Ask the Right Questions and Every Path Ends on the Middle Finger

Lex Fridman introduces Ryan Hall as a systems thinker who is willing to work outside the rules of the game, and he asks how Hall actually solves an opponent, in jiu-jitsu or in mixed martial arts. Hall says he does not have an A-to-B-to-C process. A lot of the serious thinking on conflict, he argues, is already in books people quote and then ignore: Miyamoto Musashi's *Book of Five Rings*, Sun Tzu's *Art of War*, Carl von Clausewitz. You can tell what someone really believes, he says, by watching their actions, whether or not they know they have an internal operating system.

Systems are strong because they work the way they work, and weak because they are readable. If Hall is running a system you have not been read into, especially one that preys on a weakness you tend to show, he can look shockingly effective. The trouble starts when you have read the same books. Then the system makes him deeply predictable. He sees this constantly in jiu-jitsu: people think they are doing the art when they are doing an expression of it, Marcelo Garcia's game, a Renzo Gracie lineage, old Gracie Barra, the Gracie Academy classic, or the Art of Jiu-Jitsu's approach.

His analogy is the iPhone. As an end user he can change the font and pick a blue screen over a purple one. Edward Snowden says the phone is listening; Hall believes him and still has no idea what the machine can do. A car is the same: turn the wheel, press the gas, add oil, change a tire, and beyond that, nothing. He has spent enough years around good grapplers, including a long stretch at Marcelo Garcia's academy, Osvaldo Alves-style jiu-jitsu, and a deep dive into leg locks, to feel the sting of that gap. "I understand your system better than you do," he says, and if it is the other way around, that is a huge issue. Early on he tried to run systems built by people like Rafael Mendes and discovered he was doing a worse, fully predictable version of someone else's game.

Fridman boils it down: you want to be more NSA than end user. Hall agrees. The cook-versus-chef version of the same point is the one he likes best. Give him ingredients and he can make a couple of dishes, not well. A chef knows what extra time in the oven does, what a spice does to taste and texture, and can turn the same pile into Chinese food or Italian food or lasagna. The chef is not locked to a cuisine, because the chef understands how food interacts. Transcend the named systems, Hall says, and you stop being the person who can only change the bells and whistles.

Hall uses rock-paper-scissors to show how little has to change before choice becomes fake. The honest version is rock, paper, scissors, shoot: both people commit together, and a hundred wins in a row would be luck. Change the timing so one person goes on "scissors" and the other waits for "shoot," and the second player is looking at the first player's throw. If you lose that version, Hall asks, whose fault is it? Yours. The game was built on parity. Take the parity away and you are not competing.

He compares it to investigating someone's background before you meet them, then walking in and saying you too love the New Jersey Nets, a sentence nobody in their right mind would have offered when he was growing up. The line only works because you already know. Information, in that sense, is choice. If you understand how games are structured, he says, you start to see huge holes in the thinking behind them, and you can create the illusion of choice while keeping the outcome on a path you designed.

Hall then runs a classroom trick he uses all the time. He asks Fridman to pick three fingers: thumb, pinky, middle. Then two: middle and pinky. Then one: the middle finger. Then one more: the pinky. They play again, and the tree still collapses onto the middle finger. Fridman, narrating for listeners, says that no matter which trajectory they took, the decision tree ended on the middle finger, and both of them felt like winners the whole way. Hall's phrase for it is the illusion of choice. If he is asking the questions, and he asks the right ones, there is no correct escape.

He changes the prompt and asks Fridman to put fingers *down* instead. The positions get awkward. Middle and pinky, thumb and pinky, middle and index, ring and index. Hall still arrives at a good outcome sometimes, but it is no longer guaranteed. He has gone from a situation where he literally cannot lose to one with very low probability, and the second Fridman realizes what is happening, Hall would never be allowed to win, because the ball is truly in Fridman's court.

That, Hall says, is what he is trying to put into play almost all the time. Can he develop the skill, the understanding, the discipline, the patience, and the courage to ask the proper questions and wait for the proper answers? In a perfect world, if he can, he wins, period. Fridman supplies the academic name. Game theory models interaction under a set of rules. Mechanism design is the design of the rules themselves: you build a system whose decision tree feels like a game to one party while the final outcome is already spoken for. The best traps, Hall adds, do not look threatening. Lions do not roar when they want to eat you. They roar when they want you to move back. When they want to get you, they are slinky, "furry water."

Firas Zahabi, Hall's MMA coach and the head coach at TriStar, once gave him a thought experiment he still finds salient. Imagine an impossibly huge million-person bracket with no cheating, decided by coin flips. Someone will have an unbroken string of victory through that bracket. At the end, aliens arrive and offer to flip a coin for the Earth, and that person walks in and says they are good at this. Effectively they were lucky. That is not the same thing as a perk in a video game. Entire fight careers, Hall says, can be spent inside that kind of streak.

Plenty of striking exchanges are dead toss-ups. A fighter gets five knockouts in a row, loses two, and the commentary asks what happened. Nothing happened. They were always flipping a coin. Heads, heads, heads, tails, tails, heads again, and the story we tell ourselves is that we are back on track. Human beings see signs in almost anything, then build a narrative of control. The same story gets told about the stock market and other distributed human systems: we see patterns without understanding the system that generates them. If you can see the system, that is incredibly valuable. Then you have to ask what system sits above all the systems, and which parts you can actually put your hands on.

Strategy without skill is dunking a basketball with no trampoline nearby: a great idea you cannot execute. Hall thinks fighters get so concerned with winning early that they never develop the skills they do not already have, and they take the barbarian route, putting on the one ring and taking the damage they need to take to hit the other person. Watching Israel Adesanya, Floyd Mayweather, Stephen Thompson, and Conor McGregor at their best, he says the point is that it is not a trade. They hit you; you do not hit them. That takes years. Absorbing damage that would put a lesser fighter down is an attribute people lionize, but Hall did not come here to be unable to remember his name at forty-four. The line he likes is Patton's: nobody wins a war by dying for his country.

The finger game makes Fridman think of Sam Harris: maybe the universe is that same tree, and we always end on the middle finger. Hall will not pretend he is equipped to settle whether free will is real. He starts somewhere cheaper. There are facilitative beliefs and beliefs that are not. "I think the world's out to get me": true or not, what next? Probably not a facilitative belief. Imagine you believe there is no free will. Now what? Does that justify every impulse, or is the belief in free will, even if it is only a belief in your ability to work hard and improve your situation, the one that lets a life go somewhere?

He is nowhere near smart enough, he says, to adjudicate these systems on a cosmic level. He can say what happens in a hotel lobby. Walk toward a stranger thinking this will be interesting, and you get one conversation. Walk over thinking "what's this asshole want," and you get another. You might even be right that the person is not polite. That still does not help you navigate toward a positive conclusion. Confidence is the fighting version of the same lever: believing you can win does not guarantee a win, but most people would argue it helps. Depression, as he describes it, is a negative, unfacilitative belief projected onto reality until you cannot get out of the loop.

Fridman pushes the other way. If humanity truly internalized that free will is an illusion, we might stop loading so much blame onto individuals and start seeing society as an ant colony, a mechanism for progress rather than a scoreboard of personal virtue. Hall calls that a reasonable suggestion and a seismic shift, and he will not pretend to know whether it would be better or worse. He would rather make people live in the world they say they want. If free will does not exist and things just move forward, when do you stop a figure like Hitler, who decides which node is the right one, and what counts as a good idea? A ripping change in how the species experiences itself, he says, is an incredibly high bar to prove and an incredibly dangerous idea, because it always comes back to who gets to do the choosing.

Hall says he has been encountering only two kinds of people lately, and he does not mean political parties. One has a libertarian tilt in the plain sense: you do you, what you are up to is not my concern. The other believes what you are up to *is* their concern. History, as he has watched it, is full of people who want to tell you what to do, and much thinner on do-no-harm. He finds it bizarre how many people are so confident their position is not only right for them but right enough to enforce on others. Spreading the idea that free will does not exist may or may not be damaging. He will not claim to see all the implications. He does think it is strange that the species got this far and now wants to take a sharp turn.

Fridman fills in the historical pattern: one way to attain power is to say everything you people are doing is wrong, build a following, and overthrow. Identity politics, in the critique he associates with Jordan Peterson, can start with good intentions, seek power, and become the same kind of devil it was fighting. Over generations civilization still accumulates lessons. The worry, if you strip free will all the way down, is a nihilistic or communist-flavored system in which the individual is no longer something to protect. Fridman puts Hall's own point back to him: it may not matter whether free will is an illusion. The belief that it is real protects the individual from the group. Hall agrees. That is the big issue of history. "There's more of me than there is of you. Deal with it." All the terrible things happen on that imbalance.

Leave-me-alone, though, does not survive a second why. Hall agrees that respect for another person's lived experience should fall out of the same principle, and he also agrees the slogan is not actionable. You use public transit, police, firefighters, shared infrastructure paid for by taxes that are already a social bargain, and some of that distribution is unfair. Empathy takes compute cycles. To avoid hurting the person in front of you, you have to spend energy imagining what it is like to keep a gym alive during a pandemic, to have full-time employees whose livelihoods sit in your hands, to be injured and unable to fight. Muhammad Ali's line, which Hall likes, is that service is the rent you pay for your time on Earth. Giant cosmic problems are mostly beyond any one person. What he can do is not litter, be decent to the people in his circle, and take the next correct step in the ring. He cannot control winning a fight. He can control each individual step.

Fridman asks the beginner question: how do you get good at jiu-jitsu? Hall feels lucky to have found a good room for the time and place, and he drilled a ton. The honest version of his fast rise is that he crammed more hours into a few years than most people put into two or three times that span, so "progressed quickly" can masquerade as something other than volume. There is no way around putting in the hours. Then he adds the part he thinks people skip: your choices matter. Jiu-jitsu has exploded in positions, techniques, strategies, and rule sets, which is cool, and it is also carrying a metric ton of suboptimal material, including things Hall himself taught. Looking back five years, three, two, one, there are movements he would not teach that way and movements he would never do at all.

His most politically incorrect target is the shrimp, the hip escape almost every white belt learns in the first warm-up. "I don't think the shrimp is a real move," he says. It is a giant public seizure: you scoot your butt away, push your hands out, expose your face, and lie on your side because someone told you to, until you look like a shrimp. He did it because a credible person told him to, the way you might drink unleaded gasoline for a while if the right person handed you the can. The next best thing for his life, he realized, was to stop. The shrimp itself is not the whole problem. The problem is using it as a foundation. Junk does not only waste time. It hangs on you like an albatross and changes how you think about movement from then on.

A reasonable beginner cannot easily call bullshit on a teacher, because you cannot call bullshit on what you do not understand, even if you suspect it. What you should never set aside is the obligation to understand why you are doing what you are doing, and not once. Why shrimp? To make space. Why make space? To get away. Why get away? Because he is dangerous. Why is he dangerous? Ask far enough and you discover that three-quarters of the time you did not even need to move; you were acting in the other person's interest. The reevaluation is exhausting and confidence-destroying. It is also how Hall reads his own 2007 ADCC debut as an alternate. He beat a world champion who had a three- or four-inch height advantage and a thirty-five-pound weight advantage. That should not exist. If you are the champion and that happens, you are doing it wrong: not because Hall was so right, but because the variance in the champion's game was large enough for someone of Hall's level to walk through. He won with a 50/50 heel hook, an early public signal that people can attack the legs. John Danaher, the Renzo Gracie team, Craig Jones, and others later exploited a glaring hole in jiu-jitsu's operating system: almost no lower-body defense, which meant almost no understanding of how you yourself would be assailed.

In the absence of knowing what to do, Hall tries to polish what he has, like sharpening a knife he is not yet sure how to use. Building a whole system on guesses produces suboptimal choices. What he has tried to do, over and over, is reevaluate and seek out the best, most articulate people he can find, famous or not. The other half is social. He has always tried to treat people with respect, show appreciation, and be the training partner nobody fantasizes about beating up. Even if he were internally a scumbag, he says, he would still be incentivized to be a nice guy. A black belt who enjoys rolling with a blue belt is much more likely to roll again, and that is a selfish return on investment: shake the hand, say thank you, ask about the thing that just happened, and accept that they are under no obligation to explain.

He was never terribly popular in high school, and he was never going to be rude. If you said hello, he said hello back. He held the door. Simple things go a long way, and they also keep you from pissing off the person who can currently beat the crap out of you and who is, at that moment, the closest thing you have to a font of knowledge. Fridman notes the tension: you have to keep asking why without becoming the white belt who saw something on YouTube and is now an annoying critic. Hall's test for himself is role. "I'm the one that's asking for a favor here." If he spars with Raymond Daniels, Daniels is doing him a favor, not the other way around. Same with Stephen Thompson and Georges St-Pierre. If he walks into Firas Zahabi's gym, he is a guest. Short of being asked to break his neck, the answer is yes, sir.

The internet, he thinks, scrambles that hierarchy. People announce a beef with Joe Rogan; Hall's reaction is that they do not have a beef, they are some random goof Rogan does not know. Access is new. Growing up as a sports fan, he could not send Derek Jeter a message. What still works is ordinary politeness, which he blames on a semi-Southern childhood in Virginia: yes sir, yes ma'am. Fridman says the most successful people he has met, Elon Musk among them, treat him almost as if he were the superior, a genuine beginner stance that does not look like an act. Hall's test of quality is older than any of that. You can tell someone of quality because they treat the king and the janitor the same way, with a dignity they could probably get away with skipping.

Hall's new back-attacks instructional quotes Galileo: you cannot teach a man anything, you can only help him find it within himself. Asked how to learn how to learn, he does not claim a complete answer. He has never had tons of friends, so he has had a lot of time to think. He reaches for Marcus Aurelius: the trick is figuring out what is in your control and what is not. Passing someone's guard is not fully in your control. Breathing is. Calm is. Stepping forward across the line is. No one can stop him from doing those. That is why he knows that every time he walks into the cage he will walk out with his head high. He will fight with everything he has. He cannot promise he will win.

A lot of what people call the basics, he says, are not basics. They are level six, with so many baked-in assumptions that you cannot see why anything is happening. First principles, for him, start with the things you can touch: posture, breathing, where you place frames and limbs, whether you can put a foot on the floor. He has been lucky enough to train with people he grew up watching, Stephen Thompson, Georges St-Pierre, Raymond Daniels, Firas Zahabi, Marcelo Garcia. Sometimes they can tell you what they are doing. Sometimes the genius is intuitive, and they do not see the world the way you do. At Marcelo's academy he wanted A-to-B-to-C. Marcelo would say it feels a little like this, and then just do it. Hall could not yet separate performance from understanding. He thought that if he understood, he could do it, and he could not explain why Marcelo was so elemental, able to summon the thing at will.

Josh Waitzkin, the chess prodigy, push-hands champion, and author of *The Art of Learning*, sat him down and told him he was doing it wrong. He was missing the brilliance in front of him, trying to replicate what Marcelo said instead of seeing the root it grew from, and then getting frustrated that Marcelo would not explain it in Hall's preferred language. The question to ask a genius, Hall decided, is not why they will not talk the way you want. It is where they are coming from. Marcelo did not teach him anything, and in so doing taught him to find it within. The sentence he keeps is that all learning is self-discovery and all performance is self-expression. Marcelo was a master of letting what was inside out. Jiu-jitsu starts as a science and is expressed as an art. It does not matter if you can articulate what you know. It matters if you can do it. Articulation only matters when you are teaching in words. Knowing an idea and never drilling it, Hall says, means it will never see the light of day.

Fridman, working through Hall's instructional, had once thought two hooks on the back meant victory. For his body and his game the hooks were almost supplemental, useful for points, while upper-body control did the real work. People panicked more when the hooks were not even in. A lot of that, he says, he stole from judo, where ground fighting is built around aggressive, fast choking and around transitions. Judo's rules force urgency. You get a short window, so you have to finish the conversion while the opponent is still in the air. If you can put in a submission or a controlling grip during the throw, Hall asks, why would you wait? Stop thinking of standing and ground as two phases. Standing comes before, ground comes after, and everything of value happens in the transition.

Jiu-jitsu people, he says, quote the *Art of War* and then fight like barbarians. One line he actually wants: the only thing you can be sure of attacking successfully is something that is undefended. Even a well-defended fighter has moments all day of being borderline defenseless. If you can see the moment and seize it, you should not only expect to succeed, you should be damn sure, and even if you somehow are not, you will not be countered. That is the trick of almost all conflict: showing up while the other person is taking a nap. Shrink a five-year war with its lulls and its monthly battles into five or fifteen minutes, and you get the same microcosm. There are still moments when the other person has gone for a snack. In a horror movie, that is the character who will not be coming back.

Hall likes the idea that the opposite of a profound truth can be another profound truth. The seatbelt from back control is the example. Jiu-jitsu solved it so well, and the people who used it were so prolific, that the community stopped thinking: one arm over, one arm under, choking arm on a certain side, usually the right arm over the shoulder if that is your finishing arm. Hall's gym started playing with the other side. Walk the hands over, sit on the opposite side from the choking arm, and the position is at least as effective. In some cases he calls it more powerful. That does not make the standard side bad. It means a tactic that evolved from an idea got frozen into a commandment.

Then he goes after a deeper foundation: the guard. Why does jiu-jitsu put you on your back with your legs as a wall? No other martial art developed the guard in the same way. When Hall wants to stand up, he does not want his feet on the other person; he wants them on the floor. When he wants to generate power, the first thing he does with his feet is anchor them. That does not make being on your back a terrible idea. It clearly works. The question is whether the focus assigned to it is disproportionate to its effectiveness. If the foundation is never questioned, decades of innovation get built inside a closed system that the people inside mistake for open space. You do not realize you are in a box. Answers would become immediately apparent if you looked left, but you will never look left, because you do not know left exists.

Most of the innovation he sees is not innovation. It is changing the color of a car or polishing a window: the same guard, plus a lapel grip. Not bad, and not a fundamental change. The big seismic shifts come from "that thing we thought was right was wrong," not from making the triangle a little better than it was taught. Inventing the triangle, or being the first person in a world of jabbers to throw the rear hand, is on a different list. If you are only an end user of your own life, almost everything is unexamined. If you can alter the source code, Hall says, you are Neo in the Matrix. He learned the scale of that gap the hard way. In 2007 he was a decent purple belt and Marcelo Garcia beat him up very gently. In 2008 he won no-gi worlds at his belt and Marcelo beat him up the same way. In 2009 he medaled at ADCC, won the trials, crushed people, and Marcelo was worse. In 2010, as a brown belt, the same. He could not tell which year Marcelo had been kinder. He only knew Marcelo was so far above him that the difference in Marcelo's level was invisible. In the Matrix analogy, Hall was not even an Agent, the best version of someone playing by the rules. He was a regular person who had gotten out, still losing to Agents, while Marcelo was operating outside what other people perceived the rules to be.

Fridman turns the systems talk toward artificial intelligence: a possible new kind of species that lives in information, on another timescale, not embodied as a single walker. Hall does not claim to understand the implications. What unnerves him is radically altering what it means to be human in a very short period, with no real opt-in. People in the developing world are being enrolled in the shift without a vote. The historical pattern he trusts more than any particular motive is power differential. If you are so powerful that there is nothing the other party can do to stop you, that is a recipe for something really not great. Countries that can contend with one another behave one way. An alien species that can reach Earth is a different story. Stephen Hawking's caution makes sense to him. An adult hitting a child is on the table only because of a gross imbalance; nobody in their right mind calls it a great idea, and the imbalance is still what makes it possible.

Nuclear war, he says, was constrained by mutually assured destruction and by cooler heads, and also by the knowledge that you could not hit the other side without being hit. What happens when you can get anyone and remain more or less untouchable? That is the colonial pattern he does not want to re-run. Fridman answers that ethics can grow inside a power gap: parents do not usually beat other people's children, even though they could, and a system orders of magnitude smarter than us might not want to crush the ant colony. Hall is not sure. The awful actions are on the table, and that is enough to be scared. Value alignment, in Fridman's terms, is the research program of making sure the more powerful system still wants what humans want. Hall's AlphaZero example is the speed of the jump: a program teaching itself to beat Stockfish in hours, and with that much of a gap, the ability to manipulate would be significant if the incentive pointed that way.

Near the end Fridman asks how love wins over hate, on social media and in a life that includes complicated people. Hall's preface is that he recently failed to make pancakes, so he is not a sage. What he keeps coming back to is the inability to see humanity in other people, including in enemies. Demons do not populate the world, and neither do literal angels. He has watched people get hoisted by their own petard on Twitter, amusing and awful in equal parts, often for doing the thing they were raking others over the coals for a week before. He has had text-message fights with close friends in which both people were furious because they had misread tone and reached for the least charitable reading. If he cannot see the humanity in others, he asks, how can he see it in himself? Leave-me-alone is only step one. Step two is how can I help you. Everyone, he has read, is a great lawyer for themselves and a judge for others. That is a terrible way to live even if it is an understandable one. Better, if you have to err, ten guilty people go free than one innocent person burn.