What is the research that would be most helpful for AI safety while still having value if it turns out not to be useful for safety research?
-
-
Replying to @The_Lagrangian
@The_Lagrangian safety wrt what? threat models are important. (who attacks? what is attacked? how can it be attacked?)2 replies 0 retweets 1 like -
Replying to @allgebrah
@allgebrah safety as in value alignment for general artificial intelligence a la MIRI1 reply 0 retweets 0 likes -
Replying to @The_Lagrangian
@The_Lagrangian I find it hard to talk about safety of the intransitive kind but maybe that's just my infosec background1 reply 0 retweets 0 likes -
Replying to @allgebrah
@The_Lagrangian do MIRI have a threat model or a justification for not having one?2 replies 0 retweets 0 likes -
Replying to @allgebrah
@allgebrah they have specific models that look basically like self-threat, check out their research summary page1 reply 0 retweets 1 like -
Replying to @The_Lagrangian
@allgebrah I don't have time to read this but it seems like it might answer your questions https://arbital.com/p/AI_safety_mindset/ …1 reply 0 retweets 1 like -
Replying to @The_Lagrangian
@The_Lagrangian neat, thanks! skimmed it, does look like it answers the question (or at least tries to). will give it a full read later1 reply 0 retweets 1 like -
Replying to @allgebrah
'(·) Retweeted '(·)
@The_Lagrangian also looks like some of my (mostly private) notes on calculating bug severity could be relevanthttps://twitter.com/allgebrah/status/686996450757230592 …'(·) added,
1 reply 0 retweets 0 likes
@The_Lagrangian but need to formalize first (assuming motivation), then check for usefulness and whether somebody else already did that
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.