AI labs are quietly testing whether frontier models can break major cryptographic protocols, says Scott Aaronson
141_1337 · reddit · 2026-10-08
According to quantum computing and cryptography expert Scott Aaronson, leading AI labs are now quietly evaluating whether their latest internal models can break major cryptographic protocols — extending dangerous-capability testing into territory that would directly threaten global encryption assumptions. The report circulated as a screenshot on Reddit; full details have not been made public, but the source is a highly credible figure in the field.
Related event: Aaronson Says AI Labs Are Secretly Testing Crypto-Breaking Models(3 posts)→
More from Models
- StepFun's Step 5 Preview goes live in Nous Portal, free for one week — alexcovo_eth · 2026-10-09
- Small-model evals won't stay a moat for AI monitoring platforms, argues engineer — Shahules786 · 2026-10-09
- LightOnOCR-2: 0.8B Apache-2.0 OCR model punches above its weight — IgorCarron · 2026-10-09
- MathArena gave 4 agents $500 each to write blog posts; only Opus-5.5 delivered — ChrSzegedy · 2026-10-09
- Goodfire builds cybersecurity monitors for Kimi K3 and GLM 5.3, 50x faster and cheaper — CatAstro_Piyush · 2026-10-09
- Anthropic bans sustained abuse toward its models in extreme cases — Angaisb_ · 2026-10-09