Digital minds: when thinking leaves the head

Drag the slider to the year 2020 and stop. One node is left on the screen: digital gold. That was all that existed then — the thesis that bitcoin is a harder store of value than bullion, and a man the market took for an eccentric with too much cash. Pull the slider to the right. Out of digital gold grows digital energy, then digital property, finally digital capital. Four metaphors, one thesis, six years — and you have just watched a thought ripen, at the pace it actually ripened. This is not a price chart. This is a chart of a mind.

The tool that does this is called Minds. So far it has done one thing — and done it honestly: it remembered. This essay is about the second thing, which it doesn’t do yet, but which is already within arm’s reach. And about why it — not the pretty chart of the past — is the most literal proof of this book’s thesis.

The map only remembers

Minds, in its current form, is an atlas of the evolution of thinking. It takes one person and reduces them to what is most durable in them: a dated graph of ideas. Nodes are concepts, edges are genealogy (“this thought evolves into that one,” “this one reframes the previous one”), and the year slider reveals the whole the way it accumulated — year by year. This is powerful, but it has one limit written into its very nature: the map remembers. It shows only what was actually said.

And the most interesting question lies just past that limit. Not “what did he say,” but “what would he say” — to a question that was never asked. A year from now. If he were alive today. If he collided with a situation that isn’t anywhere in his archive. The move from “what did he say” to “what would he say” is the move from archive to simulation — and that is the second thing this essay is about.

The digital mind — and why it is apparent

Once we have someone cataloged — a graph of theses, dated quotes, a genealogy of changed minds — we also have material we can work on, not merely look at. You feed this to a model and ask it for something the data doesn’t hold outright: a predicted answer, a simple decision played out “on the example of the past.” That is how a digital mind comes to be: not a new entity, but a pattern of thinking lifted out of the traces and pushed further than the traces themselves reach.

Hence the honest name: an apparent mind. Built from the surface, and the surface is not the whole person. Technically, the simplest version doesn’t even require training a model from scratch — it’s enough to put the dossier into the prompt and ask for an answer in that pattern. This cheap route has an advantage that cannot be overstated: you can see the sources the model leaned on, instead of trusting the fog of weights. It is a simplified example of what is possible, not an oracle — and that is exactly why it is a good example.

There is one more move in this, easy to miss. Every opinion and every decision someone voiced is a mappable node. The more traces you plug in, the denser and more complex the structure becomes — and the denser the structure, the less the model has to guess. The atlas and the simulator are not two tools; they are two ends of one scale. At one end you map what was said. At the other — on what has been mapped — you infer what could have been said.

This closes the definition of the era

Let’s return to the sentence this book opens with: in the post-cognitive era we share memory, knowledge, and increasingly reasoning itself with our tools. Read it again, because hidden inside is a list with three items.

The atlas ticks off the first two. One person’s memory and knowledge — what they thought, when, in what order — lie lifted out of their head onto a screen, versioned and checkable. This is the extended mind shown literally. But the third item — reasoning — the map does not touch. Reasoning is not the replaying of old answers; it is the production of a new one.

And that is exactly what the digital mind does. It takes the externalized memory and starts to think on it — outside the head of the original, sometimes long after the original fell silent. This is composite cognition in its purest form and precisely that fourth wave in which the machine stops handing over information and starts to think with you — with the difference that the material for thinking here is a specific, named mind. So Minds does not illustrate the book’s thesis. Minds is an instance of it: first it proves memory can be lifted out of the head, and then — that on the lifted-out memory you can go on thinking.

The limits are the message, not a footnote

Here a warning has to be set in capital letters, because without it the whole idea turns against the book.

The solution is not a declaration, but mechanics. The rule from chapter 7 goes:

“Don’t trust, verify” is the inverse of the blind-trust threshold (…). The habit of checking instead of believing may be the most valuable one you can carry into the age of models that sound convincing.

A digital mind passes this test only when it shows its uncertainty as data, instead of smoothing it away: this was really said, and that the model filled in; here I have a hundred quotes, and there one, and an old one. The atlas already does this — the verified:false flag is literally written-down uncertainty. It is the same discipline that chapter 5 prescribes for you toward yourself:

Every meaningful allocation — of money, but also of time and reputation — written down before the outcome: thesis, stake, review date, and one sentence on “what would convince me I’m wrong.” (…) memory will always rewrite the story in your favor; paper won’t.

An apparent mind built on this principle does not pretend to be a human. It shares the human’s pattern — with a visible seam between what is known and what was filled in. And that is why the word apparent stands in the title, rather than hiding in a footnote.

This is not an oracle

The temptation is obvious: a handful of people whose bets came good are easy to tell as prophets — and a digital mind that speaks in their voice is even easier to turn into an oracle. This book is openly non-tribal, and a tool under its roof has to keep that stance through mechanics, not a promise.

So it keeps it. Unverified theses stay marked as unverified, instead of being promoted to fact. Price targets are drawn with a dashed line labeled “forecast, not fact.” Next to a figure who sells GPUs hangs an explicit conflict of interest whenever that same figure claims that “every country needs its own AI.” Myths are kept as flagged entries, not erased. This is the difference between an Interpreter and an advocate — the same one chapter 5 named outright: the real test of caste 4c is not “am I against this,” but “do I have a thesis, a stake, and a review date.” Without those three, contrarianism is a costume — and a digital mind without flags is a costume of someone else’s certainty.

An invitation

The best way to understand this is by clicking. You can touch the atlas right now: pick a mind, move the slider, open a node, and step back down through the lens to here, to the chapter that explains what it is you’re looking at.

And if one thought is to stay with you, let it be this. The atlas you’re holding is proof that memory and knowledge can be lifted out of the head. One step further — a simulation that reasons on that memory — is not science fiction; it is a cheap, doable-today move that needs only public traces and a model reading a dossier. That a mind can be mapped this way and pushed further in a few days is not a technical detail. It is the shortest proof of the hyperliquidity of intent — of how instantly an intention turns today into a working artifact. Minds is a map of a few minds; but it is also a machine for making such maps — and, more and more clearly, for thinking onward with them. Every opinion is a node. Every node is a piece of structure. And the structure, once it grows dense, begins to answer.