FinalFormal2

[Question] How did LW up­date p(doom) af­ter LLMs blew up?

[Question] What’s the sta­tus of TDCS for im­prov­ing in­tel­li­gence?

[Question] What sub­jects are un­ex­pect­edly high-util­ity?

[Question] What are the ar­gu­ments for/​against FOOM?

[Question] Steel­man­ning OpenAI’s Short-Timelines Slow-Take­off Goal

[Question] How does con­scious­ness in­ter­act with ar­chi­tec­ture?

[Question] What’s the con­sen­sus on porn?

[Question] How would you go about test­ing a poli­ti­cal the­ory like Ne­ofeu­dal­ism?

[Question] Does GPT-4′s abil­ity to com­press text in a way that it can ac­tu­ally de­com­press in­di­cate self-aware­ness?

[Question] Are ex­pla­na­tions that ex­plain more phe­nom­ena always more un­likely than nar­rower ver­sions?

[Question] What’s the prob­lem with Orac­u­lar AIs?

[Question] What are some sub­jects with un­ex­pect­edly high util­ity?

[Question] When does an AI be­come in­tel­li­gent enough to be­come self-aware and power-seek­ing?

[Question] What’s the prob­lem with hav­ing an AI al­ign it­self?

[Question] Just don’t make a util­ity max­i­mizer?

[Question] Pre­dic­tions for 2050?

Short Post: Discern­ing Truth from Trash

[Question] Is ve­g­anism morally cor­rect?

[Question] Could we set a re­s­olu­tion/​stop­per for the up­per bound of the util­ity func­tion of an AI?

[Question] Why is in­creas­ing pub­lic aware­ness of AI safety not a pri­or­ity?