Martin Shkreli on prison, beating Bloomberg, and being hated

Martin Shkreli is the co-founder of DL Software, the parent company of Godel Terminal, a Bloomberg competitor for financial data. He previously founded the biotech companies Retrophin and Turing Pharmaceuticals and the hedge fund MSMB Capital Management. In this episode of Summation, Martin and Auren discuss: The structural reasons fintech is less competitive than other industriesHow drug discovery actually works (and doing it from prison)Why AI labs are now out-recruiting hedge fundsWh...

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Auren Hoffman?

  Hello fellow data nerds. Welcome to Summation, formerly known as World of DAS. My guest today is Martin Shkreli. Martin is the founder of DL Software, the parent company of Godel Terminal, a Bloomberg competitor for financial data. He previously founded the biotech companies Retrophin and Turing Pharmaceuticals and the hedge fund MSNB Capital Management. Martin, welcome.

• 0:23Speaker 01

  Hey, thanks a lot for having me.

• 0:25Auren Hoffman?

  Super excited to dive in. Talk us through why Bloomberg has been just so successful over the last 40 years.

• 0:30Speaker 01

  So many people have an opinion on this. Everyone has a right to an opinion. The best opinions probably come from users and users who have a little more holistic view of Wall Street, whether they've managed a fund or they've been an investor in different software companies. If you take all of that and you put it together, the simplest answer is they have a really good product. That's kind of what people forget. The second really important answer, I don't know how much I should say about this, but having done a lot of industries, invested a lot of industries, startup in a bunch of industries, the fintech software industry is not as competitive as other industries.

• 1:05Auren Hoffman?

  What do you mean by the fintech software industry?

• 1:08Speaker 01

  Financial information software. This is software that banks and hedge funds use.

• 1:12Auren Hoffman?

  The data that it's based on.

• 1:14Speaker 01

  That sort of dovetails in a little bit. The reason I mention this is if you look at sort of the first part of the answer, which is, you know, why is it a good product? There are four major terminals. We're trying to become the fifth and we're trying to be the best of the fifth. We think we're going to do that. But of the four major terminals, only one of them has any sort of messaging system at all, any kind of messaging. Mail, direct messaging, message, voice. The only one is Bloopers.

• 1:38Auren Hoffman?

  And that's. Why is that? Because that's so obvious. Like you could code that up in a day.

• 1:42Speaker 01

  It's a mix of different reasons. One of the major issues in software development in general, coming as an outsider, is the role of product and how companies canvas for what they do next. Product people, unless they've been in the seat, you really need people that have been in the seat. And then hedge funds and target investors, target customers of the space, they don't want to talk to you. They want to talk about stocks. They want to focus on other things. You have to be more jobs Ian than YC Ian. In the sense that you have to sort of decide what they want. If you have been in the job, I spend most of the day on Wall Street. I know people who they spent their entire day just talking to people. There's, there was a great lady on Wall Street. I won't say who, who was a very successful hedge fund manager. She virtually just spent the whole day calling up people, including people like me, hardworking analysts and portfolio managers and just, what, what do you like?

• 2:27Speaker 01

  She just basically was a fund of funds manager and as she cherry picked everyone's ideas, the people she talked to were good. She would invest more in that idea and she would trade ideas. If you know Wall Street, it's an extremely social place. Investor capital and PE is the same. Everyone's chatting, everyone's networking. If you make a system that you want people to use for this industry, you have to include that social piece of it. Why didn't FAX that do it? Why didn't these other guys do it? I, I spend less time worrying about those questions and I spend more time, let's implement it. The, the first feature we made at Goodell was that, and lo and behold, it was like the most popular thing. Everyone said, I, I love being able to interact with other investors. That gating mechanism of price is a good thing. Bloomberg's fairly exclusive in terms of its price point. Funny thing about Wall Street is $25,000 is not exclusive, but it is a gate.

• 3:09Speaker 01

  Almost any amount of money is a relatively good gate. Anything not a tiny amount, anything in the thousands of dollars. I think automatically who's going to use that? Only enterprises or very wealthy people. So I think that the features are really great and, and a lot of competitors stopped at sort of a base case of features. Bloomberg just went a little further. I think that the, they haven't gone a lot further. The reason Bloomberg is ahead is they went a little further than those other three, but somebody could easily go a lot further. The trick though is it's a very big engineering task. Lots of data, lots of data maintenance. Then procuring new forms of data that the street cares about and showing it to them. You know, you still have to have, I think, and this is sort of the big debate in our office. Do you need the base layer substrate of every bond, every stock, every CDS, every mortgage, every everything, every country?

• 3:55Speaker 01

  Before you can then go to that next, okay, let's further what those other guys have done. My bet is yes. Again, healthy debate for my coworkers that we should just make the killer feature. I'll give you a really good example of a startup that did that called Visible Alpha. Visible Alpha did one thing. This is why the, the. Financial information software industry is such a great space. Visible Alpha were a bunch of hedge fund people and they said, you know what stinks about investing is that I want to know what Wall Street consensus is for Tesla car deliveries. There's no way to get that. If you go on Bloomberg, if you go on FactSet, the best you can get is consensus for revenue, consensus for earnings, things like that. They made the product and they sold it to S and P Global for 500 million. I don't know if 20 million went to the company.

• 4:35Speaker 01

  That kind of ideation, quick turnaround, Wall Street uptake, the sales cycle on Wall Street is fantastic. They love.

• 4:41Auren Hoffman?

  And now you've got, you know, you've got prediction markets, you can maybe grab some of that data. To me, one analogy, you can tell me where I'm wrong here. One analogy to Bloomberg is Netflix.

• 4:50Speaker 01

  You're always right.

• 4:51Auren Hoffman?

  So, you know, Netflix, I love Netflix. I pay 20 bucks a month for it. They have obviously lots of people that pay 20 bucks a month for Netflix. They can aggregate content across all those people. They have the weirdest stuff on there that I would never use. They have all these odd documentaries, the craziest things, right? They have to cater to everyone's taste, a little bit of everyone's taste that's out there. I assume Bloomberg is similar. Most of the data sets they have, like probably by most, like 98% of the data sets, like the individual Bloomberg user would never use any of those. Right. But that's that, that one data set they really need, Bloomberg needs to have to keep that person happy.

• 5:29Speaker 01

  So far it's been just having the few more data sets than the rivals. They're, they're sort of a lazy company. They don't do tons of innovation, but they do stay a step ahead of the rest. It takes years for Bloomberg to launch a new product. It's kind of sad.

• 5:41Auren Hoffman?

  Oh, yeah.

• 5:42Speaker 01

  It's very slow. Do think that that data thesis you have is a good one. What's interesting, I think their margins are much higher than people think. While they spend a lot of money on data, they, I think, are resting more comfortably than, you know, it might seem.

• 5:55Auren Hoffman?

  Netflix has more competition in some ways and they gotta stay more ahead.

• 5:59Speaker 01

  I think so. And the margins are a little bit lower or maybe even quite a bit lower. You have, one of the things there that's really interesting, Netflix probably has this to some extent too. Bloomberg is very tough on licensing. This is kind of like our opportunity, people kind of hate the company, but love it at the same time. Have this two-year agreement that's very vigorous. They don't tell you, hey, your two years up for renewal. If you have 10 licenses, you pay them 250 a year and your hedge fund shrinks a little bit. You want eight licenses now. I mean, you're shit out of luck. It's not AWS. And that just locks them into some very weird dynamics. We're trying to change that a little bit where one of our machines, we have this really crazy AI. It's a vector search in essence for tick data. Nobody's ever made anything like it. It has to run on 8x H100s. This is a very, very, very expensive product.

• 6:44Speaker 01

  In fact, Bloomberg really couldn't afford to do it without a new line item. They don't want a new line item because if you're paying 25K, you say, well, wait, there's more stuff. I'm not getting this, right? We told one hedge fund, if we wanted this for your entire firm, it would cost 7 million bucks over maybe 200 or 300 people. They can't innovate in some ways because of that. They have a price point. And when you lock yourself into one price point, Netflix has this too, where they're hoping you watch as much content as possible. You know, on Bloomberg, some people don't use it. Some people are locked into their system. Goldman might have 10,500 licenses that they're not using. You have these numbers games that sort of help. If all of them were using the maximum compute and the maximum data possible, I think Bloomberg would have to raise its price.

• 7:27Speaker 01

  I like the idea of more of a, some base level fee, but then also like add-ins that more

• 7:32Auren Hoffman?

  Yeah, add-ins, right.

• 7:33Speaker 01

  Honestly replicate.

• 7:34Auren Hoffman?

  Or if you want to do analysis on it, then there's compute. You could just charge for costs of the compute.

• 7:39Speaker 01

  That's what's interesting. Quants don't use Bloomberg. Venture doesn't use Bloomberg. You know, so Bloomberg missed the two biggest trends in Wall Street's history. In the 90s, the quants started to emerge. They needed data. They needed software. They needed all the stuff that wasn't available. They built it themselves. They would have loved to use Bloomberg instead. Unfortunately, they dropped the ball, ran for mayor, he didn't tell you where his priorities were. They just missed that wave. Then the second biggest wave was the rise of privates. I worked for a Tiger, I worked for a Tiger Cub hedge fund. I had this disagreement with my boss that we should do privates and that, you know, companies like Facebook and so forth were probably worth investing in, even though they were private. And, you know, he complained about the liquidity and the issues like that, which made sense. But the rest of the street started embracing privates and it became this huge, huge field and VCs.

• 8:23Speaker 01

  You know, the biggest VCs back in the day were a billion dollars. Now they'll do a billion dollar check. It's a really crazy industry and PitchBook sort of came in to fix that, as did some other guys, Crunchbase, PreQuinn, others, to embrace the sort of private companies. There's still not great solutions there. We're working kind of hard on that. So I think they're this comfortable lumbering company, but you do have to be really creative. I think that's where like Visible Alpha did a killer job. I, you still can't get that data, I think, on Bloomberg. S and P did a great job of buying them. And there's other things like that. So we're working on both building that baseline substrate where somebody could use it for their day to day and replace their Bloomberg, but also the new killer features where they say, okay, this is not available anywhere else. I need this.

• 9:05Auren Hoffman?

  Now, by the way, we're, we're investors in DL Software, very happy investors. Just putting it out there for the audience. There's multiple reasons why we wanted to invest DL Software. One is you have a great product. People like the product. The market's obviously just massive, right? You could just like on Bloomberg, that's the obvious massive market. Like, one of the other things, and I would like to get your opinion about people like you. You've got a big megaphone. You're well known, you're on X, you're out there. The potential of the CAC for you to acquire a customer is much lower. We've invested in a few other folks like that, that they're doing on random software companies, but they've got a big megaphone and you could just see their CACs are just lower than, than others because they're out there more.

• 9:50Auren Hoffman?

  I'd love it to get your thoughts about like those types of entrepreneurs.

• 9:54Speaker 01

  They're certainly well aware of it and, and guys like software engineers come work here and they're brilliant guys. They're great at software engineering more than they are at business. They say, oh, Martin, you know, you can snap your fingers and get a million customers. I say, that's not how it works. You know, I, I like being well known.

• 10:08Auren Hoffman?

  Yeah, of course.

• 10:09Speaker 01

  It helps a little bit. Where it really, really, really helps is this stage of the company where it's depressed.

• 10:15Auren Hoffman?

  Early stage recruiting.

• 10:16Speaker 01

  Recruiting for sure. Palmer, Lucky and I have had some great conversations on that where you get the true believers. That's really helpful and they, they really want to work hard because they like you. They're coming to the company because they, they like what you talk about and they like how you do things. And, and you're really seeing eye to eye. What's really depressing about a software startup is you're Work really hard, you build software and you have $63,000 in revenue and you're just banging your head against the wall. How are we going to break out? And what's amazing about DL is we've been, I wouldn't say profitable, but very close to break even since we started this product. What's amazing about that is I can kind of titrate my noise level on the product based on that. Whereas if you want to compete with Bloomberg, what I would do without, if I wasn't who I am, for better or worse, mostly better.

• 11:00Speaker 01

  I would put my head down, I've raised like 25 million, put my head down for two years and not release a peep until two years passed. That's not really a viable approach for a lot of entrepreneurs. You're supposed to iterate, I think, in software. And if you pop open the product, two years from now, nobody likes it, then you lost opportunity to sort of course correct.

• 11:17Auren Hoffman?

  There are so many graves littered with people who took that second approach because they didn't get the customer feedback. They spent six months developing such an important thing that no one really cared about. And so having some sort of market reaction is helpful.

• 11:31Speaker 01

  It's kind of like the old school entrepreneurs where you don't start the company unless you know who your first 10 customers are going to be. And that's not at all how things are done today. Like you go boldly into the unknown and thankfully, most of the startups in this space are people who did it. There's a company called ROGO, it was a guy who worked at one of our customers and I think they've done relatively well. When you have that experience, it helps a lot that you can get your first couple of customers. In my case, it could be phone calls I make to enterprises or it could be social media. That's kind of the other thing is like, we started off as a, as a, almost like a prosumer product. And now because we've gotten that initial revenue from our very loyal prosumers, we can now approach enterprise and do price increases and start to think about, you know, a different TAM and a different market side.

• 12:12Auren Hoffman?

  I assume, by the way, the first customer was you, like you were like, oh, I need this product, I need to use it. A lot of times people build them and they don't even use it.

• 12:19Speaker 01

  Talking to like companies like S and P Global or others, they've even started talking about acquisitions. They said, why are you guys special? I said, It's just one reason. We have great software engineers and everything. We have great product people. We have all this stuff. It's really well done, but there's one thing that sets us apart truly, which is that we actually do invest ourselves a little bit. And I almost want every engineer to come here and have a little portfolio, whether it's on Robin Hood or whatever. I'm at the point where I'm always like, if you come in here and you don't, I'll give you the money. I want you to trade because one of our, our front end guy was, we have this tremendous front end engineer who's never invested in his life and he's addicted to investing now. He's like become from 10x to 100x because he now knows like what he's coding and.

• 13:00Auren Hoffman?

  You've got to be like, you should be thinking as an investor, you should be thinking, okay, I, I bought this random company because I think it's undervalued. Great. Okay. Here's my assumptions on it. Then maybe things change. It goes up and it changes the way where it's no longer undervalued. Now you think it's overvalued. You have to have some sort of interval of relooking at that. Get everyone in the, in the organization all the way through thinking like an investor, because that's who your customers are.

• 13:24Speaker 01

  Yeah, exactly. I mean.

• 13:25Auren Hoffman?

  The beauty of that is it's fun.

• 13:26Speaker 01

  It's easy and fun. Everybody likes.

• 13:28Auren Hoffman?

  Compliance legal tech or something. First of all, it'd be hard to put them in their shoes. I can't just do it.

• 13:33Speaker 01

  And actively cross the practice law.

• 13:35Auren Hoffman?

  Yeah.

• 13:35Speaker 01

  The delta there is at Bloomberg, you're specifically banned from trading. And I mean, look, maybe the news organization has a hit piece coming out on your favorite stock. At the end of the day, like that's a huge trade off. That's not worth it. I'd much rather have a, a house of investors that love the markets and they are the antenna. I'll take a stupid example that we don't like, the, the whole politicians trading thing. Some people try to make products around this. I think it's, it's nonsense. That was a thing for a while where you wanted to front run Congress's trades. That kind of antenna where people are really talking to the street. That's more of a retail thing, but people are talking to the street, asking what actually do people care about? That's the kind of insight we get that nobody was going to share that with, with S and P global. No offense to what, what a great company and a future acquirers, but nobody's going to really share that.

• 14:17Auren Hoffman?

  So are you going to acquire them or?

• 14:19Speaker 01

  They're a great, they're such an interesting company, Aaron. It's.

• 14:21Auren Hoffman?

  I'm a huge fan of like older school data companies. All of them been around for over 40, 50 years. They're all really interesting things.

• 14:29Speaker 01

  Moody's is very similar and I won't say which company. One of them told us, Martin, we, we are a hundred year old company. We stopped growing about 50 years ago organically and we need to buy companies like yours to exist.

• 14:39Auren Hoffman?

  By the way, these are very good acquirers because they've got a great channel.

• 14:42Speaker 01

  Distribution, yeah.

• 14:43Auren Hoffman?

  You can just stuff the channel with new interesting products. So they're good acquirers. They're disciplined. If you look at these companies historically, they've been fairly disciplined about how they acquire the companies. They can pay X more than a private equity firm because they can stuff the channel and all of a sudden that company is going to grow so much faster.

• 15:00Speaker 01

  And they can afford to make a mistake, even though they're generally very disciplined at 150 billion market cap or whatever. If they take a flyer on a billion dollar deal on a kind of speculative AI company, they kind of want to be in the space, it's a couple months of cash flow.

• 15:12Auren Hoffman?

  Absolutely.

• 15:13Speaker 01

  They're not going anywhere. All these guys, ICE, NASDAQ, these are all great partners. The best thing for us is normally you have to make a lot of noise to get noticed by these guys, but we show up on their customer bill. We sell, basically resell their data. So for us, every day we talk to these guys. So the relationship.

• 15:28Auren Hoffman?

  And by the way, for our listeners who might not be familiar, ICE is not the immigration service. ICE is owns the NYSE. They own a bunch of data companies. Before you did this, you were in the drug discovery world. There's some narrative out there that AI is going to automate or streamline drug development. You keep reading about that. You've been more public about disagreeing with that. Walk us through why, because to me, I'm hopeful that it will, but walk us through why it won't.

• 15:54Speaker 01

  It'll, it'll help a little bit, but you have to be nuanced on where. I had a baby recently. My, my partner is.

• 15:59Auren Hoffman?

  Congratulations.

• 15:59Speaker 01

  Thank you, thank you. My partner was one of the first people at OpenAI and she was also international biology Olympiad champion two years in a row. Very fond of her and I love.

• 16:08Auren Hoffman?

  Your baby is going to be so freaking smart. I feel so bad for everyone else competing with your baby.

• 16:14Speaker 01

  He's a darling. I brag about Chelsea all the time, but we talk about this nonstop because if there's something to be done at OpenAI even, she could probably marshal that effort. So what do you, what would you do? We, we have some ideas here, but I think that just like what we talked about a minute ago, you really have to talk to pharma people. You know, ask pharma people if they think AI is going to change the world. Most of the people in the drug discovery world, most of the people who are talking about this are the entrepreneurs and investors. If you actually talk to a chemist at Merck or a drug discovery person at Pfizer, they'll sort of shrug their shoulders and say, our bottleneck starts in the clinical trial world. The drug discovery part is sort of the tail that wags the dog. It's the most important part in a sense because everything comes from it.

• 16:54Speaker 01

  But if you're going to spend a billion dollars developing a drug, 900 or more of that is going to be the testing in the humans, which it doesn't feel like AI plays a role there at all.

• 17:02Auren Hoffman?

  Don't you think it can make you a better prediction as to what will go through the trials?

• 17:07Speaker 01

  But you know, just to start with the trials, how are you going to speed that up? That's mostly a money thing. You got to pay the doctors, you got to get the data, collect the data, pay the patients. It's not like direct, but you're paying the university and things like that. It's time consuming. It takes a year or two to recruit, you know, a thousand people with cancer or something. Then you got to do the placebo control, you got to do confirmation confirmation studies, so forth. A lot of waiting in the drug business and animal studies are a good example. You have to do these so-called two species rule. You have to try your drug at very high doses in two different species and sometimes for quite a long duration. However, many months you want to do in humans, you got to do in animals and you got to do histopathology. You got to dissect the little guys and see that they don't have some abnormal, you know, liver or something.

• 17:48Speaker 01

  That's like a two-year process that you do sort of concurrently with the human trials, but it's a bit staggered. So when you do the chemistry stuff, we've been using AI in chemistry for a long time. There really isn't much there. So I can say this as a chemistry patent award winner, made many, many different drugs. Chemistry in a lot of ways is a slightly more sophisticated jigsaw puzzle. If you have the crystal structure for the protein you're interrogating, any chemist worth their salt, pretty quickly can come up some ideas and having AI is instead of having a drill, I have power drill. It really is a vanity thing. That's almost Beblin and vanity for drug companies. And then target discovery is the hardest thing in drug discovery. It's ideation. Drug discovery is fundamentally an idea business. This is kind of what Vivek Ramaswamy and I both both saw as sort of investor investee,

• 18:33Speaker 01

  and also as competitors sought to disrupt where what makes a great drug is somebody synthesizing all the publications in the world and saying, Hey, you know, nobody's ever tried to make a Bruton's tyrosine kinase inhibitor for CLL, a rare leukemia or lymphoma. And the combination of those two ideas and then trying it out, that is the search space that you need to navigate as a drug discovery person, knowing that, you know, it's a 20 to 50 million dollar, you know, first roll of the dice. I think AI might, might be able to do that, you know?

• 19:02Auren Hoffman?

  That's one of the things is when I go into these labs and I see these super smart PhDs who are doing stuff like, it seems like a lot, very high percentage of their job is like manual labor, fixing things, moving things, cleaning things. To me, it's just funny. I was like, wow, so this person, they, they didn't even get like a PhD until they're 10 years after college or something like that. They spent all this time in school doing all these things and a postdoc. And now it's like they're, they're Starbucks barista.

• 19:36Speaker 01

  In a way, that's totally true and it's part of being a scientist, I think. One of the interesting things that correlates to the last question you asked is in my drug discovery career, I think I've taken 10 drugs from idea to human clinical trials, largely my idea, sometimes a close colleague. Here's the craziest thing, I've never set foot in a lab. And most of the time I email Wuji, the Chinese contract research organization. Wuji, it's all PDFs. If you go to Merck or Pfizer, like Pfizer headquarters here, you're not going to find a single beaker. You know, a lot of it is paperwork. And even crazier, while I was in jail and the COVID thing was happening, we, we started to talk to Wuji about making COVID drug. I was doing this from prison, which is even crazier. It was extenuating circumstances on many fronts.

• 20:20Speaker 01

  Wuji was emailed our chairman of the board, who's kind of keeping the company warm for me. He would copy and paste that into the jail email. It's very esoteric, ancient looking thing. And he'd say, Wuji has a protease assay they're spinning up. I'd go way back with the Wuji people and they said, you know, they'd love to do something with us. What do you think? And I'd ask questions with the assay and so forth. At the end of the day, a really smart AI agent could do the bulk of this work. Now, some, somehow a robot or human has to, like you said, do the barista job. But the thermometer checking, the sipping the coffee at the end.

• 20:50Auren Hoffman?

  It's hard.

• 20:51Speaker 01

  It's pretty much, you know, something a human can say, oh, results came back great, or I don't know about this, I'll try it again. If you know the space, you can largely do it over email and pen. That'd be a very interesting company, but entrepreneurs seem to want to focus on things that are like, on a scale of one to 10, let's make something that's already a nine to 10. I think that if you've ever run a drug company, you know where the bottlenecks actually are. And those are very tough problems as they sort of should be.

• 21:14Auren Hoffman?

  You, you mentioned jail. You famously spent some time in prison. For those of us who haven't, like, what is it like day to day? What are the surprising things?

• 21:22Speaker 01

  It's a very sad place. So civilization and its gradient and its non-intentional and maybe sometimes intentional trends. Prison in large part has always been a place for mostly very poor people. And then people with sort of mental health issues touches you very deeply because the vast majority of people in prison are minorities. Asked a friend, I actually have to catch up with him. He's really a guy with this great heart. And I said, so what do you enter here for? And he said, poverty. And I said, I said, you know, it's such a poignant thing. You know, he said, look, the answer, the real answer is drugs and guns. But the, the quiet answer is I grew up in a, in a really poor neighborhood. You know, nobody knew, you know, how to make money. The only way to sort of be an alpha male, if you will, and be a provider into success. That's what it looked like.

• 22:05Speaker 01

  You know, a great, great CEO, actually, I've had a lot of arguments with him. But Ken Frazier, the former CEO of Merck, is this tremendous, maybe one of the greatest black CEOs of all time, Ken Chenault, American Express. Those role models don't make it to the poor neighborhoods. And it's very sad, you know, single parent households. Almost everyone I met in jail, single parent or no parent household. You know, really fascinating. Mark Andreessen says that I make prison look desirable because I think he's a big reader. And I said, I read 500 books in prison, which I've read maybe two since coming home, cover to cover, and same, roughly same duration. There's no cell phone bothering you. You're just going to sit and read. And he said, that sounds like paradise.

• 22:42Auren Hoffman?

  I was going to ask you about that. Like, I assume that's a massive pastime is just reading.

• 22:47Speaker 01

  Everyone spends their time differently. I mean, for me, I, I really enjoyed that and, you know.

• 22:51Auren Hoffman?

  And you can get access to books. If it's a normal book, you can get access to.

• 22:55Speaker 01

  I had people sending me, you know, I had a staff. I had an intern who is now a Thiel fellow. And her resume is Martin Scully Prison Intern. She's, she's made it to the Thiel Fellowship. I'm so proud of her.

• 23:05Auren Hoffman?

  Can you play like settlers of Catan with other prison mates or poker?

• 23:09Speaker 01

  Yeah, you play cards, you work out, you eat, you waste time. I mean, trying to waste as much time as you can. For me, I said, well, I'm not going to waste this time. This is, you know, really productive time. I learned a lot about so many different things and it was enjoyable. You know, you miss your family, you miss everything. You think about the mistakes you've made. You know, you try to grow. I spent a lot of time thinking about that, even though I don't think I committed an actual crime. Nobody's perfect and everybody's made tons of mistakes.

• 23:32Auren Hoffman?

  Of course, yeah.

• 23:33Speaker 01

  Think about just slight adjustments you can make to your life. What am I really trying to prove to people? Am I driven by that? Or am I driven by my actual goals that I want? I was able to think through some of those things and it definitely helped. You know, a timeout.

• 23:46Auren Hoffman?

  It's like people who go to those 10 day silent retreats and come, come out with a new theory of themselves.

• 23:51Speaker 01

  But I am really alarmed at this trend of putting people like me in prison because, um, you know, a good example is, I think his name, I'm going to butcher it, is Anthony Lewandowski. He was the guy who sold auto, I think, to Waymo. They arrested the guy and his, his story is actually almost exactly the same as mine. He was at Google and he was going to go to Uber. This was a, you know, a big problem for Google. So they basically dialed an indictment. And the way that works is if you're Google, you're paying every law firm on the planet, you know, a huge amount of money. You go to the one you're paying the most, whatever big law firm it is. You say, hey, guys, we, we were just the victim of a crime. I said, really? What is it? And they said, well, our guy's just left and is taking trade secrets. What can we do to this guy? And they said, oh, I got this. I just came from the prosecutor's office. They all used to work for prosecutors offices generally, the assistant US attorney.

• 24:36Speaker 01

  I'm going to call my guy who still works there. And in, I think, 72 hours, they got the guy indicted. The same thing happened with Goldman Sachs and so forth. These guys are looking for crime.

• 24:44Auren Hoffman?

  Legitimately, there could be somebody who leaves your company, steals a bunch of trade secrets, steals a bunch of things that, that are important to your company.

• 24:53Speaker 01

  Absolutely.

• 24:54Auren Hoffman?

  And tries to bring them to a competitor, right? They can go to Bloomberg and, and bring stuff.

• 24:58Speaker 01

  This company's called, there's a name for this company. I think it's called Anthropic, but go ahead.

• 25:01Auren Hoffman?

  So, right. I mean, just whatever it is, like, I'm sure legitimately it could happen. The company can come after them. They could sue them.

• 25:07Speaker 01

  That's the first step.

• 25:09Auren Hoffman?

  In some ways, you are stealing stuff. If you're stealing information, information is valuable. If someone does steal something. They should go to jail, right?

• 25:16Speaker 01

  What degree of criminality is it, right? If you're a Chinese citizen who comes to America, gets a job at Google, sends the source code, all the source code to home, yeah, you gotta arrest that person, especially before they flee. But, you know, in his case, the facts didn't really merit this.

• 25:30Auren Hoffman?

  Yeah, I don't know his case specifically, but there are, there are people who steal things.

• 25:35Speaker 01

  Hundred percent. I'm not excusing. It's a crime for a reason, and it should be. You know, you can get a declaratory judgment. You can get a restraining order, right? There's, there's tons of things you could do before you say, let's, let's throw a guy in jail. The shift has been from, let's use the law sparingly. The long arm of the law is very powerful. Let's use it when we need to use it. As opposed to, can we, can we increase the jail population? Can we ruin some lives? I think that Sam Bankman is actually an example of this. You know, Sam probably should, should do some jail time, but he got a 25 year sentence. He had a financial institution that failed. There have been a lot of financial institutions that have failed in America. The amount of people that have gone to jail for them is close to zero. You know, I'm not saying he's some saint. I'm not saying he didn't mess up big time. I'm just saying there's, there's this reaction.

• 26:20Auren Hoffman?

  In his case, you think there's a severity. You say 25 is too severe. You should get some, but, but it might be too severe.

• 26:27Speaker 01

  Cause remember, this, this guy's now going to be banned from all kinds of professional associations. He is going to have the stigma of a felony. You know, there's, there's already going to be a lot of punishment. And I think we're, we're going to talk about Elizabeth Holmes maybe in the same way. But SBF, this guy's an MIT grad. He was Jane Street alum. You know, this, this is not the average person you find in prison. Take it from me. It's extremely unusual. Well, he's probably still could be a very productive member of society. What bothers me the most about it, Orin, is, and it's kind of hard to put a finger on it, but, but here it is. Is that the causality is reversed. I think I went to prison and I think Sam got his gigantic sentence. Maybe even Elizabeth Holmes as well, not because of the conduct, because of the reaction and the social media. That's what propels and gives the impetus to a prosecutor who says, well, I should be doing something about this.

• 27:12Speaker 01

  I'm in charge of this district. There's gotta be a crime somewhere. I want to be the knight in shining armor. I want to be the good guy. I want to be the person that took down Donald Trump in the case of Letitia James or something like that. Another foe of mine, you make your job about looking good. Instead of actually cleaning up the streets and making it a better community and things like that, these are the choices you make. Let's put a very productive person in prison. Maybe they made a terrible mistake. Maybe they messed up. There's other remedies, you know, actually locking somebody in a place that's sort of inhumane.

• 27:40Auren Hoffman?

  Back in the day, like cut off your finger. They would flog you in the street. There's something to be said about that type of public punishment.

• 27:48Speaker 01

  We got that on Twitter.

• 27:49Auren Hoffman?

  But maybe it's less harsh than putting someone in prison for 15 years or something.

• 27:54Speaker 01

  Like I said earlier, for some communities, it's kind of like an expected term style. Applying that to business writ large, like Nicola and other high profile failures. If you wanted to make a crime out of every failure, I think you could. Because at some point, you could probably find an email or a comment. They did this to Binance. You know, Binance had some junior person that said, oh, LOL. I think we're sending money to the Taliban. CZ had to go to jail and did four years for this. You know, they had some issues at the company. I'm not saying they didn't. This was much more of a vindictive, like, let's take this guy down a peg kind of thing. As opposed to, you know, is this the crime of all the things happening in this district or the state? Is this the crime that really needs to be policed right now? While we have other major issues. I think in some ways prosecutors like being famous. They like getting attention.

• 28:38Speaker 01

  If you just take down the next drug dealer down the block or the child predator down the block, of which there's gigantic amounts, that's not going to get 20 CNN reporters asking you questions. If you take down somebody like Trevor Milner, Elizabeth Holmes. Business failures, they come with problems, right? When you have a spectacular blow up, your next job is not going to be great.

• 28:57Auren Hoffman?

  A lot of people would say, okay, we had this huge financial crisis, 2008. None of the major people went to jail and caused all this havoc. It caused this problem. A lot of people are almost taking the opposite side of you. Like, we're not sending people to jail. Maybe they're too smart and they're getting out of it. They didn't put it in writing. They were actually doing something bad. They were taking advantage of people. They're taking advantage of the grandmas. They were doing really bad things for society. If we put a few more of them in jail, it would stop people from doing this or at least reduce the number of people. You hear that argument all the time, both on the right and the left, right?

• 29:34Speaker 01

  It's so totally not easy, right? I mean, any of this stuff is past my pay grade. I certainly can just see that there's some flaws. If you're too lenient, you know, you're disrespecting the victims of the crime. In my case, I had a hedge fund that made about five times its money. I'm the first person ever to be, to be, you know, make that criminalized. Now, if you tell a lie while you roll the dice and instead of you gave me a million dollars and I said, I'm going to, I have this sophisticated quant program. I'm going to make you a lot of money and I go to analytic city and put it on red and I come back and I say, hey, Orin, I made you five X. Was there a crime committed? I think yes, that should be punished, but at the same time, you know, that's, that's not what happened here. And I, I think that there is something to, in all's well that ends well.

• 30:14Speaker 01

  This goes back to the point, if Elizabeth Holmes sold her business for five billion to a lab company, what would the result here be? I mean, you'd have one very pissed off lab company, perhaps so much of her, her jury was the Wall Street Journal and a book that was written about her more than it was what actually happened at the business level. We don't know the full story. We know the Wall Street Journal reporter's story who did a thorough investigation and I'm sure did a great job, but having been in the seat and been in the arena, it's just very hard to know exactly what happened there and how it happened.

• 30:41Auren Hoffman?

  To me, that one is more pernicious than some of the other ones because there are people who relied on these tests. They got these tests, they relied on them, the tests were faulty, she knew they were faulty and they were making medical decisions based on those tests. They're, she's essentially putting people's lives at risk. And I believe that she, she knew that she was doing that. She knew she was putting people's lives at risk. That to me seems a much more pernicious problem than, than some of the other ones.

• 31:10Speaker 01

  I think so too. I would also say that diagnostics aren't perfect. Even Roche or Siemens, the best diagnostics are not perfect. If you're a doctor, you do a second test, you do a third test sometimes. And there are people that need to, to take those gambles. For example, I think we want the Theronos product. We all, we all kind of would like that, right? How do we incentivize that? And she shouldn't have had to lie, if she did lie about all this, which a jury, you know, found, but she was also exonerated of some, some of the, the claims, I believe. It is very hard to win in federal court. I won a few counts as well. It's just basically impossible. It's, the trial is almost like performative. You could almost argue the fake it till you make it for the products that we really need may be one of the only ways to, to actually get there. If the VCs cut you off at series A.

• 31:54Speaker 01

  If you never have the push to get it across the finish line, longevity to get it there obviously doesn't excuse fraud.

• 32:00Auren Hoffman?

  There's a line though, right? There's a line of being optimistic and then there's a line of faulting your revenues or something, right?

• 32:06Speaker 01

  Yeah, if you do that, that's instant jail, right? This comes to public companies all the time, right? There's, there's quantum computing where there are public companies.

• 32:12Auren Hoffman?

  Right. You're a bear on quantum computing, right?

• 32:15Speaker 01

  Yes, it's a very interesting field. My partner is not a huge expert. She taught with Scott Aronson, who is the quantum expert at MIT. And as a compute company, OpenAI has huge interest in compute. I've always had this interest in quantum computing, especially vis a vis Bitcoin. I think it's all largely a scam. There are publicly traded companies that are saying things that are palpable lies. And they're, they're a hundred percent lies. They're not going to go to jail because they're safe harbor statements. They're saying things like, well, one company is a gumption to say, we expect we'll beat Nvidia by within two or three years, which is totally insane. And quantum, quantum computers don't have algorithms to do that. You know, they have one algorithm that can speed up decryption. That's real, but they can't do matrix multiplications or list sorting or like anything like that. We're so far away, we're decades away.

• 32:58Speaker 01

  The CEO that said that, his, the founder of the company simultaneously was saying it's 30 years away. That, that's a crime right there.

• 33:04Auren Hoffman?

  The reason why people love quantum computing, it's, it's kind of easy to fall in love with. First of all, it's a fantastic name, right? It's a great branding, right? You like quantum, you like computing, put them together. You could see why we all fall in love with the story. We want it to happen, right?

• 33:21Speaker 01

  Especially now because Nvidia is a $5 trillion company. And that is enough to get people's imagination going and what's next? If, if it's a 10 billion market cap as a public company, that's still X upside. If, if it's the next Nvidia, so it's worth taking a flyer on. I'm actually really excited about optical computing. It's ironic because it's a lot easier to do than quantum and it's a lot more needed right now. But it's, it's still sort of underground and not well known, but I think it's gonna happen.

• 33:47Auren Hoffman?

  We need a better name on optical, it just doesn't sound catchy.

• 33:51Speaker 01

  Yeah. Photonic computing.

• 33:52Auren Hoffman?

  Photonic computing. All of a sudden, every company in optical computing would have a 5X valuation.

• 33:58Speaker 01

  I think you're right.

• 33:59Auren Hoffman?

  They called it photonic.

• 34:00Speaker 01

  And Nvidia is doing it a little bit themselves. You're hearing every day on Wall Street about co-packaged optics. There are lasers going to chips. All the interconnects are optical. All the data in the world is moved on fiber. We're actually more optical than we think. It's just the processors that don't do it yet. What I discovered, which is really ironic, is the matmul, the matrix multiplication. I did some GPU for a few years before it really got good all solidified. We tried to compete with 11 labs. We tried to do a few things and it didn't work that great, but I got to know GPUs really well. And 98, 99% of the GPU time is spent doing matmul. There is one way to do matmul really fast. Light does them naturally.

• 34:38Auren Hoffman?

  Let's say you are bearish on a company. You think it's overhyped. You want to short it. In today's world, shorting just seems super risky. A lot of these stocks are memes. How do you do it in a way that you feel good about it in this weird market that we're in?

• 34:52Speaker 01

  I take you back to 98. Number one answer, it goes back to every question in finance. Simplest answer is don't be greedy. Everyone that messes this up, including myself, is just too greedy. Go back to 98, 99. The two most powerful hedge funds in the world are Soros and Tiger. They made 30, 40% a year since the 80s. They both fold. In 2000, Soros quits. Lee Robertson quit, started the Tiger Cub movement. The fund I worked for was up 40% that year in Kramer's hedge fund. And the fund I would go to work for next was a Tiger Cub that was up over 150% that year. I think some of it's luck. At the timing, you sort of like weighed out, like, when is it really the most fever pitch, most euphoria?

• 35:33Speaker 01

  Maybe you pick on the worst zip code and the worst neighborhood or the worst house and the worst zip code or something. Part of it is like, just don't be greedy. Instead of making it a 10% position, which could blow up your whole fund or your portfolio, why not a 1% position? And the answer is, well, I want more money. I want to make more bigger returns.

• 35:50Auren Hoffman?

  You're weighting it by the Kelly criterion or something like that, right? And then the question is just the Kelly criterion is based on some sort of percentage. Then the problem is, okay, you may just be over optimistic about that percentage. Maybe you should just do the Kelly criterion and then divide it by five.

• 36:03Speaker 01

  A lot of people do half Kelly or quarter Kelly criteria, and I think that's fine. I think factor analysis is the only way to invest these days. If you're short of high momentum stock, you must correspondingly be long something high momentum. What I would do is I was, I went long talent here as, as sort of a hedge to quantum, even though nothing to do with each other, they share this high flyer kind of quotient or factor.

• 36:26Auren Hoffman?

  Yeah, like a Tesla two, like all those things.

• 36:28Speaker 01

  Yeah. And then you, well, the problem with those two is now you have market cap quotient or market cap factor that you're not hedging. Now you're going to need to hedge with small caps or other large caps.

• 36:36Auren Hoffman?

  Yeah.

• 36:36Speaker 01

  You sort of have this balancing act.

• 36:37Auren Hoffman?

  Or S and P or something, right?

• 36:39Speaker 01

  And then that's what's called output. What most people find is after you do all that, your returns go to zero.

• 36:43Auren Hoffman?

  Yeah, totally. And then I made less than a bronze or something.

• 36:48Speaker 01

  That factor happens today and the last few days, that factor happens to have come all the way back. Quantum stocks have gotten decimated. It's, it's this very unpredictable thing, but the basic answer is if I gave myself more time and I gave myself more leeway, I would have made a lot of money. Most people know I panicked and capitulated because I did make it a 10% position.

• 37:05Auren Hoffman?

  And I guess when you naked short, you have unlimited liability. You could cap that and you just pay a little bit more to cap it and it's all good.

• 37:12Speaker 01

  You can cap it or you can just say, look, at 1%, what are the odds a stock will go up 100X on me? And I can't get out, that's very unlikely.

• 37:19Auren Hoffman?

  Yeah.

• 37:19Speaker 01

  And, you know.

• 37:20Auren Hoffman?

  And you just, you just know once your losses are X, you just get out, right?

• 37:23Speaker 01

  Yeah. Greed will always be the one thing that causes financial problems.

• 37:26Auren Hoffman?

  We're humans, we're greedy. And even if we're not greedy, when things are going, well, there's a FOMO and there's other things. Yeah, hubris, those are so uniquely human.

• 37:35Speaker 01

  This is why quantitative trading is so great. Even though quantitative investing, the machines are not as smart as us for now. That's obviously going to change. They're extremely unemotional. So whenever I have a stop loss, I'm sitting there saying, no, no, I don't want to get out. This is admitting failure. You actually have to train yourself to the opposite. You get excited. You're like, ah, I'm glad I got stopped out of that.

• 37:55Auren Hoffman?

  Yeah, Buffett famously went into tech stocks in 2000 because all he was missing out in a way, even Buffett, right?

• 38:03Speaker 01

  I think that many people did get lured into the 2000s or tech trap. I think Buffett was actually the one who sort of stayed away from it. But look at AI, I mean, you're going to see lots of people sort of FOMO into AI. And you know, the VCs that have and haven't, and you feel the most FOMO at the top. And that's when the company gets a maximum valuation, they raise the most amount of money. This happened in the dot com bubble, it'll never stop happening.

• 38:25Auren Hoffman?

  Bubbles are not a bug, they're a kind of a feature in a way, right? It fuels a lot of innovation that happens. Obviously, that means you're often at the top, you're spending money that's ROI negative at the top. But on your way there, it can really help the world. Sometimes it doesn't help those companies, it doesn't always help those investors who are doing it. But laying all that fiber in the late 90s was really good. Even if all those companies went under, it was really good for the country.

• 38:53Speaker 01

  Yeah, it's funny, when Jensen Wong said, "Quantum's 30 years away," all the quantum stocks dropped 50% in one day. It was a great day for me.

• 38:59Auren Hoffman?

  Oh my gosh.

• 39:00Speaker 01

  And then he went on, I think it was GTC, and I didn't even know they were public. There is something beneficial to bubbles in the sense that they get a lot of attention. And they force people to sharpen their lens a bit and say, is this real? Is this not real? And nuclear is one, for me, I don't know nuclear from a hole in the wall. But when I see these companies like Oclo and others, like amazing valuations, I actually start to think, okay, well, maybe energy is important. And of course it's important.

• 39:24Auren Hoffman?

  Yeah, totally.

• 39:25Speaker 01

  You know, it makes you sort of think about it. And energy policy today in the White House is very important. What should we do about quantum computing? Is there something that's important there? They think that, you know, maybe there's little pieces. But, you know, the important thing right now is do matmoles really fast. So we'll see how PA with Nvidia will look. But I think that optics, you know, could be one way to do it. And then other than that, you know, I think Odell You know, we're getting close to a public launch where we're gonna sort of start advertising and marketing. We haven't done any marketing and we have a good revenue run rate.

• 39:57Auren Hoffman?

  It's amazing.

• 39:57Speaker 01

  And we're very lucky, but we're now gonna start turning to demand generation. It's, it's something that I'm very nervous about because selling drugs is very easy. You almost have to do almost know of demand generation. And the doctors know, your job as doctor is to know what drugs are available. You kind of give a little bit of a hint. You know, software is wildly the opposite. You know, there's a perfect solution for your problem, you just don't know what it is.

• 40:21Auren Hoffman?

  You can continually make it better. E- e- even if something was good for you in the past, it might not be anymore. Just if you just think about how I'm sure we're all moving between all the models all the time. Just because you were using OpenAI two months ago, maybe you're using Anthropic today. And maybe you'll go back to OpenAI a month later.

• 40:38Speaker 01

  Unfair things.

• 40:39Auren Hoffman?

  Things change, right? So software changes, whereas like the drug is kind of like set.

• 40:43Speaker 01

  Yes, absolutely. You can't change it after. I do think what's funny about that, just, just thinking aloud, is the moat's cut both ways. There's companies out there making better products than the incumbents. I'm sure some people feel this way about OpenAI. OpenAI is so dominant. A lot of people on the Twitterati and things like that, in the valley especially, think, oh, man, Anthropic's so great. If you look at the actual dollar share and the actual mind share, I did this with traffic. So you just use similar web. OpenAI is about 67% of all the traffic related to AI writ large, everything. I included chatbot, character AI.

• 41:14Auren Hoffman?

  Such a great, such a great brand. Brand's important.

• 41:16Speaker 01

  So the first mover, great brand. So, so what do you do? You can make the best product, but you have to fight on every front. You have to fight on the sales front, marketing front and so forth. I'm not used to that. It's, it's kind of a different vibe, but I'm, I'm having fun doing it. One of the things I want to do, this is like very tricky, is I was talking to Joe Lonzel about this a little bit. Godell is the only place where you can find the day-to-day stock price of OpenAI and SpaceX. And we're, we're the only guys who get it from NASDAQ private market and NASDAQ. They are the largest kind of hand-to-hand broker of these private transactions, but there's no order book. We want to make an order book for the first time ever where you can trade SpaceX or any other private, within reason, and oral in an actual order book where you make your market order, your limit order, you know, you cancel your order, et cetera.

• 42:01Speaker 01

  There's obviously a lot of open questions about how, how in the world could you do that?

• 42:04Auren Hoffman?

  How do you do that? And how do you do that without the company's permission for every time you do it?

• 42:08Speaker 01

  And you're not on the cap table, obviously, in this. I've come up with a really creative. Solution or us testing this, but the biggest, the biggest issue in our case is regulation. Well, Lon still said something really funny that the, the chair of the SEC on a podcast. And he said, well, I can buy Orin Coin 69 tonight and lose a million dollars and there's no regulation, no crime is committed. But if I dare try to buy a hundred thousand dollars of Anduril, oh, well, you know, there's got to be a hundred million regulations. So clearly something's wrong with that. And whether you do it offshore where there's the SEC, you know, can't reach you, uh, or you try to do it onshore and, and work with regulators who don't want to work with you. And ultimately these assets are so big. In the old days, we didn't care because the stock market's

• 42:51Speaker 01

  so big, who gives a crap. But now it's like, no, some of the biggest companies are private.

• 42:55Auren Hoffman?

  Not only the biggest, but the ones that have the most mind share.

• 42:58Speaker 01

  Most interesting company.

• 42:59Auren Hoffman?

  Stripe, Databricks, Anthropic, OpenAI, you go down to SpaceX, you just go down the list. Like these are, these are the companies we're talk about even more sometimes than Microsoft, Nvidia, even more than some of the Mac seven.

• 43:12Speaker 01

  It in 2005, the hedge fund I worked for, her biggest investment at the time was like Motorola. Facebook was taking off and like, we can't buy it because you know, we're not Peter Thiel. You know, it's, it's, and the anointed few who got into that round, no shade to any of the, you know, great friends who, who were early there, but you know, you had to go through, you know, the SV sort of hoops. And even if you're fidelity, you're still only getting invited to like series D or something. And yeah, you know, I think that.

• 43:39Auren Hoffman?

  Nowadays, really series G, I type I think.

• 43:43Speaker 01

  Right. So, so the private markets, they just use sort of the public markets in a way for exit liquidity. I think that's wise, but I also look at Figma, a lot of the privates end up being overvalued private. And when the public reality sort of hits you, these marks were really, really insane marks. That part of Wall Street is so important. If you type in Anthropic or Anduril into Bloomberg, it says not found. We have like a full toolkit for that. And one step further would be actually letting people trade. We've, we've heard from some of the bigger places, the citadels, the Jane Streets, the bigger hedge funds, that they would love to do this. It's something that we, we hope to roll out at some point, but it's a very challenging sort of regulatory thing. Nice thing about it is the SEC cares a lot less about Citadel, Jane Street trading with each other than they do about the average Joe sort of losing $500 on something.

• 44:26Auren Hoffman?

  Speaking of Citadel, I've heard you speculate a few different times that. Potentially Ken Griffin, or maybe a Michael Bloomberg could actually be the richest person in the world. Walk me through the logic of that. Obviously, they're extremely wealthy.

• 44:37Speaker 01

  When you have a private company, Forbes doesn't know how much it's worth.

• 44:39Auren Hoffman?

  Right.

• 44:40Speaker 01

  The only question is, okay, what's your income and multiply that by 20.

• 44:43Auren Hoffman?

  And Ken Griffin owns like 100% of Citadel?

• 44:46Speaker 01

  Pretty close. I've heard, I've heard different things, but it's, it's a substantial majority.

• 44:49Auren Hoffman?

  It's a lot. Like that's more valuable than almost any other financial institution, right?

• 44:53Speaker 01

  Yeah, Goldman Sachs, maybe, you know.

• 44:55Auren Hoffman?

  Yeah, some Big Bang, JP Morgan, maybe. Yeah. Imagine owning 100% of JP Morgan.

• 44:59Speaker 01

  Right, 700 billion. What's amazing is Elon's giving everyone enough of their money. You know, that's the craziest thing is that.

• 45:05Auren Hoffman?

  Yeah.

• 45:06Speaker 01

  You know, it doesn't, it doesn't even matter because, you know, unless you're Elon, he's in a class of his own. All this sort of was viable before Elon went on this like literal rocket ship to, to a trillion dollars. But excluding Elon, if Griffin makes, because he has sort of two firms now, he's got a, a market maker too. If the market maker makes to him personally 5 billion, instead of Dell to him personally, say is another 5 billion. That's 10 billion bucks, times 20 is 200 billion. He's right there in the thick of things. Most people just, just treat his cash flow as a one time.

• 45:35Auren Hoffman?

  He's way more liquid than most gazillionaires, right?

• 45:38Speaker 01

  And Citadel's still growing. What Ken did was, was amazing. He looked at the hedge fund business and didn't think about it as, how much cash can I make this year? He much more thought about it as, I'm an entrepreneur, I have to create a business, I have to get market share.

• 45:49Auren Hoffman?

  He built a company.

• 45:50Speaker 01

  Yeah.

• 45:51Auren Hoffman?

  Yeah.

• 45:51Speaker 01

  Almost every hedge fund is just one guy who likes to trade.

• 45:54Auren Hoffman?

  What I admire about Citadel, as opposed to many of these other hedge funds, is let's say Ken Griffin retires tomorrow, or God forbid, something happened to him, or something like that. It's still an amazing business. Now, is it less amazing because he's no longer running? Of course, he's a genius and he's, he's a one of a kind person, but it's like Apple. It's still a great business, even if like the founder leaves, it's still going to be an incredible company going forward.

• 46:19Speaker 01

  If Ken were here, truth be told, I think he might tell you it's, it would be better because he's super humble, one. Two, he's got an incredible team that has learned the culture. His culture is so amazing. One of the things I've, I've tried to use in my own playbook, and it, it required me hitting my head against the wall a few times, is Griffin has this unrestrained, aggressive behavior. Towards hiring people. Most of the hedge fund world, you have people coming to you begging for a job. You total your thumbs. So I think about it. This guy is just the second somebody becomes available. He takes a private jet, flies out, wines and dines them. You gotta come to sit at all. Here's why. People shake their hands and say, that sounds great, Mr. Griffin. You seem like a great guy.

• 46:54Auren Hoffman?

  One of my friends, John Arnold, has a story. John, he was.

• 46:58Speaker 01

  This is very famous.

• 46:59Auren Hoffman?

  Right. Energy trader at Enron, the good part of Enron, not the bad part of Enron. He just gets all out of the blue from Ken Griffin. It's like, what the, you know, and he's there like with a private jet, like right away.

• 47:11Speaker 01

  Ten guys with them, you know.

• 47:13Auren Hoffman?

  Yeah.

• 47:14Speaker 01

  I mean, no.

• 47:15Auren Hoffman?

  Like what? I mean, he wants to win. He clearly wants to win. I so admire Ken. I mean, how could you not admire him, right?

• 47:20Speaker 01

  Well, what's interesting is, so again, I know this from my, my partner, but I think I, I've heard it independently. Quant firms have an issue where the AI firms pay more.

• 47:29Auren Hoffman?

  Yeah.

• 47:30Speaker 01

  There's a friend of mine who just left OpenAI, he's a very senior guy, and he came from quant. Now, if you're Ken Griffin, you're used to getting the best and you having people leave to go to Anthropic OpenAI. And they, they recently sent, I can't say who, but they, they sent recruiting interests to, to one of, one of the people I just mentioned. I talked to that person. I said, I don't think they can afford you. You know, it's just kind of the craziest thing in the world, if you think about it, but.

• 47:50Auren Hoffman?

  Other, which I love. That's new. You, you had some of the best minds in the whole world that were going into hedge funds, starting, let's say in the late eighties. And then it almost decimated physics, like physics really hasn't changed since then, last 40 years. What new things in physics have happened? Like nothing, right? Decimated so many of these other interesting areas because you're still solving interesting puzzles. You're doing fun stuff.

• 48:12Speaker 01

  Make another dollar, you know?

• 48:13Auren Hoffman?

  Money. Why not go into this area? Now, a lot of those same people that could go to AI lab, that there's many more opportunities where they could do great work, really innovate and still make money. Making money is still important to people, right? You still have the opportunity to do that, which I think that's, that's a good trend for society.

• 48:31Speaker 01

  I think so. The people that really like hard puzzles, the stock market's a really hard puzzle. It's a lot of fun.

• 48:36Auren Hoffman?

  Yeah.

• 48:36Speaker 01

  It'll be attractive to a lot of people, but I also think that. Trying to make machines think is also a really great puzzle. The startup world, like you said earlier, has a lot of functions. And the AI bubble, if you call it that, one of the functions is that. There's a lot of ways to think about this. There's math AI companies now, as you know. I kind of want to make an encryption, sort of like lab, if you will. Idea of like, let's bring back the labs in some ways. OpenAI was not meant to have a product necessarily. It just wasn't meant to have. Just do nonprofit research in AI. And the opportunity became so good, they had to do it. There's something really attractive about having a pool of large capital and saying, we have a 10 or 20 year horizon. We're just going to sit here and do this. It often won't be that profitable. AT&T didn't make any money from the labs, even though they invented many things.

• 49:19Speaker 01

  So it turns out to be not a great business move, but depending on who's running it, I do think it can be great. So Bitcoin to me is like a very interesting one. I have this Brainiac partner, he's a million times smarter than me, but the

• 49:31Speaker 01

  the entire Bitcoin and the entire crypto world rests on a simple thing called elliptic curve cryptography. And elliptic curve is a very easy thing to understand. So it's y squared equals x cubed plus seven. And that's the curve that protects Bitcoin. And it's modulus of large prime number. So a talented high school student or even a mediocre high school student can understand this. Back in the day, we had these like graphing calculators. So do that and the stuff. For the longest time, the best way to take a square root was a slide rule. And it was actually an analog way to do it. But the point of Bitcoin is that this 256 bit elliptic curve protects a trillion dollars or whatever it is. And the assumption is that it's unbreakable.

• 50:14Speaker 01

  But nobody's actually really stress tested that assumption. And I'm not saying nobody, I'll say to me, the mismatch between the amount of trust we place in this and the amount of work that's gone into, you know, breaking it are very, very disparate. Of course, quantum is a big threat to Bitcoin. It's an existential threat, but I think there's other, there's going to be other threats. There's, there's compute threat, there's a math threat. And I think as AI gets better at math. You know, my fee, my P brain can't, can't find the crack for Bitcoin, but maybe OpenAI seven, you know, chat GPT seven will, you know?

• 50:44Auren Hoffman?

  Quite a few years ago, there was a point where you were quote unquote, "the most hated person in the world." What was that like?

• 50:51Speaker 01

  Yeah, I think, I think this was like a social experiment for me to understand the world a little bit better.

• 50:56Auren Hoffman?

  That was the first time I ever heard of you. And all of a sudden, he's the most hated person in the world. Oh, interesting.

• 51:01Speaker 01

  And that's the solution.

• 51:02Auren Hoffman?

  I guess I'm supposed to hate this guy?

• 51:03Speaker 01

  Right, right. And that's the solution in the question is that, and, and exactly what you just muttered is, it was a reflection of, of the people writing that. It's not so much that he is the most hated person in the world. It's, I would like it if he was the most hated person in the world. And the idea of platforming and deplatforming, this concept was never, in my mind, a thing. The answer was people making news that get interviewed, people doing interesting things that get written about. And it turned 180 where it's the, the author, the writer, the editor said, "I can make this issue really important," or not really important.

• 51:34Auren Hoffman?

  Yeah.

• 51:34Speaker 01

  Because reading the New York Times, which I loved reading as a kid, I still like reading it. But I noticed over the years that like in the book section, it went from what is like consensus most interesting book to like, little bit more like social justice, a little bit more, okay, LGBT is really important. Platform those authors and highlight them more aggressively. Well, maybe we should just talk about what the best books are. It doesn't really matter what, what the, you know, the background of the author, if they're a man or a woman, et cetera, et cetera. Those kinds of things became a fever pitch around 2015 when I got, happened to get arrested. I think the two were very related. That was the end of cancel culture. It peaked and has reverted very aggressively, almost too much.

• 52:13Auren Hoffman?

  Probably peaked a few years later, 2022. You were more at the like big start of it.

• 52:18Speaker 01

  Like, I would say that the media trust peaked with me, not because of me or anything.

• 52:24Auren Hoffman?

  Yeah, yeah.

• 52:24Speaker 01

  But sort of peaked with me. It was the last time you could really trust the media. And that started to slide and slide and slide to the point where nobody trusted.

• 52:31Auren Hoffman?

  People haven't trusted for so long, but you're, you're right. Every, every year they've been eroding trust for 30 plus years.

• 52:37Speaker 01

  Every day I walk around the streets of New York and people shake my hand, take pictures. And I'm like, is this what it's supposed to feel like to be hated? Of course not. It's this coterie of people. I, I respect all the opinions, but I'm a hyper capitalist. I believe that. That you should be able to price the product that you make at any price you choose. And there's a lot of people who disagree with that and that's fine. But if you want to use your platform to say, okay, this is the person we all have to hate. Well, we all know people that have been through that. Peter Thiel is somebody that a lot of people hate, especially that class of people, the Gawker Institution and so forth. If he ever does anything, they take it out of context. The Blood Boy thing was taken out of context and, and they're just sort of like, let's make this guy look evil. And it works. They, they, they do a really good job of it.

• 53:16Auren Hoffman?

  There was something interesting where evil people in people's perception seem more competent than good people. Everyone's like, oh, Putin's evil and because of that, he's super competent. For whatever reason, they correlate evil with competence, which I don't think should be. For whatever reason, they, if people think you're evil, they also want to leave you money to invest because they think you're super competent.

• 53:38Speaker 01

  Yeah, there's, there's a little respect that comes with being an evil genius. I, I always liked WWE, believe it or not, the most lowbrow entertainment for an intellectual, I guess. But I love the idea that pathos requires that the bad guy actually has a really hard job. You have to, in essence, get people to hate you, which is not easy. How do you do that efficiently? And you go to a wrestling match or something, oftentimes the crowd's yawning. This guy's not that interesting.

• 54:02Auren Hoffman?

  Yeah.

• 54:02Speaker 01

  Then you get a guy who's really good at revving you up. To me, me and my friends and my family, we all knew I'm not like some bad guy. I, I kind of relished the idea that I could go toe to toe with Hillary. How often in life will Hillary Clinton focus on you? Not that often. And so I was like, this is a great opportunity. She would say something like, this is a disgrace that you're raising the price of this medicine. And I would say something like, you have no power in this matter. Like you don't matter. It's my choice, not yours. And the more polite answer would be something like, well, you know, there's research we have to do and a profit we have to make and this and that. And you know, I'm sorry, you feel that way. Let's sit down and discuss. And of course, the role in that, my role in that are totally disingenuous. You know, I want to tell her to F off. She doesn't give a crap. She just wants to vote. Neither of us are really telling the truth.

• 54:41Auren Hoffman?

  In fact, sometimes people play off each other and it's good for both parties to have this, these debates, just like in wrestling.

• 54:48Speaker 01

  Yeah. In fact, one of the A16Z guys, senior A16Z partner is a huge, huge wrestling fan. He's in the White House now and we sort of bonded over this because there's something about playing that role. I often said, well, people really need somebody to hate. It's almost doing them a service because they're like, I get psychic satisfaction of like, I see this guy and I, you know, I dislike him. Because one of the things about wrestling, as you may know, is eventually the bad guy becomes a good guy. And they switch, they switch seats and the bad guy does something. And the good guy becomes the bad guy often. And you know, they do something unexpected to turn the crowd and they say, as long as you have the attention on you in some ways. And again, I didn't ask for any of this attention. I didn't want it, but I didn't feel it was going away. And I also feel like it was being typecast and like misrepresented. And that's when I started streaming at a time when people didn't stream and I made huge amounts of fans.

• 55:33Speaker 01

  It was so odd, including a lot of people in the valley. I just streamed what I knew, which is like financial statements. And there were tons of people that were watching and they were like, not only is this useful, Brian Peterson, Flexport founder, he's, I learned finance from you. Watching your stuff late at night on, on the internet and, and so many stories. I mean, he's an incredibly sharp guy. Also, I've passed guests on summation. Our investor list is largely entrepreneurs who did really well that, that watched my videos. They said, well, not only is this guy, clearly he's not a bad guy. He's a nerd. Everyone made it out. You're clearly a nerd. Right. I mean, that's like, yeah. There's no hookers. There's no cocaine. There's no, it's just programming and Excel. He's not such a bad guy. And you know, it kind of sounds like he knows what he's doing. I made so many friends and fans out of that. A lot of college students, you know, people that would end up, you know, going to investment banks, things like that.

• 56:18Speaker 01

  I was really grateful that like, social media allows you to have your own megaphone and people have abused that too. Now, two last questions. What is a conspiracy theory that you believe? I am notoriously anti-conspiracy theory. I'm in the so-called camp that it, nothing ever happens. I've been wrong, unfortunately, on my old conspiracy theories. I'm such a contrarian that when people bring up conspiracy theories, I'm just like, nah, yeah, it's not real. Almost like the other way around, sort of a more abstract and general one, but I think that wealthy people have almost no influence in politics. This is something that is almost sort of disprovable, but if you stick with my thread here, I had the fortunate life of being able to become wealthy and get some attention from politicians. I'm almost certain that as soon as you leave the room, it's an old joke on Wall Street too, both parties laugh.

• 57:02Speaker 01

  The politician laughs and says, ha, that stupid rich guy is going to give me money thinking he has influence over me. The rich guy laughs and says, ha, that stupid politician is being bought by me for only X dollars. They, they both think they got over on each other and neither did. There's analogous joke, but

• 57:15Auren Hoffman?

  Well, actually, the the baldish did get something over on the rich person.

• 57:19Speaker 01

  Yeah.

• 57:20Auren Hoffman?

  Of course, the Pauls did have to waste five minutes of their time listening to the rich person to get the small donation that they got.

• 57:26Speaker 01

  I mean, the Koch, the Koch brothers became some of the biggest donors to a home looking party for decades. And they they just threw up their hands and said, this was the biggest waste of time ever. I'm not sure nothing would have changed.

• 57:34Auren Hoffman?

  Oh, I was terrible. I agree with you there, too. I think money does not drive as much as as many other things. Okay, last question we ask all of our guests. What conventional wisdom or advice you think is generally bad advice?

• 57:46Speaker 01

  This, and again, being the contrarian, it's like everything. I have a long list on this. The first one is, and this is pretty common, is people always say, oh, it's who you know, not what you know. It's networking, this and that. I've all invested in our company along with you, and Val says this really well. If you are really good at something, the network will form around you.

• 58:04Auren Hoffman?

  It used to be you'd have these what you knows and no one could find them. So the who you knows would put that them together and they would take a vig on the transaction.

• 58:14Speaker 01

  Yeah.

• 58:14Auren Hoffman?

  Right? In some ways, the who you knows, because they had this monopoly of knowing the what you knows, they became what you knows, too.

• 58:20Speaker 01

  Yes.

• 58:21Auren Hoffman?

  In a weird way.

• 58:22Speaker 01

  And still the case.

• 58:23Auren Hoffman?

  Whereas today, it's much more easier to discover the what you knows. It's much more easy to find them. The who you knows are relatively less powerful than they used to be because other people can find them more quickly.

• 58:37Speaker 01

  This is like, one of the ways to really understand this in in numbers is if you look at the population distribution of the world and you look at some elite organizations. I was looking through the employee list of Millennium and Citadel, two elite hedge funds. And the last names indicate that very large amount of their employees are reflect broader population, which makes sense because genius is somewhat distributed somewhat evenly. And you sort of want, you know, the best people. But if you're sort of a normie sort of place, you're sampling from just, you know, your city, you know, or, you know, the local, local population. Give me the best guys in New York versus give me the best guys in the world. It turns out that, you know, good 50 to 70% of Citadel Millennium are, you know, Asia, East Asian or Southeast Asian or Indian, Chinese, et cetera. And it's it's so stark.

• 59:22Speaker 01

  Take your investment bank. It more reflects demographic of New York, which is, you know, 10 or 20% of those populations, not 70%. It's very much, I think, what you said, which is the world's getting flatter and easier to connect. The other one I would say, I have this Discord channel that is a lot of fun, such a good platform, because it allows things like this. I'd love you to come in for half an hour. It allows conversations.

• 59:42Auren Hoffman?

  I would be honored.

• 59:43Speaker 01

  Amazing. It allows conversations like this to happen, but much more spontaneously, which it can be a little bit hit or miss for that reason. But sometimes we'll have a brilliant physicist, a great programmer, a great, you know, one so, and then sometimes a crazy person. And we're all having this fun discussion. People come in that Discord all the time and say, Martin, I want to get rich. I want to, you know, etc. What's your best advice for me? Answer is always like, I don't know you. You know, so the advice for you, it depends on you. You know, if you're, if you have a ton of ADHD or something, I'm going to tell you to learn how to focus. If you have too much ADHD or too much focus, I'm going to tell you to go touch grass and get to know people and become a more personable person or something like that. Advice is so idiosyncratic to the person that almost all advice is just platitudes and aphorisms that don't really mean

• 1:00:27Speaker 01

  anything. And Buffett's famous advice is protect your reputation really well. Treat every day, what you're doing today is going to be in the front page of New York Times. The people who need that advice are people who are a little unethical and kind of play games with the truth. You don't give that advice to somebody who's like Boy Scout and is super conservative. It really just depends on who the person is. I tell people, I'll get to know you. I'll hang out with you. And after five years, I can give you advice.

• 1:00:50Auren Hoffman?

  There's some probably clear advice that you shouldn't do, but everyone knows those things anyway. Don't shoot yourself up with heroin every day. Well, probably good advice, but everyone already knows that. They don't need to, they don't need to ask you that stuff.

• 1:01:02Speaker 01

  Amazing story. Druckenmiller lost, you know, $2 billion shorting stocks in the, you know, at the Soros, you're sort of part of Soros. He said, everyone asked me, what did you learn from that? He said, not a damn thing. I knew I wasn't supposed to do that.

• 1:01:15Auren Hoffman?

  Right. I knew I shouldn't have done it. I did it anyway. That's us being human. This has been amazing. Thank you, Martin Shkreli, for joining us on Summation. I follow you at Martin Shkreli on X. We're great at X. Heard our listeners to engage you there. Discord as well. This has been a ton of fun. Really appreciate you coming on.

• 1:01:32Speaker 01

  Thank you so much.