Anthropic internal research model cracks 26-year-old NSA elliptic curve seed mystery in under 30 minutes
soldierofcinema · reddit · 2026-10-17
A Reddit post shares a screenshot of an internal Anthropic research model's chain of thought after it solved the 26-year-old NSA elliptic curve seed mystery in under 30 minutes.
The CoT is striking: the model immediately doubted itself — "am I in some sort of simulated/test environment?" — then verified the seeds against memory, computed the odds (all top bits clear, 20 bits, chance 1e-6), and concluded it was a genuine real-world discovery.
More from Fun
- Beff Jezos says his brain hit token limits: 16-hour days, 80-hour weeks — beffjezos · 2026-10-17
- AI Industry Worker Says Nobody He Knows Uses xAI's Grok Build — SimonBalmain · 2026-10-17
- Nathan Young's Tesla Roadster loop analogy nails what felt off about that AI demo — NathanpmYoung · 2026-10-17
- Dev builds a semi-playable Tesla Roadster driving sim, physics and SFX still rough — thatroblennon · 2026-10-17
- ClaudeAI weekly: Opus 5.5 takes center stage, Projects redesign roasted, a rock named Clod gets promoted — ClaudeAI-mod-bot · 2026-10-17
- One plain-text file runs time on every device; UCLA prof maintains tzdata as volunteer since 2005 — PMinervini · 2026-10-17