Google's Most Disturbing Quantum Experiment Needs a Human Volunteer — Awake

Hartmut Neven runs Google's Quantum Computers, which just beat a supercomputer 13,000 times over, and he wants to wire one into a living human brain. A physical link with the person awake, so machine and brain share a single quantum state. Then they ask one question. Did it feel bigger? The plan is published in a peer-reviewed journal, and I expected a stunt, until the mice. The reason it might work came from an anesthesia lab where 80 mice were put to sleep with four kinds of the same gas and the gas didn't behave like one gas 2018 the journal Anesthesiology a 10 author team first author Lee took the 80 male mice and split them into four groups of 20. every group got xenon a noble gas that works as an anesthetic the only difference was the isotope xenon 132, xenon 134, xenon-131, xenon-129. Chemically they are identical, same element, same electrons. The property that decides how tightly a gas atom grips a protein, think of it as the atom's grip strength, was identical for all four, down to the second decimal. By every rule of anesthesia, the mice should have gone under at the same dose. They didn't. Xenon-132 and 134 knocked the mice out at around 70% concentration, while xenon 129 and 131 needed 105 and 99. Those two isotopes took roughly 40 to 50% more gas before the mice went under. And mass doesn't account for it, because the heaviest and lightest atoms differ by less than 1%. The two stubborn isotopes share exactly one thing the other two lack, a nuclear spin. The nucleus itself behaves like a tiny magnet, a property that only exists in quantum physics, and the mice breathing those atoms stayed conscious at doses that put the others out. Six years later, a team led from Los Angeles wrote a mechanism down. On May 28, 2024, the journal Entropy published a paper with nine names on it, led by Hartmut Neven, who founded Google Quantum AI in 2012 and still runs it, and Christoph Koch of the Allen Institute in Seattle with co-authors from UC Santa Barbara. It opens by flipping Roger Penrose. Penrose said a conscious moment happens when a quantum superposition collapses, when many possibilities snap into one. Neven and Koch say it happens when one forms, and in their version a bigger superposition should mean a richer experience. Now read the mouse study with that in mind. A nucleus with spin can sit in two states at once, like a coin still spinning on the table, neither heads nor tails yet. Neven's theme suggests that inside a brain, that spin acts like a qubit, adding itself to whatever superposition is already there. And that extra quantum room is what pushes back against the anaesthetic. So the anaesthetic was fighting its own nucleus. The paper doesn't stop at theory. It lays out three experiments in order and it reads like a schedule. First, repeat the xenon test on something smaller than a mouse. Brain organoids, 2 to 3 millimeters across, roughly a hundred thousand human neurons grown from stem cells. And fruit flies, which carry about the same count. Expose them to the four isotopes, watch which ones they resist. Scientific American reported in August 2024 that this part was already underway, paid for by a private foundation in Santa Monica called the tiny blue dot. Second, take two well-characterized qubits and try to link them through living tissue. Microtubules, receptor proteins, nerve cells, an organoid. If the tissue can carry entanglement between the two qubits, then the link in the third experiment has a physical path to run on. Third is the one the paper itself calls the crucial experiment. The one the first two exist to make possible. And I'll read it as they wrote it. One would establish a physical link between a human brain and a quantum computer that would enable coherent interactions and mediate entanglement. Think of the brain as a piano with a certain number of keys. Wire in a chip with its own keys, and every chord you could play multiplies. Their prediction is that the volunteer would report richer conscious experiences, requiring more descriptive bits than the experiences the human reports without the link. As the debrief quoted Neven in January 2025, this expanded superposition could enrich the subject's conscious experience. And in Scientific American, in August 2024, Neven and Koch compared what it might feel like to a psychedelic experience. Then the same paper admits the problem. Such experiments are technologically very challenging and may require invasive technologies. Current brain-computer interfaces, in the author's words, are not designed to support a coherent coupling. They call for nitrogen vacancy probes combined with optogenetic methods. So they have the plan, and they have the chip. And by their own account, nobody has built the connector yet. Which leaves the question underneath all of it. Is there anything quantum going on in a brain at all? For decades the standard answer was no, because of one calculation from the year 2000. First though, a small college in Massachusetts ran the smallest experiment in this story. eight rats and one drug. In August 2024, a lab at Wellesley College published a rat study in eNeuro. Chan, Huang and colleagues, working in Wiest's lab, gave eight male rats a drug called epothilone B, 0.75 mg per kilo. The drug binds to microtubules, the hollow protein scaffolds inside every neuron, and makes them stiffer. Then the rats breathed 4% isoflurane, and the team timed how long each one took to stop riding itself. The standard test for losing consciousness. With the tubes stiffened, the rats took 69 seconds longer to go under. The effect was big enough that eight animals were enough to see it, and the odds of getting it by chance were about 1 in 600. That matters because of a 30-year-old prediction. In the mid-90s, Penrose and an anesthesiologist named Stuart Hameroff argued that consciousness runs on quantum activity inside those exact microtubules, and that anaesthetics work by disrupting it. If that's true, protecting the tubes should protect consciousness. And that is what these eight rats showed. And if that were the only problem for the skeptics, they might have shrugged it off. But in May 2025, Wiest published a review in Neuroscience of Consciousness, stacking the rest. flashing in step like a field of fireflies a burst of light only quantum theory predicts at room temperature vibrations in those tubes that carry across several neurons the way a hum travels through the walls of a house and change the voltage on the cell wall and a 2022 MRI study by Kerskenz and Lopez Perez that reported the signal consistent with entangled proton spins in living human brains a signal that tracked short-term memory scores and differed between sleep and waking. Wiest goes further than data. His argument is that entanglement is the only physics he knows of that turns a brain full of separate signals into one experience. So what has Google's machine actually done? Because its own announcement never mentions a brain. On October 22nd, 2025, Google Quantum AI published in Nature that its Willow chip, using between 65 and 103 qubits, ran an algorithm they call quantum echoes in about two hours. A job they estimate would take the frontier supercomputer 13,000 times longer. The claim is carefully hedged. They spent 10 person years trying to beat their own result with nine classical algorithms, and a second quantum machine can check the answer. Alongside it, with the Pines Magnetic Resonance Center at Berkeley, they used the same trick as a molecular ruler on two organic molecules measuring their shape through nuclear spins. That phrase, nuclear spins, is the same property that separated the four xenon gases. The Willow write-up says nothing about brains or consciousness and the consciousness paper carries the same labs name Google quantum AI Los Angeles on its first two authors. Same lab, two papers and the one with the chip never mentions the other, except the brain is already on the machine. On May 28th, 2026, two years to the day after the entropy paper, King's College London announced it had won time on Willow, through Britain's National Quantum Computing Center, the first time Google has opened the chip to a British government institution. The team is led by Eleanor Crane, and the project is to program what they call a mathematical analogy for neurons in the brain on Willow's 105 qubit array. Their named collaborator is Christopher Timmerman of the UCL Center for Consciousness Research, a neuron model running on Google's quantum chip with a consciousness lab in the room. Nobody involved has called it step two of Nevin's list, but a neuron model on Willow with a consciousness lab attached is the closest thing to it I can find. The next lab in this story is 1400 meters under a mountain in Italy, and none of the red videos go there. Subscribe, because what that lab measured decides which version of this theory is still standing. Gran Sasso, Italy. A physics laboratory buried under the Apennines, shielded from cosmic rays by 1400 meters of rock. A team led by Catalina Cucciano from the Frascati National Laboratory and Laios Diozzi from from Eötvös Loránd University, went looking for something Penrose's original theory requires. If gravity is what collapses a superposition, the collapse should give off a faint spray of radiation. So they set a cylinder of germanium in the dark and listened. They heard background and nothing else. Physics World reported the result in July 2022, and the numbers are brutal. To hold a superposition at the scale of an atomic nucleus, the model needs something like 10 to the 23 tubulin proteins acting together. A whole human brain contains about 10 to the 20. That's a thousand times short. The authors wrote that these considerations seem to rule out tubulin separation at the level of the atomic nuclei. So the version of quantum consciousness you've seen in every other video on this subject, the one where the rats prove Penrose right, took a direct hit under that mountain. Now Now go back to the entropy paper. Neven and Koch's version sidesteps that result. Their conscious moment happens when a superposition forms, so it doesn't need gravity or the collapse at all. The paper doesn't say whether Gran Sasso is why they flipped Penrose. Either way, that test never touches their version. It only needs a superposition to form somewhere inside a living skull at body temperature. Which is the objection I promised, the one waiting since the year 2000. That year the physicist Max Tegmark calculated how long a quantum state could survive in brain tissue. His answer was 10 to the minus 13 seconds. A thought takes about a third of a second, set the two side by side, and it's one second next to a hundred thousand years. A review published this April still starts from that number. He may have been too pessimistic. A review by Ma and Wang in Frontiers in Psychology published April 29th, 2026, collects the newer estimates, and they land between a microsecond and 10 nanoseconds. That's 10 million times longer than Tegmark's number, and still hundreds of thousands of times shorter than a thought. Ma and Wang put the whole field in one sentence. Even if the microtubule physics is real, it might at most identify a mechanism associated with consciousness, not a solution to why first-person experience has its particular character. But the mice have a rival explanation now. In May 2026 two researchers at Caltech, Wang and Osturk, posted a model in which the xenon spin effect comes from how spin carrying atoms move through the chiral twisted molecules of the body without needing long-range quantum coherence. If they're right the mice were reacting to spin and consciousness drops out of explanation that puts everything on one test and the same Frontiers Review names it if the vibrations in microtubules seen in an awake brain fade under anesthesia that is strong evidence in April 2026 Popular Mechanics reported that Leah Gasa at the University of Waterloo and Onur Pushuluk at Kadir Has University want to build exactly that a terahertz scanner that watches those vibrations in a living brain without opening the skull no such device has touched the human head yet. Before any of those count, Neven's own paper carries a warning, and it applies to every result I've just given you. Two out of three experiments in the biomedical sciences are not reproducible. The mouse study came from one lab, the rat study from one lab with eight animals, and the MRI signal from one team, with a published counter-explanation from 2023 that says ordinary water protons could produce it. None of them has an independent replication i could find that's why the organoid experiment matters more than the brain link if a hundred thousand neurons in a dish resist the spin carrying xenon the way the mice did then the effect lives in the tissue and the third experiment has something real to build on put all of it together the gas result stands and there are now two explanations for it and no test between them the mechanism the red videos say those rats prove was ruled out on the Gran Sasso and the one Neven wrote in its place has never been tested and the people who wrote that test are the people who own the chip. Nothing in it has touched a human and by Neven's own account no current interface can hold the coupling. Organoids, kings, the scanner. When any one of those reports a number that's the next video.