Anthropic and Theorem partner to use Claude for formal security proofs

Techmeme relays Anthropic's announcement that it partnered with Theorem to use Claude to formally prove security properties at the assembly level across 100-plus real-world open-source libraries. Another user, citing Anthropic research, says AI wrote machine-checked proofs for the compiled code of 123 libraries and surfaced hundreds of bugs. The user argues that a proof is only as good as its specification.