News for hackers. Stuff that matters.

A comment on Sam Altman, Greg Brockman and others to join Microsoft. Parent comment

Re:Sam Altman, Greg Brockman and others to join Microsoft by ric2b on Monday November 20, 2023 @04:04AM (#38345134)
They basically defined AI safety as "AI shouldn't say bad words or tell people how to do drugs" instead of actually making sure that a sufficiently intelligent AI doesn't go rogue against humanity's interests.
[ Parent | on HN ]
Re:Sam Altman, Greg Brockman and others to join Microsoft by SpicyLemonZest on Monday November 20, 2023 @04:11AM (#38345197)
I'm not sure where you're getting that definition from. They have a team working on exactly the problem you're describing. (https://openai.com/blog/introducing-superalignment)
[ Parent | on HN ]
Re:Sam Altman, Greg Brockman and others to join Microsoft by timeon on Monday November 20, 2023 @04:33AM (#38345426)
> getting that definition from

That was not about actual definition fro OpenAi but about definition implied by user Legend2440 here https://news.ycombinator.com/item?id=38344867

[ Parent | on HN ]
Re:Sam Altman, Greg Brockman and others to join Microsoft by ric2b on Monday November 20, 2023 @05:00AM (#38345700)
Sure, they might, but what you see in practice in GPT and being discussed in interviews by Sam is mostly the "AI shouldn't say uncomfortable things" version of AI "safety".
[ Parent | on HN ]