Quantum Computers Aren't Magic. Here's the Part Everyone Skips.

Every few months, some headline says quantum computers just broke encryption or solved a problem that classical computers could never solve. Then a physicist clarifies that it was a 5-qubit system running at near absolute zero that factored the number 15.

I got into quantum computing because I was into astrophysics and classical physics, but I’d never really explored quantum computing, and I wanted to learn about a completely different side of physics.

Here's what actually happens. A classical computer uses bits: 0 or 1. A qubit can be in a superposition of both until you measure it, at which point it collapses to one. That part everyone's heard of. What nobody thinks about or explains is why that's useful, and more importantly, why it's so hard to use.

The trick isn't that a quantum computer tries every answer at once. That's the most common misconception out there, and it's wrong. If it did that, you'd just measure it and get a random answer, which is useless. The real trick is interference. You design your algorithm so that wrong answers cancel each other out (destructive interference) and right answers reinforce each other (constructive interference). So when you finally measure, the correct answer has the highest probability of showing up. That's the thing about quantum computing: it's all probability, never certainty.

The reason quantum computers aren't solving anything big right now is that qubits are hard to maintain and hard to scale. Your phone's processor can sit on your desk and get warm. A qubit needs to be isolated from basically everything (vibrations, heat, stray electromagnetic fields) because any interaction with the environment collapses the superposition. This is called decoherence, and it's the problem a lot of people miss when talking about quantum. Right now, most quantum processors operate at temperatures colder than outer space. The error rates are still high enough that you need error correction, and error correction requires many physical qubits per logical qubit; estimates range from hundreds to thousands. So when a company says they have a 1,000-qubit machine, that doesn't mean 1,000 reliable logical qubits doing useful computation. It's more complicated and less impressive than the headline suggests.

That's not me saying quantum computing is fake. Shor's Algorithm, which can break RSA encryption, is mathematically proven to work on a quantum computer. Grover's Algorithm gives a real boost in speed for search problems. These aren't hypothetical, but Shor's needs thousands of error-corrected qubits to break real-world encryption keys. We're not there. We're years away, maybe more.

What frustrates me is that the actual progress is interesting enough without the hype. People are solving real problems in error correction, in qubit design, in the algorithms themselves. The gap between "a quantum computer exists" and "a quantum computer threatens global security" is enormous, and crossing it requires engineering breakthroughs that aren't guaranteed. The theory is solid, and some early applications in chemistry simulation and optimisation look really promising.

Next time you see a headline about quantum computing, ask what problem it solved, how many qubits it used, and whether those qubits were error-corrected.