▲ 10 points
back
4 comments
> The challenge is no longer ‘a decade away’
A March 2026 paper from Google Quantum AI, the Ethereum Foundation, and Stanford demonstrates that Shor’s algorithm for secp256k1 can execute with ≤1,200 logical qubits and ≤90 million Toffoli gates — and on superconducting architectures with 10⁻³ physical error rates, those circuits could run in minutes using fewer than half a million physical qubits.
Where we are: ~1–2 logical qubits (effectively) Where we need to be: ~1,200–2,330 logical qubits. In physical qubits: ~1,500 physical now vs. ~500,000–1,000,000+ physical needed (depending on error rates)
Sounds like a decade to me.
There’s no short-term upside to quantum computing research.
It doesn't demonstrate, it claims. They went to a great deal of effort to not demonstrate what they'd actually done, so that no-one could challenge them on it. This seems to be a requirement for quantum stuff, otherwise someone invariably comes along and points out the holes in your claims.
I am more concerned with classical methods revealing cryptographers to have been dangerously misguided to rely on unsolved mathematics to secure their secrets, and worse, to become a standard economic unit of account.