Quantum computing's real breakthrough, and the hype it does not justify
The milestone that mattered was not a bigger chip. It was proving that a quantum computer's errors shrink as you scale it up, the thing researchers chased for thirty years. Here is what actually changed, and why your encryption is still safe for now.
Quantum computing crossed a genuine milestone: Google's Willow chip showed that its error rate goes down as the machine gets bigger, not up. That "below threshold" result, published in Nature in late 2024, is the thing physicists had been chasing since the 1990s, and it turns quantum computing from a physics question into an engineering one. But the honest part matters too. Today's machines are still small and noisy, a genuinely useful quantum computer is roughly a 2030 story, and breaking modern encryption would still take on the order of a million qubits, orders of magnitude beyond today's machines. The breakthrough is real. The "quantum breaks everything tomorrow" headlines are not.
Quantum computing is the strangest beat in technology, because "a historic breakthrough just happened" and "it still cannot do anything useful" are both completely true right now. The trick is understanding which breakthrough actually happened, because it was not the one the headlines usually pick. It was quieter, more technical, and far more important than another record qubit count.
What actually happened?
In late 2024, Google's Quantum AI team ran an experiment on a 105-qubit chip called Willow and demonstrated something called below-threshold error correction. Published in the journal Nature, it is the result that matters, so it is worth explaining plainly.
They took a grid of physical qubits and used many of them together to protect a single unit of reliable information, a "logical" qubit. Then they made the grid bigger, stepping it up from a 3-by-3 array to 5-by-5 to 7-by-7. The key finding: each time they scaled the grid up, the error rate roughly halved. Bigger machine, fewer errors.
That sounds mundane. It is the opposite. For decades, adding more qubits added more noise faster than you could correct it, so scaling up made things worse. Crossing "below threshold" means that relationship finally flipped: past a certain quality of hardware, adding qubits now makes the machine more accurate, not less. Physicists had been chasing that crossover since the 1990s. Willow is the first convincing proof it works on real hardware.
Why is error correction the whole game?
Because a qubit is a coward. Unlike the solid 0 or 1 in your laptop, a quantum bit holds a fragile blend of states that collapses if you so much as look at it wrong: heat, vibration, a stray magnetic field. Left alone, qubits make errors constantly.
The fix, in theory, is to spread one piece of reliable information across many physical qubits, so the group can detect and correct its own mistakes faster than they appear. That is quantum error correction, and it is the entire ballgame. A machine full of flawless, error-free qubits is science fiction; a machine with enough noisy qubits arranged so their errors cancel out is an engineering roadmap. Willow's result is what tells you the roadmap is real, which is why people who understand the field cared far more about it than about any headline qubit number.
Did Google not say it beat a supercomputer?
Yes, and this is where honesty matters. Google reported that Willow ran a specific benchmark in about five minutes that would take the world's fastest supercomputer an almost comic ten septillion years, that is a 1 followed by 25 zeros, far longer than the universe has existed.
It is a real result and a stunning number. It is also close to meaningless for you. The benchmark is a task deliberately designed to be easy for a quantum chip and brutally hard for a classical one; it does not do anything useful. It proves the hardware is genuinely quantum and genuinely powerful, not that it can yet solve a problem you would pay for. Treat "beats a supercomputer" as a physics demonstration, not a product.
So can I buy a quantum computer that does something useful?
No, and not soon. Today's machines run on the order of a thousand-plus noisy physical qubits, and after error correction that leaves only a handful of reliable logical ones, nowhere near enough for real work. The useful applications people actually want, new materials, better batteries, drug discovery, hard optimisation, are widely put at roughly 2030, not now.
The roadmaps are at least concrete. IBM has publicly committed to a large fault-tolerant machine it calls Starling by 2029, designed to run on 200 logical qubits, and treats the machines before it as stepping stones rather than the destination. That is the honest shape of it: the science turned a corner, but the genuinely useful product is still years of hard engineering away.
Should I worry about my encryption?
Not today, and this is the myth worth killing. Breaking the encryption that protects your bank and your messages would take a fault-tolerant machine on the order of a million physical qubits. One closely watched estimate was recently revised down from about 20 million qubits to under a million, part of why the mood has shifted, but today's best machines hold only around a thousand. We are still orders of magnitude short.
What changed is the timeline, not the present danger. Progress like Willow's has moved expert estimates for "Q-Day", the day a quantum computer can break current encryption, from "decades away" to "years away", with some serious voices pointing at the end of this decade. The sensible response is already underway: new post-quantum encryption standards were finalised by the US standards body in 2024, and organisations are migrating to them now. The real near-term risk is "harvest now, decrypt later", where an adversary stores your encrypted data today to crack once the hardware exists. For individuals, there is nothing to do but let the services you use finish upgrading. There is no reason to panic, and every reason for the people running critical systems to move.
So has quantum computing turned the corner?
The science has. The product has not. That is the honest headline, and it is more interesting than either the hype or the dismissal. Below-threshold error correction converts quantum computing from an open question about whether it can ever work into a very hard engineering problem about how fast we can build it, and that is a genuinely historic shift. It just does not mean a useful quantum computer is sitting on a desk, or that your secrets are exposed. Both of those are years out, and anyone selling you certainty in either direction is selling something. For more from the lab, see the Science section, and how it connects to AI.
For the first time, we can see whirlpools on the Sun's surface. They may explain why its atmosphere is so strangely hot
The world's largest solar telescope has taken the sharpest images ever of the Sun's surface and caught something long predicted but never seen there: tiny plasma whirlpools, only about 20 km across, the same wave-curl you get in ocean waves and clouds. They may be part of the answer to a decades-old mystery. Here is what was found, and what it does and does not prove.
JWST found a 'black hole star' in the early universe, and it might solve a real puzzle
The James Webb telescope spotted an object 660 million years after the Big Bang that looks like a giant star but is really a black hole wrapped in so much gas it disguises itself. If the interpretation holds, it helps explain one of astronomy's genuine mysteries: how monster black holes appeared so impossibly early. Here is the careful version.
A blood test can now help diagnose Alzheimer's. Here's what that really means
Diagnosing Alzheimer's used to require an expensive brain scan or a spinal tap. In 2026 a simple blood test crossed from the research lab into the clinic, sharply improving how accurately doctors can identify the disease. It is a genuine turning point, and also not a crystal ball. Here is what these tests do, and what they do not.