Sampling and Certified Randomness

The famous quantum advantage experiments did not solve a useful problem. They drew random bitstrings from a pattern of odds that is very costly to copy on a normal computer. That ability has one believable near-term product: certified randomness. This chapter uses a circuit you can run to show exactly what "sampling" from a quantum pattern means.

What did the quantum advantage experiments actually do?

In 2019, a 53-qubit superconducting processor ran a task called random circuit sampling. Here is what it did, step by step:

Take 53 qubits (quantum bits).
Apply a fixed series of gates (the basic steps of a quantum program). The series was picked at random once, then kept the same.
Measure all 53 qubits to get a 53-bit string.
Repeat about a million times.

Each output string is a sample from a pattern of odds that the circuit sets up. In math this pattern is called a probability distribution. Working out that pattern on a normal computer seems to require simulating the full quantum state. That has an exponential cost.

Read the claim carefully. The machine sampled from a distribution faster than the best known normal simulation of the same sampling. That is all. It did not factor a number, plan a route, or simulate a molecule. The task was picked because it is what quantum hardware does naturally, and because it is hard for normal computers. It was not picked because anyone wants its output.

The original paper guessed a normal supercomputer would need 10,000 years to match the 200-second quantum run. Within a few years, better normal algorithms cut that to days, and then lower still on large clusters. The demonstration is still a landmark of hardware control. But its exact lead over normal computers keeps moving. It is a textbook case of filter question 5 from chapter 1: the classical baseline doesn't stand still. Full claim tracking, with sources: quantum advantage reality page.