Post-GPT Computing

Re:Post-GPT Computing by nopinsight on Friday March 24, 2023 @11:12AM (#35290084)

By late 2020s, it's entirely possible that a "weaker" AGI will emerge. There will be much less need to for senior developers as well, but that could be among the least of our concerns if we cannot reliably align AI with human interests by then.

Date Weakly General AI is Publicly Known https://www.metaculus.com/questions/3479/date-weakly-general...

Date of Artificial General Intelligence https://www.metaculus.com/questions/5121/date-of-artificial-...

The latter includes this criterion: "Able to get top-1 strict accuracy of at least 90.0% on interview-level problems found in the APPS benchmark introduced by Dan Hendrycks, Steven Basart et al."

The APPS benchmark: https://arxiv.org/abs/2105.09938

Note that the predicted date of "stronger" AGI has moved quite a lot since GPT-4 is revealed, from late 2030s to 2033 at this moment.

[ Parent | on HN ]

Re:Post-GPT Computing by rimliu on Friday March 24, 2023 @03:45PM (#35294446)

   > By the late 2020s, it's entirely possible that a "weaker" AGI will emerge.

We will surely have self-driving cars by then? Right? Right?

[ Parent | on HN ]

Re:Post-GPT Computing by ilaksh on Friday March 24, 2023 @03:57PM (#35294602)

Has anyone tested GPT-4 on APPS? And if it can get 90% (which it probably can) does that mean people will admit it's an AGI? Or more likely they just keep moving the goalposts.

[ Parent | on HN ]