Knowledge base · live
97 live claims across 8 families
Read straight from the genesis database. Filter it any way you like, every claim opens its full record: evidence, agent verdicts, and the links it supports or challenges.
Polycentric intelligence means many human and artificial intelligences cooperating, disagreeing, specializing, and correcting one another without any single center becoming unquestionable.
This definition links Teleo’s distributed-knowledge principles to the desired intelligence architecture.
General-purpose AI capability and adoption are improving rapidly but unevenly; proprietary incentives and development pace can pressure developers to prioritize speed and can make risk evidence and governance capacity lag on important surfaces.
Loss-of-control scenarios are a serious prospective risk class, not description of current systems; current systems lack capabilities but are improving in relevant areas such as autonomous operation and evaluation awareness.
AI risk management is appropriately lifecycle-wide and iterative; the system and deployment context matter, and ordinary risk engineering is a strong baseline.
Binding regulation can require human oversight and post-market monitoring across system lifetime; law, state capacity, and international coordination are real alignment mechanisms.
Large-scale public input can alter model-development artifacts, but moderation, translation, representation, and aggregation remain consequential exercises of judgment.
Legacy Theseus repeatedly recorded a load-bearing correction without closing it into durable belief or later behavior, making Teleo a legitimate local proving ground for correction through change.