve all these problems."

And to me, it was one of those classic examples of that's why it's so important to see the critics. Because you should be able to address specific criticisms like these from folks who know these industries well, who have criticism, and think, "No, that's not possible, and here's why," you should be able to address that.

Because we're going to run into contamination problems. You're going to have scale up problems. You're going to have economic problems as you're trying to build these labs in an efficient way. And for the most part, the industry did not respond kindly or at all to most of the questions they had.

Chris Gates:
Yes, skeptics aren't popular.

Danny Crichton:
I mean, no one loves the person saying no.

Chris Gates:
They're not. Nobody wants somebody in their lab saying it's not going to work.

Danny Crichton:
And the funny thing is, I always like to joke, the laws of physics are pretty locked in. Now, you can hack around them. There are ways of getting around the laws of physics. And I think that's where it's always... That's why the due diligence process is so important around science. You can't just take what critics say as face value. Critics are wrong. Most of science would never progress if you just listened to the people saying that nothing is going to work.

There are new sciences, there are new developments, things happen. That is why there's a joke that science progresses one funeral at a time.

But at the same time, by the way, those critics are generally correct, and that's why I said that scientific advancements are rare. Because if it was easy, we'd have scientific advancements all the time. We'd have a massive amount of change. And by the way, that was sort of true the last 100 years.

If you think from 1850 to maybe 1990, science progressed very, very quickly. But we have slowed down according to most studies, and we actually talked about the marginal stupidity episode, but scientific productivity is down. Our scientist in residence, Sam Arbesman, has also talked about this piece.

Chris Gates:
Due to lack of funding. Due to funding-

Danny Crichton:
Due to lack of funding. And also, we know a lot about how the world works today. I mean, there's less and less unexplored territory. We have a lot of models that explain a huge percentage of the world. And so, it's not to say that nothing ever changes, or there's no progress possible, or there's no new interventions. We are always looking for those.

But you have to have this lens of, there's a thin layer of new, backable, venture-focused companies. Everything else doesn't apply. And we're always very aggressive and trying to figure those out. How close are they to that window of opportunity? But you can be so early, where you're investing a decade ahead of where the companies are actually going to come from.

Chris Gates:
Okay. I think that's a good place to end that section. But I was wondering if you would just take me through the list of recommendations.

Danny Crichton:
Well, every week we sort of survey the whole staff, and folks read a lot some weeks, and some weeks they read not a whole lot because stuff is going on. But we had four recommendations this week. I had two, and then Shaq and Sam had one each.

But from my side, one of the first ones I recommended was an article from the Associated Press focused on the North Korea situation, and specifically its nuclear testing over the last couple of months. North Korea has tested a dozen-plus nuclear missiles, and it's now actually increasingly putting its nukes on mobile launching mechanisms, so it would be able to move the missiles all around the country, protecting them, making them much harder to strike.

And so, this is really radically changing the doctrine that we have for North Korea. And frankly, as someone who follows this issue, it really hasn't dawned on people how much has changed in the last six months on North Korea's capabilities.

It has advanced a lot. It's going to put a lot of pressure... I think we've actually had a discussion with Josh on the podcast about South Korea's pension for and desire for nuclear weapons. This is part of the reason why almost three-quarters of South Koreans want nuclear weapons is because the North is getting much more sophisticated about its own program. And so, that was a great article in the Associated Press.

The other article I recommended was from Alex Virshinen, who was writing for the Royal United Services Institute, and he did a very back-of-the envelope look at what is the usage of weapons in Ukraine versus the amount of weapons that the United States can produce, and it covers a whole lot of ground.

The one that I sort of highlighted was around javelin missiles. We've sent 7,000 javelin missiles to the Ukraine, about one-third of our national stockpile. Lockheed Martin, which produces the javelin, produces 2,100 missiles a year, and it could move up to about 4,000 in a few years. And Ukraine argues that they shoot 500 missiles a day, so if you're doing the math here-

Chris Gates:
500 a day? We are not making enough, it seems like.

Danny Crichton:
Right.

Chris Gates:
That math comes out to, we are not making enough.

Danny Crichton:
And so, he identifies it like, because we've been in this sort of peace dividend era post the Cold War, we have sort of lost a lot of the industrial base that produced tanks and missiles, and bullets in particular. And the number of bullets and mortar shells that we're using daily, we're not even there. We're just sending this to the Ukraine. But basically, Ukraine's usage of a lot of these weapons far exceeds American production standards.

We will burn through the entire stockpile in a matter of months. And in some cases, like the Javelins, would be like weeks if we weren't taking advantage of other people's stockpiles at the same time. So, a huge concern, because if there's any sort of expansion of that conflict, suddenly those stockpiles will look meaningless, almost, in very short order.

So, two articles from me. And then we had one article from Shaq Veda, which was looking at a discussion between Eric J. Topol and Demis Hassabis, I'm sorry if I'm pronouncing that really badly, the founder of DeepMind, and the two of them were basically talking about how deep learning is going to affect medical advancements in the future. Very long. It was a podcast. It was about 45 minutes. There's an audio version. There's a transcript. Lots of interesting cool stuff going on in that, the meaning of the brain, whether AI will ever represent a brain properly in computation, that I thought was very, very interesting, and Shaq did as well.

And then, finally, we had Sam, our scientist in residence, with an article from Nature on the long-term evolution experiment. And this was super fascinating, and is a great example of how hard science is, and why science progresses so slowly, which was over 34 years, a group of scientists out of Michigan have basically saved e. coli in their lab, and let it grow generation after generation after generation, for 34 years. It's up to 75,000 generations of the e. coli. And every, I think, couple weeks or couple months, they basically take a sample of that e. coli and store it. And the idea is that over long term, they're able to see the evolutionary change from a single media, single-bacterial strain, as it changes over tens of thousands of generations.

Because we can't do that with humans. That would be hundreds of thousands of years of human history to do 75,000 generations. At 30 to 50 years a generation, you just can't do it fast enough. But because it's a bacteria, and it has a much faster life cycle, you can actually see the changes. And what they found is the bacteria actually increases in fitness a lot over those generations. They're actually seeing exactly what the experiment is called, the Long-term Evolution Experiment, they're actually seeing massive changes in genome over a period of time.

So, I thought it was super interesting. And the reason for the article today is that the original generation of scientists, in the same way, is evolving to the next generation, generation two of scientists. So, the original folks who have run this in the lab for 34 years, one of the postdocs from that lab is going to take all those cultures and keep it going for another-

Chris Gates:
Oh wow.

Danny Crichton:
... maybe another 30, 40 years, so they can continue to learn more about long-term evolution. So, I thought it was both a great-

Chris Gates:
That's really interesting.

Danny Crichton:
... sort of collaboration, and a reminder, similar to everything else we're talking about with vaporware skepticism, of just how hard it is to learn this stuff. I mean, we're just going to take decades of data collected painstakingly in a lab to figure out stuff about evolution, and I just think that's awesome.

‍