DeepMind paper: LLMs can't sample randomly on their own — pass random numbers from code instead
rohanpaul_ai · x · 2026-10-27
A Google DeepMind paper, 'The Illusion of Stochasticity in LLMs,' finds LLMs cannot make truly random choices themselves but can convert random numbers supplied by a program. Across Gemini, Qwen3, and OLMo sampling 1024 times from simple distributions, every statistical test failed; temperature and top-p tweaks didn't help. In a security game where true randomness limits attackers to 75% success, Gemini's picks let them win 96.6–99.9%. Practical takeaway: generate randomness in code and pass it to the agent rather than asking the model to be random.
Related event: DeepMind Paper: LLMs Cannot Sample Randomly Without External Seeds(2 posts)→
More from coding & agent
- Cursor launches Agent SDK: agents as code with Markdown prompts and TypeScript tools — soleio · 2026-10-27
- Vibecheck, the 'usable formal methods' hackathon, sells out two floors in SF — spikedoanz · 2026-10-27
- How do you verify AI agent outputs beyond unit tests and LLM-as-judge? — Practical_Green5775 · 2026-10-27
- AI agents may soon become the only way to run code, not just write it — IgorCarron · 2026-10-27
- Hermes Agent used as browser QA tester finds 8 issues in a bug-seeded site — techlatest_net · 2026-10-27
- Open-source FableCut lets AI agents edit a real video timeline via remote MCP — FastPresence9799 · 2026-10-27