Stephen Wolfram was explaining why consciousness and free will should not be analyzed by hypothetically rewinding the universe to a prior state. Green lasers were cascading across my peripheral vision. He said too many things break down when we speculate about conditions that violate the laws of physics. Sonic audio waves from Grimes’ live DJ set were pulsing through our words.
The All-In Summit ’23 was about to come to a close after two days of ideation. It’s a strange word to define an event, “the formation of ideas or concepts”. But it’s the only word I can think of to capture the event, to borrow Stephen’s language, in an irreducible manner.
This cosmic EDM scene wasn’t what I envisioned when I told Martha we should apply to attend the All-In Summit. But uncertainty was one of the conceptual threads of the event. The nature of life is unpredictable. And possible outcomes are greater than we can imagine.
As Stephen told the summit, “the interesting way to lead life is to just let time exist and understand that you don’t know what’s underneath the universe”.
As a fan of the pod and a first-time founder in the midst of a seed-raise, I came to the Summit to celebrate uncertainty. That state that connected my own life of market risk trading, data and analytics, and founding a startup. Reflecting on the Summit, it was the conceptual threads connecting the speakers that struck me most.
Particularly because the speakers had a wide range of backgrounds and ideas. Day one went from Ray Dalio to Elon Musk with Stephen Wolfram and Jenny Just, among others, in between. The breadth of content on day two was equally vast, spanning Larry Summers to Mr. Beast.
The days were surrounded by theme parties. Casino Royale in a bank vault, Barbie on the Santa Monica pier, and yes: the intersection of Grimes and Blade Runner. Tongue-in-cheek gatherings reminding us we are here to enjoy life, which is easy to do playing Skee-Ball in Barbie gear.
The All-In Pod, for those outside the audio loop, consists of four self-described “Besties” from Silicon Valley. Jason, Chamath, Sacks, and Friedberg. The reason 1,800 people flew from all over the world to a podcast gathering is because the show - which can be serious, helpful, controversial, and inspiring - is above all else, genuine.
The All-In Pod somehow attracts many cultures while being a reprieve from culture wars. And so four friends become a different kind of 1,804 friends.
My aim isn’t to review the All-In Summit in full. There were some talks that didn’t resonate with me as much as others. A normal distribution of randomness. Other attendees will have had different perspectives and experiences. Fortunately, all the talks will be published by the All-In Pod.
My intent is to accurately reflect some of the ideas presented at the Summit. Where I’ve used quote marks around speaker language, I am quoting them directly. But the information I’ve captured from the event traveled from their minds to my notes before being processed and filtered by my own fallible memory and biases.
Errors and omissions will be present, for which I am wholly responsible. Feedback is welcome and as my understanding expands, edits and corrections will be made.
I am writing to process the event. To understand why certain concepts connected across speakers with diverse ideas. To appreciate what compelled the hosts, the speakers, and the audience to come together. And to reflect on their various incentives that combined to form a shared motivation.
I’m writing because there’s something about the ideas - these conceptual threads - that matter to Point Focal. And maybe to your business too. But I can’t quite identify the connection yet. So I’ll look to find it by writing these words.
Bear with me dear reader.
From the mind who created the largest hedge fund to the mind with the largest YouTube following, there was a surplus of success on stage by many measures. And yet across the speakers there was a real appreciation for the limits of their knowledge. A humility about uncertainty.
Ray Dalio said sometimes we overlook the really big things.
Jenny Just’s Peak 6 trades hundreds of thousands of option contracts every day. And she doesn’t know which way the market is going.
Stephen Wolfram said, “it’s a better feature to know that we do not know the answer.”
Elon Musk didn’t know exactly what he was purchasing in Twitter. And he doesn’t know whether Russia will take Starlink down.
Larry Summers said, “There are two types of economists. Those who know they don’t know and those who don’t know they don’t know.”
Mr. Beast started a dark chocolate candy business for children. He said that turned out to be stupid because he didn’t know children don’t like dark chocolate.
And David Sacks didn’t know he was creating a podcast that would reach millions of people.
It’s a manifestation of Feynman’s axiom, “the more you know, the more you realize how much you don't know”. Some of our most brilliant minds were first and foremost explicitly humble about their ability to make accurate predictions.
One got the sense that this respect for uncertainty was responsible for both their success and for their understanding that there is an element of randomness inherent in their path to success.
This is a liberating thought not because it frees one from the hard work of innovation but because it illuminates the notion that there are many possible positive outcomes from hard work and innovation, most of which we cannot foresee.
Ray Dalio, speaking about geopolitics, said “structural problems created by elites cannot be solved with one side fighting the other side with irreconcilable differences. A strong middle is needed to bring the country together. A bipartisan engineering process that takes control of the extremes”. He framed the middle as a great reformation where if he were President he would establish a Manhattan type of project where important changes are engineered by a bi-partisan cabinet.
Jenny Just, speaking about market economics, said “financial volatility has been decimated. We’re in an odd place with a dot-com, AI feel… inflation may be slowing… but it may not be slowing… there is sector rotation… there is push and pull”. She said, “we’re right in the middle… without knowing what is going to snap this market one way or another”.
Respecting uncertainty, Jenny said she was not making a call on what exogenous event is coming. But from this economic middle state, she is long vol.
Vinod Khosla, an absolute giant in venture capital, said “investors bounce between fear and greed and never live in the middle. There is the hype cycle, followed by inflation, and then deflation. And it is difficult to see the middle. When the internet bubble burst in 2002, there was a dot com bust. But in reality, the internet didn’t change. Internet traffic stayed the same”. So within the hype-inflation-deflation cycles, there was fundamental value being added in the middle. And if you are adding value, fundamentally, you will be ok.
He said “you cannot time the market so it doesn’t matter at what price you enter. But upon deflation, respond quickly”.
This process describes active risk management. People ask Point Focal if our analytics generate alpha. That’s the wrong way to think about uncertainty and measuring risk. A better question is whether our analytics preserve capital. Can our models of investor behavior help asset manages respond quickly to deflation.
Whether you’re investing in startups or securities, exits are more important than entries.
Elon Musk talked about bringing Twitter to the middle. He is after the 80% of the world that live in between the extreme edges. He said “Twitter, I mean X, has been so far left for so long that a move to the middle feels far to the right”.
With his 157M followers and critics, and as someone with a knack for building businesses, it is telling that Elon wants the X experience to be in the middle.
Larry Summers said, “the thing about stuff that mean reverts is it mean reverts”.
The mean is a representation of the middle. Humanity, politics, markets, and companies mean revert. So we have Vinod Khosla reminding us that we cannot time markets, and we have Jenny Just reminding us that we do not know why financial volatility will not go to zero, only that it won’t, and here comes Larry to put a finer point on the things we know and the things we know we don’t know: the thing about stuff that mean reverts - like markets and financial volatility - is they mean revert. Larry said, “there’s never been a time when inflation was above 4% and unemployment was below 4% when the economy did not go into recession”.
Larry concluded by reminding the audience never to forget: “the best test for our country is whether we are the place to which people from every part of the world want to come”. That place may be on the edge of democracy, the edge of technology, and the edge of aspiration. But surely it is the middle of understanding.
Humility about uncertainty and mean reverting to a middle are concepts that translate to market behavior. More directly, Ray, Jenny, and Larry all spoke about demand and supply - market representations of greed and fear that can be observed and analyzed to manage risk.
Ray talked about debt and money creation which lead to the rise and fall of reserve currencies. “When debt and debt service payments rise relative to incomes, consumption gets squeezed and society realizes they have to print money.”
“When a government has a deficit they sell bonds. Buyers buy the bonds because there is an attractive return. When buyers realize they’re not getting a good return they will sell the bonds for equities or gold.” Gold is “the only asset you can have that is nobody else’s liability”.
Ray is describing a macro form of demand and supply. Economic forces that comprise greed and fear, mean reversion, computational math, and investor behavior. He talked about regime cycles - military, education, capital markets. Long term cycles that lead to regime changes, war, and peace. And like Jenny’s view of financial volatility, while Ray expects a new world order to emerge, he cannot predict who - what country or countries - will be at the forefront.
With respect to cycles, Ray said more important than the repetition is the cause and effect. This is what Ray is studying. These are his principles. Which led to my favorite line from Ray:
“Every conclusion I have is a function of measuring statistics and evaluating them as growth rates”.
This is risk management. This produced Bridgewater. And it is the framework Ray is using to analyze the world beyond Bridgewater.
Jenny said, “at any time, we will have three to four-thousand positions in equity stocks. But we're not high frequency traders. We are servicing customers and managing inventory”. Demand and supply.
The key to Peak 6 success is “managing inventory top-down, bottom-up, all day, every day”. If we don’t manage inventory well, we’re no different than a grocery store. The inventory goes bad, and we lose”.
Peak 6, consistent with Ray and Larry, do not make predictions. They do not call tops and bottoms. Because “they’re really hard to call”. Peak 6 monitors the cost of options - another proxy for the emotional state of the market. “When markets get stretched and become risky, the cost of options tend to get expensive. Today the cost of options is very inexpensive.” Which brings us back to Jenny’s observation that financial volatility has been decimated.
Like Ray’s view of geopolitics, conditions indicate a market regime change is coming, even if its form is unknowable.
Jenny provides another insight into demand and supply. These economic forces are oblivious to gender. When Jenny setup the first options floor off of an exchange she made a bet on her success: “that women could trade as well as men”.
The bet turned out well.
$100 invested in the S&P 500 in 1997 is worth $700 today.
Berkshire Hathaway? $1,200.
$100 invested in Peak 6 in 1997 is worth $3,000,000 today.
Jenny continues to bet on women through Poker Power, where she is teaching one million women to play poker. A game the Besties share a passion for that teaches discipline, risk assessment, negotiation, and decision making. And how to act on imperfect information.
Today women makeup 10% of professional poker players. This is a statistic Jenny is measuring. Through Poker Power she will be influencing its growth rate.
All of this ties into Jenny’s belief that the next “market correction” may come in the form of women becoming 50% of investors. Demand.
I won’t try to recap Bill Gurley’s presentation on regulatory capture. It was a master class in how special interests stifle innovation. It’s a presentation that was 20 years in the making across Bill’s professional life as a business analyst and venture capitalist. His pace was incredible. The content is shocking. And understanding the impact of regulatory capture is profound. Bill’s talk has already been released here: 2,851 Miles. I promise it is worth your time.
Vinod Khosla’s view of the current state of venture capital was an extension of Bill’s presentation. In forty years Vinod “has not seen one example of a large innovation that came from a large company or a large institution. You have to go back fifty years to the early 70’s when Bank of America put debit and credit on a plastic card”.
“Apple’s iPhone is not an exception because it came from a founder-led company, distinct from a big company.”
Vinod posed rhetorical questions to the crowd: “Where are we going to get driverless cars? From Waymo or GM? Where are we going to get reusable rockets? From SpaceX or Lockheed Martin? Electronic transportation? Tesla or Avis?”
Vinod advised founders to “always be looking for off-ramps - where else can we go”? And he advised investors, when considering if an investment is still a good investment worthy of more capital, “to focus on the quality of the CEO”.
Stephen Wolfram’s thesis committee included Richard Feynman when Stephen received his PhD in particle physics from Caltech in 1980. I did not know, dressed in Rick Deckard’s jacket, that I was merely two degrees of separation from Albert Einstein. Had I known, I would have enjoyed wondering what Einstein would have thought of Grimes and her light particles in that Blade Runner moment.
Stephen’s discussion with David Friedberg, who runs the All-In Pod’s “Science Corner”, was a torrent of information. Stephen posited that despite complete knowledge of computational rules that define physics, we cannot infer all possible outcomes.
“Even with simple rules, consequences are hard to work out. In general, it’s not possible to predict output even with known rule sets and equations. Within science, there is a fundamental inability to say what will happen.” Friedberg responded by noting that we should therefore be careful about using simple heuristics.
Stephen suggests “the computational universe, the set of things we could compute, is vast. And surprisingly, even a very simple program can do very complicated things. It’s a bit of a trick that nature has figured out”.
“The number Pi is an example of a simple rule with complicated output. The output appears to be random, but we know it is created from a simple rule.” This type of observation relates to Taleb’s notion of being fooled by randomness.
Stephen explained that “artificial intelligence is limited to the computational universe - all the possible rules. So while we’ve given large language models four billion webpages that we humans care about, the LLMs can therefore only make things that are like what we care about which is a tiny part of the computational universe. LLMs can only construct stuff from the limit of what we’ve done”.
“The reason LLMs work is because they are telling us through science how we construct language - noun, verb, noun, etc. - assembling sentences that make sense. The logic is a formula of language producing reasoning that is impressive. LLMs have noticed patterns of language that can be reused. It’s a linguistic user interface.”
“We have a small set of inferences, bullet-points, and the LLM can turn it into a big report that you send to someone who can use an LLM to reduce it back into points. LLMs have become a transport layer for language.”
Friedberg noted that our “rate of communication is a bandwidth limiter”. Stephen responded, “while the structure of each of our brains, our neuron composition, is different, LLMs can package their communication”. And he added questions. “What else can we learn from AI? What LLM science and language semantics have we not worked out yet?”
Stephen said “we may learn that it’s simpler to understand human language than we thought. There are 50,000 common words in English. And we have 50,000 numbers that represent the probability of the next words. We’re asking questions in the space of how many word combinations can reveal meaningful answers. And the number of word combinations are greater than the number of atoms in the universe”.
“And yet when we consider the vector of numbers and the things that correspond to their arrays we can contemplate the idea of inter-concept-space. How many words might exist in between cat and dog? A huge amount of stuff. And AI is finding the inter-concept space in between words.”
Stephen’s work and his talk revolve around his concept of computational irreducibility.
Computational irreducibility: Computations that cannot be sped up by means of any shortcut are called computationally irreducible. The principle of computational irreducibility says that the only way to determine the answer to a computationally irreducible question is to perform, or simulate, the computation.
“As computationally bounded humans, we necessarily observe aggregate behavior. When we look at gas molecules, we see randomness because we are observers with certain characteristics. If we were different, we would make different observations. We are not capable of seeing gas differently. So our heuristic is the gas law: temperature. We sum up the gas molecules and take an average. We have to take the average because what is underneath is computationally irreducible. And we are computationally bounded.”
Stephen says “our brains are full of branching behavior. And we assume our branching brain perceives the branching universe persistently in time. Yet at any moment we are made of different atoms and we perceive ourselves as the same. We are computationally bounded observers in time and inevitably observe the laws of physics that we can observe, based on our nature”.
“There are likely things that we take for granted about he universe and about us as observers. Aliens will observe different laws.”
Stephen asked us to “imagine if our minds were each a small point in the matrix where an aggregate consciousness as a species could emerge”. He suggested “the future of civilization should be to expand into space. Expand our understanding and our consciousness”.
Friedberg suggested that “humans are the unconscious computer”.
In conclusion, Stephen said “the universe itself is a computer. It is computing. And time is the progress of that computation. The question is what do all of these atoms of space make. And the answer is they make Einstein’s equations”.
For Point Focal, first and foremost, this notion of humility about uncertainty is core to our active risk management philosophy. We can observe, analyze, model, and forecast. But we cannot know. Because even simple rules can create not merely complex output, but complex output we could not imagine.
To help asset managers lose less money, we must combine this ethos of humility with adherence to “every conclusion being a function of measuring statistics and evaluating them as growth rates”.
Jenny Just goes all-in when she sees opportunity. But going all-in is a function of understanding probabilities produced from our computationally bounded minds.
Market behavior information expands the set of computational possibilities for modeling securities prices. It is the market inventory of U.S. equities. Demand and supply. Like Peak 6, though from a different perspective, we have to understand inventory levels top-down and bottom-up all day, every day.
We’ve created market behavior derivative statistics. We can measure them and evaluate their growth. We can observe amplitude scores mean-reverting. Like Jenny, we don’t need to call tops and bottoms to be effective.
By expanding what we know about the computational possibility of equities, we improve our bounded understanding of market behavior, which improves risk management.
This is what we’re dong with the ideas from the All-In Summit. What will you do?
Despite the two-day blitz through a mind expanding lineup of speakers, the best session may have been the last. Limited to the Besties and their 5th protagonist, Brad Gerstner, it was a letting down of the guard. Like the calm aftermath that follows an intense storm, it was the Besties in reflection.
One reason the All-In Pod works is because each of the Besties is flawed. We know this because we are constantly reminded of one Bestie’s flaws by three other Besties. They understand each other’s flaws perfectly. And they each own their own flaws.
J-Cal, Chamath, Friedberg, and Sacks explained why they produce the pod and what matters to them. Their friendship. And their families. They said all that truly matters when one looks back at their life is their loved ones. So spend more time with your family. Make time for your friends. And tell them all you love them. It is all that matters.
After an exchange of vulnerability, Sacks said, “I think all these emotions on stage are a bug, not a feature”. And in so doing, he expressed, I believe with self awareness, his own flaw, an apparent lack of empathy.
But then the other Besties, recognizing one of their own might need rescuing, explained how Sacks had been up until two in the morning before the final day of the Summit. He had been writing and publishing an extensive thread on X in defense of Elon.
During Elon’s talk, a video-conference using Starlink from his jet, Elon explained the impact of providing internet to Ukraine with Starlink, the only communication network that is functional in Ukraine. He explained the costs to Starlink. And the existential risk to Starlink’s business that has resulted from becoming a war target. Elon explained how not responding to a more recent Ukraine request for extended-territory internet via Starlink was an attempt to keep the war from escalating. For both decisions, to provide Starlink and not to provide Starlink, Elon has been attacked from all sides.
So Sacks went from his best Barbie party beach-themed Ken mode to write a 1,200 word geo-political piece into the morning hours to defend his friend.
My favorite part of the discussion with Elon was Elon’s laughter. Like the pod, it was genuine. Regardless of the topic or its gravity, Elon seemed to be able to find humor in life. And if the man leading energy transformation and space travel can find laughter amid the madness, maybe we can too.
Sacks may think emotions on stage are a bug, but he could not sleep until he had defended his friend. And in the end, there is nothing more computationally irreducible than a friend. In a vast universe we aspire to understand, friends - Besties - are what matter.