Scott analyzes comments on his previous post about POSIWID, showing how the phrase's ambiguity leads to multiple contradictory interpretations while promoting conspiracy thinking. Longer summary

Scott responds to comments on his previous post about the phrase 'The Purpose of a System is What it Does' (POSIWID). He examines various interpretations offered by commenters and argues that while some contain valuable insights, the phrase itself is problematic. He shows how POSIWID can push people from balanced views toward paranoid conspiracy theories, and demonstrates how different commenters interpret the phrase in contradictory ways. Scott argues that the phrase's ambiguity allows people to smuggle in unwarranted assumptions and that there are clearer ways to express any valuable insights it might contain. Shorter summary

Scott analyzes OpenAI's new deliberative alignment approach and explores different possibilities for who should ultimately control AI systems as they become more powerful. Longer summary

Scott discusses OpenAI's new paper on deliberative alignment, which combines constitutional AI with chain of thought reasoning to create more thoughtful AI responses. He explains how the process works by having AI models reflect on moral questions using a specification document. The post then explores different possible approaches to AI chains of command, including prioritizing companies, governments, specifications, moral law, average citizens, or humanity's coherent extrapolated volition. Scott expresses concern that we're heading toward either corporate or government control of AI systems, while acknowledging there may be better alternatives. Shorter summary