The Exchanges

Q the big things that we've talked about is how many backward revisions there are to everything from non-farm payrolls to GDP that they've become so unreliable. And so it's very difficult for people that are transacting in market to know what to do. Have you guys ever thought about that? Because I'm sure that you have a much more accurate sense of where the economy is than many other people.

A We have, and I feel a bit rueful, you know, with you asking the question, because I feel like on some level we should have done it. And the thing that makes it kind of tricky is because, well, two things. One, Stripe is not like a full cross-section of the economy. You know, we're more biased towards online, we're more biased towards innovative companies, you know, whatever. It's kind of net that out somehow. And, you know, there can be these stories where, I mean, during COVID, like, the online economy was doing great. The offline economy is sort of a different story, so the interpretation can be a bit tricky. But then just the second thing is the Stripe business is growing so quickly and changing so fast that, again, you know, it's not necessarily representative of the economy. And even if Stripe is way up year over year, you know, you have to be a bit hesitant in drawing conclusions from that. Having said that, I think, in principle, you could draw, you know, some conclusions. And, you know, one thing we did look at was just inflationary data over the last couple of years. And I think you can construct, and the team did construct, a pretty reliable kind of leading indicator for inflation. And so we would like to share that openly because, you know, I think it's, I think it's a public good for there to be better and more reliable economic data.

AI assessment note: “We have, and I feel a bit rueful, you know, with you asking”

Q A holistic view of it. Patrick, any thoughts on, um, what we're seeing in defense tech and, uh, saving money through Doge?

A Well, obviously what, um, what, you know, Andrew and others are doing is, uh, is pretty amazing. Um, but, you know, we're, we're obviously not defense experts, uh, but sort of just bringing the credit card merchant, uh, perspective, uh, to bear here. Uh, you know, we naturally just go and look at the, the time series, um, and the sort of the data around it. And I guess I'm struck by, and again, maybe I getting some of the details wrong here. Um, this is outside of our zone, but as far as I can tell, the cuts proposed Over the next couple of years for the Defense Department are of approximately the same magnitude as the reduction in the defense budget that occurred between 2010 and today, and so it's not like this is some unprecedented transformation in DOD budget. We, we, we've done this. And then secondly, you know, as far as I can tell, one of the most ecumenical, uniformly shared, bipartisan issues in Washington is the inefficiency and the profligacy of defense procurement. You had James Fallows writing a book about this in the late eighties. You had, you know, Augustine's Laws and their whole book about this. Just everyone seems to, you know, fervently believe, um, that defense procurement is monstrously inefficient. Now, you know, it's possible to Make budgetary changes without fixing that, but obviously the prospect of meaningful improvement there, uh, seems, uh, seems li…

AI assessment note: “it's not like this is some unprecedented transformation in DOD budget.”

Q think they are perceived by the people buying them more like securities and more like Bitcoin? They do, uh, to steel man the other side of the argument, uh, they do trade with a ticker symbol. They are traded, uh, on major platforms like Coinbase and Robinhood. And people share charts about them, and so where do you stand on it? You're in the finance business. MemeCoin's good, MemeCoin's bad.

A I'm facing with Dave. I mean, I, I, well, They seem to me to be maybe analogous to, to gambling, um, which, you know, I don't know that we want to ban gambling. Uh, like if you're able to do it responsibly and you understand what you're getting into and so forth, like, I guess that's fine. But as you say, judging by the tweets that I see, there are awful lot of ticker symbols and charts And prognostications about future price trajectories, and so forth, that lead me to think that people are placing somewhat more weight on the asset and security value of these, uh, as compared to the, I don't know, some numinous, intrinsic, uh, aesthetic value.

AI assessment note: “They seem to me to be maybe analogous to, to gambling”

Q Is there a version of a network effect inside of Stripe for their customers where if I just allowed you guys to just be integrated into my GL somehow, and you gave me some kind of phantom bank account, why isn't it just a ledger entry if I'm just making a payment from me to somebody else that's also on Stripe?

A The thing we really want to solve is all the calculation. The ID verification, the risk, like, those are the things that are actually expensive, um, if you, if you look at this flow, it's where companies lose their money today. Having said that, you're right, um, you know, we're, the fraction of money movement on Stripe where, you know, the two counterparties are both part of the Stripe network is obviously growing, and so I think that'll be another way we can reduce fees over time. Although again, I actually think the biggest part of that is it's gonna be because we reduce fraud. Like, both counterparts are known And like, I talked to a company, a payroll company recently, and they were describing, you know, how big a deal it is for them that, you know, people sign up, you know, fraudulent companies, whatever, and, you know, they can lose millions of dollars in a single attack. And so having some kind of trusted node rather than just routing an account number, that would be a really big deal for them.

AI assessment note: “the fraction of money movement on Stripe where, you know, the two counterparties are both part”

Q Where does it stand in the spectrum of different tools that folks would use to solve these life sciences problems? There's cell models that are being developed by some, then there's these protein models. Where does this fit?

A This kind of landscape of, of, um, of foundation models in biology, it's obviously very new, so it's, uh, um, it's a bit of an open question, sort of how exactly people are going to find, you know, ways to, to, to use it and applications for it. Part of what I think is cool is that proteins and RNA and, I mean, phenotypic expression and everything, you know, all these things sit on top of the DNA. Like in some sense, the DNA encodes everything because, you know, the, the whole, Organism comes from the DNA. And so I guess the question would be, and, you know, we don't really know yet, is DNA all you need? And with EVO one, we saw some, you know, encouraging suggestions that, for example, you can build really good protein structure prediction models out of a DNA foundation model, even if you don't train on a lot of, you know, protein structure data. So, but I, I'd say it just, it's a really exciting time and it's kind of an open question. And I don't know if you analogize EVO II to, I don't know whether it's GPT-II or III or something, but, you know, I, I think we're going to see a similar Cambrian explosion of, of applications over the next couple of years. The thing we're really excited about at Arc is, is training, yeah, cell state models and trying to better understand, you know, how cells, you know, what, what causes them to change states, and so we're thinking a lot about t…

AI assessment note: “you can build really good protein structure prediction models out of a DNA foundation model”

Q X, some of your own excess capital and other people will do the same and keep funding arc. And then if there's something that arc creates or innovates on, if it can generate some amount of money, it would just kind of flow back. Is it meant to be self-sustaining or is it always just going to be via patronage from successful folks that just want to keep it going?

A John and I are, you know, we are ourselves very committed to it, and, um, and we're, we're kind of underwriting it in that regard. But, well, one, we're just lucky where there's a growing donor pool of other people supporting it. I think it's just better for an institution if it's not kind of beholden to the whims of, you know, one donor or one group of donors or something. So I think that's just a much healthier structure for us. I think there's also a larger group of people who are just becoming interested in science and realizing, I mean, you know, Jason was on his, uh, his Moral pulpit. My pulpit is that, you know, all is not well in basic research in the US today. And again, the way to see this, it's just talk to the scientists themselves, and they tell you how kind of inhibited they are and, you know, the, the, the kind of problems caused by the strictures and structures around them. And we don't see ARC as, you know, the answer. Hopefully it can be sort of one Point in the space, but then, you know, there's other people doing cool stuff. You know, Brian Armstrong, of course, started a company in the longevity space, and Yuri Milner and others started Altos, and, you know, this is the people, this, uh, the Chan Zuckerberg Institute, and so, you know, people are trying different things, but no, our, our Arc is, um, you know, we're, we're very happy to support it, and then …

AI assessment note: “it's possible that Arc over the long term becomes self-sustaining”