Science cannot prove that the Earth is not flat

Lex Fridman introduces Michael Stevens as the creator of Vsauce, then one of the most popular educational channels on YouTube, with more than fifteen million subscribers and more than 1.7 billion views, and as the maker of three seasons of *Mind Field*. He starts not with a viral riddle but with psychology: human behavior is messy, far from the rigor of theoretical physics, so how do you get from the twentieth-century tangle to something precise? Stevens's path is to find psychology's physical foundations. If moods, feelings, and actions are mechanical behaviors of atoms in a brain, can you find the correlations, even if chaos and the uncertainty principle make it nearly impossible to know every quantum state's position and velocity?

Reach that point, he says, and you can start talking about consciousness in physical and mathematical language, including the old puzzles of self-reference and what kind of structure can think about itself thinking. Lex offers panpsychism, the idea that consciousness is a fundamental piece of reality, almost a law. Stevens is guessing, not claiming expertise or divinity, and his guess runs the other way: consciousness is probably something that can be fully explained within the laws of physics. Bodies, brains, and the quantum world are rich enough that he would be surprised if there were no room for it there.

Why we should be conscious at all is, he thinks, a strange and important question that the next few thousand years may have to answer on faith. His own conjecture borrows Julian Barbour's time-capsule states: in the configuration space of possible universes, some arrangements contain memories of others, so you have a scratch on your arm and, in the same state, a memory of the cat that made it three days ago. Those states make you feel that time happened and that the future is unknown, because they do not contain information about what comes next. Either such states are more plentiful, or they are rare and the anthropic principle does the rest: you can only ask why you are conscious from inside a state that lets you ask. Things without consciousness do not pose the question.

Lex hears a deterministic universe in all of this: fixed laws, initial conditions, the same unrolling every time along our line. Stevens calls that a very useful way to think, the kind of picture that got people to the moon. He will not say he believes in determinism in that absolute form. Maybe it is true. He is not going to live as if it were, and he does not much care.

The gap between perception and reality is, in his speculative view, a question with no experiment attached. Everything you experience is an event in a brain: looking at a cat, you cannot prove the cat is there, only that the perception of cat is happening inside you. "I am only a witness to the events of my mind." The useful inference is that the cat is literally there, with feelings and motivations, and should be petted, fed, watered, and loved. That, he says, is how you should live. Whether we are in a simulation or a brain in a vat, he does not know.

He does not really care about the ontology either, except as a door into physics, psychology, consciousness, and philosophy. At the end of the day the cat still needs to be fed or it is a dead cat, and neglecting a Tamagotchi does not feel like neglecting an actual animal, a psychological difference that is itself worth studying. The simulation hypothesis is useful now because virtual reality and games keep getting more complex, unlike a Newtonian audience asked to imagine a clock complicated enough to make love. You cannot prove the hypothesis either way, which is exactly why it is a fantastic way into the deepest questions, including the moment a little robot friend becomes someone you would not cancel plans with.

Lex calls him a great scientist for touching questions others flinch from. Stevens is not sure he is one. Science is a way of knowing, *Mind Field* did run experiments with hypotheses, and many of the questions he investigates are not scientific questions at all. He would rather describe himself as curious, and he hopes the curiosity is contagious. Asking a good question, even one society files as currently out of reach, is in Lex's framing the biggest step of the method; Stevens meets him there by building a deliberately absurd theory of force.

Suppose, he says, that he does not believe in forces: when he pushes a massive object, a ghost leaves his body, enters the object, and gets lazy around massive things, with laziness proportional to mass. Prove him wrong. You can never find the ghost, so the theory collides with no observation and is not scientific. The moment he asks whether he can see the ghost, why there should be one, and what he would expect if there were not, he is designing tests, watching predictions against results, and thinking scientifically. A picture of a black hole is a photograph, an image, a sensory experience. Science is how we got it, how we understand it, how we believe it, and how we reduce uncertainty around what it means.

He is impressed by scientists every day and still thinks revolutions happen only under special, scary circumstances. If you defend the ghost theory by ridiculing every rival, ego walks in and you go nowhere; the question of what a force actually is still has to be asked, just in a politically mature way that treats everyone as learning together. Galileo had problems with the authorities. Descartes, in Stevens's telling, more or less agreed and chose to write other things so as not to get in trouble. Knowledge still inched forward. Revolutions are difficult in every field, and certainly in science.

In a Reddit AMA Stevens named "Is the Earth flat?" as one of his favorite questions ever answered. Living in New York in his late twenties, before the idea was as mainstream as it later became, he wandered into Flat Earth Society forums and read how members answered the usual objections. He could not tell how many actually believed the planet was a disk and how many were trolling. That, he realized, was the fascinating part: how do we know anything, and what makes a belief good rather than merely tenable? A flat-Earth believer can answer every criticism in an ad hoc way without the rebuttals forming a cohesive, non-contradictory whole. The video leans on Occam's razor and on Newton's flaming laser sword, the demand that a claim be distinguishable by experiment.

He also hands the idea a real concession from special relativity: space contracts along the direction of motion, so to a particle approaching Earth near light speed the planet is flattened, and to that particle Earth is flat. Wild ideas are still thinking, still communication, still a human act, and Lex likes the topic because nobody gets hurt in the conversation while scientists, in his test, roll their eyes too fast without sitting with the arguments. Stevens goes on the record: the Earth is not flat, it is a three-dimensional spheroid. He does not know that in the sense of proof. "Science doesn't prove anything. It just reduces uncertainty." Could it actually be flat? Extremely unlikely, which makes the notion ridiculous if you care about probability, and still not a mathematical theorem. A rocket-eye view of a ball does not rule out a simulation, a brain in a vat, or an elaborate ruse. The leftover doubt is exciting, and practically useful once you are talking about quantum mechanics or string theory.

The thought experiment of walking on a disk accidentally teaches gravity, relativity, and geometry: on a sphere you fall toward the centroid, on a disk you get pulled at a weird angle near the edge, and anything much larger than the Moon wants to be a sphere anyway, so you cannot have a teacup the size of Jupiter. Lex notes that this episode drew more dislikes than usual. Stevens has not looked at the numbers and guesses two audiences: people who disliked a title they took as propaganda, and people who wanted him more dismissive. Either way he does not want to lose the unorthodox by calling them idiots. How idiotic are they, really, when the Earth is not a sphere at all but an oblate spheroid, a fact he chased in *Which Way Is Down*, where down does not point at the center and the whole universe tugs? In small rooms we pretend gravity is uniform. It is not, and satellites already use the resulting torque to align without fuel.

Asked where artificial intelligence stands and what it would take to build human-level or superhuman intelligence, Stevens starts at the foundation: he does not know what intelligence means, and that is his biggest question at the moment. How do you measure it in a machine, or even in a human, once you stop taking the fuzzy thing for granted? A chatbot you would have a beer with might pass a Turing test, and he is impressed by almost everything, but if someone then pulled the curtain on wires he would not feel like a murderer for destroying it, more like totaling a really cool car. Give it movement, or have it scream and resist as if in pain, and the category flips almost immediately from physics to psychology. Even if you do not believe it has the qualia of suffering, not wanting to hear it cry is a good instinct, and a dangerous one, because robots can work it. Facebook and YouTube recommenders are robots of that kind.

Existential risk worries him, especially weapons. One famous picture of empty skies is that civilizations hit a destructive technology, radioactivity among them, and do not survive the hurdle, an idea he remembers from a YouTube channel whose name he cannot recall. On AI specifically he is not very worried, because plenty of other people already are, and the worries often sit in the road. In a bonus scene from his trolley-problem episode he asked an autonomous-vehicle engineer what a car should do in a contrived swerve-and-kill-a-child case, and the engineer asked what a human would do. Resist progress over every little issue and you get a stumbling block. Lex pushes back with Elon Musk and Sam Harris, and with nuclear-deterrence people, all of whom say the worried are too few. Stevens does not know the right number. People have feared new tools since Plato objected to writing on the grounds that it would atrophy memory, and the conversations are still worth having because we learn about ourselves.

A conscious, self-replicating AI is something humans already make every day. It is called a baby. He does not want a world that refuses births because the child might become mean, or that answers an eighty-three percent genetic chance of psychopathy with "let's not have them live." He wants an answer in which that life still deserves dignity, decent treatment, and rights, knowing he is inside an American value bubble, having been born in the Soviet Union, where nation and society could outrank any individual. He also wants to retire the fantasy of a pure human organism without technology. Glasses, writing, spoons, and cups would not look like inventions to a landing alien; humans and their tools are one organism, which is why the question of where you end and the car begins gets less obvious the longer you sit with it. Worry is warranted and the merger is inevitable. It may even be silly to picture ourselves as separate intelligences rather than as one living creature, civilization and every organism on Earth intertwined.

Musk's fandom, an endearing nerd with an uncommon following, excites him, and the power comes with a question he would actually ask if they met, not about rockets or tunnels: how do you feel about the responsibility? Fans that large are almost bigger than the person, and the psychology of becoming a symbol, of people taking you earnestly and defending you, can be dangerous for society and for the individual. Lex throws it back: Stevens is a funny, curious guy with a gigantic educational following. He answers with humility. He is not at the bench; he is sharing the work of real scientists and celebrating how they think. He wants it always clear that we do not have all the answers and will probably never hold an equation you plug tomorrow into. Believing in science as a faith makes people elitist and hard. It should humble you and make you feel small, weak, and in need of other people. When episodes open with "this is Michael here," Vsauce and Michael are different things in his mind. He is plugging into a socket millions already share, hoping to encourage curiosity, responsible thinking, and being all right with doubt.

Lex is about to talk with Cristos Goodrow, YouTube's vice president of engineering for search and discovery, and asks what Stevens would want to know, and how he would change the algorithm. Stevens thinks of it as a mirror. It reflects what people put in, individually and in aggregate, and we do not always like the face. Recommendations are not even "what you want" so much as what you will click, watch some of, and stay for. He doubts YouTube cares if you last a second, as long as the next move is another video; closing the app is the problem, because this is not a twenty-dollar-a-month subscription. The question he is actually curious about is whether they optimize for your longer-term development as a person. Short-term dopamine and cat videos can end with the feeling of waking from a drug and needing to quit. Making you feel like a better human after the session might be the better retention strategy.

He has been lucky enough not to worry much about the algorithm. Look at the thumbnails: he does not go wild, and on *Mind Field*, partnering with YouTube, he often asks them to pull back and stay mysterious. If everyone else is doing crazy Photoshop, his move is aligned lines and a title that is just a word. The Vsauce channels have cultivated an audience that would rather watch a video called *Stains* than *I Explored Stains, Shocking*. Other audiences want the second kind, and most people, he thinks, want what the algorithm already favors: mainstream celebrity and news. That is the difference from Netflix. Nobody opens Netflix to find out what happened in the world, but they open Twitter, Facebook, and YouTube, and they will click Ariana Grande reading a high-school love letter. A video titled *Law and Causes* with math lines looks like a whole lesson you do not have time for on the bus. Before you get furious at YouTube, admit it is reflecting human behavior.

The system is a black box even inside the company: employees know what they tweaked, then the model learns, and sometimes they shrug and wonder whether to change how much watch time matters. He misses the days when a like from him, say on a Potassium video, put that video on his subscribers' home and could send a hundred thousand views. Following someone should mean seeing what they like, letting them curate, the way Twitter and Facebook sometimes do, annoyingly. Education, unlike old television, is a rising tide: watch Numberphile and you become more likely to watch Vsauce or DONG, because the two shows are not airing at the same hour. Collaborations take work. Comments, likes, and subscriptions should be easier and more powerful. Cat videos and even the dumbest vlog are still human, as a Unamuno line about cats calculating jumps more often than crying is meant to remind him, and he does not claim a better right to the slot than they have. Carpool Karaoke is an extremely important part of being on Earth. He just wishes the people who find him also found the next small channel he had found.

Does he think about his own death? "Yeah, every day. It's really scary." Mortality makes him realize he is destined to become a notion, a memory. Life extension is exciting and still cannot protect you from an accident; a saved connectome, a digital rebuild of consciousness, could be lost with the medium it sits on. The cool part, for him, is that someday there will be no Michael who can step forward and say that is not what he meant. People will have to guess, and the version that lives on as memory may not even be who he wanted to be. There is power in that, and power in letting future people run the show. He is glad to get out of their way and say it is their world now.

When he says memory, he does not mean a name, a birthday, a Wikipedia photograph. All of that can vanish. He hopes people keep asking questions and teaching concepts in ways he found useful, and if they never know he tried to popularize them, that is fine. If Earth were burnt to a crisp today, the universe would not care, and Jupiter would not say oh no, and that could happen. We can still launch things into space to extend how long the record lasts. Recording the world, learning, writing stories, and preserving them is, to him, the essence of being human. "We are autobiographers of the universe and we're really good at it," better than fossils, better than spectra, collecting much more detailed memories, much better data. That should be the species' legacy, and he hopes something like it is his, while knowing you cannot fail to have an effect and that the effect might look really small, which is all right.

One of his favorite lines, from *Tess of the d'Urbervilles*, is that the measure of a life depends not on external displacement but on subjective experience. If he is happy and the people he loves are happy, can that be enough? If so, excellent. Fridman leaves him, and then the listener, with Einstein: the important thing is not to stop questioning.