Analytic

Yudkowsky and Soares insist the anti-extinction coalition stay single-issue: "We don't think it should be packaged together with any other position. Adding on any other ask risks human extinction if the bigger package fails." They unbundle even their own positive programme ("you don't have to agree with us about later steps"). That is a direct objection to how LivingIP tells its story, which ties safety to a positive build (collective superintelligence, "and we are building it"). The objection exposes a real trade-off: one negative ask ("humanity should not go extinct") gives the widest coalition and the easiest common knowledge, but nobody can picture themselves inside it; a positive picture recruits builders, but every term in it is one more thing a potential ally can reject. Our answer has to keep the two separable: back the single ask wherever it is the live question, and offer the collective-AI story as a part people may take on, never as a condition of joining the coalition against extinction.

Alignment dynamics