A one-shot cron job prompt got Claude to crack ciphers and advance a 1932 number theory bound
IgorCarron · x · 2026-10-17
- Security researcher elderplinius (amplified by IgorCarron) shares how he pointed a freshly-reset $200 Claude account at unsolved problems via a typo-laden one-shot cron job prompt that ran daily.
- Results: autonomous solutions to uncracked codebreaking challenges, more efficient matrix multiplication methods, and a computer-verified new bound on a number theory problem open since 1932.
- The story doubles as a playbook: a minimal prompt plus a cron schedule can let a model autonomously grind on open research problems for days.
More from coding & agent
- HeyGen demos 3-step AI video pipeline: ChatGPT script, MCP avatar, auto-editing — HeyGen · 2026-10-17
- DeepSeek launches standalone Search & Fetch APIs at ~$0.70 per 1,000 successful requests, failures free — teortaxesTex · 2026-10-17
- OpenAI's new Python and Node SDKs make browser and computer use much easier — OpenAIDevs · 2026-10-17
- .NET 10 Lambda ARM64 base image ships with 170MB of leftover Yum cache — DanielLockyer · 2026-10-17
- Backend migration from Python to Go now takes one engineer just one week — rakyll · 2026-10-17
- Audio founder says Claude one-shot decompiled rival algorithms, moat gone — ChrisGPT · 2026-10-17