Quantum Computing Explained Simply: What It Is and Isn't

You hear big claims about quantum computers. They will break the internet. They will cure diseases. They will make AI smarter. Some of those claims have a real idea behind them. Many do not.

This page gives you a simple, honest picture. It covers three things: what a quantum computer is, what it is not, and what it can really do today. Each claim links to the evidence, so you can check it yourself.

A normal computer stores everything as bits. A bit is like a light switch. It is either off (0) or on (1).

A quantum computer uses qubits. A qubit also gives you a 0 or a 1 when you read it. But before you read it, it holds two numbers called amplitudes. An amplitude is a number that says how strongly the qubit leans toward 0 or toward 1. Square an amplitude and you get the chance of that result.

Here is a small example. Say a qubit's amplitude for 0 is about 0.71. Then 0.71 × 0.71 is about 0.5. So there is about a 50% chance you read 0. The same goes for 1. Our lesson on qubit vs bit shows this step by step.

A coin spinning in the air is a common picture for this. It helps a little, but it stops working fast. A spinning coin is really heads or tails at every moment. You just do not know which. A qubit is different. Its amplitudes can be negative. That one fact is what makes quantum computing work.

The secret is interference. Think of waves in a pool. When two wave tops meet, they add up into a bigger wave. When a top meets a dip, they cancel out and the water goes flat. Noise-cancelling headphones use the same trick on sound.

Amplitudes behave a bit like those waves. A positive amplitude and a negative amplitude can cancel. Two positive ones can add up. A quantum program is a list of steps called gates. A good program arranges the gates so wrong answers cancel and right answers add up.

Then you measure. Measuring turns each qubit into a plain 0 or 1, at random, by those chances. If the program did its job, the right answer is now the most likely thing you will see.

This is why a quantum computer does not simply "try every answer at once". If it did that and then you looked, you would get one random answer. That is useless. The hard, clever part is the canceling. Only some problems have a known way to set it up.

Qubits can be linked so that their results depend on each other. This is called entanglement.

Picture two sealed boxes. Open one and see a 0, and you know the other shows 0 too. Open one and see a 1, and the other shows 1. That part sounds like two matching socks packed by a friend. But entangled qubits do more than matching socks. Tests show no "packed in advance" story explains all their results. The math of amplitudes does.

Entanglement is what lets many qubits act as one big system. Without it, a quantum computer would just be a pile of separate coin tosses. You can build the simplest entangled pair below.

The H gate put qubit 0 in an even state: half a chance of 0, half a chance of 1. The CX gate then flipped qubit 1 only when qubit 0 was 1. That tied the two qubits together.

Count the results. There are four possible outcomes for two qubits: 00, 01, 10 and 11. You should only see two of them, 00 and 11. Each shows up about 512 times, because half of 1,024 is 512. The other two never appear. Each result on its own is random. But the two qubits always match.

You can change this circuit and run it again in our Lab.

Here are the common mix-ups, in short.

• It is not a faster normal computer. It will not make your spreadsheet, games or web browser faster. It helps only with problems that have a special quantum method.
• It is not a replacement for your laptop. Every quantum chip needs a normal computer to control it. See what is a QPU.
• It is not breaking your passwords today. As of 2026, no quantum computer has factored anything larger than toy numbers like 15 and 21. See can quantum break RSA-2048?
• It is not judged by qubit count alone. Qubits make mistakes. A smaller chip with fewer errors can do more. Our guide to reading hardware specs explains why.

We cover more of these in quantum computing myths.

Qubits are fragile. A tiny bit of heat, a stray signal or a small shake can nudge them. Then the amplitudes drift, and the result goes wrong. This is called noise.

So builders go to great lengths. Some chips are cooled to a few thousandths of a degree above absolute zero. Others hold single atoms in place with lasers inside a vacuum. You can see the main designs in types of quantum computers.

Even then, every gate has a small chance to fail. Long programs pile up those small chances. The long-term fix is error correction. It spreads one reliable qubit across many real ones. Our page on logical vs physical qubits explains how.

Here is the honest list, as of 2026. Each item links to a page with sources and dates.

• Beating normal computers on test problems: yes, narrowly. Quantum machines have beaten the best known classical methods on made-up benchmark tasks. But no quantum computer has yet beaten normal computers on a problem anyone would pay to solve. See quantum advantage.
• Error correction: early, real progress. Error correction that gets better as you add qubits was shown on real hardware in December 2024. Machines with hundreds of reliable qubits are still only plans. See fault tolerance.
• Drug discovery: not yet. Real hardware has simulated only molecules of a few atoms. See drug discovery.
• Optimization: no shown speedup on any real-world task. See optimization.
• Machine learning: no shown speedup for normal data. See machine learning.
• Learning and research: yes. You can run small circuits on real chips over the cloud, and on simulators for free.

The theory says quantum computers should one day help with a few big things. Simulating molecules and materials is one. Breaking some kinds of encryption is another. Both need large, error-corrected machines that do not exist yet.

If you want to try it, open the Lab and build the circuit above yourself. If you want a guided path, read how to learn quantum computing. It maps our courses from zero math to error correction.

A good first stop is the Quantum Computing for Developers course. It starts from bits and builds up to your first real experiment. You can also start with single topics, like superposition or Grover's search.