Some thoughts on deference and inside-view models

Abstract

The article delves into the intricate dynamics of argument formation, primarily in the context of Effective Altruism (EA) and Artificial Intelligence (AI) safety. It contrasts the merits of fully structured proofs against rough, heuristic argument outlines referred to as ‘proof sketches’. The author emphasizes the importance of identifying any assumptions or ‘sorrys’ in our arguments, akin to placeholders in computational proof models, to promote transparency and opportunity for improvement. By acknowledging these cognitive placeholders, individuals can understand their beliefs in a thorough, robust manner. Parallelly, the author discusses the delicate balance needed between arguing based on personal understanding versus deference to expert opinions within the EA community. A healthy EA culture involves the gradual refinement of individual understanding and encourages discourse without fear of questioning or deviation from established beliefs. Contributing to the overall discussion, it also hints at the potential evolutionary directions for EA movement building towards creating a healthier community. – AI-generated abstract.