TLA+ goes viral, but modeling is only the start

The internet discovers TLA+. Now what?

Boris Cherny's viral tweet put TLA+ in the spotlight. This Reasonable post explains TLA+ as a language for describing system behaviors and temporal properties like 'never two leaders at once.' TLC model checking only explores finite instances, and the model isn't the implementation. The team turned 16,000+ TLA+ spec/property pairs into 3,000+ machine-checked Verus proofs, aiming to connect specification, proof, and Rust code. The post doesn't disclose accuracy or latency numbers for this agentic pipeline.

Why it matters: Boris Cherny's viral tweet using Opus 5.5 to model the Claude Agent SDK in TLA+ sparked this practical intro, which extends into a loop connecting temporal specs, proof systems, and AI agents. Hits H and K — the tweet-to-tutorial arc is novel and the post adds concrete toolcha...