YT YouTube2Text

YouTube2Text → Transcripts → Alec Litowitz, Founder of Magnetar Capital and Qstar Capital

Alec Litowitz, Founder of Magnetar Capital and Qstar Capital — Transcript

by The Alpha Exchange with Dean Curnutt · 15,674 words · 2,161 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00A lot of people think about the left
0:01tail and we talked about that already is
0:03I don't care about small losses where I
0:05get feedback back. I learn and I go
0:07redraw. I care about catastrophic
0:09losses. Right? So, but let's think about
0:12the right tail because anybody who's
0:15very familiar with decision-m could say,
0:17Alec, it sounds like you're saying say
0:18fail or fail safe. There's words for
0:20that. But here's where mine differs a
0:23little bit which is that I also believe
0:27that when you're experimenting and this
0:29will be very very well understood by
0:31investors who sit around and might be
0:32listening going but wait a minute you
0:34know power law venture there are certain
0:37times where I don't know the answer and
0:39so an opportunity pops up and well
0:42should I take it should I not and my
0:43answer is really simple if it's not that
0:46expensive meaning it doesn't wipe you
0:48out so I'm out of money I have no more
0:50chance to iterate experiment fermenting
0:52and it's irreversible. You should take
0:54it.
1:00My guest today on the Alpha Exchange is
1:02Alec Lawitz. He is the founder of
1:04Magnetar Capital of Qstar Capital, a
1:07philanthropist and now an author. Alec,
1:10it's great to have you on the podcast
1:11today.
1:12>> I appreciate you having me here. I been
1:14I'm a I'm a fan of the alpha exchange
1:16and I was laughing thinking about what
1:19you called it because I I think I spent
1:2130 years hoarding as much alpha as I
1:24could. Um and here we are having an
1:26alpha exchange. So I hope I make the
1:28transition well.
1:30>> Yeah. Well, listen, we're going to talk
1:31about markets. We're going to talk about
1:33your your book, The Adaptability
1:36Quotient. uh really fascinating and I
1:38would say timely read for an environment
1:42in which uh it would be almost
1:43impossible to say the pace of change is
1:45not remarkable and probably accelerating
1:48from here. So it's uh and I think that's
1:51we'll get into your motivations for
1:52writing the book but I think that's a
1:54big part of it. Let's start with the
1:57investing side of things. So you've been
1:59in this business for 30 odd years. Um,
2:03you founded Magnetar in I want to say
2:052005 or is it 2006?
2:08>> It was 05.
2:10>> 2005.
2:11And um, you know, kind of the
2:13prefinancial crisis period. So things
2:16were about to get absolutely uh haywire
2:19in in markets. As you look back on your
2:22time in markets and you think about the
2:25characteristics that make for truly
2:29differentiated investors,
2:32what do you think that looks like with
2:33the benefit of of so many years of doing
2:35this of of seeing so many people of
2:38experiencing so many different cycles of
2:40markets and volatility? Is there a
2:43thread you would say that kind of ties
2:44together the success of investors in
2:48this business?
2:50Yeah, it's a it's a great question and
2:52it makes me look back and realize I'm
2:54somewhat old because I have that 30
2:56years of having seen lots of different
2:58regimes and and working alongside lots
3:00of incredibly talented people. I think
3:04what I would say is is probably that I
3:06think most people think about this not
3:08quite right. I think a lot of people
3:10would probably have an inkling or or an
3:13intuition that it's an IQ contest and
3:16that people think that the highest IQ
3:19would would win out all the time. And I
3:21think the analogy that I like to think
3:23about is from the time I spent between,
3:27you know, when I was at one of the
3:28original partners at Citadel to founding
3:30Magnetar05,
3:31I used to do iron man triathlons. And
3:34you know, you try as best you can to be
3:36really good at swimming, biking, and
3:37running. And when I started doing open
3:41water swims, you know, the goal was
3:43always your stroke. You know, it's to
3:45think of IQ as like your technique and
3:47how good a swimmer you are and how big
3:49your engine is. And the reality is if
3:51the if the course is straight and the
3:53wind isn't blowing and everything's
3:56perfect and it's a very stable
3:57environment, then probably the person
3:59with the best stroke and the biggest
4:00engine works. That's sort of your IQ.
4:03But I've been in a lot of iron man swims
4:06and half iron man swims in the open
4:08water and it is rarely calm stable.
4:11There's often buoys you have to turn
4:13around. There could be multiple buoys.
4:14The wind and the current could be
4:15moving. And so you're always stuck with
4:17this question of well how do you adapt
4:19to that right? And if the wind and the
4:22current is blowing you have to do
4:24something not natural. You have to lift
4:26your head up. And when you lift your
4:28head up you interrupt your stroke. And
4:30and why don't people like doing that?
4:32because you've been trained to keep your
4:34head down and be efficient and but you
4:37can be incredibly good. You can have an
4:39incredibly high IQ and just swim faster
4:43into the wrong direction, you know, when
4:45the world is changing really quickly. We
4:47you alluded earlier to the fact that
4:49this world is changing faster than maybe
4:51we've ever seen. And so I in my time
4:54over 30 years in different periods, I've
4:56realized IQ alone doesn't cut it. That's
4:59actually not all that matters. What
5:02matters a lot is understanding that
5:05there are different conditions and I
5:07like to think about three types of
5:08conditions that you have to swim in and
5:10not all of them are perfect for someone
5:12who has the highest IQ and so that the
5:14first one is a world that is more
5:16stable. Um I think of that as a world of
5:18risk. I think probably a lot of the
5:20conversations you have on the alpha
5:22exchange rightly so are about the world
5:24of risk and that is a world I I think a
5:27lot of your um listeners are familiar
5:29with where I I describe it as I know the
5:32possibilities and I know the
5:34probabilities and that's a world where
5:35you can do expect expected value mean
5:37variance optimization
5:39sure you have to think about fat tales
5:41and other things but but we can make
5:42bets and play that game and that's a
5:45great environment and that's where IQ
5:47helps a lot um it's calculation
5:50Then there's another
5:53side where I don't know any
5:54possibilities and I don't know any
5:56probabilities and that's black swans.
5:58Nasim Taleb has written some phenomenal
5:59books on that and that is survivability
6:02and convexity. I think the most
6:04important type of environment is the one
6:06that's neither of those and it
6:07ironically is the one that gets the
6:09least written about and it's the one
6:10that we live in almost all the time and
6:12that is a world of uncertainty.
6:15It was sort of well written about by a
6:17gentleman named Frank Knight. Often it's
6:19called Knightian with a K uncertainty.
6:21That's a world where we know the
6:23possibilities but we don't know the
6:25probabilities. And before someone sits
6:27and goes well I've not heard of that and
6:29therefore how much can it apply you know
6:31examples that are easy ones are you know
6:33hire someone I I know that they might
6:35work out they may not could be mediocre.
6:37Like I know the possibilities I don't
6:38know the probabilities. Go meet someone
6:40for the first time. Date someone for the
6:42first time. It could go great. It could
6:43go terrible. Maybe it's okay. If I know
6:45the possibilities, I don't know the
6:46probabilities. A lot of life is that
6:48way. And a lot of markets are that way.
6:50And I think they become the most crucial
6:53times for the markets because a lot of
6:55the money that gets made by the best
6:58type of investors are ones that can
7:00navigate all three of those
7:01environments. And within uncertainty,
7:04the job is to resolve it. And in my
7:07career, while sometimes in a in a world
7:10of risk where the market has a
7:12distribution, the market has
7:14probabilities, I just happen to think
7:15there are different ones. Typically, the
7:17edge is smaller, there's a lot of people
7:20looking at it. Um, I think the great
7:23investors are the ones that can sit and
7:24say in a I know when the world is like a
7:27risk world and I can survive and and and
7:30prosper in that by having better
7:32estimations than others. I want to
7:35prepare for a black swan and so I need
7:37to survive and have some convexity and
7:39understand where my model may be wrong
7:41and I need to be able to survive those
7:43kind of outlier situations but also and
7:45more importantly the biggest money gets
7:48made when there's periods of uncertainty
7:50where something's changing and industry
7:53um you know I could you know maybe we'll
7:54talk about it later but you know energy
7:56when hydraulic fracking and horizontal
7:58drilling came along rapidly changed
8:00we've seen the production function of
8:03lending change in the great financial
8:04crisis. Now we're seeing AI. When you
8:06get to these periods and you have
8:08uncertainty, the big money comes from
8:10resolving the uncertainty, going through
8:12a process and knowing how to take that
8:15uncertainty, ask the right questions,
8:17test appropriately, get feedback loops
8:20to make better decisions. So my end
8:21answer to you is that it's not just
8:23strictly IQ. It's the ability to survive
8:26and adapt during moments of change that
8:29we see as an investor in the market very
8:32often and capitalize on those situations
8:36because often the conversion of
8:38uncertainty into risk is where the
8:39biggest biggest money gets made in my
8:41opinion. So great if you can make money
8:43in risk, great if you can survive black
8:44swans, but really great if you can do
8:47all three and prosper when everybody
8:50else is wondering what's going on and
8:51you can actually resolve what's going
8:53on. If you go back to let's say the
8:56global financial crisis because it's in
8:58some ways the biggest test of
9:00survivability you know for banks for
9:02investors we saw um you know such in
9:07incredible consequence to potentially
9:10mising your exposures um you know
9:12[clears throat] to being in the wrong
9:14instruments. Um so those were
9:17you know careerending in some ways or
9:19firm ending events and yet and perhaps
9:23you know we look back on these times
9:25with more clarity than they really had
9:27but by March09
9:30you know the market turns right and so
9:33my question is around the folks that
9:36might have gotten that risk event right
9:39going into it who saw perhaps the
9:43systemic uncertainty the systemic risk
9:45of the banks. Um, and then being able to
9:49ultimately pivot towards something and
9:51maybe it's not exactly March09, but
9:54something where you finally see the
9:56government is pulling is putting such
9:59weight and capital behind this thing.
10:01Um, where failure is not an option and
10:04that it's ultimately going to pivot. Is
10:06that some example of kind of this
10:09complex pattern recognition or seeing
10:12things change from your perspective? I'm
10:14just trying to kind of understand it a
10:16little bit better.
10:17>> Sure. I think it's a very good question
10:20and I think it goes to you know what is
10:22AQ right or adaptability quotient in my
10:25eyes and uh I I think what I often think
10:30of as the most important part of AQ uh
10:33even though it's got sort of multiple
10:35pieces to it three pieces in particular
10:36three phases I call it one of the most
10:38important ones is is the first one which
10:40is metacognition so so very often people
10:45forge their identity with a thesis
10:48And especially if you're right, so you
10:50go into the crisis, you have a different
10:53opinion, you feel like you're
10:55investigating and getting feedback loops
10:58from reality, and you just are sitting
10:59around going, I don't think people see
11:00what I see. It's incredible. Like, I
11:02have these data points. I'm testing it,
11:04probing, and nobody sees what I see, and
11:06you turn out to be right. It is very
11:08easy at that point to make that your
11:11identity. That your identity was you saw
11:13this problem, and and you were right.
11:16And what AQ teaches you to do is to have
11:19strong opinions weakly held. You can
11:22have conviction, but how you hold it has
11:24to be weak and it has to be provisional
11:26at all times. So it is it is this
11:29razor's edge as you know being a great
11:30trader is sort of like that confidence
11:32and humility. Like I'm confident enough
11:34to put on a trade. I'm I have enough
11:36humility to know constantly probing
11:38where am I wrong? Where am I wrong?
11:39Where am I wrong? I'm not trying to
11:41prove that I'm right. I'm trying to
11:42prove I'm wrong. So the smartest people
11:44were able to sit and go, I have data
11:46points. I see what's happening and then
11:49constantly in a state of revising based
11:51on new feedback loops. I I I have a
11:54phrase in my book that I say it's better
11:56to make decisions right than make the
11:58right decision. Right? So at any given
12:00moment, I'm not trying to have made the
12:03right answer. I don't care what my
12:04answer is. I care about my process. My
12:07identity is tied to my process of of
12:09adaptability. I'm in a constant state of
12:12wanting to update. If the data says no,
12:15then I don't update. But I'm not in a
12:17constant state of wanting to be right.
12:19You nobody's ever right. I like the
12:20Volta Voltater quote, which is
12:25um you know, uncertainty is challenging,
12:27but certainty is absurd, right? So at
12:30the end of the day, I'm not I I think
12:32the the great investors back to that and
12:33what people who played both sides, you
12:36know, did it well going in, survived and
12:38come out are people that are in a
12:40constant state of questioning and even
12:41when they get it right, they know that
12:44is a provisional I just got something
12:45right that moment in that environment,
12:47but if the wind changes or if the
12:49current changes, I'm in a constant level
12:51of going, okay, is my is my new model
12:55right or do I have to go and update that
12:58model that just won that last
13:00environment. The environment is changing
13:01again. And I I'm going to be honest, I
13:04think that's why this world is really
13:05difficult. We're undergoing maybe we'll
13:07talk about later multiple changes that
13:10are epic. Um you know, I talk about that
13:12in the book. I you know later maybe
13:14we'll talk you know if you want about
13:16both the fourth industrial revolution
13:18and the second cognitive revolution I
13:20talk about it. So there's so much
13:22change. It's a very very good question.
13:24How do you how do you constantly get it
13:26right? And the answer is you won't. But
13:28you can you can put yourself in a
13:30position to be more right than not. Luck
13:32is infatuated with the efficient and the
13:34efficient people that are great
13:35investors are in the constant state of
13:37operating in a certain way. And unlike
13:39IQ or EQ, I think AQ is learnable. I
13:42think it's a learnable skill. And the
13:44learnable skill is know when your model
13:47no longer is touching reality properly.
13:49Know how to update your model and act
13:52before the market forces an action on
13:54you. That's that's what I think the key
13:55to a great investor is. Well, as I read
13:57the book, uh, this strong opinion weekly
14:00held, I said this to myself so many
14:03times, s
14:07it it, at least at face value, it feels
14:10a little contradictory. And I wanted you
14:12to kind of dive into it a little bit
14:14more because if I have a strong opinion,
14:16I'm convicted. It feels to me like I
14:19shouldn't have the ability to let that
14:22go very quickly. that my prior is strong
14:25and that my updating is kind of up
14:27against that strong prior. Just kind of
14:30g let's explore that a little bit more.
14:32It is a challenge, right? Um uh and uh I
14:36understand the language. My my I'm the
14:38son of two psychoanalysts, but my mother
14:40who before she became a psychoanalyst
14:42was a linguist. So I'm pretty careful
14:44about language and I I could see the
14:47concept of like aren't these against
14:50each other? So let's let's sort of
14:51unpack it in a way. I think hopefully
14:53it'll help you understand um and that is
14:55that so if you go through my process and
14:58maybe it'll help to just to for for
15:00listeners to understand the three of
15:02them because that'll tell you like where
15:04does a strong opinion come in and where
15:07do you start to test it and so the first
15:10phase of of in my book of of
15:12adaptability of building your
15:14adaptability quotient is metacognition
15:17and and that's important for a variety
15:19of reasons. A lot of people's decision
15:21processes don't start with that. But the
15:23reality is if you're examining a system,
15:25if you're examining the facts on the
15:27field, reality, etc. You're a system.
15:30And so if you think you're seeing
15:33reality, nobody sees reality. You bring
15:35biases. You have your own process. You
15:36have your, you know, the functioning of
15:38how a brain works. I go a little bit
15:39into the book without going too much.
15:41And the first important thing is what I
15:44call adaptive optics, which is you have
15:46to understand your lens and how when you
15:48see facts, you're seeing them with a
15:50bias. So if you optimize for facts that
15:53don't correlate with reality, you could
15:55be perfect. You you'd have a great
15:57swimmer. You're just going in the wrong
15:58direction. So the first part is
16:01clear your mind. Don't walk in with a
16:03bias. It's a bit of a beginner's mind
16:05type of an environment. The second phase
16:07is simulation because if the world's
16:10changing, you're seeing these facts and
16:11your old model's sort of not you're sort
16:13of, you know, we're all feeling this
16:15right now. Wow, things are changing.
16:16What does that mean? Is that going in
16:18this direction? It's very easy to feel
16:20confused. And the first thing you have
16:22to do actually is put aside your old
16:24model and turn around and simulate a
16:27bunch of possibilities. And this is
16:29where we get to strong opinion weekly
16:30held which is the goal of the end of the
16:32second phase is to you know if the first
16:35part of it was what's your relationship
16:37to yourself the second part is what's
16:40your relationship to possibility what
16:42are the possibilities what could
16:43possibly explain the things that I'm
16:45seeing and you don't get to leave things
16:46out you don't get to come up with a
16:48theory that explains two of the four
16:50things you're saying it's got to explain
16:51them all and there's a there's a word
16:53for that which is abduction which is a
16:56lot of people in venture and other
16:57places talk about first principles. You
16:59break things down. You start noticing
17:00patterns and you come up with what a
17:02strong opinion weekly held is. It's it's
17:05a um induction, you know, to the best
17:08possible answer. And so the end of phase
17:10two is not certainty. It is just given
17:13what I'm seeing, what is my best best
17:17hypothesis. Okay? So the reason why you
17:20hold it strongly is because you're going
17:22to go test it. It's just the best among
17:24the possible
17:27answers that you think you have, but
17:29it's provisional. That's the point of
17:31phase three. So the strong part is that
17:34you are willing to go from there to
17:37actually experimenting. The weak part is
17:40I'm not tied to it. It's the best
17:42version of what I could see as an
17:44explanation, but I'm wedded to this
17:46loop, not to the answer that the loop
17:49has in this given moment. And so the
17:51strong opinion is I I it was, you know,
17:53I could have called it the strong opin
17:55strongest opinion. It's the strongest of
17:57the ones that I have. Now I'm going to
17:59go hold it weekly when I go test it. And
18:01phase three is I'm going to go clash it
18:04with reality. I'm going to go
18:05experiment. This in venture world is a
18:08minimally viable product. An AV test.
18:10Let me figure and Musk is testing
18:12spaceships, right? So it's I'm going to
18:14go see if this theory holds, right? I
18:16can have um you know heat shield and I
18:19can have places where there is none,
18:21places made of metal, places made of a
18:23carbon fiber composite and I'm going to
18:24see which which one works best. I have a
18:26theory, but I'm going to go test. And
18:28then you're not done at phase three
18:30because you get feedback. And when you
18:31get that feedback, you go back in revise
18:33your hypothesis. That one that was held
18:35strongly, you revise it instantly and
18:38now it's stronger and you go test it.
18:40And so it's just this loop. And all it
18:42means is in phase two, you know, there's
18:45a trade-off between exploring and
18:47exploiting, which we can go into. I I'll
18:49leave that if you want to go into it,
18:50but ultimately the goal of phase two is
18:52to come up with the strong opinion
18:54weekly held. And it means that you have
18:57to have it. It is the best one you've
18:58got, but you're not wedded to it. What
19:01you're wedded to is experimenting and
19:03testing it and being open to whatever
19:06your feedback is. You don't get to
19:07decide the feedback. The world decides
19:09the feedback.
19:10>> Yeah. And we'll talk about this in the
19:12context of markets and some of the
19:14investing that you've done. But one of
19:16the tensions that you describe in the
19:18book is the resisting the temptation to
19:22decide too early. Uh that uh this
19:26beginner's mind can be an advantage.
19:28Sometimes expertise is almost a
19:30disadvantage. So going in there and just
19:32with a kind of beginner sense of
19:35exploration, but allowing yourself to
19:38take it all in without necessarily
19:40deciding too quickly. There's got to be
19:42some tension there between that, which
19:44is time's on your side, versus the
19:46world's moving fast, maybe your
19:48competitors are moving fast. Maybe just
19:51a a little bit on that. I know we'll
19:52we'll touch on that in the context of
19:53some of the specific investing you've
19:55done. you're highlighting very well the
19:58tensions that are purposely present in
20:01the world and they're present of course
20:02in any design uh of a decision-making
20:05apparatus and so it's it's good to
20:07highlight them. Let me analogize for the
20:09listeners because let's take something
20:11abstract and make it really really
20:13concrete. Um when people are going out
20:15to a restaurant wherever they live or
20:17they're going to go on a trip if they're
20:18going to travel somewhere they're in a
20:20constant state of this struggle between
20:22explore and exploit. Okay, it's this
20:25question of, well, I heard there's some
20:26new restaurants. Should we go to that
20:29new one, but we have the ones we love.
20:31Um, so I go look at the new ones listed
20:33in a magazine. I'm like, maybe that
20:34would be a good one, but is as good as
20:36the one that we love the best. So,
20:37you're like exploring. And then at some
20:40point, you make a decision. We're going
20:41to either go to the one we love or we'll
20:42try the new one. That's exploiting. Same
20:44thing when you travel. Should we stay at
20:46should we go to this city or that? Do we
20:47stay at this hotel or that? most days
20:50people are deciding you know do I keep
20:52researching or do I stop and exploit
20:54something so there's a tension that
20:56exists and for people like me that were
20:59born into a family of like peel the
21:00onion peel the onion peel the onion a
21:03little bit more intellectual there's
21:05this you know and if you're riskaverse
21:07there's this desire to just keep peeling
21:09the onion back like why why exploit why
21:11unless someone's putting a gun to your
21:12head just keep keep researching keep
21:14researching but you're absolutely right
21:15that you know to be a great investor and
21:18to be in the to be a founder, you
21:20realize there's pressure. There's
21:21capital, there's time, there's
21:23competitors. The world's not stopped and
21:25moving. And so there's a function and a
21:28way that I get past that and I talk
21:31about it in the book. And the way to
21:32think about that is and and this is the
21:35release valve of it, which is listen,
21:38what you're trying to do when you come
21:40up with a strong opinion weekly held is
21:42to avoid when you go experiment things
21:45that stop you from learning effectively
21:47that kill you, right? that that are
21:48knockouts or maybe for your audience,
21:51you know, you want to avoid a very big
21:52left tail. But not all failures are big
21:56left tails. So what I try to do is think
21:59about what paths that I could go test
22:02when I go test it would eliminate my
22:04ability to continue testing it. Those
22:06are not viable because I lose my chance
22:09to resolve the uncertainty through
22:11learning and learning and learning. And
22:13I call those typically and the worst
22:15type of errors are, you know, without
22:16going too much detail, type two errors
22:18where I don't think there's a signal,
22:20but there is. Um, you know, that's
22:22Blockbuster saying streaming, yeah, I
22:24don't think it's that big a deal. Um,
22:26those things kill you. Type one errors
22:28where, you know, which is sort of the
22:30the smoke signal goes off, but there
22:31isn't a fire. That costs me money. It
22:34costs me time to do the experiment, but
22:35I get to go back and relearn and revise.
22:38So the way that that the the the
22:40insight, the answer as to where you stop
22:42and where you stop exploring and going
22:44and exploit and experiment with whatever
22:47the best version is at that time is you
22:50try with a great partner. Play devil's
22:51advocate. Have a steelman army. Talk
22:53about things with people. Think through
22:55things. Come up with the best answer you
22:56can that doesn't kill you when you go
22:59experiment. And once you've done the
23:01best job you can of avoiding the really
23:03really bad left tails, making other
23:06errors, those aren't errors. Those
23:08aren't failures. Those are just feedback
23:09loops. They make you better, stronger,
23:11and get you closer to the answer. So,
23:14I'm completely fine. You know, a lot of
23:16people in the market in the investing
23:17will know this as like making smaller
23:19investments, probing, you know, those
23:21are like probes. It makes me focus more.
23:23I go talk to the company, etc. So, think
23:26of this as, you know, there is this
23:29tradeoff. There is reality that banks
23:31and you have to go execute. the time to
23:33go down and actually start experimenting
23:35is when you've effectively avoided the
23:38most disastrous ones if you can and
23:41allow yourself to make other mistakes
23:43because that is learning. That's
23:44actually what happens. So that that
23:46that's how I think about it.
23:47>> And so it seems to me that type one
23:49versus type two at least in markets
23:51maybe it's in in other in life in
23:54general is is about sizing in some ways,
23:57right? the the type two error is really
24:00going to come down to some version of
24:02catastrophic loss that maybe comes from
24:05sizing being too big. It sounds to me
24:07like you're a big proponent of again
24:10experimenting with maybe little
24:12allocations to different strategies,
24:15just trying things out to try to get
24:16them up and running so that they can
24:18give you feedback. Is that a way to
24:20framework?
24:21>> No, I think that that's a very fair
24:22assessment. I mean, you're you're 100%
24:24right, at least in my mind, on the right
24:25track. Think about it this way. Remember
24:27our three versions. If you're in a world
24:29of uncertainty, you don't know
24:30probabilities. If let's do a poker
24:32example for lots of people probably on
24:34the show or listening to the show. So,
24:36you know, if you go poker in my mind is
24:39is is more of a world of risk. Other
24:41than if someone's bluffing, that's 19
24:44uncertainty. I don't know how you bluff.
24:46If you go play someone new, are you
24:48going to bet everything on probabilities
24:50and play or are you going to play GTO?
24:52Because you don't you have to resolve
24:54the uncertainty. You sit, play, and then
24:56begin to get tells. You begin to know
24:57how the person bluffs. Then you scale up
25:00the trading and investing. When you get
25:02that answer, remember I think the world
25:05is more uncertain, especially a world
25:06like we're in right now. So, I'm not
25:08going to make huge bets, which might cut
25:11off my learning when I haven't resolved
25:13the probabilities. We can get into
25:15details. someone on, you know, some of
25:17your listeners might say there's
25:18something called subjective bays where
25:20you throw in probabilities, but I I talk
25:22about it in my book and I can just tell
25:24you that there's dangers of doing that
25:26in many ways. And so my answer back to
25:29people is that, you know, you go in, you
25:32do these probes, you resolve that
25:34information, and over time you start to
25:36get clarity around the probabilities.
25:38That's when you make your bigger bets,
25:40when you're getting, you know, true big
25:42answers. It's back to, you know, the
25:44equivalent again of like product market
25:46fit. I'm starting to see the answer. I'm
25:47starting to get feedback loops. Okay,
25:49now I deploy a lot of capital to scale.
25:51It's really that same process um
25:55at large. There's a statement you make
25:57in in the book and I I took so many
25:59notes in this book because I thought
26:00there was some just brilliant
26:03insights here. Um this so I want you to
26:06just reflect on this since you wrote it.
26:10If pattern recognition is the seed of
26:12human intelligence, pattern editing may
26:15be its highest expression. And I think
26:17this uh you know kind of ties back to
26:19understanding when the maybe the rules
26:22have changed uh when the ground
26:24underneath us is is shifting. I'd just
26:27love to get a little bit deeper on on
26:30that as you you know that's a I think a
26:32big part of why you wrote the book and
26:34sort of what AQ is all about. I did
26:37write it, by the way, and and and
26:38whether for better or worse, I wrote it
26:40before AI really came along. So, that
26:42wasn't written with AI at all. Um maybe
26:45it would have been a lot faster process
26:46if I had had AI, but I I didn't get to
26:49do that. So, um you're literally right
26:51that I did write it. The point that I
26:54was making there about editing is really
26:58goes back to whether and again we could
27:00apply it to investing but I think it
27:02applies to life broader which is that
27:04the most successful people they don't
27:07just learn faster they let go faster
27:10they're willing to edit their thoughts
27:11faster. I I again back to the quote from
27:15another quote from my book it's better
27:16to make decisions right than make the
27:18right decision. Right? I don't care if
27:20I'm wrong. If I have a group of people
27:22that I hire and I could build a firm
27:24that the people around the table help me
27:27edit constantly, give me ideas I didn't
27:30have, come up with answers that I didn't
27:31have, that I've done a really good job
27:33of hiring people. I don't care if I'm
27:35the one with the answer. Why would I
27:36care about that? The goal here is to is
27:39is to be closer to reality, to be able
27:41to make some reasonable level of
27:44forecast. I say forecast, not
27:45prediction. A prediction is a certainty.
27:47It's this will happen. forecast is
27:49there's a range of outcomes. As you and
27:50I both know, the best investors don't
27:52sit with spot forecasts. They think
27:54about a range in a distribution. And
27:56even if they can't get it spot forecast,
27:58they think about the shape of the
28:00distribution. Our entire prior
28:01conversation was don't think about the
28:04shape of the distribution and don't get
28:06caught in the left tail while you're
28:07revising uncertainty and and getting
28:09probabilities. Same thing's true here is
28:11is you know, it's one thing to go
28:13through a process and humans don't like
28:16uncertainty. people don't like it. And
28:19I'm, you know, one of the things in my
28:21book that I'm trying to explain to
28:22people is uncertainty is the greatest of
28:25things. And you know, if if for those
28:28that, you know, are scientific bent that
28:30are listening, you know, there's a the
28:32father of uncert of information theory,
28:34Claude Shannon, who said that if I tell
28:36you something you already know, you
28:38learn nothing. The most important and
28:40valuable thing is the thing that's
28:42orthogonal and completely different.
28:44It's the one with the most potential
28:46value to you. You don't have to agree
28:47with all the thing that is different,
28:49the otherness. But the most potential
28:51for you to learn is the complete
28:53opposite and is otherness. So when I say
28:55that, you know, it's one thing, you
28:57know, to have a pattern, but sticking
28:59with it without the willingness to edit
29:01it, is to sit and say, I know I've got
29:04it. And and in a world that we're in
29:07right now, that is possibly the most
29:09dangerous thing you can do. It's just
29:12being able to edit yourself. the
29:13willingness to do it, the willingness to
29:15separate your identity from being right.
29:18I keep trying I think I say it in the
29:20book like make your identity that you're
29:22an adapter. You know that that's what
29:25you know I I I have a a section in the
29:27book where I talk about resistors
29:28adopters and adapters. You know in this
29:30world of extreme change resistors sit
29:33and go well this is a bubble I think
29:35it'll pass and I'm going to just stick
29:37with my old model. Adopters and that
29:39sounds good. adopters sit around and go
29:42hm this is something big I'm going to
29:44adopt the new technology
29:47but I like to say that AI is an example
29:49of the biggest maybe of these changes AI
29:52is a tool at the interface but it's a
29:55regime change at the system level
29:57adapters not only adopt the tool they
29:59sit step back and go what does it mean
30:02at the system level what does it mean
30:04that the cost of knowledge is going to
30:05zero what does this change and they
30:08begin to see the world differently and
30:10they're willing to edit their prior
30:11opinions and prior positions because
30:13they see it at a bigger level, at a more
30:16abstract level. I get it. I'm going to
30:18use this tool. It's an incredible tool.
30:20But what does it mean for me, for
30:22society, for this position, for their
30:23investment? And you know, if I said to
30:25you in the past, you know, let's take
30:28the following things, right? Let's take
30:30real estate. Let's take um energy and
30:33power and utilities. Let's take
30:36semiconductors. Uh let's take um
30:39software I I might have software debt. I
30:42might have said these are these are
30:43unconnected. I got a pretty diversified
30:45portfolio. They're all correlated now to
30:48one thesis to one hypothesis. The
30:50artifact was you know we we can talk all
30:53we want and I could talk extensively
30:55about my history of portfolio
30:56construction risk management in a
30:58different world. That world isn't this
31:01world. So maybe now I need to diversify
31:04by hypothesis because it means something
31:08different that that was the artifact of
31:10you know the current if you use the
31:12current set of tools to do risk
31:13management from the prior world you have
31:16the artifact of diversification you
31:18don't have the function the function's
31:20different right you need the function of
31:22it not the artifact and so if I sit with
31:25my prior opinion and I don't edit it I
31:27have the artif I have the artifact of
31:29the prior opinion I don't have the
31:31function of the updating to well what do
31:33I do now what is current now what
31:35matters now what's right now and in the
31:37markets as we know that's all that
31:39matters and I think it's true in life
31:41and I I think we're on the precipice of
31:42some you know maybe we'll talk about it
31:44later some you know part of the reason
31:45for writing the book is we're on the
31:46precipice of this mattering way more
31:50than markets in my opinion
31:52>> as you talk about updating priors and
31:54and editing based on the flow of
31:56information coming your way I can't help
31:58but think of our our mutual friend Ross
32:00Ross Stevens I I had a chance to have
32:03him at my macro minds conference and we
32:05talked a lot about his background. the
32:06three of us are all University of
32:08Chicago uh folks and his deep background
32:11in basian statistics and this and you
32:14talk about him a little bit in in the
32:15book of you know being really patient
32:18approaching things with this beginner's
32:19mind but then allowing
32:22you know kind of being as Ross says a
32:24little less wrong each day trying to get
32:26yourself a little less wrong and is that
32:29I mean is is a large part of AQ linked
32:32to
32:33a craft like basian statistics in your
32:36you I consider myself a pretty basian
32:39person and I and and so I I'd say let me
32:42say two things um and they're not at
32:44odds. The first is that I love being a
32:47being a basian thinker, but my whole
32:50conjecture is it doesn't work in a world
32:52of uncertainty because there are no
32:54probabilities. And so you can have a
32:56basian type approach, but you have to
32:59change what you do. If you're making
33:00decisions in a world of risk and you go
33:02to a world that becomes uncertain and
33:04you use basian logic, it doesn't work
33:06because there are no probabilities. So
33:08the key then becomes first, how do you
33:11get the probabilities without blowing
33:13yourself up? How do I resolve them? Once
33:15I get them, I can be really basian. I
33:17can optimize like a basian, but I have
33:19to get them first. And black swan says
33:22you're never going to get them. It's
33:23unresolvable. Some uncertainty is
33:25resolvable. Some isn't. So being basian
33:28is I'm trying to get there. I want to
33:31convert the uncertainty into
33:33probabilities and then I can go with
33:34things. But there are times where that's
33:36harder. what I was referencing Ross and
33:39I, you know, close close friend of mine
33:41and I have incredible admiration,
33:43respect for what he's built and how his
33:44mind works and otherwise it's, you know,
33:46I consider him to be very high AQ. Um,
33:49and one of the things that he does that
33:52I talk a little bit about in the book is
33:54not just his sort of patience um, and
33:57trying to get a little better, which I
33:59know is true. one of the things he does
34:01as well as anybody and frankly I had to
34:04learn this lesson uh the hard way a long
34:07time ago and that is that you know
34:09sometimes you're trying to resolve
34:11uncertainty and you're doing it
34:12appropriately you're running through the
34:13way I explain to people in the book that
34:15I think is right although I'm adaptive
34:17and so someone on the someone listening
34:19might make me better at my own process
34:22um but one of the things that happens is
34:23sometimes you bang up and nothing's
34:25getting resolved right you're just not
34:28getting any feedback loops back. And one
34:31of the things Ross does really well is
34:33he he checks his experiment. He checks
34:35his thinking. Am I testing it the right
34:37way, but there are times where you're
34:39just not getting the feedback loop back.
34:42And a lot of times at that point, people
34:44have this desire like sunk cost like,
34:46well, I'm already into this. I've spent
34:48a lot of time on it. Let me keep banging
34:50up against the wall. But sometimes you
34:52have to sit and go, I should probably
34:53move on to another problem. I'm not
34:54gonna go build a business around that, a
34:57trade around it because I can't. Maybe
34:59someone else can't. I can't resolve that
35:01uncertainty. And instead of treating it
35:03like risk and going and making a bet, I
35:05have to be comfortable with the fact
35:06that I'm going to leave that one. I
35:08can't resolve it. I'm going to move on
35:10to another place where I see an
35:12uncertainty and go try to resolve it
35:13because I'm not getting any feedback
35:15loops here. that ability to to you know
35:18I I sometimes when I'm interviewing
35:19people I say is is you know is cleaner
35:22version of is quitting a copout or is it
35:24is it a skill right which is it you know
35:27there is a right way to quit and a
35:29reason to stop when you just can't
35:31resolve something and you have no
35:32business being in that trade or in that
35:34business and then you move on and try to
35:36find something that is and that that's a
35:38different kind of patience so that kind
35:40of patience is I'm going to I'm going to
35:42wait and resolve it and improve and get
35:43better but it also is the patience to
35:45sit sit and go and the wisdom to sit and
35:47go, "This time I'm not getting anywhere
35:50and I'm going to cut that capital off
35:52and go deploy new capital somewhere else
35:54rather than trying to run down a dry
35:55hole." Well, let's go back a ways. We're
35:58going to go back to your uh early days
36:00at Citadel. You're a freshly minted JB
36:04MBA, I think, from U of Chicago. and uh
36:06you you landed Citadel in the very very
36:09early days and you very quickly
36:12take responsibility for the risk our
36:15business and um I'd love to just you
36:18talk a fair amount about this in the
36:20book in terms of you know the the nature
36:22of that business um it's obviously got a
36:24lot of asymmetry there's information
36:27aspects to it but it is hard to scale
36:30very difficult to scale almost by
36:32definition and so you really approached
36:34it differently I'd love for you to kind
36:36of take us through that and then of
36:38course how it relates to this adaptive u
36:42thought process.
36:43>> Yeah, I I I always laugh when I think
36:46back to that time because I was pretty
36:48friendly minted. I had spent six months
36:50at JP Morgan doing investment banking
36:52and um that was which which is a great
36:54business just wasn't right for me. Um I
36:56I need very fast feedback loops and and
36:59uh I went to work at uh what was going
37:02to be called Citadel. We we it was
37:03called Wellington at the time. We had a
37:05name the firm contest and Citadel won it
37:07that year. But when I went in there,
37:10what Ken was really doing to be honest
37:12cuz you know I didn't know what riskar
37:14was, never traded a day in my life. I
37:16mean I was pretty pretty raw. I got
37:18there in February of 94. Ken was worried
37:20about he very correctly called that
37:22there would be an M&A boom and he was
37:24worried at the time about convertible
37:26bond arbitrage and cash takeovers.
37:28Right? So at the time, there are some
37:30provisions that are different now, but
37:32at the time your short would get hurt
37:33and then your conversion would collapse
37:35and you would lose a lot of money. And
37:37so he said to me in February of 94,
37:39right when I started, you know, by
37:40April, you have to be in 10 risk deals.
37:42What do I know? I don't know. I don't
37:44know that much about this business or I
37:45didn't know anything. So long-waited
37:47story, I start researching and really
37:49thinking I had no priors. I didn't have
37:52any there was very there was not a book
37:54around. There was a pamphlet. I could
37:55talk to some Wall Street people and
37:56there were some very very two very very
37:59helpful very kind people to me that I
38:00acknowledge in the end of my book from
38:02Bear Sterns who helped me a lot but when
38:05I was looking at the industry I saw the
38:07following I went back did a little
38:09research tried to investigate the facts
38:12on the field and I found out that 92% of
38:14all risk deals go through well that's
38:16pretty good so you know if I just play
38:18all of them you know 92% go through the
38:21market treats the deals like 87% % go
38:25through generally. So if I play all the
38:28deals and appropriately size them, then
38:29there's like an extra 5% edge in there
38:32because the marketplace is paying you
38:34for being a liquidity provider. The long
38:35only people sell because they don't know
38:37how to assess the deals going through or
38:38not and you can play them all. And
38:39that's where a lot of people might do
38:41that or um you decide which trades. I
38:44started I did something different. This
38:46is where sort of like well you know how
38:49do I analyze this different than other
38:51and I looked at the deals and tried to
38:53figure out well which deals break I mean
38:55some break so which ones break and it
38:57turned out deals break for two real
38:59reasons one is financing the deal just
39:01falls apart because they couldn't get
39:02their financing and the other was
39:04regulatory typically antitrust and I
39:07realized well when markets fall apart
39:09for financing it could be idiosyncratic
39:11it could be when the market bumps right
39:12when the market fails and the capital
39:14markets close up and the financing I'm
39:16not too I don't think I'm going to fresh
39:17out of this law, you know, JDMBA to
39:20predict that. But my favorite class in
39:21law school was antitrust. That was my
39:23favorite class. Great professor. And
39:26what I did was I said, "Well, let me let
39:28me think about that problem because
39:30here's the thing about antitrust. When a
39:32deal gets announced, it goes to the FTC
39:34or the DOJ. They're not experts on these
39:36industries per se. So what do they do?
39:38They call up customers, competitors.
39:40They have to do an analysis of the
39:41industry and figure out whether the deal
39:42is anti-competitive or not." I can call
39:45people. I mean, I love antitrust. I kind
39:48of know the right questions to ask. So,
39:50I started thinking, well, wait a minute.
39:52Why don't I become a specialist in the
39:53regulatory stuff and the complex deals?
39:55I'm not going to necessarily play
39:56everyone because unlike the market where
39:59a deal is announced, it trades at a
40:01certain level and the market has a
40:03probability. So, that's a game of risk,
40:05right? I know the possibilities, I know.
40:07I went and said, well, where's the
40:08uncertainty? The uncertainty is in two
40:10places. Pricing when the deals blow up
40:12because of financing. pricing when the
40:14regulatory gets complicated. I can't
40:16resolve one. I I think I might be able
40:18to resolve the other. So, I went into
40:19there and said, "Okay, I'm going to how
40:21I'm going to do this process. I'm going
40:22to follow these comp companies when
40:24they're going through through this
40:25complex regulatory process." And then
40:27getting to the question that you asked
40:29around scaling that, I did something at
40:31the time which was, you know, pretty
40:33novel at the time. Um, and that was that
40:35now everybody knows there's like GLG and
40:37there's all these expert networks, but I
40:40was thinking about the production line.
40:41A a deal gets announced. It could be
40:43across all these industries. I have all
40:45these analysts. They're supposed to get
40:46up to speed on the deal. I've taught
40:48them riskb you know get on the
40:49conference call, read the merger
40:50agreement, think, etc. And the question
40:53was, well, how do I if a deal gets
40:56announced in a particular sector, you
40:58know, as an example, you know, a classic
41:00example would be, you know, when when
41:02Boeing was buying McDonald Douglas, you
41:05know, there was there were only three
41:06widebody aircraft manufacturers in the
41:08world. Those two and Airbus usually at 3
41:10to2 the deal breaks. So, was it going to
41:12go through or was it not? My analyst is
41:14not an expert on that. So, what I wound
41:16up doing is I wound up building an
41:18expert network captive to Citadel. I had
41:22hundreds and hundreds of people. In
41:24fact, I had I hired a woman one time to
41:27do some research for me on a bank merger
41:29in Florida. She was so good I hired her
41:31full-time and her job was when a deal
41:33got announced to line up the right
41:36consultant so that we could get up to
41:37speed on the industry and start getting
41:39in touch with customers, competitors,
41:40etc. So what we wound up doing is we
41:43built at scale effectively our own
41:45consultant network captive to to
41:47Citadel. What wound up happening is that
41:49I was, you know, we became pretty pretty
41:51good at at, you know, people always
41:52thought, well, why do why do you guys,
41:54you know, we had billion-dollar
41:55positions in the 90s. Those are pretty
41:57big positions in Riskarb in the 90s. And
42:00I, you know, I lost money in four of my
42:03first, 1100 trades. And the answer, you
42:06know, which is, I'm not, it's not a
42:07gloating thing. It's a process at work.
42:09It was just, wow, we were pretty good at
42:11resolving that uncertainty. And you know
42:15there's some great stories around that
42:16of like well once you think about well
42:19yeah great you can make those phone
42:20calls but why is somebody going to
42:22answer the phone and so thinking and
42:24really really rigorously around how do
42:26we build a network that we can get a
42:28hold of people that we can get in touch
42:29with customers and competitors and begin
42:31to do the work that the the government
42:33was doing. That was not typical work
42:35that people by themselves, one or two
42:38people did at a hedge fund was build a
42:40production line for the resolution of
42:42uncertainty of antitrust risk. We did
42:45that and it acred huge benefits. We
42:47debated at one point whether to make a
42:49commercial and do like a GLG. We decided
42:51that you know we were making enough
42:52money on Riskar internally to keep it
42:54captive. But but that was an example of
42:58going into something nobody told me not
42:59to do it. You know, everybody else
43:01called a former DOJ lawyer in DC and
43:03said, "What do you think? Is this deal
43:04going through?" They'd say, "I think
43:06there's a 70% chance it goes through.
43:08What am I going to do with that? I mean,
43:10if I put it on and it breaks, can I call
43:12my investor and go, well, my my lawyer
43:14said it was 7030. That was pretty good.
43:16It was trading like it was 50/50." That
43:17wasn't it for me. And so, that's where
43:20the marketplace offered a risk bet and I
43:22wanted to find out where the uncertainty
43:24was, go resolve it. And I find again as
43:27I alluded to earlier that if you're the
43:29first person to go in and systematize
43:31resolving that uncertainty, you get
43:33enormous gains from doing that. Rather
43:35than trying to predict, you know, long
43:37short is a slightly different game. You
43:39know, there is a distribution of where
43:41the stock's going to go in earnings, you
43:43just have a slightly different one. In
43:44antitrust, there was no marketplace for
43:47the probability of that deal, right?
43:49There was I mean, the market traded it
43:50somewhere, but it traded based on the
43:52fact that it had no model for it. I just
43:54tried to build take that and resolve it
43:56into probabilities and and and bet
43:59against a deal or bet for a deal which
44:01we did both. Mark Mitchell, a famous U
44:05of Chicago professor has um paper that
44:09he wrote probably in 2000 or so, the
44:11risk characteristics of risk arbitrage.
44:14Essentially, you know, illustrating the
44:16tail risk, the short put kind of
44:18components. And we know that these
44:20deals, they they jump to default in some
44:23ways, right? They break, right? That's
44:24the term. And so that almost sort of
44:28conjures up an idea of it just happens.
44:32And so that you almost don't get to
44:34observe, you know, it seems your your
44:37process is about observing things and
44:40allowing those to inform you. And is it
44:43this all of these conversations and this
44:46deep and wide expert network? Are those
44:49the is does that become the set of
44:52information that allows you to really
44:53kind of create you know the the
44:56mechanism to update your prior where if
44:58you're just in the deal and just waiting
45:00for something bad to happen you don't
45:02really get that information.
45:03>> I think that that's right. I think about
45:05the following and and I I'm not you know
45:08pouring cold water on anybody. I I I
45:10read that obviously I you know I was
45:12insatiably reading whatever I could
45:13around our business and otherwise.
45:16Um, I was a practitioner, right? I I've
45:20played I don't know 10,000 mergers. I
45:22mean, I've read 10,000 merger. There are
45:24people who have been around a long time.
45:25I've been around a long time. I've
45:26played a lot of Risk Aarb. I know a lot
45:28about Risk Aarb. Um, if there was one
45:31thing I might argue that I actually am
45:32an expert in in the whole world, that
45:34might be it and nothing else. And so,
45:36what I would say is is that when when I
45:38started doing Riskarb, I didn't know
45:40what was important and what wasn't. And
45:42so yes, I focused on antitrust, but I
45:44kept track and built a database before
45:46people were this is in their middle 90s
45:48of all these aspects around mergers,
45:51merger agreements that I provisions etc.
45:54And so a lot of people that do those
45:55studies and it was a great study and a
45:57great research report. They're doing it
45:59from data uh that that is available. I
46:02had my own proprietary database that
46:04were that came at the time from I had to
46:05read the merge agreements and capture
46:07data. I had to talk to CEOs and I began
46:10some of the data was useless. some was
46:11incredibly valuable. And so where I had
46:14the advantage was that whereas someone
46:16might have x number of data points and
46:18they're discerning a pattern from it, I
46:21had a lot of data points that I was
46:24keeping track of. And there were a lot
46:25of times where I'd work with the quant
46:27team, you know, at Citadel and then at
46:28Magnetar and I'd ask a lot of questions
46:30like, hey, you know, what's the answer
46:33to this problem, right? and even
46:35portfolio construction simple questions
46:37maybe like you know if I if I if a deal
46:40is trading at a 95% probability of going
46:42through in the market and I think it's
46:4498
46:46but another deal is trading at a 50%
46:48probability and I think it's 70 which is
46:51better obviously I can get into
46:53confidence intervals etc but you know
46:55there were lots of questions that you
46:56could ask when you have enough data
46:58points and enough information that might
47:00seem like an easy question but when a
47:01deal is a collar deal you have to
47:03constantly be reassessing ing what the
47:04value of the collar is. If it's a quanto
47:06option, not only on the stock, but on
47:08the currency, that's harder to
47:09calculate. I mean, you know, we can go
47:11in into a infinite loop on this topic.
47:14Actually, I've spent a lot of time on
47:16this one and always always something
47:17that one misses, but I would sit and say
47:19that going back to your original
47:21question.
47:22I had a lot of data points that I was
47:24sitting on and and it doesn't mean that
47:26there isn't jump to defaults. But I find
47:29that and and that there isn't true
47:31uncertainty that's unresolvable. But I
47:33do find that very often if you're very
47:37very maniacal and looking constantly for
47:40where you might be wrong, you know what
47:42I did in risk was the following and and
47:45it's very basian. the operating part of
47:47a basis the equation uh for you know bay
47:50theorem is not p it's the not hypothesis
47:54all you're doing is trying to figure out
47:55how you're wrong ask anybody and I have
47:58trained a lot of people and they've gone
47:59on to build huge businesses you know
48:01we're I think that the people who are my
48:03mentees it's it's way way over hundred
48:05billion dollars under management that
48:06they created those businesses not that
48:08they're working for people in that but
48:09that they're they're running those asset
48:11management businesses and probably they
48:13would all tell you you know not the
48:15easiest thing working for I was I'm a
48:16pretty intense guy. Um I hope I'm fair,
48:19but I demand a lot. And constantly just
48:21saying, "Listen, where are we wrong?
48:23Where are we wrong? Where are we wrong?
48:24Where are we wrong?" And someone says,
48:25"Well, I just spoke to someone
48:27yesterday." I'm like, "Yeah, but that
48:28new piece of data came out. Call them
48:30back again." And people are like, "What?
48:31But I just spoke to the CEO yesterday."
48:33I'm like, "Yeah, but that came out and
48:34now I'm not sure it's right anymore, so
48:36call again." So, you know, being
48:38relentless about that, constantly
48:42finding any little piece of data and not
48:44necessarily saying I can ignore it. It's
48:46an outlier. We did all our work already.
48:48I was in a perpetual state of trying to
48:50figure out whether that marginal new
48:52piece of information was a new line. And
48:54whether it meant that something had
48:56changed, the regime changed, something
48:57happened. That was that's what you know
49:01what we built really well at both
49:04Citadel and Magnitar. Well, let's talk
49:06about another um big investment on um
49:09Magnetar's behalf, and that's Corore
49:11Weave. And I think this really
49:14speaks to a different way of seeing an
49:17opportunity. I'll let you run with it,
49:20but u my understanding Corweave's a
49:23crypto miner um in a cryptobear market
49:27uh but it's got some really interesting
49:29assets and you guys are able to see
49:31something there that not a lot of others
49:33did. But why don't you walk us through I
49:36just would love to get inside that
49:38conversation with you and your you know
49:40your business partners kind of how that
49:41whole thing materialized
49:43>> and I should be clear here you know so
49:45so I left you know three years ago um
49:49from running day-to-day magnetar I'm
49:51very close there I still remain an owner
49:53and investor um and I'm very close to
49:55the people running it so some of this
49:57overlap with me some of what was after
49:58but I'm I'm very close to them but let
50:00me let me step back for one second I'm
50:02not avoiding it we'll get back to core.
50:04But I think coreweave again is is about
50:06a process rather than a trade. So let me
50:09step back for a second. I'm going to go
50:11fairly quickly. I'm I'll try to be um
50:13talk slowly, but I'll be very quick with
50:16it to just give you evidence of a
50:18pattern of thinking which I think is
50:20more valuable to people than just one
50:22trade. And so um go back to what we
50:25talked about earlier around uncertainty,
50:27risk, and and black swan.
50:30When we built Magnetar, we did something
50:32somewhat unusual. When we launched
50:34Magnetar, we said to everybody, source,
50:36evaluate, structure, risk, manage. And
50:38most people in the hedge fund world back
50:40then would have said, "Source, what are
50:41you sourcing?" You know, risk our deals
50:43are announced, converts, you get a call
50:44from a bank. Long, short equities, you
50:46just see that those are your sector
50:47names. What are you sourcing? And what
50:50are you structuring? Most of that stuff
50:51just comes to you.
50:53We were focused on going and finding
50:56places where uncertainty was and
50:58resolving it. and when you find it
50:59sourcing it. So let me give an example
51:01and then we'll go right to Corey. In
51:032005 we launched and right around 20056
51:08a new technology came out which was
51:10horizontal drilling and fracking and um
51:14it it made a it was a regime change in
51:17energy and regime changes I consider
51:19this will be important for core I
51:21consider regime changes to have four
51:22components. One is the production
51:24function changes.
51:27Two is things become abundant that were
51:30scarce and things that were scarce
51:31become abundant. Three is you get
51:33bottlenecks and four is that the old map
51:37doesn't explain what's happening and
51:39that change is irreversible. So let's go
51:41to energy and then we'll go right to
51:43core wheat. Energy production function
51:45clearly changed. We we weren't
51:46vertically drilling. We were
51:47horizontally drilling. What we went from
51:50importing a lot of oil and gas to being
51:52as we are today the number one producer
51:54of oil and gas in the world. What became
51:56scarce was not the the the carbon. What
52:00became scarce was because we had never
52:02had it. We had minimal buildout of
52:05infrastructure, both rigs upstream to do
52:08the drilling and also midstream to get
52:10the oil from where it was newly found to
52:12where it needed to get to the hubs. And
52:14so what became scarce was the capital
52:17formation. I don't know if anybody was
52:18around like I was yours old enough but
52:20we had MLPS maybe five maybe10 billion
52:24dollars a year of capital formation in
52:26energy going to hundreds of billions. So
52:28what we did is we stepped in and said
52:30you know an adopter might go wow um I
52:32believe that this change is real like
52:34what's going to be the price of oil. We
52:35stepped in and said well wait a minute
52:37we're adapters what does this mean for
52:39the ecosystem and what what is a where
52:42is the bottleneck and where do we fit in
52:44to help resolve that and what can we not
52:46resolve? So we went in and said, "Well,
52:49you know, mid-stream assets, you know,
52:51we don't know where the commodity is
52:52going to go." Um, mid-stream assets on
52:55with taker contracts, so you get paid
52:57regardless of used or not or volumes.
52:59You know, if you're a midstream and you
53:01need to build a pipeline and you need
53:03capital and you turn around and you have
53:05a takeaway a takeer pick contract of BP,
53:07you know, very good credit, then you
53:10have an asset that's cash flowing and
53:12you're going to build it. What if we
53:14lend to you? We can be at various places
53:18in the capital structure. If we're if
53:19we're too worried, we can be senior. Um
53:21if later, you know, if we're confident
53:23that, you know, we we could be more in
53:25equity or otherwise. But I sit and say,
53:28you have a a core fundamental piece of
53:30collateral that cash flows. I want to
53:33contain that. I want to get paid back on
53:34that. Maybe I'll do a convert or a
53:36preferred, but I want upside on your
53:38stock. Um, you know, I want to right
53:40tail as well as that. And because you're
53:42one of the first people to go in there,
53:44you source it, you structure that, you
53:46try to structure it in a way that's
53:47forgiving to the commodity. I run a
53:49scenar, you know, simulate a number of
53:51scenarios. How do I avoid a left tail if
53:53crude goes to 20 or 150? And so take
53:57that, leave that aside and go, well, I'm
54:00going to fast forward to core week. And
54:02now, yes, you're right. We were at the
54:05time, and I think a lot of people
54:06thinking about the production of
54:08Bitcoin, right? And you know, there's
54:10Bitcoin miners. It's an asset. It
54:11generates cash flow. Obviously, it's
54:13dependent on the level of Bitcoin and
54:15the operations and the cost of
54:17electricity etc. But um more importantly
54:20was that again I don't want to speak on
54:22behalf of of Magnetar because I'm not a
54:24spokesman for Magnetar but um you know
54:27being close to to all of them um in fact
54:29you know I was with Dave yesterday um
54:32that
54:33think about the same thing I I have if I
54:36have GPUs and I need to build a data
54:38center it's a little like a mid and I'm
54:41going to call it a mid-stream asset. I
54:43have a contract with Microsoft very good
54:45credit for the offtake of those GPUs and
54:49I see the production functions changing.
54:52We're a fourth industrial revolution is
54:54going on right now. But what is it
54:56producing? It's producing knowledge and
54:58when the cost of knowledge goes to zero
55:01that changes everything. You know I say
55:03AI is a tool at the interface. It's a
55:06regime change at the system level. So
55:08now we get to step back and go great
55:10there's a production function change
55:12scarce abundant again capital formation
55:15here is going to be scarce right so
55:18we're going to step in just like we did
55:20before negotiate and structure something
55:22where we we can protect ourselves around
55:25the left tail and get upside
55:27participation it's this it's the same
55:30game obviously it's a different
55:31disruption a different regime shift but
55:34what we became and designed our entire
55:36firm around from In the beginning it it
55:38wasn't that later we did it. We designed
55:40our whole firm around how do we go and
55:44go into places where others don't like
55:45to go because it's not resolved yet and
55:47figure out a way to sort of zoom out
55:50think about the system as a whole figure
55:52out where we could be helpful in that
55:55system provide value and make sure that
55:57we in a fair way get the appropriate
56:01riskreward that we can architect both by
56:03having sourced it and structuring it in
56:05a way that great here are the things we
56:07can resolve here's the things that we
56:09can't resolve if We have our own
56:10sourcing and structuring. Can we within
56:13the structure get rid of the things that
56:15we can't resolve versus creating
56:16investments and having to go as a pool
56:19continuously resolve how I hedge all of
56:21that when I combine it all. This can be
56:23a lot of it can be done directly through
56:25negotiating and structuring. So to me
56:28these trades are emblematic of you know
56:31what we tried to do and what we did
56:33again and again and again at Magnetar
56:35which is we tended not for the most part
56:37to go to where the risk was and say we
56:40can do it a little bit better. You know
56:42long short has a little bit of edge but
56:44sharp ratio is edge times the square
56:45root of n. So I diversified in a number
56:48of independence over large n and I can
56:50have something if I use leverage. We
56:52tended to be not levered or very very
56:54little leverage at Magnetar and turn
56:56around and go places where we think the
56:58money is once that uncertainty is
57:00resolved into risk. The marginal person
57:03comes in with a much lower cost of
57:04capital the return gets squeezed out and
57:07the money we've made is having been a
57:10participant early on from the
57:12uncertainty coming down to the risk
57:14level. That's that's how we think about
57:16it. That's what was going through our
57:17minds. I mean, obviously that you could
57:19do a whole whole episode on on just one
57:22trade, but that was the gist of the
57:24structure. Um, and and we did it, you
57:26know, I gave you two examples, I gave
57:29you a third with RCARV. It's it's how we
57:31focused and what our focus is at at
57:33Magnetar.
57:34>> It seems like a lot of your thinking and
57:36and maybe AQ itself is related in some
57:41ways to optionality being trying to be
57:44long optionality.
57:46um your efforts to cut off the left tail
57:50uh in risk arb your core
57:54financing certainly has a convertible
57:56component to it which embeds call
57:57optionality
57:59and so I I'd just love to sort of hear
58:02you talk out loud about that is there a
58:04connection [clears throat] there between
58:06you know adaptive thinking and trying to
58:09keep yourself long optionality
58:12>> 100% and uh this is a pretty detailed
58:17part. I think it's the part of my book
58:19where if people read it and they want to
58:20muscle through, I think it's meant to be
58:22approachable. You can tell the audience,
58:23but there is a slightly more dense part
58:26um relatively um when I get into sort of
58:30decision theory and sort of chapters 8,
58:319, and 10. And I put something in a
58:34footnote so that I don't, you know, if
58:35people want to read it, they can read
58:36it. But it is a pretty important point
58:38around optionality. And so, let me let
58:40me unpack that and a little more and
58:42because you're so right and on to
58:44something. So, let me zoom back for a
58:46second, you know, and I know given the
58:48nature of how you think and your
58:49audience, this this I think will be a
58:51fruitful fruitful little piece. If I
58:54don't I'm modeling the world and it's
58:57uncertain and I know there are different
58:59possibilities, but I don't know the
59:00probabilities. I'm going to go imagine
59:03things, try to make sure that I avoid
59:05the worst case scenarios when I go
59:07experiment. But I have a decision tree,
59:09right? And I have one I have a line lit
59:12through the decision tree. Meaning I
59:13think this is my strongest guess as to
59:16you know going right left right left.
59:18You know if if I I have a you know as
59:19you know an example of a restaurant I
59:21built in there. It's like what what is
59:23the food? What is the decor? What you
59:24know you have this line through all
59:25these different decisions. That's my
59:27best answer. But the goal when you go
59:31into an experiment is to try to preserve
59:34the entire tree. And the reason why is I
59:36don't know which the best part of it is.
59:38I'm going I have what I think is the
59:40best line through the tree, but thinking
59:42as a basia, like I don't want to prune
59:44anything. Now, I have to prove prune the
59:47ones that are going to kill me because
59:49if I go in and I land on one of those,
59:50game's over. I can't even go back to the
59:52tree. So, the first thing you want to do
59:56by simulating things, even if you pick a
59:58strong opinion, weekly out, it might be
1:00:00the wrong line. One of the nodes might
1:00:01be wrong, but I want I don't want to get
1:00:03it lost. So now you go in and one of the
1:00:05things that happens is
1:00:08let's say you wind up with a scenario.
1:00:10So to two a lot of people think about
1:00:12the left tail and we talked about that
1:00:14already is I don't care about small
1:00:16losses where I get feedback back I learn
1:00:18and I go redraw. I care about
1:00:20catastrophic losses. Right? So but let's
1:00:23think about the right tail because
1:00:25anybody who's very familiar with
1:00:27decision-m could say Alec it sounds like
1:00:29you're saying say fail or fail safe.
1:00:31There's words for that. But here's where
1:00:33mine differs a little bit which is that
1:00:36I also believe that when you're
1:00:39experimenting and this will be very very
1:00:41well understood by investors who sit
1:00:43around and might be listening going but
1:00:44wait a minute you know power law venture
1:00:47there are certain times where I don't
1:00:49know the answer and so an opportunity
1:00:51pops up and well should I take it should
1:00:54I not and my answer is really simple if
1:00:56it's not that expensive meaning it
1:00:58doesn't wipe you out so I'm out of money
1:01:00I have no more chance to iterate exper
1:01:02experimenting and it's irreversible, you
1:01:05should take it. This is a little unusual
1:01:07in decision-m and why is it that I say
1:01:10if it's a not that expensive option on a
1:01:12big right tail, you should take it. Very
1:01:14easy answer. Here's why. It's because if
1:01:18I don't take it and it's irreversible, I
1:01:20have pruned the tree. If I later find
1:01:23out that was the best node and path,
1:01:26it's not there anymore. I didn't take
1:01:27it. If it's so expensive that I ruin my,
1:01:31you know, through the system, I can't
1:01:32take it because I got to keep learning.
1:01:34But if it doesn't cost me much to take
1:01:36it, even though I don't know if it's the
1:01:38right path, if it's irreversible, Bezos
1:01:41talks about one way, two-way decisions,
1:01:43if it's irreversible, take it precisely
1:01:46because it keeps your optionality open
1:01:48that later when you refined and you go
1:01:50back and go, "Oh my god, that was on the
1:01:52best line of the decision tree, it's
1:01:54available." So slight level of more
1:01:56granularity to what is slightly unique
1:01:59about my method. It's a little unusual
1:02:01relative to what you'll read about in
1:02:03the literature whether you get into
1:02:04basian more risk worlds or whether you
1:02:06get into uncertainty in other people's
1:02:08models. Um the way my mind works is
1:02:11there are times where you do place bets
1:02:13on that even though you haven't resolved
1:02:15it if you just are going to miss that
1:02:17opportunity right and you know I think
1:02:19about that a little bit um you know I
1:02:21invested in SpaceX you know very early
1:02:23on and one of that was you know he only
1:02:26raised $9 billion ever for that business
1:02:29right it there was a feeling of
1:02:31irreversibility I mean it turns out
1:02:32there were a couple other chances but
1:02:34there weren't many he did a bunch of
1:02:35secondary trades but but primary trades
1:02:37he didn't do much So I I didn't know how
1:02:40big the right tail was. Obviously the
1:02:41left tail is my investment and you know
1:02:44I think he's you know an incredible
1:02:46entrepreneur and um we can you know
1:02:47people can debate about lots of things
1:02:49about Elon but in my opinion he may he
1:02:52he he exhibits a lot of high AQ in his
1:02:54decision-m experimenting and probability
1:02:57and feedback loops. So I wanted to
1:03:00participate alongside that even though
1:03:02you know uh there was no there was no
1:03:05Starlink in the sky. I knew about
1:03:07Starlink and I thought it would be a big
1:03:08opportunity but you know I didn't know
1:03:10about AI data centers in the sky and you
1:03:12know other things like that. So you know
1:03:14that optionality I wanted to preserve
1:03:16that optionality and I I didn't want to
1:03:18lose it. So I don't know if that answers
1:03:19part of your question.
1:03:20>> Well I want to finish with one or two of
1:03:22the kind of investments or areas of
1:03:24focus for you at QAR. Um before before
1:03:28we do that, if there were one or two
1:03:31just big picture messages that you want
1:03:34folks to come away from the adaptability
1:03:37quotient, how would you how would you
1:03:39frame that out?
1:03:40>> Most of what we have talked about here
1:03:43is about investing which you know
1:03:45obviously I've had a career of that. Um,
1:03:48but the reason why I wrote the book um
1:03:51was for a different reason. And and so
1:03:53let me tell you the reason why and then
1:03:55tell you what or your audience what I
1:03:58what I think I would want them to take
1:03:59away. Um and so you know I spent a
1:04:0330-year career as we just described sort
1:04:04of like thinking about risk and
1:04:06uncertainty and playing with various
1:04:08points of uncertainty. Energy had a
1:04:10regime change. The great financial
1:04:11crisis was a regime change you know etc.
1:04:13Um so when I left Magnitar I did it
1:04:16because I had an opinion that um before
1:04:18chatbt was announced that data compute
1:04:21and material science were converging and
1:04:22that we were going to have in in you
1:04:26know biology terms a punctuated
1:04:28equilibrium a moment of unbelievable
1:04:30change across a variety of different
1:04:32industries energy and transition
1:04:33synthetic biology maybe a new financial
1:04:35rail system quantum fusion space
1:04:37robotics AI defense I mean it's all
1:04:39changing incredibly rapidly I wanted to
1:04:41be present for at and I wanted to be I
1:04:44wanted to map it and I didn't want any
1:04:46constraints. I wanted to give you know
1:04:48the way to test whether Magnitar was a
1:04:50business was for me to leave um and to
1:04:52see whether and they prospered and been
1:04:54you know done an incredible job great
1:04:55team um but it also as I started to face
1:04:58this uncertainty I was like looking at
1:05:00it going man how do I resolve this how
1:05:02do I map this and I realized oh I I kind
1:05:05of have been mapping this kind of
1:05:06uncertainty forever but when I saw this
1:05:08one the prior uncertainties I described
1:05:11all centered around an industry
1:05:13investors it was a small group of people
1:05:15or a reasonable
1:05:16group of people. The center of gravity
1:05:20of regime shift has changed to society
1:05:22in this one. And the better thing for me
1:05:25to do would be to hand people the source
1:05:27code. This is how you survive and thrive
1:05:29in a world of uncertainty. I've done it
1:05:31before many times in investing. This
1:05:35one, you know, everybody's an
1:05:37entrepreneur right now, right? You
1:05:38either are one, your industry is
1:05:40disrupted and you're one or you're a
1:05:42young person whose career path is
1:05:43disrupted. like what do you do facing
1:05:45all this uncertainty? So the first thing
1:05:47is like that that that's what I saw. The
1:05:49second thing is is that you know why why
1:05:51is it so urgent now and then what's the
1:05:53takeaway? The urgency is is that we've
1:05:55been through industrial revolutions
1:05:57before. So the first industrial
1:05:59revolution and the second industrial
1:06:00revolution we manipulated atoms to
1:06:04offload work, right? We made hammers, we
1:06:06made tools, we even had tools make
1:06:08tools. We did it at scale production,
1:06:11you know, in factories, etc. And that
1:06:13was great. We offloaded labor. We went
1:06:14to do other things. The third industrial
1:06:16revolution, we manipulated bits, right?
1:06:19And bits were storage, retrieval, and
1:06:21computation. But in all of those cases,
1:06:24it it changed how we work, right? And
1:06:28humans were always in the process at
1:06:29some point, but it changed how we work.
1:06:32And when you manipulate a bit, you know,
1:06:34if I have a spreadsheet and I run
1:06:36through some numbers, every time I run
1:06:38it through, the numbers come back the
1:06:39same. If I use the same inputs,
1:06:42this is different. This is the fourth
1:06:45industrial revolution. At the at the
1:06:49physical layer, we have power, GPUs, we
1:06:53have land, shell, craze we're seeing in
1:06:55the marketplace of what is rising. we
1:06:58you know you know depending on how time
1:07:00we could talk about bubbles or not but
1:07:01but what I was going to say is that it's
1:07:04important to think about the fourth
1:07:06industrial revolution as what is what is
1:07:08how are we producing knowledge but what
1:07:11is being produced
1:07:13is tokens and someone on you know
1:07:16listening will say but Alec tokens are
1:07:17bits but they're different and here's
1:07:20why they're different a bit doesn't
1:07:21matter when it's next to another bit
1:07:24context any all the prior revolutions
1:07:26were fixed and contextless. It didn't
1:07:28matter if the hammer's next to the
1:07:30screwdriver, didn't matter if a bit's
1:07:31next to another bit. Tokens, where the
1:07:34word is placed matters and therefore it
1:07:37has meaning. And that means for the
1:07:40first time in history, we have a partner
1:07:43in how we think. We have a partner in
1:07:45meaning making. When I use Google and I
1:07:47Google a lot of information, I get all
1:07:49human thought and I have to compress it
1:07:51and make meaning out of it. That's not
1:07:53what happens now. I use AI, it
1:07:56compresses for me, right? And so I just
1:07:58And so this is something different is
1:08:01the second part. And now what do I want
1:08:02the message to be? The message to be is
1:08:06it's a very dangerous moment that we're
1:08:08in right now. And it's an incredibly
1:08:10optimistic moment. And I'm going to take
1:08:12social media and I'm going to take AI
1:08:14and combine them as this experiment
1:08:16we're running. And we've already run one
1:08:18for 20 years. And that is you can use
1:08:21both of these tools which are both
1:08:23regime changes. They both change the
1:08:25environment of how we think. Not how we
1:08:28work and offloading prior ones but how
1:08:31we think. I can use it to augment how I
1:08:33think and connect with people or I could
1:08:37use it to atrophy how I think and
1:08:39connect to people. And that is a massive
1:08:42massive import. It's sort of I was going
1:08:44to leave your audience like there's
1:08:46alpha exchange and then there's like
1:08:47life alpha. This is life alpha. Okay,
1:08:50this is what is critical for people to
1:08:53understand. We just had an experiment
1:08:55where you gave up attention in return
1:08:57you were supposed to get connection. And
1:08:59if you used it to connect to your
1:09:01longlost, you know, buddy from high
1:09:03school that you didn't talk to before,
1:09:04that's great. You expanded the set. If
1:09:06you used it to do a reunion with your
1:09:09fellow, you know, fraternity brothers,
1:09:11great. You expanded the set. But if you
1:09:13use it to replace and make an artifact
1:09:16instead of the function. If I turned
1:09:18around and said, "I used to date. Now I
1:09:20swipe left and swipe right. Oh, I used
1:09:22to go on vacation with Dean, but now he
1:09:24shows me pictures as a vacation and I
1:09:26thumb up and like it." That is a
1:09:29illusion of connection. And that's why
1:09:31we have an epidemic of loneliness
1:09:32because people aren't using it to to to
1:09:35expand the function. They're using it as
1:09:37an artifact. It can be used well, but
1:09:39it's not being used well. Now we go to
1:09:41AI. Are we going to run the same
1:09:43experiment on cognition? Right? At the
1:09:45end of the day, I can use it to say,
1:09:47"Play devil's advocate with me. Where am
1:09:49I wrong? Where help me think about
1:09:51possible scenarios that I haven't sat
1:09:53through?" But if I ask it for an answer,
1:09:55if I just slowly give it away, it
1:09:58atrophies my cognition. And my point is
1:10:00is that and my book is really about
1:10:02agency. It's about you are sitting on
1:10:05the greatest machine ever made better
1:10:09than an AI AI. It's a you are a pattern
1:10:12processing like superior being and AI
1:10:16and social media are geared in a certain
1:10:19way. I call it the double whammy. If you
1:10:21just step back for a moment, you know,
1:10:23hopefully like this irony. What is what
1:10:26does social media do? Engagement forces
1:10:29it. If you're optimizing for it, that's
1:10:31the reward function. It gives you the
1:10:33extremes. What does AI do? What does an
1:10:36LLM do? It gives you the mean answer
1:10:38that everybody would give. So you're
1:10:41getting the worst of the polarized
1:10:43extremes and you're getting the middle
1:10:45of the distribution. Like it's sort of a
1:10:47if you let it think for you and connect
1:10:50for you and do the you're going to get
1:10:53double whammedi. So my point to people
1:10:55is it you don't give that away bit you
1:10:59know all of a sudden you know a tail
1:11:01risk. You give it away bit by bit by how
1:11:03you interact with it. So keep agency,
1:11:06keep what is uniquely yours. It'll help
1:11:08you become a better investor too, by the
1:11:10way, if you apply it. But this is your
1:11:12life. And if enough people make the
1:11:14choice, like we've done with social
1:11:16media, we're just slowly going to fade
1:11:18away to where we have this false
1:11:20illusion of connection, but we don't.
1:11:22False illusion of cognition, but we
1:11:24don't. And we become avatars. And I'm
1:11:28very fearful. I I'm kind of trying to
1:11:29yell and scream that we There's a part
1:11:33of this that is a world of abundance,
1:11:34but it's not all abundance. How you use
1:11:37it is everything. How you think about
1:11:39making decisions is everything. You are
1:11:42the keeper of choice. You are the keeper
1:11:44of agency. Don't give that up lightly. I
1:11:47would say don't ever give it up. And
1:11:49what my book is is basically a repair
1:11:51kit that says, I'll give you the way to
1:11:55compress and be active in thinking even
1:11:58a world that's confusing and that it's
1:12:00that's uncertain. I'll give you the code
1:12:02on how to do your own compression. Don't
1:12:03offload it. Do it yourself. I I'm not
1:12:06going to do it for you. I'm just going
1:12:07to, you know, what's inside the ears of
1:12:09the best investors, the best
1:12:10entrepreneurs. This is what we're
1:12:11they're doing. It just wasn't written
1:12:13down in one book. There's some great
1:12:15books, but I tried I don't you you'd be
1:12:17the judge, but I tried to write it down
1:12:19in one book. This is how to preserve
1:12:21your agency. It doesn't mean don't use
1:12:23these tools, but do not let them take
1:12:26over for your agency. That that would be
1:12:29a major major loss. I think that's a
1:12:31great way of uh kind of summing up what
1:12:34is a you know, there's so much insight
1:12:37in this book, but this idea of retaining
1:12:39agency is a very very strong um you
1:12:44know, strong recommendation or almost
1:12:46warning as you say at a time as we
1:12:49started this conversation amidst just
1:12:52just incredible change. Just in terms of
1:12:55QSTAR and you're investing there, are
1:12:57there one or two investments that you're
1:13:00particularly excited about? When we
1:13:02talked last uh you had kind of this
1:13:04barbell approach of, you know, true tech
1:13:08and then much more inerson stuff. You
1:13:11mentioned your restaurant, things like
1:13:12that. what's a uh an investment that
1:13:14you're currently, you know, either have
1:13:16made or just an area of u of study for
1:13:20you that you're thinking about right
1:13:21now?
1:13:22>> Yeah. No, I think well, you you you
1:13:24recall correctly. So, let me let me zoom
1:13:26out for a second and and describe QSTR
1:13:28as as our family office, just our own
1:13:30capital um and team of, you know, about
1:13:3215 people here. I I I would say the
1:13:35mission of Qstar um is
1:13:39to compound human flourishing. that
1:13:41that's our goal and we think that
1:13:43there's two aspects to that. One is
1:13:45capability which is what is the best
1:13:47what is the most that humans can do that
1:13:49tends to be more technology focused um
1:13:51you know capability um that could be
1:13:53healthcare could be technology um you
1:13:56know it's just h how do we achieve more
1:13:58as humans what tools and otherwise get
1:14:01built and investing in those um you know
1:14:04could help us live longer live better
1:14:06etc. So the second part is what we call
1:14:09belonging which is well great there's
1:14:11all those advances but what what makes
1:14:13it have any meaning and that is
1:14:15connectivity and having each other um
1:14:18and understanding you know that that
1:14:20despite lots of differences you know we
1:14:23really have a lot of the same hopes,
1:14:25dreams, fears etc. And so the part on
1:14:28the belonging side you know so we tend
1:14:30to call the capability side
1:14:32technosentric and the belonging side
1:14:35anthroentric. And we have teams that
1:14:37focus on those two aspects. And so
1:14:39thematically, where do we think the
1:14:41world's going? And then as typical you
1:14:42would expect of us, you know, bottoms
1:14:44up, what's the most forgiving asymmetric
1:14:47investment we can make across those. So,
1:14:49um, you know, on the on the anthrop
1:14:51side, you know, we think, you know, we
1:14:53have a we're minority owner and a major
1:14:55league soccer team. We think a lot about
1:14:57hospitality. Um, we think and that that
1:15:00could include, you know, music or
1:15:01anywhere where people gather. We have a
1:15:03a pretty large parcel of real estate
1:15:06that we've bought and assembled in
1:15:08Chicago and we're going to be building a
1:15:10cultural arts hub that doesn't really
1:15:12exist in Chicago. Um, you know, it's
1:15:15sort of sort of like you could think of
1:15:16like what happened in Dumbo. You can
1:15:18think about Miami Design District. This
1:15:20is a maybe a smaller version, but
1:15:22thinking about what we can bring to
1:15:24gather people together in a community
1:15:26that, you know, I think because of the
1:15:29Chicago fire, we just don't have a grid
1:15:31system that has as has um supported
1:15:34building something like that. But when
1:15:36you accumulate a five city block area,
1:15:39you now control all the blocks, you can
1:15:41do something you might not have before.
1:15:43So that more on the an on the anthro
1:15:44side and on the just as an example what
1:15:47we're looking at on the more capability
1:15:50and technology side I mean one of the
1:15:52things going back to the formulaic piece
1:15:54that we talked about on the production
1:15:57side of the fourth industrial revolution
1:15:59um we think a lot about the electric
1:16:01grid and that being a bottleneck and the
1:16:04fact that it's old and we have problems
1:16:06and so you know um I'll I'll keep this
1:16:09big picture rather than too technical
1:16:11but it's great that we solar and we have
1:16:13wind in addition to, you know, carbon
1:16:15based um forms of energy, but they
1:16:18create challenges when they come into
1:16:20the grid for things that need to be 59s,
1:16:24you know, like a a data center needs to
1:16:26be up 99.999%
1:16:27of the time. Um same thing with some of
1:16:30the reshored heavy industrial device,
1:16:34you know, if you're a manufacturer um
1:16:36and you're using special equipment, you
1:16:38you can't have it go down in the middle
1:16:39of the run. So there's a concept um not
1:16:42of grid following. So uh solar and wind
1:16:46work in a certain way that can be good
1:16:49overall for us but can be destabilizing
1:16:52to the grid for businesses like this.
1:16:55And then there's a new business or a new
1:16:57area I I guess maybe it's not so new but
1:17:00called grid forming not grid following
1:17:02which is building products that resolve
1:17:04some of those problems so that we can um
1:17:07not have this technology build disrupt
1:17:10you know what we need as typical
1:17:13consumers and otherwise. So investing in
1:17:15people solving that problem is
1:17:16interesting to us as a, you know, this
1:17:18is a bottleneck. It's a problem. How do
1:17:20you resolve that problem? How do you
1:17:23repair that, you know, we could use a
1:17:24new grid, but in the meanwhile, how do
1:17:26we repair the existing one to support
1:17:28some of these things without, you know,
1:17:29how do we get the good without harming,
1:17:31you know, individuals? And I think
1:17:33that's an interesting area that we've
1:17:35spent some time on as an example.
1:17:37>> Well, Alec, this has been a pleasure to
1:17:39host this conversation. Congrats on the
1:17:41book. Obviously a tremendously heavy uh
1:17:44lift putting it together. I really and I
1:17:47mean this really enjoyed reading it. I
1:17:48think it's a um it's at the right time
1:17:52too and you come back to this idea of
1:17:54retaining agency and this um you know
1:17:57method for trying to think through
1:17:59things in a structured way. So kudos to
1:18:01you for writing it and thank you so much
1:18:03again for being a guest.
1:18:05>> Well I appreciate all of your efforts. I
1:18:07you know I'm a I said I'm a I'm a fan. I
1:18:09I I think it's incredible for you to
1:18:11take your time and democratize a lot of
1:18:13the knowledge that um you know you give
1:18:15to people. I'm hoping that whatever I've
1:18:18said is is helpful to people and and um
1:18:21making it through um safely uh the
1:18:24environment that we're in. It's it's
1:18:26it's it's important to recognize that
1:18:28it's a challenging one and um you know
1:18:30I'm I'm just I'm normally quiet but um
1:18:34hopefully um speaking out a little bit
1:18:36more than I normally do um to give
1:18:38people tools um so hopefully if it helps
1:18:40someone one listener even or people a
1:18:42little bit then then I'll be I'll be
1:18:44thankful and hopefully um you know
1:18:46valuable to the audience. So appreciate
1:18:48you for having me.

About this transcript

This page contains the full transcript of Alec Litowitz, Founder of Magnetar Capital and Qstar Capital by The Alpha Exchange with Dean Curnutt, generated from the public captions YouTube serves with the video. The transcript has 15,674 words across 2,161 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.