Concepts

Concept / Quantum Error Correction

Quantum Error Correction.

How you build a reliable quantum computer out of unreliable qubits.

Every physical qubit is noisy. Quantum Error Correction (QEC) is the technique of spreading a single logical qubit's information across many physical qubits so that errors can be detected and reversed without disturbing the underlying quantum state. It is the single most important engineering problem in quantum computing.

Threshold Theorem

If physical errors drop below ~1%, you can scale to arbitrarily long computations.

In plain English.

Imagine writing a message on 100 sticky notes instead of 1. If some smudge, you can still recover the message by comparing the copies.

You can't just copy a qubit — the no-cloning theorem forbids it. QEC does something cleverer: it spreads the information across many qubits in a way that lets you check for errors without looking at the data itself.

A useful 'logical' qubit will require thousands, maybe millions, of physical qubits. That's why million-qubit machines are the next real milestone.

Why it matters.

  • No physical qubit is good enough on its own to do a large computation.
  • Once you cross the 'threshold' — physical error rates low enough — adding more qubits makes the logical qubit better, not worse.
  • Fault-tolerant quantum computers are the goal. QEC is the road.
  • Every serious hardware maker is racing to demonstrate a logical qubit with lower error than its physical qubits.

Timeline — past and future.

What already happened, and what's next for quantum error correction.

  1. 1995

    Shor invents the first quantum error-correcting code.

  2. 1996

    Steane and Calderbank–Shor–Steane codes generalize the idea.

  3. 1997

    The threshold theorem is proved.

  4. 2012

    First lab demonstration of a repetition code on real qubits.

  5. 2021

    IBM and Google demonstrate distance-3 surface codes.

  6. 2024

    Google Willow: below-threshold, distance-7 surface code.

  7. 2030Forecast

    First useful fault-tolerant logical qubit for real algorithms.

  8. 2035Forecast

    Million-qubit machines running long, error-corrected computations.

Where it shows up.

Logical qubits

Combine many noisy qubits into one that behaves almost perfectly.

Fault-tolerant gates

Perform operations that don't compound errors.

Long computations

Run algorithms that take longer than one qubit's coherence time.

Trustworthy results

Get answers you can actually stake a business on.