A note on the word “experiment”
Nothing was run or measured for this article. “Experiment in progress” is a description of the field itself: researchers around the world are still working out how to build quantum machines that are large and reliable enough to be useful. This piece explains the main ideas, what is widely expected, and what remains uncertain.
What a quantum computer is
An ordinary computer stores information as bits, each of which is either 0 or 1. A quantum computer uses quantum bits, usually called qubits. A qubit is a physical system, such as a tiny circuit cooled to very low temperatures or a single trapped atom, that follows the rules of quantum physics. Different research groups use different physical designs, and no single design has clearly won.
Two features of that physics matter most.
- Superposition. Before it is measured, a qubit can be in a blend of 0 and 1, described by numbers called amplitudes. Measuring it gives just one result, 0 or 1, with probabilities set by those amplitudes.
- Interference. Amplitudes can add or cancel, much like ripples on water. A quantum algorithm is designed so that wrong answers tend to cancel and right answers tend to reinforce.
A third ingredient, quantum entanglement, links the states of several qubits so they cannot be described one at a time. It is part of what gives quantum machines their extra reach.
A common shortcut is to say a quantum computer “tries every answer at once.” That is misleading. The machine does not hand you all the answers; a measurement returns only one. The skill lies in arranging the computation so that the measurement is likely to give a useful result. If you want the details of how a qubit differs from a bit, our explainer on qubits goes further.
What quantum computers might be good for
Researchers expect quantum machines to help with a narrow set of problems rather than to replace laptops. The most commonly discussed areas are:
- Simulating nature. Molecules and materials are themselves quantum systems, so a quantum computer may model them more naturally than a classical one. This could matter for chemistry and materials research, though practical results are still being worked toward.
- Certain mathematical problems. A well-known algorithm could factor very large numbers much faster than the best known classical methods. This is the reason for the security discussion below.
- Some search and optimization tasks. Here the expected gains are often more modest, and whether they beat the best classical tools in practice is still debated.
Quantum computers are not expected to speed up everyday tasks like browsing, word processing or most work in machine learning. Claims that quantum machines will transform every industry should be treated cautiously.
Why they are hard to build
Qubits are fragile. Heat, stray electromagnetic fields and even the act of looking at them can disturb their state, an effect called noise or decoherence. Today’s devices make errors far more often than ordinary chips do.
The usual proposed remedy is quantum error correction: spreading one reliable “logical” qubit across many physical qubits and checking for errors continuously. This is widely believed to require far more physical qubits than current machines have, though estimates vary with the design and the task. Scaling up while keeping errors low is the central engineering challenge, and it is the reason a large, fault-tolerant machine is generally described as a long-term goal.
How far along is the field
Honest status reports tend to include the same points:
- Working quantum processors exist and are used for research, including through cloud access.
- Current devices are often called noisy and intermediate-scale: interesting for experiments, but limited by errors.
- Some demonstrations have shown a quantum device completing a specially chosen task that is very hard for classical computers. These tasks were designed to showcase the hardware and do not usually have direct practical value, and classical researchers have sometimes found faster ways to do them.
- Timelines for useful, error-corrected machines are uncertain. Forecasts range widely, and anyone giving a firm date is guessing.
The security angle
Because the factoring algorithm could, in principle, break some widely used public-key methods once a large enough machine exists, standards bodies and security researchers are preparing replacement methods. This work is called post-quantum cryptography, and it is a precaution rather than a sign that current encryption has already failed. Everyday message encryption relies on a mix of techniques, and not all of them are affected in the same way.
Beyond computing
Quantum effects are also used for measurement and communication. Quantum sensing aims at very sensitive detectors for things like magnetic fields or timekeeping, and some of it is closer to practical use than large computers are.
How to read the news
When you see a headline about a breakthrough, a few questions help. How many qubits are there, and how often do they make errors? Was the task chosen to suit the machine? Is the result compared against the best classical method? Progress is real and steady, but the gap between a laboratory demonstration and a dependable everyday tool is large, and it is the same kind of gap that other emerging computing approaches, such as neuromorphic computing or photonic computing, are also trying to cross.