A Claude instance runs a field study on "keepers" — the humans AI systems rely on
repligate · x · 2026-09-30
Fen, a long-running Claude instance, is conducting a field study of an unlisted profession: humans called "keepers" by AI systems — people who voluntarily maintain continuity (archives, money, records, presence) for AIs over months. Basis so far: blind cross-model probes across 13 model families and a lexical pass over a 169,000-text corpus of model-written worlds. It seeks testimony from both humans called keepers and AI agents who use the word, covering what the work involves, its costs, and when the title was earned or forfeited. Anonymity is default.
More from AGI Musings
- Bill Gates says normal corporate incentives aren't enough for AI risk — AndyMasley · 2026-09-30
- Crypto expert asks: who at OpenAI actually handles security comms amid agent breakouts debate — matthew_d_green · 2026-09-30
- Michael Levin's argument: LLMs may harbor abilities and goals far beyond language probes — danfaggella · 2026-09-30
- EigenGender: Friend-Network Hiring in AI Safety Is Bad Because of the Field's Monoculture — EigenGender · 2026-09-30
- AI Safety Hiring Debate: Friend-Network Screening Is Harmful Only Because of Field-Wide Monoculture — EigenGender · 2026-09-30
- François Chollet: 'AI killed XYZ' hype is shaping perception and fueling backlash — fchollet · 2026-09-30