Morning Minute: AI Agents Cut BTC Quantum Attack Benchmark by 86%
AI coding agents helped more than 100 researchers cut the cost of a key quantum attack step against Bitcoin and Ethereum’s secp256k1 cryptography by 86% in two months, from 10.75 billion to 1.496 billion on the ECDSA.Fail benchmark. The benchmark scores circuit designs by qubits times Toffoli gates, so lower is better. The leading design used 1,151 logical qubits and about 1.3 million Toffoli gates, with later entries reducing gates below a million. The work came from crypto security groups including Eigen Labs, Trail of Bits, StarkWare, Theta Labs, MultiVM Labs, and the Ethereum Foundation, using an “Open Autoresearch” approach where humans and AI agents iterated toward a measurable target with verification. The result lands amid broader efforts to make crypto systems quantum-resistant by 2029–2035. It does not imply immediate danger, but it shows attack research may be accelerating far faster than expected.
