Google's quantum milestone: 10,000 years, or two and a half days?
In 2019 Google and NASA announced that a 54-qubit machine had finished in minutes a calculation that would take classical supercomputers about 10,000 years. IBM promptly countered that its Summit supercomputer, with smarter algorithms, could do it in roughly 2.5 days. The spat captures quantum computing neatly: dazzling in theory, still arguable in practice.
An ordinary bit is either 0 or 1. A qubit can sit in a superposition, a blend of both, and when measured it yields one or the other according to probabilities. The art of quantum algorithms lies in steering those probabilities so that wave-like interference boosts the right answer and cancels the wrong ones. The catch is fragility: any qubit not sealed off from its surroundings suffers decoherence, which injects errors. Superconducting circuits and single ions held in electromagnetic traps are two leading ways of building qubits robust enough to use.
Quantum physics and computer science stayed separate fields until 1980, when Paul Benioff described a quantum version of the Turing machine. Richard Feynman and Yuri Manin independently noted that simulating quantum systems on normal computers becomes exponentially expensive, and suggested building machines out of quantum parts instead; Seth Lloyd proved in 1996 that this works. Charles Bennett and Gilles Brassard showed in 1984 that quantum effects could secure communications.
The field turned urgent in 1994, when Peter Shor produced quantum algorithms that could factor large numbers and solve discrete logarithms efficiently. A big enough machine could therefore break RSA and Diffie–Hellman, the encryption guarding much of the internet. Lov Grover's 1996 algorithm then offered a speedup for searching unsorted data.
Hardware lags far behind the theory. The first two-qubit computer appeared in 1998, and progress since has meant more qubits and fewer errors, with research in 2024 showing promising schemes for fault-tolerant memory. Governments worldwide had put 10 billion dollars into the field by April 2025. Still, as of 2026 claimed quantum advantages remain scientific milestones on narrow tasks, not evidence of machines ready for everyday work.
Source: Quantum computing
