←back to thread

Problems the AI industry is not addressing adequately

(www.thealgorithmicbridge.com)
197 points baylearn | 2 comments | 05 Jul 25 10:33 UTC | HN request time: 0.465s | source
Show context
bestouff ◴[05 Jul 25 11:10 UTC] No.44471877[source]▶
Are there some people here in HN believing in AGI "soonish" ?
replies(5): >>44471902 #>>44471982 #>>44472003 #>>44472071 #>>44472107 #
impossiblefork ◴[05 Jul 25 11:45 UTC] No.44472071[source]▶
I might, depending on the definition.

Some kind of verbal-only-AGI that can solve almost all mathematical problems that humans come up with that can be solved in half a page. I think that's achievable somewhere in the near term, 2-7 years.

replies(2): >>44472097 #>>44473375 #
1. whiplash451 ◴[05 Jul 25 15:22 UTC] No.44473375[source]▶
What makes you think that this could be achieved in that time frame? All we seem to have for now are LLMs that can solve problems they’ve learned by heart (or neighboring problems)
replies(1): >>44475522 #
2. impossiblefork ◴[05 Jul 25 21:03 UTC] No.44475522[source]▶
Transformers can actually learn pretty difficult manipulations, even how to calculate difficult integrals, so I don't agree that they can only solve problems they've learned by heart.

The reason I believe it can be achieved in this time frame is that I believe that you can do much more with non-output tokens than is currently being done.