Riskgaming

We need to go deeper with the inception of deep geothermal energy

Description

Historians survey the past and the Twitterati (X-erati?) process the events of the present day. But what does it mean to search the future for clues of what’s to come — and how much longer will we have to wait for it?

In this episode of “⁠Securities⁠”, ⁠Danny Crichton⁠ welcomes ⁠Lawrence Lundy-Bryan⁠, research partner at Lunar Ventures and the publisher of “State of the Future”, a ⁠Deep Tech Tracker⁠ whose distinguishing feature is its extraordinarily wide remit to investigate the interstices of science and technology and find the morsels of innovative goodness that will power the planet in the years ahead. Also joining is ⁠Lux⁠ Capital’s own scientist-in-residence ⁠Sam Arbesman⁠, who is certainly no stranger to the crazy ideas straddling science fiction and science fact.

Lawrence shares his unique approach to identifying and evaluating emerging technologies such as deep geothermal energy. We then pivot to exploring Lawrence’s approach of finding the future through the methodology of “horizontal scanning.” What’s to come? Listen and find out.

Transcript

This is a human-generated transcript, however, it has not been verified for accuracy.

Danny Crichton:

Hello and welcome to Securities, an audio and video podcast, plus a newsletter focused on science, technology, finance, and the human condition. I'm your host, Danny Crichton. And once again, we have our own scientist in residence, Sam Arbesman joining us on the program, plus a special guest, Lawrence Lundy-Bryan, the research partner at Lunar Ventures and the publisher of the always insightful and interesting State of the Future. Lawrence, thank you for joining us.

Lawrence Lundy-Bryan:

Thanks. Thanks for inviting me.

Danny Crichton:

So, Lawrence, you've published a bunch of new articles recently. I'm curious, what's been top of mind for you and your publication with State of the Future?

Lawrence Lundy-Bryan:

Well, the aim is to try and focus on what no one else is focusing on, which isn't particularly good for journalism or media because everybody wants some more AI stuff from me. Well, I think the most interesting thing, in terms of pickup, has been the three most undervalued technologies that we scanned because everybody likes list of threes. And in particular, I think whenever I mention it, people seem to get quite excited.

And so deep geothermal energy, for me, has been something that's been particularly interesting. We tend not to focus on energy systems that much. We're not a climate tech VC, but the work is to just scan as many emerging technologies and surface what we consider to be interesting regardless of Lunar. And deep geothermal, to me, struck me as one of those technology areas that doesn't really require a breakthrough, technically. There's no supply side that needs to change. Most of the existing drilling equipment is sufficient.

We don't really need to change the demand side. People want clean, abundant technology. We're going through a transition. All it would require is a change of political will or action or framing. And I really like those technologies because it would just take something changing, whether it was what's changed with the nuclear movement in small molecular reactors. They went from everybody saying, "Nuclear is terrible," to, "Maybe we should consider nuclear," have a gut feel that deep geothermal might go through a similar journey soon.

Danny Crichton:

So let's talk about this. You do this concept of horizontal scanning, so there are thousands of potential changes in the world, thousands of new technologies that might come along. How do you take that huge funnel of potential ideas, flesh it out, and then get it all the way down to deep geothermal energy?

Lawrence Lundy-Bryan:

Well, I started by doing interviews and asking as many experts as I could find in different areas. That's synthetic biology, DNA data storage, deep geothermal. What I realized is asking people is the wrong approach. And what's better is writing down the wrong thing and then asking experts to tell me why I'm wrong. So actually, my approach has been to write stuff wrong on the internet and then ask experts to tell me why I've been wrong, and I've been wildly effective.

Yes, you're right, you can't cover everything, and there's always going to be biases in the process, the things I'm interested in, the thing my interviewees are interested in. But my approach was to be as broad as possible. Don't have any filters, speak to people and just ask as many people as possible, "What's interesting to you at the moment? What technology areas would you like to focus more on if you had time? What is the thing that you cared about in your 20s that work has squashed, and now you focus on something that's commercial? What would you focus on?" and just made a long list.

And from that long list, I tried to create the most simplistic methodology possible that I could just very easily explain to anybody, scores of one to five. I could have gotten the full quantitative model, and I could have done some algorithmic sourcing and looked at patents, and there's many approaches you could take. Mine was pretty simple. It was to have a simple methodology, score these five criteria one to five, and that way I would go from an extremely long list to a more manageable short list of things that are interesting, 20, 25, 30 things. And then I had some clay that I could mold, and that's when I went in and did expert interviews.

Danny Crichton:

So you went through this whole funnel. You got the scores, you lined them up, and it was deep geothermal energy.

Lawrence Lundy-Bryan:

It was just wild. That's true, yeah.

Danny Crichton:

And why is that? Obviously, you just said, "We have this political transition. It's available. The technology doesn't require a lot of think." Are there a lot of people thinking about it, though? Is there a whole community of folks building these products, or is this one where you're saying, "Look, we need a whole new crop of engineers and entrepreneurs going into this space because it is available for the taking for someone who's ready to seize it"?

Lawrence Lundy-Bryan:

Right. Part of the methodology is around... Well, I'll sort of explain it. It's viability, which is a good start, which is how technically ready is the technology? That helps the first baseline. Is it still lab, or is it being commercialized? Is it RISC-V open-source construction set architectures, which you commercialized, or is it nano mechanical computers, right, that are still in the lab? So you have a wide variety.

And then just go through a thought exercise and some desk research and understanding, what are the key drivers? What might be the restraints? What impact could it have in terms of market size, and how novel is it? And the reason I finish on novel is the novelty of the technology, i.e. is it something that if it existed, would have no competition because it does something that's not possible now? Or would it have to take market share from something else?

That sort of filter, I think, has been the most interesting. If you think about photonic computing, for example, right? It's fast, but in theory it's higher bandwidth, lower latency to do certain types of calculations. So, it would compete with classical computation, but it would compete for computation with other types of computing architectures. And I suppose nuclear fusion's the same. It's a way of generating electricity, low-carbon electricity, but it competes with a whole bunch of other technologies.

So the thing I try to focus on, what are the technologies that if it existed, creates something that's not been possible before? It has no competition and would have an extremely large impact beyond just economic, so it's from a societal perspective or cultural. Then we shouldn't focus on deep geothermal. It's just the one I like.

But another example could be brain recording, right, which is probably a better example because we can't do it today. So, for example, is there a breakthrough in imaging, brain imaging, EEG imaging or something, or non-invasive imaging, non-invasive bio signals from the ears? Is there something that gives us access to brain patterns that enables us to do a whole bunch of things that we can't do today? So that's been my methodology to get to a really small list, something that creates totally new capabilities.

Danny Crichton:

I mean, Sam, you obviously do this in your own way, which is I feel like science fiction from the 1950s and '60s, which is chock-full of ideas and concepts and thoughts that have never come to realization over the last couple of decades. It's a graveyard of good ideas. How do you think about exploring the vast futures of space?

Because I think one of the biggest challenges in venture, particularly when you're doing deep tech and hard sciences, is that sort of cutting edge of the possibilities on the frontier, right? We don't have these technologies today, to being able to evaluate... It's very different from a software as a service company where we, in some cases, are inventing fundamental science.

Sam Arbesman:

Yeah. And there's a very large space of potential things. And, as Lawrence was saying, there's probably... Did you say there was a thousand things on your list initially or whatever? I'm sure it was a very... Yeah.

Danny Crichton:

He did, yeah. That's only a thousand. You should see Sam's list. That's why his bookshelves are full right now.

Sam Arbesman:

So there's many, many. And the interesting thing is a lot of these are kind of dependent on each other. It's like there's kind of this, and whether it's this tech tree map or the way in which kind of people think about the linkages and connections. In terms of figuring out what things to focus on, or what are the area... And for me, one of the kind of touchstones that I like to think about is, there's this book co-authored by Ken Stanley and Joel Lehman called Why Greatness Cannot Be Planned.

And the idea behind it is whenever you're looking at some sort of high-dimensional search space, and they're looking at this from the kind of world of evolutionary computation. Whenever there's a high dimensional search space, sometimes actually having a specific objective in mind is not the best way to get to there. Rather what you should do is just focus on interestingness or curiosity or novelty and then build new kind of stepping stones and then combine those in interesting ways.

And so, I think you always kind of have to have a balance between, okay, these are really interesting technologies that are kind of on the horizon and are possible, that are things that we can aim towards versus things that are just so far off that we can't even aim. Going back to science fiction, warp drive sounds really cool. I think it's a little bit too far off to actually have a roadmap for that kind of thing.

But the kind of things that Lawrence has in the State of the Future, those are kind of in that intermediate stage where like, "Okay, we know what are the building blocks to get there for those kinds of things. And there are things that are worth aiming towards. And then there's the things that are very, very near-term, and we maybe don't even have to... They're not as exciting anymore."

So I think that kind of figuring out the right horizon, the number of building blocks that are required, kind of recombine things to get together versus saying, "Okay, let's just kind of have wide open, allowing basic research to kind of do its thing." When we have a very far off goal, we just kind of want basic research to do its thing. Once it becomes a little more near-term, then we can say, "Okay, how can we productively kind of move things forward and then recombine them in certain ways?" And we have DARPA or focused research organizations to help allow when things have been de-risked enough, we can say, "Okay, let's actually get some people together to do this in some sort of concerted way," which is really exciting. Yeah.

Lawrence Lundy-Bryan:

One other lens to add to that would be... So surprisingly few of the technologies in the list, and maybe this speaks to the near sort of medium-term, is how much science risk. When I went into the project, I thought the project will have a whole bunch of science risks that we still have, that we have to de-risk the science. And actually it turns out, aside from maybe AGI or whole brain emulation and, well, maybe even a quantum computer, there is a gap. We don't really know exactly how to get there. There were a few. Even space elevators, we sort of know what it would take to build it, sort of. We probably could with the money and the resource and materials.

But there were very few pure science. And what I realized, to your point, Sam, I think, is that, well of course, because you can't preemptively say, "Well, we need a new model of physics to describe X." It relies on fundamental breakthroughs that then would breach State of the Future, so there's this whole world of search space that we're just not searching because it needs the fundamental research to go into, and then it needs engineering. So that was a learning for me, that, actually.

Sam Arbesman:

Yeah. I was going to say, I think the kind of topics and things you're talking about within the State of the Future, right, once they've been reasonably de-risked where it's kind of, "Okay. There's an engineering effort, but it's maybe certain people." It's a situation of the limits of will or whatever, or there's certain political ramification. It's almost like that laying these things out there can provide almost a galvanized [inaudible 00:13:02]. So rather than just saying, "Okay, here's how it is, and we'll kind of just allow the scientific community or the market to kind of figure it out," laying it out that clearly can hopefully actually have some sort of catalyzing effect too because it is in this kind of near- to medium-term, which is pretty powerful and exciting.

Danny Crichton:

Well, I think of, as an example, the Federal Reserve just recently implemented its FedNow wire transfer program, which will allow you to move money from account to account instantly, a service that's available for decades in many countries that in 2023 we are now just getting in the United States. And to your point, there's nothing original here. There's no new science. It was just basically willpower, and it took 2019 until we actually had the willpower to do something about it and another four years to actually implement it. And now we're trying to get people to adopt it.

And I think we see this a lot across sciences, right? So in some ways there is this William Gibson quote that we quote, but the corollary of that is science is already here. It's just not equally known by everyone. We can't all agree what it means and what we could do with it. And I think we see that particularly in biology. We've certainly seen that in aerospace. What's been exciting to me is that in some of these fields, which felt in some ways moribund, have been revitalized thanks to investment, thanks to some political leadership where people have said, "Well, look, we can go back into space. Well, we can go do this. We have the technology. There's a lot to be done, and by pushing harder on that frontier, we will find other challenges, other problems that we need solving and keep that leap going on and on and on."

Lawrence Lundy-Bryan:

There's a interesting, I suppose, hypothesis here. It's probably already playing out a little bit, which is as we move into atoms... I get that you say the last 20 years have been development of bits, not entirely, but for schematically. And now a lot of the work is on new energy systems, biology, and semiconductor space, things that are atom driven. There's land that's needed, materials. You have to dig stuff out of the ground. It becomes inherently political. It is very interesting to think that can you really make predictions of a future or make 10-year bets without seriously considering the political dimension?

And so we're predominantly focused on sort of science and what I would say to be supply-side changes. But I think the demand-side change is really interesting. And how does political will change? How does regulation change? It's fluffy, and you have to read tea leaves, and it's much less quantifiable. But I sort of have a hypothesis, I guess, that you can't be a good deep tech investor in the next decade without really taking considerable care on thinking about the political landscapes, which is why I've seen a couple of funds recently hire, not necessarily an ex politician, but a policymaker, a head of policy or a, I wouldn't call it head of lobbying but effectively, head of lobbying because I think that's the right angle, honestly.

Danny Crichton:

Right.

Sam Arbesman:

That's interesting. I was going to say, I feel like there's also some interesting things there in terms of differences in timescales where certain of these advances, where it might require longer timescales. Venture requires longer timescales than other types of investing. But within the political realm, depending on who you're interacting with, it could be the timescale you're operating on is the timescale of an administration or the timescale of some sort of election cycle.

And figuring out how to handle these kind of mismatches, I think, is a really difficult thing. And you see the same kind of thing of maybe scientists want to do long-term research, but they're still beholden to grant cycles. But yeah, the political timescale and how to navigate that, I think, is a really interesting thing when thinking about deep tech advancement.

Lawrence Lundy-Bryan:

Well, yeah and-

Danny Crichton:

Well, and I think it's not just timescale, it's the complexity of the feedback loops, right? So, I think systems thinking and systems dynamics, you're going to be trying to evaluate feedback loops, right? So you mentioned climate with deep geothermal energy. That's a good example where there's a lot of force all around the world coming from a lot of different constituencies, right, developing countries worried about heat waves in the equatorial regions. You have populations in industrialized countries who are concerned about the environment. You have folks who want more reliable power, having experienced power outages and floods over the last couple of years, and all those kind of combine.

Now what's interesting is, is it today? Will those all sort of come together or intersect each other in a couple of years? I don't think anyone could have predicted that last year in the United States where we passed the IRA, the Inflation Reduction Act, which had tens of billions of dollars of climate incentives. Why? I mean there's an easy why, which is, well, there's a democratic president with Joe Biden in office. He is making this a huge issue. He gets it through Congress. It's skirted by at a very specific time. People have been asking for this for 20 years.

I think the hard part is even though the forcing function seems to get higher and higher, and those dynamics can be kind of identified, to your point, it could be very challenging to actually pinpoint, "Well, in 2027, all these things will kind of line up beautifully, and nuclear power is going to be the new hot thing, and everything's going to launch."

To me, it's a lot more stochastic in the sense that the oil shock of OPEC in 1973 triggers everyone's concern around energy efficiency and getting cars to be better. That's when we passed the first auto efficiency standards, was in a direct effect from something else that happened exogenously to the country. And so, to my mind, it's always about response. Can you move quickly when these opportunities arise? Don't let the crisis go to waste, so to speak.

Lawrence Lundy-Bryan:

Right. So a good example, we'll have two different trend. One's predictable, one's not. So demography is a good example. Another of the technologies that I've focused on is IVF or assisted reproductive technologies, just on the basis of declining total fertility rates globally and increased infertility rates. So, the increased infertility rates isn't the same as demography, but it seems to be pretty consistent over an extended period of time. Now, it intersects policy and the unpredictableness quite well in that it's a known statistic. I didn't find something new. Nothing changed, really. These numbers are out there. Everybody knows.

And so I say, "Well, I think that this is really important, and soon people will care enough for things to happen." Right? But is soon 2023, 2027, 2030? When will there be a cascade effect? When will something happen? I think I try to skirt around that by just saying, "Is it 2020 to 2025? '25 to 2030, or 2030 plus?" Because I came to the same conclusion you did. There's a bit of cargo cult math for me to say, "Well, it's going to be in 16.7 years." Right?

I think the fusion number, I think, one I saw recently, is the average consensus view is something of that range. But still, I think, yeah, you can't predict it with any surety, so you have to say, "What's your conviction level? What's a reasonable timeframe?" And for a venture fund, that's fine, right? It could be fund 3, fund 4, fund 5. You've got sort of 10-year windows, so actually, we have quite a lot of plausible deniability. There's a big range for us, which means it is actually doable.

Danny Crichton:

What I think is interesting is, I wrote a piece last year on prediction, but Philip Tetlock, who's most famous for Super Forecasters but wrote a much more rigorous academic book called Expert Political Forecasting, Expert Political Opinion. Before that he did hundreds of psychological studies and has kept database of predictions from people over the course of decades. And so he's like, "Okay, in 1992, what is the GDP of China in 15 years?" And now we've crossed 15 years, and he's cross checking who was right and who was wrong. And the key theory that he really comes up with is basically, I would call it, cognitive agility, but this idea of the fox, not the hedgehog, not someone who is a specialist or someone in-depth but someone who really just engorges on an enormous amount of information, has super flexibility in thinking, and is sort of looking at it holistically is almost always more correct.

And actually, what's striking is so much of, particularly politics, but also in science, engineering, we're used to the idea that experts predict the future. So you ask an AI expert, we're like, "When will AGI arrive?" And Tetlock's work basically says they will be the worst at identifying when it comes because they're so involved in the processes. They're so narrowly tailored that they're not going to realize that a bunch of other factors are actually going to determine whether AGI arrives when it does.

It will affect based on the number of tips that arrive. It'll be based off of tabs. It'll be based off of political and social dynamics. It'll be profits in the business sector that allow them to reinvest to create the technology in the first place, and so you have to really know business cycles, not really the technology scale. So, there's all these other factors, and so that flexibility and agility really becomes the fore of prediction here.

But I want to move the conversation a little bit to part two here because one of the unique things with the three of us is we all have sort of odd jobs in venture. Lawrence, you were talking about a bunch of firms are hiring heads of policy. I actually still find that an odd job because it's not usually a lobbying position, which would be a lot more sane and clearly defined. Often then, that's much more of trying to understand the policy landscape and help people make decisions. But we have a scientist in residence, an editor, and a research partner, all of whom are trying to figure out the future. Lawrence, from your perspective, how do you sort of fit into the broader remit of a venture firm?

Lawrence Lundy-Bryan:

Badly, badly. When I spoke to Sam a while ago, I feel like it should be probably horizon scanning because I think research covers a multitude of sins. And everybody that has a title like research does something very different to me and maybe different to each other, actually. Research generally is data research. It's probably more closely aligned to algorithmic sourcing and data infrastructure for funds, which is a really important job, but not what I do.

And so I think actually, yeah, my job is to look at areas that the fund wouldn't otherwise look at because they're too busy, and you have path dependency. "Hey, we're good at machine learning. We've invested in lots of machine learning companies. We get lots of machine learning deal flow." And you end up becoming specialized. It's just how it works. But how, as a deep tech fund, you avoid just the natural dynamic of path dependency of just specializing in a few areas because you made investments in that space.

So my sort of role is explicitly to try and sit outside of that and to only look at things that we don't otherwise look at, which is pretty novel and definitely not what researchers do in other funds. But just more generally to the point, I think there should be more experimentation around what a venture fund is. I think it happens maybe on the financing side, the instrument, right? We had tokens as an instrument and equity and the debt, and there's evergreen funds and 10-year funds like this.

There's financial innovation, but it's still pretty structured around a partnership model that was created back in the origins. And most of the job titles are exactly the same 40, 50 years later. That's surprising to me. So I would love to see more interesting job titles, more interesting people joining venture and to experiment. I don't know, Sam, you've been doing it longer and thinking about it longer than me. What do you think?

Sam Arbesman:

I definitely agree with you. I actually have this spreadsheet where I've been compiling a list of kind of like outlier titles and roles within venture because they're so rare that they can fit on a single spreadsheet. There are not many. I definitely agree there need to be more. I mean, going back to what we were saying with Philip Tetlock and kind of this mental agility, I think there's a similar kind of need at an institutional organizational level where, and what you were saying, Lawrence, most people are specialized in certain areas. They know a lot about those things.

But we need for an organization at an organizational level to have the space for people to kind of play intellectually, look on a longer time horizon, explore ideas that are maybe not being looked at. And so I definitely think that kind of outlier, more kind of generalist approach is really complementary to everything else that an organization and certainly a venture firm does.

Yeah. My role is often very upstream from ideas or from the things that are investment, where it's like, "Okay, I'll find certain communities or areas or topics or fields that might not be ready for investment now, but maybe in several years they'll be ready for it." And once that happens, we'll be able to hit the ground running because we've kind of been laying the groundwork, and I've been exploring these different topics. Yeah.

And one of the other ways I think about this is also... One of our partners, Josh Wolfe, talks about this idea of randomness and optionality, kind of promoting the conditions for just finding weird things. And I almost feel like sometimes my job is almost being this optionality machine, just kind of finding interesting things that kind of can be brought into the orbit of Lux, but we need more of that. I agree that venture is... Venture as an industry has kind of, ossified might be too strong a word, but I would say stabilized. And having more of these kind of outlier roles can almost act as sort of like a simulated annealing kind of thing or kind of like jumping out of its current local maximum or whatever it is.

Lawrence Lundy-Bryan:

Do you have job titles? Sorry, Sam, do you have job titles that you would like to see filled that you haven't seen yet? I was thinking why are there not more chief economists. That's a banking thing, but I don't know. There was a head of crypto economics at a fund I saw once.

Sam Arbesman:

Oh, that's interesting.

Lawrence Lundy-Bryan:

That's interesting, but I don't know. Are there others?

Sam Arbesman:

Yeah, I'm not sure I've seen any. And there's certainly interesting titles. I'm not sure I can think of a title that is missing a person within venture funds. And in truth, I feel like a lot of these titles are almost kind of placeholders for people doing more generalist kind of things that are valuable to the fund. And then the actual title is sort of like a way of just signaling that, "Okay, someone's doing something a little bit different." That's a really good question. I can't think of any specific titles that are missing people. I'll have to think about that more. I like that question a lot.

Danny Crichton:

Well, I think what you're all getting at is venture firms try to become more process oriented, right? So it started as an apprenticeship business. It was very unprofessional in the sense that there wasn't a lot of norms in how to do the work, right? It was basically people being like, "Let's invest in these companies." And at least in the United States, when the RESA Act, Employment, Retirement, blah, blah, blahs Act in 1978 really opened up venture capital for pensions and a bunch of other asset allocators, we suddenly had to create a profession.

And so over the last three decades, I think, people wanted to show that venture could produce reliable returns week after week, year after year, decade after decade. And that meant process. That meant sourcing networks. That meant theorems. That meant, "We're going to look at all 5,000 companies doing exactly the same thing and identify... We have this magic sauce to identify the core thing."

But I think, particularly since we're all in the deep tech, hard sciences world, the process doesn't work this way. It requires a lot more stochastic opportunities thinking. You're looking for edge scientists in a lab who just discovered something new that connects three other things that probably took two decades and a Nobel Prize to get to. And now that's ready for commercialization. It's going to do something in the real world.

Or, as you mentioned earlier in the podcast, Lawrence, you have some political dimensions. So clean tech, for the last 10 years, has really been a challenge. No one's buying it, at least in the United States, really tough. Utilities can't buy clean energy products. They have very strict kind of stock constraints from Wall Street that prevents them from kind of investing in their infrastructure. And then the IRA passes, and all of a sudden it's ready to go. There's tens and tens of billions of dollars of subsidies, incentives, and everything you know about that world is wrong. And so, to my mind, it's constantly being prepped, prepared, learning, looking for stochastic opportunities, looking for the edge, and that's what these roles do.

And it hearkens back to, I think, kind of the origins of this field before it professionalized, and everything became software and algorithms and searching a million LinkedIn profiles for the word stealth. Real thing that firms do, which is why, now that you have the counter factor, which I always enjoy, which is now every founder who wants to get a new round of fundraise just adds stealth to their profile, so all the LinkedIn trackers click up and down Sand Hill Road, and everyone kind of shows up with stealth.

Lawrence Lundy-Bryan:

Right. One thing I've been wanting to add to this is, it's funny, then, if this is true. I mean, we could take it as a hypothesis that there needs to be some optionality or some inefficiency in the system in a venture firm. It should be a competitive advantage. What you do at Lux and what I do at Lunar, it should give us some competitive advantage. We would hope that it uncovers a thesis before others. We find the best founders, et cetera, et cetera. And of course the feedback loop is so slow.

I always think about it like it's a bet. We could be right? And I hopefully take it from Sam and Lux that it's a good bet. We're taking a bet that me spending a day reading is a valuable use of time, but we won't know for sort of 10 years. I can see why everybody is tended to efficiency, but we don't really have the data, and we're sort of on the frontier. We don't really have the data that this sort of optionality works or works at scale or is repeatable, or it's just luck. It's hard. It keeps me awake.

Danny Crichton:

I agree, Lawrence. We are important people. Yeah.

Lawrence Lundy-Bryan:

Yeah, I think... Yeah.

Danny Crichton:

But look, I think what you're getting at is there's a difference between quantitative and qualitative investing in venture. As a bunch of funds have scaled up, they focus on quantification. They try to count the number of deals they look at every week, the number of things that are getting sourced and put into the CRM. And I think we're in this world where we're trying more qualitatively... We're saying, "Okay. Yes, there are thousands of companies, and anyone can go to LinkedIn and track them down."

But no one's looking at these labs. No one's looking at, in your case, deep geothermal energy. That's not going to show up unless you're really doing, in your case, horizontal scanning or network-based, I keep calling it stochastic opportunity creation, but essentially bopping around in the right places where there are thought leaders on the edge of science, on the edge of technology, and listening and going, "Okay, you're not going to build what you're talking about building, but that is an amazing idea, and now I'm going to start looking for this. It's in my prepared mind. I'm going to start looking around to try to find exactly what's getting built from someone else who may actually have the wrong idea but is going the right direction." And that's how it all comes together.

And that, to me, is the messiness of venture that it gets so hard to do, particularly as we see a 

lot of SaaS investors moving out of SaaS into deep tech and hard sciences as the SaaS world has sort of struggled in the last two years. It's been interesting as people kind of arrive as tourists, and they're like, "Well, we can just do the exact same process we did." And it's not true. 

The kinds of metrics you saw in the SaaS world don't apply in technology.

I was just having a lunch with a couple of DC types just looking at industrial policy. And one of the things I really emphasized is like, "Look, there's so many steps involved in building one of these companies." It's not just the early scientific discovery. Then you have to figure out how to turn into a company. Then you got to figure out how do the revenues line up with the timing for the technology commercialization, so you don't hit some sort of valley of death at some point in the next five or six years.

And then even there, there's a scaling challenge of, "Okay, now we have a great company, a great product. How do we get enough buyers as we build out this huge factory to make sure that we're not committing capital to some large physical plant that isn't going to pay itself back?" And so my mind is, there's constantly this kind of stochastic calculus that goes in throughout the whole period. And that's what makes it so fun because everything else is just an Excel spreadsheet. This actually requires thinking.

Lawrence Lundy-Bryan:

It's risk at capital. That's where it should have been. We sort of got away from it. We created toolkits and checklists, but this is TRL 6, 7s. It's science. It is messy. And if it was planned, if you could easily scorecard it, that would already be in existence, and science would be a very clear linear process of progression. But it's not, so it seems consistent that the venture capital industry to find those would equally be non-linear.

Danny Crichton:

Well, look, if we could plan scientific progress, the Soviet Union would've been much better off in the real world. But on that note, Lawrence, Sam, thank you so much for joining us.

Lawrence Lundy-Bryan:

Thank you.

Sam Arbesman:

Thank you very much.

continue
listening