Running Codex for 5 days to enumerate every published AI safety idea
AaronBergman18 · x · 2026-09-29
Aaron Bergman ran Codex for five full days with semantic search to enumerate as many published AI safety ideas as it could find online, cross-referencing them against Coefficient Giving's funding wish list.
- Coefficient Giving is a philanthropic funder focused on reducing catastrophic AI risks, listing high-priority initiatives spanning biosecurity, global catastrophic risk opportunities, and Navigating Transformative AI
- The fund explicitly says the list isn't comprehensive and it wants to fund more work, including from existing teams
- The experiment was meant seriously: systematically inventorying safety ideas and matching them to available funding
More from coding & agent
- Steal this prompt: Opus 5.5 plus Runway MCP for polished motion design videos — notiansans · 2026-09-29
- Pluto launches in beta: a personal agent with inbox, memory and a computer — Rasmic · 2026-09-29
- Anthropic adds official eval-building and hillclimbing workflow to Claude Code skill — ClaudeDevs · 2026-09-29
- Anthropic engineer: don't run Sonnet at max effort — use Opus instead — edwinarbus · 2026-09-29
- Hugging Face agent attack postmortem: allowlists gate where agents go, not what they do — kimmonismus · 2026-09-29
- PR Council MCP: Open-Source Multi-Agent PR Review Playground for Agentic Engineering — mostly_deterministic · 2026-09-29