News for hackers. Stuff that matters.
Re:A.I. Poses ‘Risk of Extinction,’ Industry Leaders Warn by mitthrowaway2 on Tuesday May 30, 2023 @12:38PM (#36126991)
The statement isn't about LLMs. It doesn't refer to them even once.
Re:A.I. Poses ‘Risk of Extinction,’ Industry Leaders Warn by EamonnMR on Tuesday May 30, 2023 @02:06PM (#36128282)
But if LLMs hadn't captured the popular imagination I doubt this would have been written this year and gotten the attention of enough prominent signatories to frontpage on HN.
Re:A.I. Poses ‘Risk of Extinction,’ Industry Leaders Warn by mitthrowaway2 on Tuesday May 30, 2023 @02:16PM (#36128406)
Maybe. It's happened before. [1] And several of the signatories have expressed views about AI risk for many years.

That said, the renewed anxiety is probably not because these experts think that LLMs per se will become generally intelligent. It's more that each time we find out that the human brain does that we thought were impossible for computers to do turn out to be easy, each time we find that it takes 3~5 years for AI researchers to crack a problem we thought would take centuries[2], people sort of have to adjust their perception of how high the remaining barriers to general intelligence might be. And then when billions of investment dollars pour in at the same time, directing a lot more research into that field, that's another factor that shortens timelines.

[1] https://news.ycombinator.com/item?id=14780752

[2] https://kotaku.com/humans-triumph-over-machines-in-protein-f...