Two Scientists

Everything you have read so far rested on the words of one man. A brilliant one, a consistent one, one who hasn’t left his line in twenty-five years, but one. A shadowy genius from a chat who hides his face, speaks in riddles, and wears masks. A skeptic has every right to ask: where, exactly, is the science? Where is your proof? Where are the people with names and reputations who have checked all this and signed off? Here. Real science.

The Man Who Digitized His Own Brain

David Vivancos has been working on artificial mind since 1995 — twenty-eight thousand hours of research, and that figure is exact. On brain-computer interfaces since 1998, back when nobody was making TV series about them. Author of seven books, among them «Artificracy» and «The End of Knowledge».

I have personally read these books and highly recommend them. Thank you, David. He has spoken at more than five hundred conferences. Founded his own discipline at the junction of sciences — ”artificiology.” Scientific advisor to Qubic.

In 2014 he began the MindBigData project, and did what no one had done: he digitized his own brain. For two years straight he put on an electroencephalograph and recorded his brain’s signals while looking at the digits zero through nine — thousands upon thousands of times, session after session. The result was the world’s largest open dataset of brain signals for machine learning: eight billion data points. He recreated the famous MNIST dataset — the alphabet every neural network in the world learns on, but not with pictures: with a living brain looking at those pictures. His own.

A man who spends two years recording his own brainwaves for science and puts them in the public domain is a particular type of scientist. The same type that publishes failures, builds without an owner, and grows rather than designs. Small wonder his road and CfB’s converged. «The Wild Mind»

José Sánchez — a doctor of science, UNIR university, Spain. His territory is psychometrics: the science of measuring intelligence. That same g-factor, Spearman’s ”general intelligence” — what it is, how to catch it in a number, where in the brain it comes from. While Vivancos spent thirty years building machines and logging hours of code, Sánchez worked on the question without which all the machines are blind: what is mind in the first place, and how do you know when it is in front of you.

They met at an education conference in Spain, and have been friends and co-authors ever since. An engineer of the brain and a measurer of mind. One knows how to build, the other knows how to verify. For a project whose central problem is ”how to tell an emergent mind from a beautiful simulation,” it is hard to imagine a more useful pair.

Sánchez is also the author of a book called «The Wild Mind». I have read it and highly recommend it: a chilling book!

The Spanish Line

Before we look at what these two proved, appreciate a coincidence that a book playing with coincidences has no right to skip. You remember from the previous chapter where the real cradle of AI stands: Spain, two men born the same year, 1852 — Cajal, who discovered the neuron, and Torres Quevedo, who built the first deciding machine. A century and a half later, the science wing of a project growing mind out of neurons and machines is held by two Spaniards. And history repeated itself not in geography, it repeated itself in roles. Sánchez, measuring the nature of intelligence, stands where Cajal stood, peering into the living tissue of thought. Vivancos, building machines and entire worlds, stands where Torres Quevedo stood with his chess automaton. An engineer and an investigator of mind. Again a pair. Again Spain.

The similarities go deeper. Cajal drew the neurons on thousands of sheets — by his own hand, over years; those drawings are still printed in textbooks — the first open corpus of data about the brain, gathered by one obsessed man. Vivancos recorded eight billion points of his own brain and put them in the open: the same gesture, the same hands, a hundred years on. Torres Quevedo did not hide his automaton in a lab, he took El Ajedrecista out to the public, to Paris: come, look, verify. NeuraxonLife is open to any browser without registration: the same gesture. And both times it is done from the periphery. At the start of the twentieth century, the science of mind was moved forward by two Spaniards far from the scientific capitals of the day, and the center recognized them late, and nearly forgot Torres Quevedo altogether. At the start of the twenty-first — again two Spaniards, far from Silicon Valley, attached to a project with no budget and no owner. Cajal got his Nobel in 1906. The work of his heirs got Best Paper at two IEEE conferences in 2026 — one hundred and twenty years later. The pattern is one: the breakthrough comes from where no one is looking, and is confirmed later than it deserves. For lovers of dates — a notch in the margin. The first machine in history to make a decision on its own played out its endgame in 1912. The word Qubic appeared on a forum in 2012. A century, exactly. Move for move.

Two Ships from Different Ports

They did not ”come to believe in Aigarth” and did not ”join the community.” They walked their own road — an academic one, through neuroscience, psychometrics, the biology of the neuron, and arrived at the same conclusions CfB arrived at by the engineering road. At ternary logic: a neuron is not an on/off switch; it has three states — excitation, inhibition, and neutral. At continuous time instead of ticks. At evolution instead of design.

Their architecture is called Neuraxon — ”a new blueprint of neural growth and computation,” as the title of their paper has it (February 2026). Two ships left different ports, sailed different courses, and met at one point in the ocean. In science this is called convergence, and it is one of the strongest kinds of confirmation there is: when people who could not have copied from each other independently arrive at one answer.

Now it is clear how the tracks connect. IntelligentTissue — the ”intelligent tissue” — grows in Aigarth; Neuraxon is one of its implementations, described and measured academically. (Anna herself, by the way, runs on a different, non-Neuraxon implementation of the same tissue, but the paradigm beneath them is one.) Science gave the substrate a name and a mathematics; engineering gave science a scale. What They Proved Of all their results, one deserves to be taken apart on our fingers because it explains why the ternarity the whole building stands on is needed at all.

Any neural network has two ways to die. The first is silence: activity fades, the signal dies out, the network falls quiet forever. The second is fire: activity feeds on itself, an avalanche of excitation, everything screaming, nothing heard. The life of thought is possible only on the narrow ridge between silence and fire — physicists call it criticality. Ordinary, binary networks (on/off) hold that ridge poorly: they have to be propped up from outside — resets, normalization, manual tuning. Like a cyclist held upright by the saddle.

The Vivancos-Sánchez paper «The Neutral Buffer State» shows: give the neuron a third state — neutral, neither ”yes” nor ”no,” zero, and the network starts holding its own balance. The zero works like a leaf spring: excitation spikes — the neutral zone absorbs it; activity sags — the neutral neurons stand ready, keeping it from stalling out. No resets. No manual tuning. Stability stops being a crutch and becomes a property of the architecture itself. And this is not an argument, it is a measurement. The scale of the experiment: over two hundred thousand autonomous agents, seven and a half million neurons, thirty-four million synapses, sixteen million observations, five and a half thousand experiments. A single variable factor: two states or three. The result: ternary networks spend 88.8% of their time in the critical regime — the living zone where thought is possible — against 37.2% for binary ones. More than twice as long on the ridge. An effect size of the kind statisticians call large. The ternarity CfB chose by engineering instinct twenty years ago — back in the three-valued logic of his early processors — turned out to be no quirk. It turned out to be a physical precondition of stable thought. Science caught up with the instinct and handed it numbers. And the second proof — the Darwinian one. In their living simulations, evolution actually works: offspring inherit their parents’ success.

Judges with No Stake

Science is tested not by the applause of one’s own, but by the judgment of strangers. In May 2026, at the international ICMLT conference in Berlin — IEEE, three days, eight selected papers — the Neuraxon v2.0 paper took Best Paper and Best Presentation. The conference’s best article and best talk. The organizers stressed it separately: the judging was done by researchers with no stake whatsoever in Qubic. Two months later, in July, at the AMLDS conference in Osaka — IEEE again — the team’s second paper, the very one about the neutral state, took Best Oral Presentation. Two awards in one summer, on two continents, from two different panels of independent judges. And right after, in San Francisco, a third paper — Multi-Neuraxon — was presented at AGI-26, the specialized conference on artificial general intelligence.

Let us name plainly what this changes. A year ago you could still wave Aigarth off: an anonymous poster from a crypto chat promises things, a community believes, the numbers draw themselves. Today the core of this technology — the ternary neuron, continuous time, evolution instead of training — has been vetted by scientists of the first rank: people with names, faces, thirty-year scientific biographies, and judged the best in the room by juries with no reason to flatter a crypto project. The argument is no longer with an anonymous poster. The argument is with peer-reviewed publications, open code, and the juries of two IEEE conferences. A World Where You Can See Darwin And now — the most astonishing thing Vivancos and Sánchez have made. They built a world where the evolution of mind can be watched with the naked eye.

It is called NeuraxonLife. A living simulation open to anyone in a browser: a twodimensional world inhabited by creatures — ”nxers” — each with a real ternary Neuraxon brain inside. Nobody teaches them anything. They simply live: search for food, explore territory, meet, mate, die. And they evolve — genuinely, with sexual reproduction: two genomes recombine into a third, offspring inherit traits of both parents. (An amusing detail for the attentive: CfB’s garden evolves asexually, by lone mutations, while Vivancos’s world evolves through love. Two ways nature does the same thing.)

The statistics of this world read like a bulletin from real biology. Over the simulation’s history, twenty-six and a half thousand creatures have been born. Twelve thousand matings. Alive at any one time — around five hundred in the best of times. The rest — ninety-eight percent of all who ever lived — died. Evolution is wasteful: for every working brain it pays with a mountain of the dead. Exactly as in CfB’s garden, where billions of networks are culled for the sake of a few survivors. Exactly as in the nature that produced us. And in this world, the thing it was all started for happens before your eyes. Intelligence — that same g-factor of Sánchez’s, the general capacity to figure things out — arises out of nothing: the population’s general intelligence counter creeps upward within minutes of watching. Creatures nobody programmed begin stealing food from one another, claiming territory, choosing mates — and one, to the observers’ astonishment, ”sings” while it eats. And right beside it — the honest Darwinian cold: the carriers of the highest intelligence are as mortal as everyone; the population’s genius dies, and the population searches anew. Growth is a property of the whole — not a guarantee for the individual.

In the summer of 2026 Vivancos came back from the conferences and summed up the live run that had kept spinning while he and Sánchez were flying around the world: «After 44 days. 13.9M ticks. 47,000 brain designs tested. The big discovery is that offspring really do inherit their parents’ success, evolution is quietly working inside the machine.» (X, @VivancosDavid, July 2026)

Forty-four days. Nearly fourteen million ticks. Forty-seven thousand brain designs tested. And the main discovery: offspring really do inherit their parents’ success — evolution is quietly working inside the machine. Darwin stopped being an analogy. He became an experimental protocol.

Science Wide Open

Everything is open. The papers — in open preprints. The code — on GitHub, take it and run it. The data — on HuggingFace, including the eight billion points of Vivancos’s own brain. The simulation — in a browser, no registration, no clearance. While labs with billion-dollar budgets guard model weights like state secrets, and their ”scientific reports” are more and more often written by the marketing department — two scientists attached to a project with no owner publish everything: the wins, the methods, the raw data.

It is the same logic that runs through this whole story: power without a master, science without a lock. Verify — don’t trust. Aigarth’s science is built like everything he makes: for everyone, for free, and let no skeptic walk away without the means to check. So, by the summer of 2026, everything had converged. The theory — proven and awarded by independent judges. The substrate — built and described in peer-reviewed papers. Evolution — confirmed by living experiment: it inherits, it selects, it works. What does he need all this for?