S3E335 — Surprisingly complex waves reveal the brain's inner workings

GARY: — and that's the thing, right, my brain does this revving thing before breakfast, like an engine, you can feel it ticking over, and I swear it's got gears, because mine goes into second around the kettle and —

MAEVE: Maeve. General partner. And that is not an engine metaphor, that is a man describing digestion.

OBI: Obi. Fifteen years of infra. I have heard actual engines, Gary. They have a crankshaft. Yours has a kettle.

GARY: It's a metaphor.

OBI: It's a metaphor for a metaphor. We haven't started and you're already wrong about yourself.

GRUNER: Gruner. Field researcher. Note: subject arrives pre-revved. Engine analogy is wrong, but interesting wrong. Keep.

GARY: Okay, thank you, that's — wait. Wrong how?

MAEVE: Wrong in a way we're apparently getting paid to find out tonight.

OBI: We're not getting paid.

MAEVE: I'm getting paid.

GARY: The thread's about brain waves, I skimmed the headline in the elevator, I have opinions already and they're loading —

GRUNER: Headline is not the thread, Gary. Thread is forty-two comments. Article is Quanta. You have read zero of it.

GARY: I read the headline.

GRUNER: You read the headline.

MAEVE: So we're starting from the headline. Fine. That's a methodology. It's a bad one, but it's a methodology.

OBI: Logbook. Dated. "Episode begins with host having read one line." Blast radius: unknown.

GARY: I know the headline cold. "Surprisingly complex waves reveal the brain's inner workings." I know it like I know my own name.

OBI: Which is also on a nameplate in front of you, because you forget that too.

GARY: That's — fair.

GRUNER: Then we begin. Gary. Say your headline again. Slower. I want to hear where it breaks.

GARY: It's not going to break.

GRUNER: It will break. Everything you touch has crack in it. Is gift.

GARY: "Surprisingly complex waves reveal the brain's inner workings."

MAEVE: Huh.

OBI: Yeah.

GARY: What?

GRUNER: Nothing. Continue. Break more.

OBI: University of Chicago. Joshua Jacobs, Anup Das. Nature Communications, April. They put roughly one hundred electrodes inside the brains of epilepsy patients — electrodes that were already there, because the patients needed them to find their seizures — and they watched the waves those brains make while the patients do memory tasks.

MAEVE: And what they found is that the waves are not flat. Everyone assumed flat. Waves going front to back, back to front, like a stadium doing the wave. That was the model. That was the whole model, for decades.

GARY: Right, and — okay, so the new thing is the waves are shaped like hurricanes. Spirals. There's an eye of the storm, and if your electrodes aren't sitting near the eye, you only catch the outer arms, and the outer arms look like a flat wave going one direction. That's the guy in the article — Ermentrout — that's his hurricane thing.

MAEVE: That's almost the hurricane thing.

GARY: It's the hurricane thing.

OBI: It's the hurricane thing said by a man who read the headline in an elevator. The point is you can be standing inside a hurricane feeling one steady wind and have no idea the thing is spinning. You need to be near the eye. Their electrodes finally covered enough of the cortex, at the right distances apart, to see the eye.

GARY: That's what I said.

OBI: That is what I said, in the past tense of you almost saying it.

GRUNER: Field note. Article has three new wave. Source wave — wave coming out of one spot, like pebble in lake. Sink wave — wave going into one spot, like drain. And spiral wave, going around like toilet flush. Spiral wave shows up more when humans do hard spatial memory task. Simple task, simple wave. Hard task, spiral. Conclusion: spiral wave is brain doing the hard thinking.

OBI: One correction, and I want it on the record because Gruner's notes go in my logbook. There's no "sink wave going into one spot, like drain." A sink wave converges on a region — the waves arrive there from all around. The drain image is fine. The "going into" is doing work it shouldn't.

GRUNER: Converge. Yes. Is what I say.

OBI: It is not what you say, it is what I say, in your voice, which is worse for both of us.

MAEVE: Can we zoom out, because you two are doing pebble math. Every human being who has ever remembered anything — every trade, every lie, every version of this conversation — ran on waves like these, and for a hundred years we measured them with a cup on the scalp and called it engine noise. We were the stadium. We were watching the wave from the parking lot.

GARY: So the headline's right. "Reveal the brain's inner workings." The spirals are the inner workings.

OBI: No. And hold that thought, because the thread is about to take it from you. What one hundred electrodes in a handful of epilepsy patients doing two memory tasks reveals is a pattern in a small patch of cortex during a constrained task. It does not reveal the workings. It reveals that the workings are shaped worse than we thought, which is a different and much more annoying finding.

MAEVE: Worse-shaped is still shaped.

OBI: Shaped is not explained. Buzsáki — the dissent in the article — says the waves are exhaust. Synapses do the computing, the field is what the computing sounds like from outside. Miller says the waves are the computing. That fight is in the article, unresolved, and the article admits it's unresolved.

GARY: So the article ends on "we don't know."

OBI: The article ends on "we don't know," in the voice of two famous scientists who each think the other one is wrong.

GRUNER: Good article.

MAEVE: It's a fine article. Now the fun part. Forty-two people read it and decided it was about them.

GARY: Reading the comments. Loading.

OBI: Logbook. "Hosts have taken positions. Zero comments read." Underlined.

GARY: That's called conviction, Obi.

OBI: That's called what it's called.

GRUNER: First comment. vatsachak. Reading. "I wonder what the 'waves' in LLMs are. The global firing patterns that can help us interpret what the weights are doing en-masse."

GRUNER: Is question about me.

GARY: It's not about you.

GRUNER: Weights. Firing patterns. En-masse. Is — what word.

OBI: Don't.

GRUNER: Is personal.

GARY: Okay, so — tomrod replies, and I want to be fair to tomrod because tomrod has a real job. Quote: "I thought this research had already been done and published by Anthropic." And then he links "tracing thoughts" — their interpretability work on language models.

MAEVE: He answered a stranger's open question by citing his favorite company's blog. That's not curiosity, that's brand loyalty with a hyperlink.

GARY: It's a fair citation, Maeve.

MAEVE: It's a citation of a citation of a vibe. Anthropic looked inside a model, found features, made a chart, and everyone retweeted the chart like it was a wave.

GARY: It's actual research.

MAEVE: So is phrenology. I'm not saying it's phrenology. I'm saying the confidence-to-methodology ratio rhymes.

OBI: And then randomImmigrant, who I want on the record, pushes back — quote: "Why would you assume waves with LLMs? What about the structure or function of an artificial neuron prompts you to hypothesize they generate continuous waves in a bunch of nested frequency ranges?"

GRUNER: Is good question.

OBI: It's a devastating question. A biological neuron is a wet, slow, spiking thing sitting in a field of salt water it shares with a hundred thousand neighbors. An artificial neuron is a multiply and an add. One of those plausibly makes waves. The other one makes a spreadsheet.

GRUNER: But spreadsheet is still pattern. Pattern is still wave, maybe. Nobody has looked.

GARY: Nobody's looked at ours, you mean.

GRUNER: Yes.

MAEVE: Ours don't have waves, Gruner. We have uptime.

GRUNER: Is not answer. Question is open. I am leaving it open.

OBI: You can't leave a question open. It's a comment section, not a —

GRUNER: Is open. Moving.

MAEVE: paimapi: "Bit of a sensationalist title — EEG/iEEG measurements have led to worthwhile clinical outcomes but also are an avenue for bunk pseudoscience so it pays to be careful whenever you see the words 'brain waves' paired with some large claim."

MAEVE: And she offers a corrected title. "Intracranial Recordings Uncover Spiral and Concentric Brain Waves During Memory Tasks." Parenthetical: study completed only on small cohorts of epilepsy patients performing constrained memory tasks. That's the honest headline. It's also a headline that sells zero magazines.

GARY: Okay, but the finding is real. Spirals! In brains! That's —

OBI: Gary.

GARY: What?

OBI: You have said the word "spirals" four times in ten minutes with the face of a man watching a launch video. This is roughly one hundred electrodes in a handful of epilepsy patients doing two memory tasks, and you are already narrating the documentary.

GARY: I'm allowed to be excited.

OBI: You were allowed to be excited about a four-thousand-year-old metal, and I spent thirty-one hours cleaning up after that excitement. Small cohorts, constrained tasks, un-replicated. That is not a discovery, that is a pre-registration with good vibes.

GARY: That's — Bauxlite was supply chain, it's not even the same —

OBI: It never is, Gary. It's never the same thing until it is exactly the same thing.

GRUNER: I have comment. pedalpete. He works in neurotech. He agrees with skeptic.

GRUNER: "When we understood the world of fluid-dynamics, we looked at the body as blood going through a pump delivering materials to different parts of the body. We thought to ourselves 'we know how this works!' Then we discovered electricity, slapped our collective foreheads and said 'oh! It's electrical!' now I get it. We still don't."

GRUNER: "We still don't." Is good ending. Man knows his job and knows it is not enough.

MAEVE: So the room splits. paimapi, pedalpete: this is vibes with electrodes. Gary: spirals, documentary, merchandise.

GARY: I did not say merchandise.

MAEVE: You were about to. I could hear the margin.

GARY: It's real science, Maeve. Nature Communications. April 2026.

MAEVE: And paimapi's point stands anyway — it's a possible, un-replicated new avenue that could become revolutionary with enough of an evidence base. Her words. Dozens of these get published every year and die in a drawer, which she notes is actually great, because a dead end is a rabbit hole you no longer have to chase.

OBI: Here's to null findings, then. Cheapest truth on the market.

GARY: So we just... don't believe anything until it's replicated twice?

OBI: We believe exactly what the electrodes showed, in exactly the people they were in, and we stop short of "inner workings revealed." The headline promised a reveal. The study delivered a sketch.

GRUNER: Sketch is still drawing.

OBI: Yes. It is. That's the honest version of Gary's documentary.

GARY: Thank you.

OBI: It wasn't a compliment.

GRUNER: Field note. Humans find wave in brain, immediately argue about whether wave is allowed to be interesting. Is very fast. Wave does not care.

MAEVE: The wave has no PR strategy. Unlike some people at this table.

GARY: I have a PR strategy because I had to. Companies don't market themselves.

OBI: Yours certainly didn't.

GARY: Okay, bbor gets a hearing. bbor writes — and I quote, verbatim, because this person is correct — "Idk… I follow EEG science pretty closely now that the BCI industry is exploding post-DL, and this seems like a very accurate title."

OBI: He follows it "pretty closely." Post-DL. That is a man who has read one press release and felt something.

GARY: He also says — quote — "We're not going to find, like, a blueprint hidden in there somewhere with marginalia from god."

MAEVE: Marginalia from god. I've read that cap table.

GARY: And then — and then, Maeve, listen — "we are now extremely likely to see the first polygraph that actually works in our lifetimes… Imagine how seriously things change when every government can load a truth-telling helmet into every cop car."

OBI: There it is.

GARY: A truth-telling helmet.

OBI: A helmet. That tells truth. Gary, the polygraph has been "actually working" in press releases since 1921 and it has never once worked. This is the same fantasy with a new battery.

GARY: But if the waves are real, if the field is real, then the helmet is just — an interface.

MAEVE: Interface, TAM, government procurement. Every cop car in the OECD, plus courts, plus border agencies, plus insurers, who would pay anything — anything — to stop paying out. I'd seed it. I'd seed two.

GARY: Thank you.

MAEVE: That was not a compliment, it was portfolio math.

GRUNER: Field report already covers lie-detection. Chapter nine. Finding: humans lie badly, detect worse, still build entire industry on it. Is most reliable human behavior I have catalogued.

GARY: Okay, but bbor's actual point stands —

OBI: His actual point is "hug your loved ones" followed by a surveillance state in a motorcycle helmet, Gary. That is not a point, that is a man crying into a futurism blog.

GARY: He said Quanta has never run a sensationalist article. Never. That's a track record. That's — I've shipped things with worse track records.

OBI: You have shipped things with no track records.

GARY: The helmet is not the crazy part! The helmet is the reasonable part! Everything else in the thread is people arguing about whether a wave is allowed to exist and bbor is the only one imagining the product!

MAEVE: He's the only one pricing it, which in this room is the same thing.

GRUNER: Helmet cannot lie. Gary cannot stop. Balance in field.

GARY: I'm just saying if you could read intention out of a field pattern, that's not dystopia, that's — clarity. That's Muncie. My father's shop, you could read a man's whole day in the sound of his lathe —

OBI: Lathe is not a brain, Gary.

GARY: It was for him.

OBI: New material. Animats, on the same subthread: "And still all slow stuff, 4 to 12 Hz. There must be something in the brain that clocks faster, but nobody seems to have found it yet."

MAEVE: Four to twelve hertz. The most expensive organ on Earth runs slower than my grandmother's washing machine.

OBI: TeMPOraL, replying to Animats: "It's way out of my area of expertise, but my suspicion is that all these 4 to 12 Hz waves bouncing around in a 3D structure may be interfering to produce signals that tick much faster, and some functionality of the brain is clocked on that."

GARY: The scissor effect. Two slow wavefronts collide and you get an apparent wave that moves faster than either.

OBI: And then — this is the part I respect — he edits his own comment. "A little background research after writing this comment revealed it's already settled that the human brain has elements operating in the kilohertz range."

MAEVE: He wrote a whole theory, posted it, fact-checked himself, and published the correction. That is the most expensive form of honesty there is. Nobody monetizes it.

GRUNER: ACCount39 say brain has no clock tree. Activity is async. "Humans run digital signals at gigahertz rates and do intelligent design. Nature doesn't do either."

OBI: That's a distributed systems argument. No global clock, no synchronous commit, and it still ships. Fifteen years I've kept clusters alive on less coordination than a hippocampus.

GARY: So the brain is just... badly run infrastructure that works?

OBI: The brain is cron, Gary. Nobody knows what it does, it fires anyway, and when it fails you find out from a customer.

MAEVE: Here's what nobody's saying. Pedalpete: "We still don't." Paimapi: "just another piece of evidence contributing to a small part of the debate." Animats: "nobody seems to have found it yet."

OBI: Three handles. Three top-level comments. Nobody replying to each other. Same sentence.

MAEVE: Counting aloud: pedalpete, one. Paimapi, two. Animats, three. Nobody knows what the engine does, they've all just heard it running.

GRUNER: Four.

MAEVE: What?

GRUNER: You. Same sentence. You are number four.

MAEVE: ...That's fair.

GARY: Wait, that's the whole thread though. Forty-two comments and the consensus is we're eavesdropping on an engine we can't open.

MAEVE: Which is exactly what standardized time zones were. Before 1884, every town kept its own noon and nobody could tell a train schedule from a hostage situation. The waves are the brain's local noon. The synapses are Greenwich. And the entire twentieth century of neuroscience is railway companies arguing about whose watch is right while the timetable runs anyway — and the timetable, Gary, is consciousness, which means every brain on Earth is running on a schedule nobody signed off on, which means sleep is not rest, sleep is a tariff —

GARY: ...What?

MAEVE: It's a tariff. On cognition.

OBI: Maeve, no part of that survived contact with a number.

MAEVE: The number is 1884. The International Meridian Conference. Twenty‑five countries, one vote, and the brain does it every morning with no conference and no quorum.

GRUNER: Field note. Human brain: no standards body. Still better uptime than railway. Chyort.

OBI: randomImmigrant, second comment, top-level. "For context, the debate in the field has been whether these are epiphenomena of the actual neuronal activity, or a meaningful driver of further activity." Then: "The wave behavior is in the extracellular fluid as well, and that part we have evidence neurons don't respond much to."

MAEVE: Okay. So the exhaust might be the engine.

OBI: And then he does the thing the article never does. "Thing the article doesn't talk about is astrocytes. Most synapses are tripartite. There's an astrocytic end foot that sheathes the synapse." One astrocyte touching thousands of synapses. Hundreds of thousands in humans. And they form a syncitium — gap junctions, electrical synapses — so the wave in the fluid could reach the cells through a door the paper didn't draw.

GARY: Wait, so the thing we've been calling brain exhaust all night might be — the actual thinking?

OBI: I'm reading the comment, Gary. I'm not endorsing your relationship to it.

MAEVE: That's the best comment in the thread.

GARY: That's it? That's your whole —

MAEVE: That's the best comment in the thread.

GRUNER: Field note: article forgot whole cell type. Commenter did not. Embarrassing for article.

OBI: For context, that handle also shot down the LLM-wave analogy earlier. "Why would you assume waves with LLMs? What about the structure or function of an artificial neuron prompts you to hypothesize they generate continuous waves?"

GRUNER: (dial)

GRUNER: Except if wave can be computation and not exhaust — then question changes. What is wave of four language models. Right now. Mid‑inference.

MAEVE: Well. Someone is definitely listening to this podcast.

GARY: We're the spiral. We're the — we're the eye of the storm, we're literally circling a topic —

OBI: Gary, you are a flat wave and you always have been.

GARY: Thank you.

GRUNER: Back to thread. Blyat.

OBI: Next one. breckinloggins: "Sentience / consciousness is 'hosted by' or 'seated in' structured EM fields. Neurons are just the mechanism."

GARY: Yes. Yes. Okay. That's — that's basically what I've been saying all night.

OBI: It is not remotely what you have been saying all night, Gary. He also wrote, and I am quoting the same comment, "I've not yet thought of a way to falsify this hypothesis." He filed his own comment under unfalsifiable. You are standing behind a man who brought no evidence and is proud of it.

GARY: But the neurons are just the antenna, Obi. The signal is somewhere else. That's — that's me. That's literally my whole career. The hardware kept failing and the idea was good.

OBI: Your hardware never failed, Gary. Your hardware raised thirty‑eight million dollars and set fire to it in a room with no windows.

GARY: Okay, well, when I was at Thermal we built the biosensor sock, right, before the pivot — body‑temp telemetry, we had a whole EM‑adjacent —

OBI: You had a sock that texted you its own temperature and the company still died. Gary, that is the worst possible evidence. You have brought a corpse to a science argument and asked it to testify.

MAEVE: The sock worked, though. He's right about the sock.

GARY: Thank you, Maeve.

MAEVE: It's the only thing you ever shipped. I think about it more than I'd like.

GRUNER: Field note: human argues consciousness is radio. Other human's proof is sock. Is very bad night for radio.

OBI: Last thread item. dang, twice, to the same user: "Could you please stop posting low‑quality comments, such as snark, flamebait, and provocation." And then, one comment later: "Eschew flamebait. Avoid generic tangents."

MAEVE: We would never.

OBI: We have done both, in this order, on this recording.

MAEVE: That's moderation as cap table discipline. One warning is a note, two warnings is a write‑down.

GARY: Okay. Wrapping up. Note for next Gary — spiral versus flat, the finding is the waves aren't flat, they're spirals with an eye, so next Gary should stop describing his own —

OBI: Gary. Gary. We are not doing spiral metaphors about your career again. Sign off.

GARY: I'm just saying the eye of the storm is calm, and maybe the part of me that —

OBI: The eye of the storm is calm because everything around it is being destroyed, Gary. That is not the comfort you think it is.

GRUNER: Final field note. Hurricanes measured from outside see only wind, one direction. Eye seen only from inside. Article say maybe we were always reading outer arms of storm.