Latest / Logosofia no Dia a Dia
Can 100 Volunteers Let AI Govern Them for a Month?
What happens when a hundred volunteers hand governance decisions to a council of AI agents for a month? This episode explores Daniel’s proposed “AI agentocracy” experiment — not as a serious political proposal, but as a concrete way to test whether multi-model deliberation can help humans navigate complex stakeholder trade-offs. We trace the lineage from Denmark’s Synthetic Party to DeepMind’s democratic AI paper to Ireland’s citizens’ assemblies, then walk through a full experimental design: recruitment, decision domains, the Karpathy council architecture, guardrails, and metrics for…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.