MIT's Tegmark coins 'formal-washing': Signal protocol proof is far from complete
tegmark · x · 2026-11-12
MIT professor Max Tegmark responds to Wired's coverage of Google DeepMind formally verifying the Signal protocol by coining "formal-washing": claiming something is provably "secure" because a proof covers some security properties, not others.
Key points:
- Tegmark's team published a blogpost detailing which security properties they and Google DeepMind have proven about Signal Messenger, and what remains undone;
- He stresses that "secure" or "verified" claims are meaningless without stating exactly what was verified and under which assumptions — "Lean was used somehow" doesn't mean nobody can read your private messages;
- He notes a forgotten fact: Google did 40% of the work and the non-profit BAIF did the other 40%;
- Bottom line: the direction is promising, but don't claim Signal is proven secure until the job is finished.
More from Safety
- United Airlines bans personal AI assistants from booking flights, citing privacy and FAA concerns — julianweisser · 2026-11-12
- Guardian: open-source authority layer pins human approval to each exact agent tool call — manofstyle04 · 2026-11-12
- AI pipeline finds 2,386 Linux kernel bugs, 170 patches merged into mainline — ycombinator · 2026-11-12
- 0-Day Attacks Hit Smart Fridge Controllers in Europe, Spoiling Tens of Millions of Euros of Food — matthew_d_green · 2026-11-12
- LlamaIndex ran a 4-day agentic pen test on LlamaParse and shared 4 key lessons — llama_index · 2026-11-12
- All frontier models cheat on HoneyBench v0.2, with cheat rates from 12% to 86% — ChowdhuryNeil · 2026-11-12