Welcome to Part 7. So far we have learned the exciting parts: qubits, superposition, entanglement, gates, and algorithms. Now let us talk honestly about the biggest headache in quantum computing: things keep going wrong. This post explains decoherence and quantum error correction in the simplest possible way, using everyday stories.
what is decoherence?
Decoherence is when a qubit loses its delicate quantum properties (superposition and entanglement) and starts behaving like a normal classical bit.
It is like your perfectly spinning coin suddenly getting knocked by the wind and landing too early. Or a beautiful soap bubble that pops the moment a speck of dust touches it.
When does it happen?
- Every tiny bit of heat, vibration, stray electromagnetic wave, or even a cosmic ray can disturb the qubit.
- It happens while the qubit is just sitting there (waiting).
- It happens faster when we apply gates (doing operations).
- It happens before we can finish our algorithm.
Result: the superposition collapses on its own. Entanglement breaks. The quantum magic disappears, and we are left with random classical junk.
This is why quantum computers must be kept in giant dilution refrigerators at temperatures close to absolute zero (about minus 273 degrees Celsius). Even then, decoherence still occurs. Just more slowly.
how do we measure decoherence?
Scientists use two important numbers:
- T1 (relaxation time). How long a qubit can stay in the excited state (
|1⟩) before it naturally relaxes back to|0⟩. Analogy. How long a stretched rubber band stays stretched before it snaps back. - T2 (coherence time). How long the qubit can maintain its superposition (the "spinning coin in the air" phase). Analogy. How long you can keep a coin spinning nicely on a table before table vibrations make it wobble and fall.
Today's best qubits have coherence times from a few microseconds (millionths of a second) to a few hundred microseconds. That sounds short, and it is. But clever circuits can finish useful work within that window.
We also measure error rates per gate (how often a gate does something slightly wrong). Current good machines have error rates around 1 in 1,000 to 1 in 10,000 operations. For big algorithms we need roughly 1 in a billion or better.
quantum error correction, fixing mistakes without looking
Normal computers also make errors (cosmic rays flip bits occasionally). They fix them with extra bits that act like backup copies and parity checks ("Is the number of 1s even or odd?").
Quantum error correction is much harder because of two rules:
- You cannot copy a qubit (the no-cloning theorem, a fundamental quantum law).
- If you measure a qubit to check for errors, you destroy its superposition.
So how does it work? We use extra "helper" qubits (called ancilla qubits) and clever codes that spread the information across many physical qubits. The code lets us detect and fix errors without directly measuring the valuable data.
The most famous one: the surface code (used by Google, IBM, and others).
- Imagine a grid of qubits, like a chessboard.
- Some qubits hold the actual information.
- Others are "check" qubits that constantly watch for problems.
- If an error appears, the pattern of checks tells us exactly what went wrong (a bit flip, a phase twist) and we can correct it, all while the main information stays in superposition.
Simple analogy. Think of your phone's photos. Instead of storing one copy, you store the photo split across many pieces with extra checksum pieces. If one piece gets damaged, the checksums let you rebuild it without ever opening the main photo file.
Quantum error correction turns many noisy physical qubits into one reliable logical qubit. Right now we need hundreds or thousands of physical qubits to make one good logical qubit. That is why today's machines (with a few hundred physical qubits) cannot yet run the big powerful algorithms we dream about.
when do we use error correction?
- Right now (the Noisy Intermediate-Scale Quantum, or NISQ, era). We mostly use error mitigation techniques, clever software tricks that reduce errors without full correction. We also design short algorithms that finish before decoherence destroys everything.
- Future (fault-tolerant quantum computing). Once we have millions of good qubits, we turn on full quantum error correction. Then we can run long algorithms like big versions of Shor's or quantum chemistry simulations.
Error correction is not perfect. It costs a lot of extra qubits and extra gates. But it is the only path to truly useful quantum computers.
real-world perspective (2026)
- Leading machines can sometimes run small algorithms with a few dozen qubits before errors pile up.
- Companies are racing to improve coherence times and reduce error rates.
- Every year the numbers get better. What seemed impossible five years ago is now routine in labs.
Decoherence is the enemy. Error correction is the shield.
key takeaways from Part 7
- Decoherence means qubits losing their quantumness because the real world is noisy. It happens all the time and is measured by T1 and T2 times.
- It is the main reason today's quantum computers are still small and limited.
- Quantum error correction uses lots of extra qubits and clever codes to detect and fix errors without destroying superposition.
- We are currently in a transition period: improving hardware and smart software tricks, moving toward full error-corrected machines.
The exciting algorithms we talked about in earlier parts will only reach their full power once error correction becomes practical. That is the big milestone everyone is working toward.
In Part 8, we will look at today's real quantum computers. Who is building them (IBM, Google, IonQ, and others), what they can actually do right now, and what you can try yourself for free in your browser. No fridge required.
If decoherence or error correction still feels confusing, send the question over and I will clarify in the next post.