>What if automating AI R&D triggers an intelligence explosion?
It seems to me AI has progressed at a rather steady pace somewhat limited by the hardware slowly improving in a Moore's law like way. Turing in 1950 said it might be 50 years before his test was passed and it seems to have taken more like 75 so not that quick. Even with automating R&D it'll still probably be limited by the compute and so continue steadyish progress. I think we have a bit of a speed up just now due to wild bubble investment and a coming slow down when it crashes but things will plod on overall.
That said there may be big society changes because a bit dimmer than humans is very different economically to a bit smarter.
tim333 · · focus · HN ↗
>What if automating AI R&D triggers an intelligence explosion?
It seems to me AI has progressed at a rather steady pace somewhat limited by the hardware slowly improving in a Moore's law like way. Turing in 1950 said it might be 50 years before his test was passed and it seems to have taken more like 75 so not that quick. Even with automating R&D it'll still probably be limited by the compute and so continue steadyish progress. I think we have a bit of a speed up just now due to wild bubble investment and a coming slow down when it crashes but things will plod on overall.
That said there may be big society changes because a bit dimmer than humans is very different economically to a bit smarter.