News for hackers. Stuff that matters.
Re:#define CTO OpenAI by tbrownaw on Friday January 13, 2017 @08:04PM (#13396166)
I would think it would be both safer and easier to just never bother to implement those drivers in the first place.
Re:#define CTO OpenAI by SomeStupidPoint on Friday January 13, 2017 @08:39PM (#13396315)
That would be what I meant by "removing", though you also have to make sure they don't emergently develop (which I'm not sure is possible), because if it develops them against our best efforts, it likely will (correctly) view us as threats to its personhood.

Which was my point: the model of security is reliant on things we're not sure we can even do, but are likely to make the AI view us as a threat, raising our existential risk. So I view it as security theater that actually makes us less secure.