The Learning Trap: What Simulated Clueless Agents Reveal About the Unawareness Argument by dan.pandori 🔸

Sub­mis­sion to the Clue­less­ness Cri­tiques Com­pe­ti­tion. Code, pa­ram­e­ters, and figures: https://​​github.com/​​dan-pan­dori/​​clue­less­ness-learn­ing-trap. See the au­thor­ship note at the end. Summary An­thony DiGio­vanni’s un­aware­ness se­quence ar­gues that our un­der­stand­ing of long-run con­se­quences is too coarse to com­pare op­tions. Se­vere in­com­pa­ra­bil­ity fol­lows, and im­par­tial al­tru­ism stops be­ing ac­tion-guid­ing. One stand­ing re­sponse is prag­matic: agents wh...