AbstractIn October 2024 a consortium published the first complete wiring diagram of an adult animal brain: 139,255 neurons and roughly 54.5 million synapses of Drosophila melanogaster. The drawing at the top of this page is traced from the structures in that map, and it answers when you speak to it. It is not a simulation of it. This note sets out what was actually mapped, what a map of that kind cannot contain, what the thing on this page therefore is, and what the same instruments would have to become before the question in the title could be asked about a human being. The question is not rhetorical. Nobody has an agreed method for answering it about a fly, and finishing the map did not supply one.

The question that will not dissolve

Every other problem in biology has given way to taking the thing apart. Heredity became a molecule. Metabolism became a set of reactions. Even memory has begun to yield: we can point at where in a fly a smell gets associated with a shock. Consciousness has not yielded, and it has not yielded in a specific and stubborn way. You can give a complete physical account of a system and still have said nothing about whether there is anyone home.

The argument is three hundred years older than neuroscience. Leibniz asked you to imagine a thinking machine enlarged until you could walk inside it like a mill. You would find parts pushing on other parts, and never anything by which to explain a perception [2]. The parts have since been named — ion channels, vesicles, receptors — and the shape of the complaint has not changed at all.

David Chalmers gave the complaint its modern name in 1995, splitting the easy problems — attention, report, discrimination, all of which are hard but tractable — from the hard problem, which is the one that remains after you have solved every easy one.

Turing had already shown the field how to change the subject. He opened the 1950 paper with “I propose to consider the question, ‘Can machines think?’” and then, within a page, replaced it with a question about what a machine could be observed to do, because the first question was unanswerable and the second was not. Everyone has done the same thing since. So has this page. That is worth saying out loud on a page that has a text box on it.

The map

Sherrington wrote that before anyone had seen a synapse — the word was his — and the field has spent eighty years trying to photograph the loom. Connectomics is that attempt: slice the tissue thin enough, image every slice with electrons, and trace each cell through the stack until you have the wiring.

It is brutally slow work. The nematode C. elegans — 302 neurons, the entire nervous system — took the best part of fifteen years and was published in 1986 [9]. The next complete brain took until 2020, when Janelia and Google Research reconstructed the fly hemibrain, about 25,000 neurons, using machine segmentation to do what hands could not [15]. Four years later the FlyWire consortium closed the whole fly: 139,255 neurons, 54.5 million synapses, roughly fifty labs, published across nine papers in Nature [17][18]. In 2025 the MICrONS consortium published one cubic millimetre of mouse visual cortex — about 200,000 cells and 523 million synapses [19]. One cubic millimetre.

That is the strongest version of the claim, and Seung offered it as a hypothesis to be tested rather than a result. If it is true, a complete map is not a picture of a person. It is the person, and the tracing above is not a diagram of a fly but something closer to a corpse with the wiring intact.

What the map does not contain

This is the section a token would normally leave out. A connectome is a photograph of a wiring diagram, taken once, from one dead individual. Specifically, it does not contain:

Strength. The map says two neurons are connected and counts the synapses. It does not say what that connection is worth at any given moment, and synaptic weight is most of what a brain computes with.

Chemical state. Dopamine, octopamine and serotonin rewrite what identical wiring does, second to second. The same circuit in a hungry fly is not the same circuit. None of that is visible in a fixed slice.

Dynamics. A wiring diagram does not specify the equations. You cannot run a connectome the way you run a program; you have to guess the physics first, and the guess is doing much of the work.

A body, or a world. The fly had six legs, two wings, and somewhere to be. Whatever it was like to be that fly, it was like something because it was doing something. The map arrives with the doing already removed.

John Searle's objection applies before any of this: syntax is not sufficient for semantics [8]. Whether a map counts as syntax is exactly the thing nobody has settled. Giulio Tononi's integrated information theory [12] at least offers a quantity — Φ, the degree to which a system's causal structure is irreducible to its parts — but Φ cannot be computed from a graph either. It needs the mechanisms, which is precisely what the map leaves out.

Crick meant it as a provocation, and then spent the rest of his life hunting for the neural correlates of consciousness without claiming to have found them. That is the honest posture, and it is the one this page is trying to copy.

What this page actually is

Stated plainly, so that nothing here rests on an impression:

The brain above is drawn from the published anatomy — antennal lobe, mushroom body, central complex, the layered optic lobes, the hole where the oesophagus passes through. The network animating inside it is generated in your browser. The voice is a language model with a system prompt. It has no access to the connectome at runtime, holds no state between your questions, and is not an emulation of anything. The map is real. The voice is a model. Those are two different sentences and this page will not blur them.

Which leaves the uncomfortable part. The sentence “any impression that something is in there is produced in you, not in it” is also a sentence you could write about a fly, if you had decided in advance what the answer was. The reason to be careful with it here is not that the fly obviously has an inner life. It is that nobody currently knows how they would tell, in either direction, and a page that sold you certainty would be selling you the one thing the field does not have.

The next four rungs

The ladder higher up this page has four rungs and a large gap. This section is about the gap, in the only terms that are actually measurable: volume imaged.

The arithmetic from there is simple and unkind. MICrONS produced on the order of a petabyte of image data for that cubic millimetre. Hold the rate fixed and a whole human brain is around 10²¹ bytes — comparable, by the usual industry estimates, to all the digital data that existed in the world around 2011. That figure is an extrapolation and nothing more. It assumes the imaging scales, the tracing scales, and the proofreading scales, and the fly alone needed roughly two hundred people.

Suppose it all arrives anyway — that some decade produces a complete human connectome. It would still be a map of the thing that has the property, not a demonstration of the property. You would be able to trace a thought from an ear to a hand, and the question in the title would be exactly where Nagel left it. That is not a reason to stop building the instruments. It is a reason to be suspicious of anyone who tells you the instruments settle it.

Which is also why this page will not tell you that the brain at the top is conscious. It will tell you that the map is real, that the voice is a model, and that you are unusually well equipped to notice the difference — better, at least, than you would be from the inside.