Quantum Computing Is Real Now. The Hype Is Still Lying to You.

The Machine That Works and Doesn't

In late 2024, Google announced that its Willow quantum chip had solved a benchmark computation in five minutes that would take a classical supercomputer an estimated 10 septillion years. The number is so large it's essentially fictional, which is precisely the problem. The benchmark was designed to be hard for classical computers and easy for quantum ones. It measures nothing you would ever want to do. It is, in the most charitable reading, a proof of engineering. In the least charitable reading, it is a very expensive press release.

This is not a knock on Google's engineers, who are doing genuinely difficult work. It is a knock on the communication layer between that work and the public, which has been running at a consistent deficit of honesty for about a decade. Quantum computing is real. The machines exist, they operate, and they are getting better at a pace that serious researchers find notable. The gap is between what the machines can do and what the coverage implies they will do, and that gap is still wide enough to drive a classical supercomputer through.

What Quantum Computers Actually Do Well, Right Now

To be precise about the current state: quantum computers are good at a narrow class of problems involving probability distributions, optimization over large combinatorial spaces, and simulating quantum systems. That last one is the most important and the least discussed. Simulating how molecules behave at the quantum level is genuinely hard for classical computers because the state space grows exponentially with the number of particles. A quantum computer handles this more naturally because it is a quantum system. This has real implications for drug discovery, materials science, and catalyst design — areas where understanding molecular behavior at high fidelity could compress years of lab work.

IBM's quantum roadmap, which has been more transparent than most, shows processors in the hundreds-of-qubits range with error rates that are improving but still significant. The term of art here is fault-tolerant quantum computing, which refers to systems that can correct their own errors fast enough to run long computations reliably. We do not have that yet. What we have are noisy intermediate-scale quantum devices — NISQ, an acronym the field coined partly to manage expectations and partly because researchers needed a way to talk about the present without constantly apologizing for it.

The Error Problem Is Not a Detail

Quantum bits, or qubits, are fragile. They maintain their quantum state — the superposition that gives quantum computers their theoretical power — only under extremely controlled conditions. Vibration, heat, electromagnetic interference, even the act of measuring them can cause errors. Classical computers have errors too, but at rates so low they're effectively irrelevant for most computations. Current quantum processors have error rates that require significant overhead just to keep computations coherent long enough to be useful.

The path to fault tolerance runs through quantum error correction, which works by encoding one logical qubit across many physical qubits so that errors can be detected and corrected without collapsing the quantum state. The overhead is substantial. Current estimates suggest you might need anywhere from hundreds to thousands of physical qubits to produce one reliable logical qubit, depending on the error rate of the underlying hardware. Google's Willow chip showed meaningful progress on this ratio, which is why researchers paid attention. But progress on a hard problem is not the same as solving it.

Who Is Actually Using These Machines

The honest answer is: researchers, mostly. IBM's Quantum Network gives access to real quantum hardware through the cloud, and a community of academics and corporate research teams use it to explore algorithms, test error correction approaches, and probe the boundaries of what current hardware can do. Pharmaceutical companies have run early experiments on molecular simulation. Financial firms have explored quantum approaches to portfolio optimization, though classical algorithms remain competitive for the problem sizes that matter in practice.

What you will not find is a quantum computer running production workloads at scale. The machines are not replacing data centers. They are not breaking encryption — RSA and elliptic curve cryptography remain secure against current quantum hardware, though the cryptographic community is already standardizing post-quantum algorithms as a precaution, which is the correct and boring thing to do. The National Institute of Standards and Technology finalized its first post-quantum cryptographic standards in 2024. That process took eight years. That is how seriously the threat is taken, and also how far away it is.

The Timeline Problem

Predictions about when quantum computers will achieve practical advantage over classical machines for commercially relevant problems have been consistently optimistic and consistently wrong. In 2017, several major technology companies suggested five to ten years. We are now eight years into that window. The goalposts have not moved so much as they have revealed themselves to be further away than the original map suggested.

This is not unusual for deep hardware development. Transistor scaling followed a similar pattern of confident predictions meeting stubborn physics. The difference is that classical computing had a clear, measurable metric — transistor density — that correlated reasonably well with useful performance. Quantum computing's equivalent metrics are more complex: qubit count, coherence time, gate fidelity, connectivity, and error correction overhead all interact in ways that make simple projections unreliable.

The reasonable prediction, based on current trajectories, is that fault-tolerant quantum computers capable of running useful algorithms at scale are probably a decade away, possibly more. That is not a dismissal. A decade is not long in the context of a technology that could meaningfully accelerate drug discovery or materials research. It is, however, long enough that anyone building a business strategy around quantum advantage in the next three years should be doing so very carefully.

What to Watch Instead of the Headlines

The signal worth tracking is not announcement volume. It is error correction overhead ratios, coherence times at scale, and whether the gap between physical and logical qubits is narrowing at a rate that suggests fault tolerance is approaching rather than receding. Google's Willow results were notable on this metric. IBM's roadmap updates are worth reading for the same reason. Academic publications on error correction techniques are where the real progress gets documented, usually without the press release.

The other thing worth watching is the classical computing response. Quantum algorithms have been motivating improvements in classical simulation techniques for years — researchers trying to understand what quantum computers can do have, in the process, found better classical approaches to some of the same problems. This is not a reason to dismiss quantum computing. It is a reason to hold the competitive claims loosely.

Quantum computing is a serious technology being developed by serious people working on genuinely hard problems. It deserves coverage that matches that seriousness — which means being precise about what works, honest about what doesn't, and resistant to the temptation to make the timeline fit the narrative. The machines are real. The revolution is still in the error-correction phase.