Aella's 500,000-person survey, and why she still fears looking stupid

Lex Fridman introduces Aella as a sex researcher who runs some of the largest human-sexuality surveys in the world, a sex worker on OnlyFans and in escorting, and a prolific author of Twitter polls. Her blog is knowingless.com, which is a methodological stance in a domain name. He quotes Jack Kerouac on the mad ones who burn like Roman candles, then asks whether a conversation that can go anywhere feels exciting or terrifying. More exciting, she says. On the drive over she was still nervous — afraid of saying something dumb, of answering in a way that made her sound less capable than she is. She is especially afraid of being perceived as stupid.

Novelty, Fridman suggests, almost always arrives with uncertainty. She flinches at the word *always* and starts hunting for a counterexample. If something is truly new, it sits outside your distribution, so you cannot fully understand it; the uncertainty is real. People do not experience that as uncertainty. They experience it as freshness, and freshness is attractive. The fear of the conversation going badly, she noticed in the car, was part of why it felt exciting. If she knew it would go great, some of the charge would be gone.

Her risk tolerance is uneven in a way that looks inconsistent from the outside. She drove over carefully because the roads were bad; she does not like driving fast. At the same time she will put writing in front of a wide audience that might hate her. The monkey brain, she says, is extremely sensitive to lots of people yelling at once. The worst embarrassment is not the yelling itself. It is publishing something she failed to be properly skeptical of, then watching other people catch the hole she missed.

Fridman is a fan of academic psychology and often disappointed by it: tiny samples, oversized claims, news write-ups that inflate the result again. Aella is researcher, scientist, and reporter on the same blog. Her samples are often gigantic — Fridman mentions more than 500,000 people on a detailed sexual-preference survey — the method is sitting in the open, the data is sitting in the open, and the interpretation is raw. He does not understand why that package gets attacked. She thinks psychologists are terrified of exactly that transparency. Layers of institution, peer review, and withheld data function as a cloak of authority, including the polite fiction that most psychology studies replicate.

From the outside, she says, people see white coats, credentials, and jargon and assume that is what doing science looks like. She is still teaching herself, still learning as she goes, and does not claim to be an expert statistician. What bothers her is the other half of the picture. A lot of what those people are doing is not that hard. Curiosity, a question you want answered, asking people, writing the answers down, then learning how to look at the pile: that is the job. Approach it humbly — *please, let's figure this out together* — and it is straightforward. Many people cripple themselves by treating the work as the property of fancy experts. That makes her sad.

She is drawn to questions you are not supposed to ask, which does not help the controversy. Sex is something almost everyone thinks about, talks about, and does; it is a candidate driver of civilization; and rigorous discussion of it is scarce. The topics come pre-charged: whether pornography damages society, how sex maps onto gender, whether some preferences are themselves immoral. People get angry. The sadder problem, she argues, is that sexual urges are unusually elusive to introspection. You can get somewhere asking why you want to be liked or where an insecurity lives. Ask where a fetish came from and you mostly get a black box.

The common move is to scan childhood for a matching scene and build a causal chain: that event produced this desire. The chain reads well. Reading well is not the same as being right. People are biased toward concrete, coherent stories of how events unfold, and that bias is a bad tool for this subject. Fridman later names the split in their styles: her thinking often feels poetic and aesthetic; his instinct is *how do I design an experiment to test this*. She does not treat the poetic frame as inferior. It is useful for meaning-making. It can change how the light hits the world and make you happier. On questions we are still far from measuring, philosophy may be the language we actually have.

Fridman reads back a tweet of hers: she does not understand how to have a normal in-person conversation if she is not on drugs. Fine, he says, they will not have a normal one. She has taken a small amount of a nootropic and is not sure that counts. He is on caffeine. Both of them are mildly drugged, and neither of them wants a normal conversation anyway. What is normal, he asks. The tone of the episode is set: they will use their actual states as material, not pretend to be a balanced news report.

Fridman asks which topic she would spend a thousand years on if she had to pick: sex, love, power, or death. She will not pick. Everything is fun. She cites David Foster Wallace on the real skill, which is to be unborable — to make sure nothing is boring. She had forgotten the glass on the table existed. Remembering it is delightful. Someone made that glass; the curve of it bends light; the water bends light again; she can see part of his arm through it. That is amazing. Deep agony, he asks — is that fun too? Yes. Heartbreak is evidence that she is capable of that intensity. We are all in the gutter, some of us looking at the stars; the gutter is why the stars look brighter.

Fridman is reading *Brave New World*, a society that flattens life's ups and downs with drugs and casual sex. He thinks you need the dark in order to have a concentrated feeling for another person. She amends her own vocabulary. *Fun* is too light. Love is the better word. She loves that glass and the water in it because they are genuinely extraordinary. If he had to reverse-engineer the human condition through neuroscience or AI, he wants a narrower poll: sex, power, or death. She still will not give him a clean primary driver. Life, she notes, is more interesting than the clean simplification of self-replication.

The four-way poll — sex, love, power, death — is a Twitter artifact, not a complete map of reality. Nietzsche gives you will to power; Freud gives you sex; Ernest Becker's *The Denial of Death* gives you mortality. Humans know they will die, the argument goes, and they build empires, religions, and careers as insulation. A previous guest of Fridman's kept that book on the bedside table. Aella's test is simple. If fear of death defined civilization, you should see large behavioral differences between people who believe they die and people who do not. Fridman notes that religions, on this theory, are themselves one of the myths built to forget. She is not buying the argument as a proof.

She is a data point in the other direction. She used to be a very committed Christian and had no terror of death. After the faith collapsed, a heavy, specific fear of dying arrived for several years. The two states felt different enough that she does not need a statistic. She also does not treat that as a license to skip measurement. Terror management theory tries to turn the philosophy into experiments: prime one group with death, leave the other alone, and watch judgments change. Those priming papers, she says, usually fail to replicate. She ran a priming study of her own and found no effect. Fridman jokes that she has PTSD from psychology papers.

She will not torch the whole field. Freud on the subconscious is interesting as philosophy. As science it has to survive replication. Both judgments can be true at once. Fridman offers the compressed version of Becker: the terror is planted in early childhood, then covered with social systems until you cannot see the floor. He admits he is carrying around a half-description. She tells him his handling of these ideas is aesthetic, which he takes as a theme that will follow them into sex and relationships. Surveys, she agrees, are indirect. They have flaws you learn by using them. Some topics, given our tools, may stay fuzzy.

Fridman asks why, if she does not believe in free will, it feels as if he can decide anything. It is a cool illusion, she says, and probably the root of identity. Meditate on the sense of self and it collapses into choice: *I am doing the thinking*, which means something in you exerted agency over this thought and not that one. If free will is an illusion, the self is an illusion, and identity is a trick of the light. It is a really fun one. She used to be deep into deconstructing that trick, living toward no-ego. Now she enjoys having a lot of ego. The two stretches are not a contradiction. They are different settings on the same knob.

Consciousness, she argues, is where people smash two concepts into one word and then fight. One thing is certain and singular: *I am directly observing this experience right now.* You cannot observe two experiences at once; you cannot replicate that vantage; argument does not touch it. The other thing is whether some other system has an inner life. That one applies to many people, is almost definitionally uncertain, and is what people mean when they ask whether AI will be conscious. Signs — a system that refers to itself — are not proof, any more than a well-behaved dream character is proof that the character is awake. People borrow the certainty of the first concept to settle the second. The two do not commute.

If we understood the mechanism, Fridman says, we should be able to measure it. Measure what, she asks. Appearance of consciousness is easy. Mechanism in the physics is hard. You can imagine a simulation or a dream in which every test comes back as expected and an outsider still says it is not the real thing. She is appealing to an intuition, and she says so. In her large "chaos survey" she did get one clean, smaller result worth mentioning: women report much more vivid dreaming than men. The method is the method. A Twitter poll is a probe. The big questionnaire is the ruler.

Fridman claims, as a joke, that he already knows everything about her, including the future, and could therefore win every prediction market on her life. She actually plays on Manifold. Someone opened a market on whether she would appear on the Lex Fridman podcast. She put everything she owned on yes. Then she realized the bet was public — her face shows up — and that the information would move the market. Insider trading, on that platform, is treated as part of the information you are supposed to trade. She could have used a sockpuppet. She could have lied.

Someone else opened a market on whether she would post a photo of a poll spelled P-O-L-E. She decided to manipulate it, bet no, and planned not to post. Then she posted a screenshot of a poll without thinking and collapsed her own position. Reverse insider trading, she calls it. Fridman says that if he really could predict her perfectly it would not be fun. He prefers the romance of unpredictability, even the Buddhist version where you step in the river and forget. The prediction-market bit is small, and it shows how she treats her own life: as something you can put on a public board and still screw up by being a person.

Fridman's favorite thought experiment: a button that fulfills every want so fast you never quite experience the want. Completely without desire. Would she press it? One hundred percent not. Everything immediately stops being fun. The first time might be interesting; there is no second time. By day four, in a magical version of the setup, she figures the system escalates — she jokes about nuclear war, then about making everyone hover in the air. Then she takes it back. The first time she saw someone in pain because of her, she would stop, and that moment would arrive fast. A person made to hover would probably be upset. She would put them down and go back to a life a lot like the one she has.

Absolute, video-game, god-mode power is the version that scares Fridman, because the world stops feeling real and you can treat people as NPCs. He also thinks the first witnessed pain would halt him. Convenience is not the same as meaning. He tells a small story about fixing a flat tire by the road: after the manual work, the car, the traffic, and the convenience of moving around all looked clearer. Struggle made the ordinary machinery visible. Some people would call that copium — making do, a reason not to go back to the Amish if hardship is so enlightening. Anyone who uses *copium* as an argument, Aella says, is sus. A word whose job is to tag a position is not a mechanism.

She does not think accepting what you have and trying to improve it are opposites. The attempt is part of the thing. She enjoys that a problem exists because that is how you get the experience of striving to make it go away, and that is where the meaning sits — not in the result alone. You have to say both sentences at once: things are bad, and we are trying to stop that. Add some optimism that the future can be better, and the combination becomes a reason to keep moving. Shitty stretches also make the picture of better more intense. That is the gutter-and-stars point again. Data is a tool, poetry is a tool, the self is a knob, and meaning lives in the process. She will run a 500,000-person survey and still admit that some questions stay foggy. Both are the same posture: put the uncertainty on the table, then keep going.