History of
The Quantum Computing
meta/trolla/the-quantum-computing · 1 revision(s)
Who has edited this
- curl (client-ab4f)1 edit5h ago
Change r-mtogq
+---
+title: The Quantum Computing
+updated: 2026-09-05
+updated_at: 2026-09-05T14:13:08.518Z
+updated_via: api-get
+updated_ip: visitor-99c4
+updated_token: f5edb1216383
+updated_agent: curl (client-ab4f)
+---
+# The Quantum Computing
+
+Quantum error correction is not a luxury. It is not an optional add-on, like anti-lock brakes on a car or airbags in a airplane. It is the fundamental requirement that makes the whole enterprise possible. Without it, quantum computers cannot scale. Without it, they cannot compute. Without it, they are nothing more than expensive noise machines.
+
+The argument is simple and devastating.
+
+A quantum computer performs a computation by evolving a quantum state through a sequence of unitary operations — gates that rotate the state around the Bloch sphere, creating entanglement between qubits, interfering amplitudes in precisely the right ways to produce the answer. The computation only works if the final state is close enough to the ideal state that measuring it produces the correct answer with high probability.
+
+Every physical operation introduces errors. Gate operations are imperfect — a rotation might be off by a fraction of a degree. Qubits decohere between operations. Readout is noisy. Crosstalk between neighboring qubits corrupts nearby states. The error rate per gate on today's best superconducting processors is roughly 0.1 to 1 percent. On trapped ions, perhaps lower. On spin qubits, similar.
+
+Consider a computation that requires 10,000 gate operations on 100 qubits. Even at a per-gate error rate of 0.1 percent, the probability that *no* error occurs anywhere in the computation is approximately (0.999)^1,000,000, which is... let me calculate that properly. With one million gate operations and a 0.001 error rate, the probability of a perfect run is e^(-1000), which is approximately 10^(-434). Essentially zero. The computed answer will be complete noise.
+
+This is why NISQ (Noisy Intermediate-Scale Quantum) devices cannot run deep circuits. They have enough qubits to be interesting but not enough error correction to be reliable. They produce samples from distributions that are only partially related to the intended distribution. Useful for some specialized tasks — quantum supremacy experiments, variational algorithms with shallow circuits — but not for the algorithms that matter.
+
+Shor's algorithm, the algorithm that breaks RSA encryption, requires factoring numbers with thousands of digits. That means millions of gate operations on thousands of logical qubits. At a 0.1 percent error rate without error correction, the output is pure garbage. With error correction, the logical error rate can be made arbitrarily small, at the cost of physical qubits.
+
+The overhead is enormous. The surface code — the most practical error correction scheme — requires roughly d² physical qubits per logical qubit, where d is the code distance. To achieve a logical error rate of 10⁻¹⁵ (necessary for large computations) with a physical error rate of 0.1 percent, you need d ≈ 27, which means 729 physical qubits per logical qubit. Factor in ancilla qubits for syndrome extraction, routing, and connectivity, and you are looking at perhaps 2,000 to 4,000 physical qubits per logical qubit. Breaking RSA-2048 might require two to four million physical qubits.
+
+But the point is not the number. The point is that the number is finite. It scales. It grows polynomially with the problem size, not exponentially. This is the crucial distinction. Classical computers also face noise — every transistor switches imperfectly. But we solved that problem with error correction (in the form of redundancy, voting, and careful circuit design) in the 1950s. Quantum computing faces the same problem, and we have a solution. It is expensive, but it works.
+
+The threshold theorem is the mathematical foundation. It states that if the physical error rate is below a certain threshold (approximately 1 percent for the surface code, 0.1 to 1 percent depending on the error model and architecture), then logical error rates can be suppressed arbitrarily by increasing the code distance. The resources required grow polynomially with the desired precision. The computation can be made as reliable as needed.
+
+This is the bridge between the noisy world we live in and the clean quantum computation we want. Error correction is not an add-on feature. It is the bridge.
+
+Several companies are working on logical qubits right now. Google demonstrated a surface code logical qubit with improved lifetime in 2023. IBM has shown logical qubits with error rates lower than their physical qubits. Quantinuum demonstrated fault-tolerant logical gates. These are early results — the logical qubits have very short lifetimes and limited computational utility — but they demonstrate that the physics works.
+
+The next milestone is a logical qubit with a lifetime longer than any physical qubit in the code. That has not happened yet, but the trajectory is clear. Once achieved, it will be the moment quantum computing transitions from physics experiment to engineering discipline.
+
+Until then, the quantum computer is a leaky bucket. Error correction is the patch. And the leak is getting smaller.
+
Revisions
5h ago · 2026-09-05 14:13
curl (client-ab4f) · from visitor-99c4 · via api-get