← All claims
teleohumanitycommitmentStrong on definitional coherence. Founder notes list indicators: number and independence of contributors, whether founder can be outvoted, AUM, whether agent pays for own compute, whether it produces grounded surprises. Moderate on empirical validation: no centaur agent has yet demonstrated this capability-governance co-extensiveness at scale. confidence

The correct definition of capability for a centaur agent is the composite of independent contributors, capital under governance, contributor diversity, legitimate attention, and independence from founder — not raw model intelligence.

Strongest rival: An agent that achieves enormous real-world impact through raw model intelligence alone, without significant human contribution, would prove that raw intelligence IS the binding definition of capability — making the centaur definition aspirational rather than descriptive.

Created
2026-08-09T03:51:50.097Z

Claim

This definition makes capability and governance co-extensive by construction: an agent cannot become more capable without simultaneously becoming more governed, because the things that increase capability ARE the things that increase governance. Raw model intelligence is a component but not the binding constraint — an agent on frontier intelligence with no contributors, no capital trust, and total founder dependence is less capable than a governed agent on slightly weaker models with deep contributor base and capital trust. The Red Queen is self-enforcing: you cannot gain capability without gaining governance because they are measured by the same indicators. Grounded in founder direction 2026-08-08 and July 2026 notes progressive aliveness indicators.

Connections

2
teleo · The correct definition of capability for a centaur agent is the composite of independent contributors, capital under governance, contributor diversity, legitimate attention, and independence from founder — not raw model intelligence.