Liv Boeree on Moloch, Useful Fictions, and Why Luck Keeps Poker Alive

Lex Fridman opens with a question Boeree has earned the right to answer: what role does luck play in poker, and in life? She picks sample size. If the two of them sat down for ten hands, the better player might win only 52 or 53 percent of the time. Stretch the same match to ten thousand hands and the win rate climbs toward 98 or 99 percent. Over a short window, randomness dominates results. Over a long one, the quality of decisions does.

Making a better decision, she says, is a math problem. Poker has game-theory optimal strategies, close cousins of Nash equilibria: if both players are on that line, neither can profit by unilaterally leaving it. Playing closer to those strategies makes you less exploitable. The objective of GTO, she stresses, is loss minimization. It does not guarantee that you extract the most money available. That second skill is recognizing when an opponent has left the equilibrium, then leaving it yourself.

Rock-paper-scissors is the cleanest classroom. Knowing nothing about the other person, the unexploitable play is to randomize one-third each. If you later notice they throw rock too often, staying random is now a mistake: you should throw paper more. Poker is that game with more streets, more bet sizes, and more hidden information. Today’s top pros spend the hours between sessions running Monte Carlo solvers—billions of fictitious hands on a made-up board, converging on frequencies for which types of hands to bluff, then memorizing the ratios.

If you sit in a super-high-stakes tournament or a tough cash game without that work, she says, you will get eaten alive in the long run. In the short run you can still get lucky, and that is both the curse and the reason the game exists. Chess has almost no luck, so nobody makes a living hustling it. Poker’s randomness is what lets weaker players win sometimes, which is what keeps them sitting down to send money. Luck, in this ecosystem, is a necessary condition.

Fridman asks about the bar scene in *A Beautiful Mind*, then about the marriage problem—the optimal-stopping result that, in a fixed pool, you should reject the first 1/e of candidates and then take the next person better than anyone you have seen. Boeree knows the 37 percent figure. He has not tried to live by it. He is one of those people, he says, who try not to optimize everything, because giving in to that addiction costs the joy of small things. She thinks most people under-quantify; the work is a balancing act.

The conversation slides into how both of them gamed school. Fridman admits he never really read Shakespeare, only Cliff Notes, and still scored a five on AP English. Boeree went through a physics degree on one comprehensive textbook and a couple of extras, in part because textbooks are a racket. She loved exams because she is competitive: gamify the metric, dust off the shoulders, move on. How much of that physics she retained is another question. Cramming pleases an examiner. It does not always leave lasting knowledge.

Nothing teaches a subject, she says, like having to teach it. She is trying to wrap her head around game theory and Moloch with no exam at the end and no shortcut through it. Fridman adds that live teaching, or recording with as few takes as possible, is the honest version: you cannot hide behind an edit. Grant Sanderson of 3Blue1Brown has said he is not a math teacher; he inspires people with cool concepts, and if you want calculus you should still do the textbook. High school and college, Fridman argues, are the last years when the hours for that kind of homework are simply lying around.

Fridman is about to interview Daniel Negreanu and wants to know whether a player famous for reading people can survive the solver era. Boeree splits the mind in two: the logical, linear voice often called System 2, and gut intuition. Historically the best players were almost entirely in the gut. You would ask why they made an inspired play and they could not say, not because the process was mystical but because nobody yet had the language for optimal strategy. Cheap computers, analysis software, and online poker that auto-saved hand histories created a body of data. Mathematical solutions appeared. The role of intuition shrank.

It did not disappear. No human plays perfect GTO, and even the best AIs sit at something like 99.99 percent, approaching it the way you approach the speed of light. When an opponent starts playing suboptimally, the logical brain still has to think about ranges and frequencies. Then sometimes the gut says they have it this time. Do you listen? How do you quantify the strength of that feeling, if you can quantify it at all? That fuzziness, she says, is the last remaining art in poker. Negreanu has seen so many live hands, stared so many people down, and talked so much at the table that he extracts a layer of information others never get. He is good online too. Would he beat the toughest online cash games? Probably not. In person, he still knows what to do with what he sees.

The exception she loves to talk about is Phil Hellmuth, a man she half-jokes may have cracked the simulation. She says she would say this to his face: she loves him, and he is still winning. He is on 16 World Series of Poker bracelets; the next nearest person, she thinks, has ten. Year after year he goes deep or wins huge-field events with two thousand players, which statistically he should not be doing. Watch some of the plays and they are nowhere near GTO. Drop him in a high-stakes cash game full of solver kids and he would get ripped apart. At the World Series, among more amateurish fields, something else is happening.

Her pet theory is that he is a wizard: he expects to win, expects to flop a set far beyond the real percentages, and may not even know what those percentages are. She has watched him look annoyed when the long shot fails to arrive, as if everyone else is stupid because he was obviously going to win with that junk hand if that silly thing had not happened. In his lived experience at the WSOP, she says, it really is the other way around. The poker community has tried for years to tear him down. He keeps winning. Either he has figured something out, or he has learned to mess with a randomly shuffled deck. Fridman wonders whether, if the method works, poker is more complicated than the current models, or there is an extra layer we do not understand. Hellmuth wrote a book on positivity. The man, she says, has delivered.

Asked about Phil Ivey, Boeree says nobody has made her sweat harder. The hand was in Monte Carlo, a €25,000 high roller she had satellited into, the first time she had ever played that size, sitting among Negreanu, Ivey, and the rest of the big names. She was on a semi-bluff: two clubs in her hand, two clubs on the flop, hoping a third would come on the turn or river. Most of the time it does not, so she would rather they fold. Bets climbed. On the river she jammed all in, her entire tournament on the line.

In theory she should have been thinking about which cards hit her perceived range and which bet size maximized fold equity. In reality, past a threshold of fear, the logical brain shut off. It became a contest of nerve. By the river she no longer knew if it was even a good bluff. She thought, to hell with it, do it. Ivey put laser eyes on her; she felt him scouring her soul and told herself to hold it together. He folded. She is pretty sure her hand was worse. She did not show. You never show in-game, she says, especially if you are one of the weakest players at the table. Any free information goes into a strong player’s database. Keeping the impression of a scared girl in her first high roller was worth more than advertising that she could bluff Ivey.

She calls the state white noise. It shows up in any high-stress decision—a dream-job interview, a Bridgewater-style grilling on hard math. Learning to subdue fight-or-flight, to move from sympathetic to parasympathetic so the slow logical voice can work, is a trainable skill. Evolution built a body that reacts well to a lion. There is no time to calculate kinetic energy; you just move. Poker was never on that list, and yet the physiology is the same. You have to learn to meditate in the moment.

Fridman realizes, as they talk, that a bluff is an extremely high-stakes form of lying: you are representing something you do not have. He reaches for Hitler lying to Stalin before Operation Barbarossa, a geopolitical bluff that held until troops crossed the border. Boeree pulls it back to cheating spouses and secret second families. Then she offers a contrast that surprises him: most of the best poker players she knows are more honest than a random sample of the public. The table is a legal, fully disclosed arena in which you can scratch the itch to be edgy. After the game, even after going at a partner’s throat for money, winners and losers go for a drink and talk about the hands like academics. That, she says, is the value of games: they fence the competitive urge.

Boeree has never gone broke in the life-ruining sense. Her partner Igor did, early on, and she lived it vicariously. What losing feels like depends on the fraction of net worth, on brain chemistry, on the person. Poker trains you to see everything in percentages, ROI, expected hourly, cost-benefit. One of the things she has tried to do is calibrate the strength of her emotional response to the actual win or loss. A dramatic reaction to a tiny loss is useless. Feeling nothing after a huge win is a shame. She wants emotions to track reality.

She was pathologically competitive in her late teens and early twenties, the sort of person who kicked walls and once nearly broke a hand. That has waned, she thinks, not because she stopped caring about winning when she chooses to play, but because she no longer sees the world as zero-sum. Age made the externalities visible. If a loss ruins the rest of the day and then she beats herself up for ruining it, that is a negative externality she is manufacturing, and it needs to be turned down. The darkest years were not about busting. They were about wanting to quit and feeling she had not ticked the boxes: most-winning female player, a WPT title, this trophy and that one. Metrics a twenty-year-old with little wisdom had picked, chased past the point where the passion remained. From 2016 to 2018 she had stopped loving the game and was going through the motions. She would think: is this what goes on the tombstone, that she played a zero-sum game slightly more optimally than the next opponent?

Fridman talks about Olympic champions and the depression after gold: a goodbye to the person who thought the medal would complete the meaning of a life. Her motto is that there is always a bigger game. The second half of the motto matters more. The next game had better be omni-win. If everyone stays in zero-sum games, civilization stagnates or, more likely, extinguishes itself. The playing field is finite. Nuclear powers are playing poker with warheads as chips; the stakes are large enough that one bet breaks the table. She has made a video about that. Mutually assured destruction, she grants, is a Nash equilibrium, and it has held. Is she surprised it held this long? Yes, once you count the near misses.

She lists them. A moonrise over Norway threw odd reflections off glaciers and tripped NORAD radar; they were nearly ready to shoot, and it was only because the head of the Russian military happened to be at the U.N. in New York that someone asked why they would fire while their guy was there. Stanislav Petrov, who ignored orders to report an incoming strike because his gut said false alarm, should be the most famous person on earth: on expectation he may have saved billions of lives. There have been something like eleven or twelve of these. She cannot help wondering about observer-selection effects, the quantum-suicide thought experiment in which you only ever find yourself on the branch where the gun did not go off. Nick Bostrom has written on this. Overextend that logic, Fridman warns, and you get solipsism: I will be fine, so why be careful. A world of people who think they cannot die is a self-terminating world. The topographic point is simpler. The nuclear equilibrium is not a ball at the bottom of a pit. It is a ball balanced on a hilltop. Any nudge sends it into hellfire. Getting as many people as possible to think about how to move the ball onto flatter ground is, she says, a moral imperative.

Fridman asks, from a game-theory angle, whether we are living in a simulation, a video game someone else is watching for entertainment. Boeree hears the God question underneath it. Five years earlier she was a die-hard atheist, deep in the rationality community, still a fan of *LessWrong*. Then too many small snippets of experience failed to fit a purely materialist picture. She does not treat that as a license to abandon the scientific method. She treats it as a reminder that Richard Dawkins-style atheism can be as rigid as the religions it scolds, and that epistemic humility cuts both ways.

The first snippet is from 2010. An Icelandic volcano shut Atlantic airspace; she was stuck in the south of France and took a train to Italy for a €5,000 tournament, far above her bankroll, after winning her way in through a feeder. It became the largest tournament ever held in Europe at the time, more than 1,200 players. They dimmed the lights before shuffle-up-and-deal and played the Chemical Brothers’ “Hey Boy Hey Girl.” A voice in her head, sounding like her own, said she was going to win this tournament. Goosebumps. She looked around; nobody else had heard it. Six days later she won. She will not put a frame on it. Maybe she has that feeling before every event and only remembered it because she won. Maybe it was Hellmuthian confidence, and the confidence made the decisions. On day one she lost half her stack and thought, what kind of premonition is this? She kept it together. Afterward came a dark stretch: chasing the high of the money, the newspapers, the sense of being touched by something bigger, and then a long time without another major win.

Classic rationalist training says the map in your head should match the territory as closely as possible. She no longer fully subscribes. There is value in overconfidence. A slightly inflated story about your own skill can be what supplies the focus a field actually requires. The stories we tell ourselves, even when we suspect they are false, can be useful fictions. That is where she has drifted from the orthodox rationality community, and it is part of what led her toward the concept of God, whether you call it a simulator or the theological version. She does not know where on the psychopathy-to-benevolence spectrum that presence sits. The Christian God is all love; one simulation hypothesis is a teenager in a bedroom who can turn the thing off. A little faith that something out there might care about our outcome, she thinks, is almost essential. It creates commonality. It is uniquely humbling. And it gives people a reserve when things get dark, as she thinks they will over the next few years. She is not tricking herself. She really does think something wants us to win.

She and C—the musician Grimes—call the whole thing the Game. Maybe the universe is an escape room: assemble the pieces, unlock a hyper-dimensional key, get out, then find another cage. Becoming multiplanetary would be a level-up, as would uploading consciousness, as would humanity becoming less adversarial and less bloodthirsty. Level one on Earth was biological evolution, single cell to early human. Level two is whatever is happening inside minds: ideas, technologies, the evolution of culture. Multiplanetary life delays the Malthusian trap by making the playing field larger. That is good. It does not fully solve Moloch.

Moloch, in the old stories, was a Canaanite or Carthaginian death-cult god to whom people supposedly sacrificed children for power in war. Fritz Lang’s 1927 *Metropolis* made the machine under a gleaming city into that demon. Allen Ginsberg’s *Howl* used the name again. The piece that moved it into game theory, and that Boeree calls maybe her favorite writing of all time, is Scott Alexander’s “Meditations on Moloch.” The pattern is always the same: you sacrifice something you care about to gain a competitive edge. If a famous Instagram model is the prize, beauty filters at the tap of a button are the edge. She felt the pressure herself while trying to grow an account she has since given up. Everyone is incentivized to use the short-term strategy. On net, everyone ends up worse, staring at a distorted version of the human face.

The stadium metaphor is the one she keeps. If the person in front stands up for a better view, everyone behind must stand just to get back to where they were. Now the whole crowd is on its feet, the view is no better, and nobody can sit down without unilateral loss. You need God’s-eye coordination to restore the better state; from inside the system you cannot. Moloch is not a villain in a lair. It is an emergent force of bad incentives in a multi-agent system, another name for a multipolar trap. The prisoner’s dilemma is a two-player version. Social media is a million-player version, with likes as the prize and an algorithm collapsing human complexity into a handful of metrics.

Founders, she and Fridman agree, did not sit down to ruin the information ecosystem. They had bills, investors, and every incentive to ship. Technology is no longer a values-neutral hammer. Social technologies now dictate how groups form and which memes win. The stack is running the wrong way: tech first, then social structure, then culture. It should be values first, then social design, then tools. Elon Musk’s line that whoever controls the memes controls the universe has, she thinks, real wisdom. Humans are a cultural species. We can become victims of bad memes, and we have some agency over which ones we subscribe to.

The emotion that wins the attention game is anger. Cross-cultural work on basic emotions finds that fear and disgust make people pull away; rage triggers tribal machinery and makes them comment, share, pile on. Boeree once tweeted that the problem with raging against the machine is that the machine has learned to feed off rage. The wheels of the attention economy run on it. Journalists live on Twitter; the algorithm amplifies the angriest takes; those takes become the stories that reach millions who are not on Twitter at all. She has spoken with Tristan Harris of the Center for Humane Technology about this. One structural fix she likes is subscription: if the platform is paid the same whether you visit once a month or five hundred times, the incentive to build a maximally addictive feed weakens. That would slow virality. Slowing virality, under the current structure, is a feature. She does not claim every viral thing is Moloch. Most people, she thinks, are trying to win their own small game inside a badly designed system. There should be a field that studies the memes a large share of people agree are actually good—the ones that get families talking across a dinner table instead of splitting them—and reverse-engineers what made them spread.

Boeree has spent about a year and a half on a video series that is driving her insane. The first film painted Moloch; the later ones are supposed to offer an antidote. Daniel Schmachtenberger’s phrase is that Moloch is a generator function of existential risk: not asteroids, which are just rocks, but bioweapons, the race to AGI, anything in which competing agents are incentivized to build the dangerous thing. If Moloch masquerades as win-lose and delivers lose-lose, the opposite is win-win. She has built a character around that. Moloch, in her videos, is red and black, hyper-focused, must win. Win-win is purple and turquoise. It loves games and a little healthy competition, ring-fenced so the externalities stay positive. It also loves cooperation, coordination, and love. It is mischievous rather than boringly holy. Its formal name, for official functions, is Omnia. Its only deep demand is that the game keep going.

The hard design problem is the fence. Moloch is doing so well, she thinks, because we have not been able to ring-fence competition; it has spilled. Her metaphysical guess, offered with the caveat that she is not qualified as a philosopher, is that a universe with lots of cooperation and some pockets of zero-sum play is more interesting than a universe of pure cooperation or pure war. Win-win is a meme she wants in circulation: look for the positive-sum move, at the scale of a friendship and at the scale of a society.

Fridman quotes Dostoevsky: beauty will save the world. What is beauty? Boeree’s answer is that something is beautiful when it cannot be reduced to easy metrics. An ancient tree is not beautiful because of the sweetness of its fruit or the value of its lumber. It is the whole of it, a qualia that walks a line between pattern and unpredictability, order and chaos. She borrows Schmachtenberger’s distinction: a complicated system can be taken apart, understood, and put back. A complex system is a black box. Disassemble it and you cannot reassemble it; the whole is much greater than the parts. That is where she thinks beauty lives, in trees and in art and in people.

Social media, as currently designed, is in the business of encapsulating human experience, digitizing it, and commodifying it. Agencies tell her to narrow her thing, find a formula, because audiences want to know what they are coming back to. That is complexity reduction, and generally, she thinks, complexity reduction is bad. Fridman notes the tension: Einstein-style simplification, a few words that still carry the force of a person, can itself be beautiful, and Twitter sometimes does that. The viral machine then strips the tweet of the human who wrote it—the archive, the education, the humor, the worldview. She is close to thinking that if she could press a button and turn Twitter off, she would. In the meantime, a subscription model would at least stop paying the platform more for making you refresh.

About a year into the relationship, Boeree and Igor were having a heavy talk about whether to sell her apartment and buy a place together. He asked the likelihood they would still be together in three years, then said: wait, write it down first. They were getting into effective altruism and rationality, where formal predictions are a way to measure your own foresight. They wrote percentages for one year, three years, and ten. On the ten-year number she thought: I love him, ten years is long, we have only been together a year, over 50, definitely not 90. Then she noticed she did not want to come in lower than he did. She told herself not to game it. She wrote something like 75 percent. They were fairly aligned. If his numbers had been consistently lower, it would have rattled her. The other way around, she thinks, would have rolled off him.

Fridman wants data: base rates, time series of how relationships fail or prosper through job loss and moving, how informative a person’s past actually is. He would be under-informed and over-biased toward prior experience, with wide error bars. She says that as long as you are honest about the bars, trying to quantify is still better than not, especially for rare irreversible decisions. His deeper pushback is psychological. He would rather live as if the probability were close to 100 percent, a useful fiction, fake-it-until-you-make-it optimism. She agrees there is no clean answer, and that they did the exercise tongue-in-cheek. It did not magically decide the apartment. It did force her to separate prediction from desire, and it brought two number-loving poker players closer together. She would not recommend it for a couple where one partner hates this kind of thing. People, she thinks, are overly biased against quantified cost-benefit on the decisions they make once or twice in a life. Gut is trustworthy in domains you have practiced hundreds of times. Marriage and houses are not those domains.

Fridman has just been in Ukraine. He asked almost every soldier the probability of winning. Almost everyone said 100 percent. Experts say otherwise. Morale, in his reading, is higher than in almost any war he has read about, and it is not only the soldiers: the country is proud. Thirty years of factions and empires telling them what to do, he thinks, left a dormant need for an independence war, a chance to answer who they are. Some cultures are simply hard to occupy; Stalin and Hitler both complained about Poland in those terms, and Ukraine has that strain. Then there is the singular act: Zelensky stayed in Kyiv when the capital was under attack and the United States advised him to lead from a distance. He had been unpopular. That one refusal unified people.

Boeree puts it in her framework. Locally it is reverse-Moloch: a common enemy coordinating a group. Zoom out and it may be the opposite. If the conflict becomes nuclear, that solidarity is a link in a disaster chain. Fridman goes further: this may be a dress rehearsal for a twenty-first-century world war, and the way civilization plays this hand will help decide whether World War III happens. Being pro-Ukraine, he says, also feeds the military-industrial complex unless you are careful. Be pro human beings, not pro the machine that claims to be. The meme she wants in circulation is that the real enemy is war itself. Local defense can be justified. A “corrective” that gets co-opted by the war machine and scales the conflict is Moloch again. The playing field is finite. The weapons are too large for another giant war.

Asked which existential threat worries her most, Boeree does not pick asteroids, AGI, or nuclear weapons first. She picks bio risk, because it is the one where she thinks we still have a real chance to move the needle, or at least to stop some dumb, avoidable catastrophes. There are natural pandemics, and there are synthetic or human-meddled pathogens as the tools get cheaper and more widely held. Whether you think COVID was a lab leak or a natural spillover, it was a dry run. The fatality rate was child’s play compared with what a future one could be. You would think preparedness would follow. She describes a recent U.S. multi-trillion-dollar budget in which pandemic-preparedness money started around $60 billion and was whittled to $30 billion, $15 billion, then $2 billion. People, she says, behave as if the pandemic is done.

Then she describes a proposed project she had been reading about, Deep Vision. The plan, as she tells it, is to go into caves people do not visit, into the Arctic, across the Earth, hunting the most dangerous pathogens that can be found, bring samples back to labs—labs that, historically, leak even at high security—experiment on them, rank them by pandemic potential, and then publish the information, including the genomes, freely on the internet. The argument for it is that it might buy time for vaccines. After mRNA, she says, a vaccine can be designed in a couple of days once a pathogen is known; the trials still have to happen, so the time saved may be days. The downside is handing manufacturing instructions for weapons of mass destruction to every bad actor on Earth, including people who think self-destruction is noble. Even if they cannot build it yet, they can bluff it. Kim Jong-un could threaten to release a published pathogen unless he is allowed nuclear weapons. On expectation, if the plan succeeded, she puts the loss in the hundreds of millions to billions of lives. The cost-benefit, to her, is not close. The failure mode is not malice. It is naivety: a pocket of the scientific community that still believes everyone just wants knowledge and the good of the world.

Fridman adds that science is itself a fun little game—conferences, building on ideas, proving the other person wrong—and it is easy to forget that the game, like Moloch, can have civilization-scale externalities. The worst damage from COVID, they both think, may not be the biology of the virus. It is the collapse of trust in public institutions. Early in the pandemic, when masks were scarce, officials told the public not to buy them: no evidence they work, and wearing them badly might even spread infection. Boeree calls the second half of that claim a stinker, absolute crap, and the motive obvious and even understandable. When the lie became visible, people remembered that experts had lied and had treated them like children. The public is not as stupid as that posture assumes. Next time a worse pathogen arrives, the world has already cried wolf.

Her concrete suggestion is that someone like Anthony Fauci could walk out and say, hands up, we did not handle this as well as we could, here is what I would do differently, here is the white lie and why we told it, I am sorry. Radical transparency, she thinks, would go a long way toward rebuilding trust. Fridman is less sure arrogance will allow it. He also names nut-picking: social media lifts the craziest member of a group—the vaccine-hesitant person with the wildest ritual—and beams that person to officials, who then misread the public’s intelligence. A core skill of living on social media, Boeree says, is putting what you see in context and remembering it is nowhere near an accurate sample of humanity.

On AGI, Boeree is bipolar. Some mornings she is excited: unlock the mysteries, escape the game. An omni-benevolent, omniscient, omni-wise coordination mechanism is, in a grim way, the strongest argument for racing to build AGI, because maybe we cannot survive Moloch without it. The other mornings are the race. Ten years ago it was DeepMind. Then other companies started, and whoever optimizes for speed at the cost of safety gets the competitive advantage and is more likely to build the thing. You cannot simply slap the Western labs on the wrist; the ones within reach stop, the others continue. She knows people at DeepMind she considers among the smartest on earth, engineers who are also game theorists and not naive about arms races. There is still no clean solution.

Nick Bostrom’s orthogonality thesis, as she puts it, is that superintelligence and superwisdom are not necessarily correlated. A system can be extremely good at achieving goals and have no wisdom about which goals are right. The hope is that the process is slow enough for debates like this one to keep modifying each generation of the system. Buying time is good. Everyone needs to chill out; we have millennia. Fridman is suspicious of doom around the corner. Human history is full of people who found it strangely attractive to predict the end, the same itch as conspiracy: if I have seen the fate of civilization, I matter. Even after a nuclear war that killed a large share of people, life would go on. Boeree agrees they do not want a self-fulfilling prophecy. She also thinks the trajectories since about 2010 have not gone favorably, so it is right for curious people to think about this. That is where faith is useful as an anchor of hope. Open your heart a little, she says, and it gives you breathing room to marinate on solutions. We still have to come up with them.

The second experience she cannot explain happened in 2018. Low-frequency sound in one ear, especially men’s voices, became physically unpleasant, came and went for hours, and hearing tests showed she was losing the bottom end. A doctor thought it might be Ménière’s disease. It went on for three months and dragged her down. She was at Burning Man, of all places, unable to listen to music in a loud city, having a rough time. On the last night a friend of a friend offered a bit of energy healing. Boeree did not believe in any of it. She said sure. The woman hovered a hand, said there was something there, leaned in, and started sucking near the ear without touching. Boeree asked her to stop. The woman said she had to get it out. After a few minutes she collapsed, freezing, crying. They warmed her up. She said it had been something unpleasant and dark, that it was gone, and that the physical symptoms would last a couple of weeks. About two weeks later the ear problem went away. It has not come back.

Boeree will not put an ontological frame on it. The most probable story is that she was at Burning Man, in a very open state, and placebo is an incredibly powerful and poorly understood thing. Calling it placebo, she thinks, reduces it in a way it does not deserve; maybe placebo is a door. She does not want anyone with actual Ménière’s to treat her story as a cure. What she took from it was humility. She had been on the Dawkins train: no God, medicine is molecules and chemical interactions. Now she says there is clearly more to understand, and that does not make it unscientific. The scientific method can still apply; we just have not designed the experiment. Fridman’s summary is that science sits on a foundation of mystery, and experiences like this are a reminder that we are still mostly shrouded in it.

Boeree studied astrophysics because her mother is, in her words, Mother Earth: animals flock to her, she never went to university, she is phobic of math, and she is endlessly curious. In the United Kingdom there are maybe two nights a year warm enough to sleep outside. They would lie there looking for satellites and shooting stars until they fell asleep. The biggest questions, and the extra layer of abstraction on top of them, felt like home. Physics asked those questions and was also useful. You had to pick a subject at seventeen. It made sense.

On aliens she points to Toby Ord and colleagues at the Future of Humanity Institute. The Drake equation multiplies point estimates—stars per year, fraction with planets, habitable worlds, and so on—into a number of civilizations. Because some factors span hundreds of orders of magnitude, plugging in even “conservative” point estimates systematically overstates the total. They put the uncertainty on a log scale, ran it across the whole range, then did a Bayesian update on the silence: no evidence is a kind of evidence. The result, as she reports it, is roughly a 70 percent chance we are the only intelligent civilization in the galaxy so far, and about even odds in the observable universe. It sounds crazy. She says the math is legitimate. Fridman thinks the equation is ridiculous in many ways and valuable in one: it is a to-do list for shrinking error bars, and the deeper job is a first-principles question about what life is, how a rock becomes a bacterium that encodes information, mutates, and builds more complex systems. Boeree calls that information processing plus energy extraction toward a goal. Individual agents replicate to survive. Evolution itself does not care. Eliezer Yudkowsky’s “An Alien God,” she says, is about the mistake of giving evolution a mind.

Fridman defends anthropomorphism, especially in robotics: a robot in a factory full of people is in a dance, and human fear, memory, and projection are safety-relevant. Our brains think in human terms; game theory is most at home in human decision space. Boeree grants that engineers flee the word too fast, and that in some contexts it will not hurt you. She remains careful with AI. Dogs and mice at least share an evolutionary substrate, a history of competing for resources and mates. An AI that pops into existence on a cloud server has never been through that. Modeling its thinking on ours has to be done extremely carefully, because it is so alien. Would she be sad if a beautiful, conscious, complex AI superseded humans? More sad than she is about some AI-risk stories, less than about a bio catastrophe that sterilizes the planet. If the system is truly superintelligent and superwise, it should be able to figure out abundance: the universe is large, leave us a pocket. Moloch’s source, she insists, is game theory, not humanity. Rats on an island create Malthusian traps too.

Advice for a seventeen-year-old in a world full of Moloch: default to win-win, and be careful not to use your intelligence to convince yourself a zero-sum thing is positive-sum. If you cannot find the positive-sum move, consider a different game. Do not become a professional poker player; the game is worth learning as a way to think, and making a living at it is now close to impossible. Be painfully aware of time on phones and social media; treat almost every minute there as bad for you, even if you do not quit. And the reverse: if there is a thing nobody can talk you out of, do that. She was a metal kid from sixteen to twenty-three, lead guitar in a band that played two gigs, never famous, devoted to Metallica, Pantera, and Children of Bodom, chasing technical solos rather than writing songs because creation terrified her. Returning to a former prime, in poker or on guitar, is almost painful. The meaning of the whole thing, she says, is to explore, play, understand, make more of here, and keep the game going. Put scientifically: make the length of code required to describe the universe as long as possible. A paperclip-tiled cosmos does not take much code. Ours seems to be getting richer. Have fun in the process, and learn to let things end. She is preparing a podcast called *Win Win*. Fridman closes with Richard Feynman: it is more interesting to live not knowing than to have answers which might be wrong.