Almost, but not entirely, Unreasonable's question on Safe Exploration

From Stampy's Wiki
Almost, but not entirely, Unreasonable's question on Safe Exploration id:UgwRuEvbNo3itKnQTI14AaABAg

'Agent' training losses seem unavoidable....... young & stupid leads to old & wise..... or injury & death.

Why would this be any different for AI? it cannot know, what it doesnt know.
Set the 'safety net' and hope it's big enough....


Tags: None (add tags)
Question Info
Asked by: Almost, but not entirely, Unreasonable
OriginWhere was this question originally asked
YouTube (comment link)
On video: Safe Exploration: Concrete Problems in AI Safety Part 6
Date: 2018-09-21T12:12
Asked on Discord? No


Discussion