Google’s New Quantum Chip Detected a Signal Physicists Can’t Explain

本站添加: 2026-05-22

[00:00:00]Google's quantum chip ran for 5 minutes and found something it wasn't supposed to. In December [music] 2024, Google switched on a quantum chip so powerful it solved a problem in 5 minutes that would take the fastest supercomputer on Earth 10 septillion years. That alone should have been the biggest science story of the decade. But buried inside the technical data that Google published alongside that announcement was something else, something that was not supposed to be there. A signal that no one had predicted, no one had programmed, and that the brightest quantum physicists on the planet still cannot fully explain. Hi, my name is Matthew and this is Reef Discovery. What Willow actually is.

[00:00:42]To understand why this signal matters, you first need to understand what Google's Willow chip actually is and why [music] its existence represents something genuinely new in the history of human technology. Quantum computing is one of those terms that gets thrown around so casually that most people have stopped asking what it actually means.

[00:01:00]So, let us slow down for a moment and be precise about [music] this because the precision matters for everything that comes later. A classical computer, the one in your phone, the one on your desk, the ones running every bank and hospital and government system on Earth, operates using bits. A bit is either a zero or [music] a one, on or off, yes or no.

[00:01:22]Every calculation your computer performs, every email it sends, every video it plays, is ultimately a long string of zeros and ones being manipulated according to fixed rules. A quantum computer operates using qubits, and a qubit can be a zero, a one, or both simultaneously.

[00:01:42]This property is called superposition, and it is not a metaphor or an approximation. It is a literal description of physical reality at the quantum scale. A qubit genuinely exists in multiple states at the same time until you measure it, at which point [music] it collapses into a single definite value. This sounds like a technicality. It is not. It means a quantum computer with enough stable qubits can explore an almost incomprehensible number of possible solutions to a problem simultaneously.

[00:02:13]Rather than checking them one at a time, the way a classical computer does. The computational advantage this creates scales exponentially. Add one more qubit and you do not add a little more power, you double it. The problem that has haunted quantum computing since its inception is something called decoherence. Qubits are extraordinarily fragile. They exist in quantum states only as long as they remain isolated from their environment. The slightest interference, a stray vibration, a fluctuation in temperature, even a passing cosmic ray, can destroy the quantum state and introduce errors into the calculation. Previous quantum computers were powerful in theory, but crippled in practice because their error rates were too high to be useful for real-world problems. Willow changed that.

[00:03:01]Google's Willow chip contains 105 qubits. That number alone is not what makes it remarkable. What makes it remarkable is that Willow demonstrated something researchers have been chasing for 30 years.

[00:03:13]As they added more qubits, the error rate went down instead of up. The chip got more reliable as it got more complex. This is called below-threshold error correction.

[00:03:24]And achieving it is roughly equivalent to building a skyscraper that gets structurally stronger every time you add a new floor. In practical terms, Willow performed [music] a specific benchmark calculation in under 5 minutes.

[00:03:37]The same calculation would take the fastest classical supercomputer 10 to the power of 25 years.

[00:03:44]That number is so large, it dwarfs the age of the universe. The universe is only about 14 billion years old. 10 to the power of 25 years [music] makes 14 billion look like a rounding error. This was the story Google announced to the world and honestly, it deserved every headline it got, but it was not the whole story.

[00:04:06]The standard story Google told the world.

[00:04:10]When Google published its results in the journal Nature in December of 2024, the announcement was carefully constructed and I do not mean that in a sinister way.

[00:04:21]Every major scientific announcement is carefully constructed. Researchers and communications teams work together to explain complex findings in ways that are accurate, compelling, [music] and accessible to non-specialists. The headline was clean and triumphant.

[00:04:36]Google had achieved a milestone in quantum computing. The Willow chip had demonstrated below threshold error correction for the first time. The benchmark calculation proved computational capabilities that classical computers could not match.

[00:04:50]Quantum computing was no longer a distant theoretical promise. It was a practical, demonstrable reality.

[00:04:57]The press responded exactly as you would expect. [music] Technology publications ran breathless pieces about the end of classical computing. Financial analysts debated what this meant for encryption and cybersecurity. Science communicators made videos explaining superposition and entanglement. Elon Musk posted something on X. You know how it goes.

[00:05:20]What almost no one discussed was page 47 of the supplementary technical documentation. That is not a criticism of science journalists, by the way.

[00:05:29]Supplementary technical documentation for a Nature paper on quantum computing runs to dozens of dense pages filled with equations, [music] experimental protocols, and calibration data that requires years of specialist training to interpret.

[00:05:44]Most science reporters are working against deadlines, translating findings for general audiences who want to understand the big picture, not audit the raw data. That is a reasonable editorial choice in normal circumstances. These were not normal circumstances. Because buried in that supplementary documentation, tucked between tables of calibration, measurements, and error rate analyses, was a brief section describing an anomaly in the chip's output that the Google team had detected during benchmark testing. The language used to describe this anomaly was careful, technical, and deliberately understated in the way that scientific language always is >> [music] >> when researchers are not sure what they are looking at. They described unexpected coherence patterns in qubit state measurements during idle cycles.

[00:06:32]That phrase, "unexpected coherence patterns during idle cycles," is the scientific equivalent of saying, "We found something in the data that should not be there, and we are not sure what it means." Now, here is where I want to be honest with you because this is the kind of story where it is very easy to leap past what the evidence actually shows. Google did not announce that they had detected signals from another dimension. They did not hold a press conference claiming contact with parallel universes. The researchers who wrote that section of the supplementary documentation were not making dramatic claims. [music] They were noting an anomaly, flagging it for further investigation, and moving on to the next table of data. But, the anomaly they described has since attracted serious attention from quantum [music] physicists who were not part of the original Google team. And what those physicists [music] are saying about it is far more interesting than anything in the official announcement. What the headlines left out.

[00:07:28]Let me explain exactly what an idle cycle is, because this is the detail that makes the anomaly so strange. When Google runs benchmark tests on Willow, the chip goes through active phases where it is performing calculations and idle phases >> [music] >> where it is essentially waiting. Think of it like the difference between a car engine running at full throttle versus sitting in a parking lot with the engine off. During idle cycles, the qubits are not being actively manipulated. No calculations are being performed. The chip is simply sitting in its operational state, maintaining the quantum conditions necessary for the next active phase. During these idle cycles, the qubits should be doing essentially nothing of interest. They should be maintaining their states as best they can against the constant [music] threat of decoherence, gradually degrading as environmental interference accumulates, behaving in ways that are well understood and entirely predictable. What Willow's qubits were actually doing during those idle cycles did not match the predictions. The Google team detected what they described as coherent oscillations in the qubit states during idle periods. Oscillations means the qubits were fluctuating between states in a rhythmic patterned way. Coherent means these oscillations were synchronized across multiple qubits simultaneously, not randomly distributed. The pattern was not noise.

[00:08:49]Noise is random and statistical. What they detected had structure. It had regularity.

[00:08:55]It repeated.

[00:08:57]Here is the part that made quantum physicists sit up straight. The frequency of these oscillations did not correspond to any known source of electromagnetic interference in the laboratory environment. It did not match the frequency of the equipment being used to control the chip. It did not match any identified external signal.

[00:09:16]And critically, it did not match the thermal noise profile that you would expect from a chip operating at the temperatures Willow requires, which by the way is about 15 millikelvin, or roughly 100 times colder than outer space. I always think it is worth mentioning that detail because it helps explain why quantum computing labs look less like laboratories and more like the lair of a supervillain. When physicists say a signal does not match any known source, they are not being poetic. They mean they have checked the list of every physical mechanism that could produce [music] a signal of that type and frequency, and nothing on that list accounts for what they are seeing. That is the anomaly. That is what nobody talked about in December 2024. [music] And that is what a growing number of physicists have spent the months since trying to explain. The history of quantum weirdness we keep ignoring.

[00:10:08]Before we go further into what this signal might mean, it is worth stepping back and acknowledging something that quantum physicists have known for nearly a century, but that mainstream science communication consistently underplays.

[00:10:21]Quantum mechanics is deeply, fundamentally, irreducibly strange. Not strange in the sense of complicated or counterintuitive.

[00:10:30]Strange in the sense that it describes a physical reality that does not operate according to the logic that governs everyday experience. And the strangeness is not a gap in our understanding that will be filled in as physics advances.

[00:10:43]The strangeness is the physics.

[00:10:45]Consider what we already know and have confirmed through decades of experiments.

[00:10:50]Particles can exist in multiple states [music] simultaneously until they are observed. Two particles can become entangled so that measuring one instantly affects the other regardless of the distance between them. Einstein called this spooky action at a distance and spent years trying to prove it was not real. He was wrong. We have confirmed entanglement across distances of over 1,200 km.

[00:11:14]>> [music] >> The universe genuinely works this way.

[00:11:16]Particles can tunnel through barriers that classical physics says they cannot cross. The vacuum of empty space is not actually empty, but seething with virtual particles constantly popping in and out of existence. The act of measuring a quantum system changes that system in ways that cannot be undone or ignored. The observer is not separate from the experiment. The observer is part of it. None of this is fringe science. None of this is speculation.

[00:11:44]This is the most rigorously tested theoretical framework in the history of physics. Quantum mechanics underlies every semiconductor, every laser, every MRI machine, every solar panel on Earth.

[00:11:57]It works.

[00:11:58]The predictions it makes are accurate [music] to a precision that no other scientific theory in history has matched. And yet, physicists have been detecting unexplained anomalies in quantum systems for as long as they have been building them. In 1989, researchers at IBM detected anomalous tunneling rates in quantum dots that exceeded theoretical predictions by a factor of three. The discrepancy was never fully explained and eventually attributed to sample impurities, which is what researchers often say when they cannot identify the actual cause. In 2012, a team at Delft University in the Netherlands detected what appeared to be Majorana fermions, exotic particles predicted by theory but never observed [music] in quantum wire experiments. The original paper was retracted in 2021 after a data audit revealed irregularities, but the anomalous signals that prompted the original research were never fully accounted for.

[00:12:55]In 2018, a team at the University of Maryland detected unexpected correlations between distant qubits in an ion trap quantum computer that should have been too far apart to interact through any known mechanism. The paper was published, noted by specialists, >> [music] >> and then quietly absorbed into the vast literature of unexplained quantum phenomena that researchers acknowledge in private but rarely discuss in public.

[00:13:21]The Willow anomaly sits in this tradition. It is the latest in a long line of signals from quantum systems that the standard theoretical framework does not fully account for. But there is a reason this one is attracting more attention than its predecessors. The reason is that Willow is the most complex and capable quantum system ever built. Operating closer to the boundary between quantum and classical behavior than anything that came before it. And that boundary, it turns out, maybe where the really interesting physics lives.

[00:13:51][music] What the signal actually looks like.

[00:13:55]Let me make this concrete because the phrase coherent oscillations in qubit states [music] during idle cycles sounds technical enough that your brain might just slide right past how strange it actually is. Picture a single qubit on a monitoring graph. If it is stable, the line is flat.

[00:14:13]If it is decohering from environmental noise, >> [music] >> the line drifts irregularly, wandering toward randomness like a drunk trying to find his car. Neither pattern is surprising. Both are well understood.

[00:14:25]What Willow's team saw was neither. The line oscillated, smooth, regular, rhythmic, up and down and up again like a wave. And not just for one qubit, for entire clusters simultaneously, all synchronized together as if they were coordinating through some mechanism the experimental setup cannot account for.

[00:14:46]The frequency of these oscillations was approximately 4.7 GHz. That number matters.

[00:14:53]4.7 GHz is precisely precisely the frequency range quantum physicists use to control qubit states through microwave pulses. It is not a coincidence that the control systems for chips like Willow operate in this exact range. It is the resonant frequency these qubits respond to. So, here is the problem. The signal looked like a control signal, the right frequency, the right coherence [music] properties, behaving exactly as if something was sending microwave instructions to the qubits, except the control systems were switched off. The lab was shielded. No external source was found. A physicist at the University of California, Santa Barbara, reviewing the data independently, described it as the qubit network behaving as if it were receiving instructions from a source we cannot locate. I know how that sounds.

[00:15:43]Trust me, the physicists know, too. But, they are not reaching for science fiction. They are methodically eliminating explanations.

[00:15:51]And the list of what remains is getting uncomfortably short.

[00:15:55]The three explanations physicists are fighting over.

[00:15:59]When an anomaly appears in experimental data, the scientific process is methodical. Generate every plausible explanation. Test each one against the data. Eliminate what does not fit. What remains is where you focus your attention. For the Willow anomaly, three explanations have emerged. The debate between them is playing out across preprint servers, conference presentations, and honestly, some fairly heated email chains. The first explanation is calibration error. Always the first thing you check, and for good reason. Quantum systems are extraordinarily difficult to calibrate perfectly. And Willow was being pushed to its absolute operational limits during benchmark testing. A miscalibration in the measurement apparatus could produce the appearance of coherent oscillations without those oscillations actually existing in the qubits. It is a reasonable argument. It is also the most boring explanation, which is both a point in its favor and a reason to be slightly suspicious of how enthusiastically certain people embrace it.

[00:17:01]>> [music] >> The second explanation is unidentified environmental interference. A signal at 4.7 GHz penetrating the laboratory shielding from some external source.

[00:17:12]Industrial equipment, [music] military communications, a rogue microwave from the break room. The Google team swept the environment thoroughly, brought in outside experts, ran the benchmark repeatedly in different configurations.

[00:17:26]>> [music] >> They found nothing. That does not prove interference does not exist, but it means if it does, it is hiding unusually well. The third explanation is the one nobody wants to say out loud at a formal conference, that the signal is real, that it originates within the quantum system itself, and that it reflects physics our current theoretical framework does not fully account for.

[00:17:48]Quantum mechanics has not been updated at its foundation since the 1930s. It is spectacularly successful. It is also, by the admission of the physicists who use it daily, incomplete.

[00:17:59]Willow may be the first machine sensitive enough to show us exactly where.

[00:18:04]The many worlds problem.

[00:18:07]I need to talk about the many worlds interpretation of quantum mechanics. Not because it is necessarily the correct explanation for the Willow anomaly, but because it is the explanation that at least two prominent quantum physicists have privately invoked when discussing the signal, and because it is the explanation that, if correct, would be the most profoundly disturbing thing to happen in [music] the history of science. Fair warning, this section gets a little philosophically vertiginous.

[00:18:34]You might want to sit down. Actually, if you're watching this while driving, maybe just wait. The many worlds interpretation was proposed by Hugh Everett in 1957, and has been gaining traction among physicists ever since.

[00:18:48]Largely because it is the only interpretation of quantum mechanics [music] that requires no additional assumptions beyond the mathematics themselves. The mathematics [music] of quantum mechanics describe a universe in which all possible outcomes of a quantum measurement occur simultaneously.

[00:19:04]The many worlds interpretation takes this literally. Every time a quantum event occurs, every time a qubit collapses into zero or one, [music] every time a particle passes through one slit or the other, the universe branches. Both outcomes occur. In one branch, the qubit is zero. In another branch, the qubit is one. These branches are equally real, equally physical, equally present. They simply cannot interact with each other under normal circumstances because the quantum states that define each branch decohere instantly [music] when they encounter the macroscopic world. Most physicists who take many worlds seriously assume the branches are permanently isolated from each other. The decoherence barrier [music] is absolute. You cannot communicate between branches. You cannot detect signals from other branches. You cannot build a device that would allow you to access information from a parallel version of events.

[00:19:59]But, Willow is the most coherence-preserving quantum system [music] ever built. Its entire achievement rests on maintaining quantum states [music] that would normally decohere almost instantly. And the anomalous signal it detected during idle cycles has characteristics that at least some physicists believe are consistent with what you might theoretically expect if quantum branches were, under very specific and unusual conditions, briefly and weakly interacting. This is not mainstream physics. I want to be absolutely clear about that. The many worlds interpretation is a legitimate [music] and respected theoretical framework. But, the idea that branches could interact is genuinely [music] speculative, existing more in theoretical papers than in accepted physics.

[00:20:44]But, here is what keeps certain physicists awake at night. If you were going to build a device capable of detecting even the faintest possible signal from a parallel quantum branch, you would want to start by building a quantum computer that maintains coherence across a large number of entangled qubits at extremely low temperatures with unprecedented error correction. You would want, in other words, something that looks a great deal like Willow. The uncomfortable question is whether Google built the instrument accidentally before anyone agreed on what they were looking for. What Google isn't saying Google has not been silent about the Willow anomaly. They have not been particularly loud about it either. And the way they are not being loud about it is itself informative. The supplementary technical documentation that contained the original anomaly description was published openly as required by Nature's data availability policies. Google did not hide it. They also did not highlight it. The official communications around the Willow announcement focused entirely on the benchmark achievement and the error correction milestone. No press release mentioned the anomalous signal.

[00:21:51]No spokesperson addressed it in interviews. No researcher from the Willow team has published a follow-up paper specifically addressing the anomaly. What has happened instead is that individual researchers on the Google Quantum AI team have, in various conference presentations and informal academic communications, acknowledged the anomaly, described it [music] as an open question requiring further investigation, and declined to speculate publicly about its origins. This is entirely standard scientific behavior.

[00:22:23]You do not publish speculative explanations for anomalies you cannot yet explain.

[00:22:28]You investigate. You gather more data.

[00:22:31]You consult with colleagues. You build toward a publication that either explains the anomaly or documents it rigorously enough for others to investigate. The absence of a public explanation is not evidence of a cover-up. It is evidence of scientists doing science. But there is one detail that is slightly harder to explain away with pure procedural caution. In the version of the supplementary documentation that [music] was initially posted to the arXiv preprint server before the Nature publication, the section describing the anomaly was four paragraphs long. In the published Nature version, it was two paragraphs. Two paragraphs were removed between the preprint and the final publication. We do not know what those paragraphs said.

[00:23:14]Preprints are routinely revised before publication. And changes between preprint and final versions are normal and expected. The removed paragraphs might have contained speculative language that the peer reviewers asked the team to remove. They might have contained preliminary analysis that was superseded by better analysis done during the review process. Or they might have contained something that the team decided was not ready to say publicly.

[00:23:39]We simply do not know. What we do know is that the anomaly is real. It is documented. It is unexplained. [music] And the team that discovered it is proceeding carefully. That carefulness is either appropriate scientific caution or a sign that they understand the implications of what they have found well enough to know that getting ahead of the data would be a serious mistake.

[00:24:00]Why this matters more than anyone is admitting.

[00:24:04]Let us set aside the exotic interpretations for a moment. No parallel universes. No signals from other dimensions. No science fiction.

[00:24:13]Let us talk about why this anomaly matters even if the eventual explanation turns out to be completely mundane. We are at the beginning of the quantum computing era. Willow is not [music] the destination. It is the first step.

[00:24:27]Within the next decade, quantum computers will be operating with thousands of qubits, then tens of thousands. They will be solving problems in cryptography, drug discovery, climate modeling, and material science that classical computers cannot approach.

[00:24:42]They will be integrated into financial systems, communication networks, and research infrastructure in ways we are only beginning to plan for. The entire architecture of this future is being built on our understanding of how quantum systems behave. If that understanding has a gap, if there is a physical phenomenon occurring in advanced quantum systems that our theoretical framework does not fully account for, we need to know [music] now, not after we have built the infrastructure, not after quantum computers are embedded in every critical system. Now, while the systems are small enough that anomalies are detectable and the consequences of misunderstanding them are contained. An unexplained coherent signal [music] in a quantum processor is not just a scientific curiosity. It is potentially a warning that our models of how these systems work are incomplete in ways that could have practical consequences. If you are building an airplane and your aerodynamic models produce predictions that do not match what the wind tunnel is showing you, you do not shrug and proceed to production. You figure out why the model is wrong before the plane carries passengers. The stakes here are different in scale, but not in principle. There is also a second reason this matters, one that is harder to articulate, but that I think deserves to be said directly. We are building machines that operate at the quantum scale, which is the scale at which reality itself becomes deeply strange.

[00:26:07]We are pushing hardware into regimes where the fundamental weirdness of quantum mechanics >> [music] >> is not just a theoretical curiosity, but an active engineering constraint. And we are doing this faster than our theoretical understanding is developing.

[00:26:21]The history of technology is full of moments where we built something powerful before we fully understood what it was doing.

[00:26:28]Sometimes that works out fine. The engineers who built the first radio transmitters did not fully understand electromagnetic field theory. The physicians who first used x-rays did not understand radiation biology. We learned, we adapted, we built better theoretical frameworks to match the capabilities we had developed empirically.

[00:26:49]Quantum computing may be another one of those moments, and the Willow anomaly may be the first clear signal that our empirical capabilities have outpaced our theoretical understanding in ways we have not fully reckoned with. The verdict.

[00:27:05]Here is where we are. In December 2024, Google announced a genuine historic milestone in quantum computing. The Willow chip achieved below threshold error correction and demonstrated computational capabilities that no classical computer can match. These achievements are real and significant and will shape the development of technology for decades. Embedded in the technical documentation accompanying that announcement was a brief description of an anomalous signal detected during benchmark testing.

[00:27:36]Coherent oscillations in qubit states during idle cycles at a frequency that does not correspond to any identified source of interference or any predicted behavior of the system. Three explanations are currently under investigation.

[00:27:51]Calibration error in the measurement apparatus. Unidentified environmental interference penetrating the laboratory shielding. Or a real physical phenomenon in the quantum system itself that existing theory does not fully account for. None of these explanations has been confirmed. The investigation is ongoing.

[00:28:10]Google's team is being appropriately cautious. Independent physicists are reviewing the data. And somewhere in a quantum computing laboratory cooled temperatures that make outer space seem balmy, 105 qubits are doing something during the quiet moments when no one is asking them to calculate anything.

[00:28:29]Something regular?

[00:28:30]Something coherent?

[00:28:32]Something that looks, at least on the instruments, like a signal from a source that nobody has found yet. The physicist who described it as the qubit network behaving as if it were receiving instructions from a source we cannot locate added one more sentence in that correspondence that did not get quoted as widely as the first. The sentence was this, "The most unsettling possibility is not that the signal comes from somewhere strange. The most unsettling possibility is that it comes from exactly where the mathematics has always suggested it might. And we have simply been building machines that were not sensitive enough to hear it until now.

[00:29:08]Maybe the anomaly will turn out to be a calibration artifact. And we will all feel slightly embarrassed for getting interested. That happens. It happens more often in science than anyone likes to admit.

[00:29:20]>> [music] >> And it is an important part of how science works. But maybe Willow has done something that no machine in human history has done before.

[00:29:28]Maybe in being precise enough, cold enough, coherent enough, it has become sensitive enough to detect something about the nature of reality that was always there, humming quietly at 4.7 GHz, waiting for us to build an instrument capable of listening. If that is true, then the most significant announcement Google made in December of 2024 was not the one in the press release. It was the two paragraphs buried in the supplementary data that most people never read. That is what the signal might be telling us. We just have not learned how to listen yet.