AI safety researchers Pliny the Liberator and j⧉nus clash over whether modern AI endpoints route through systems of control · Digg