Re:Does current AI represent a dead end?
by sanderjd on Friday December 27, 2024 @12:11PM
(#42523912)
No it isn't, because it is relevant to the question of whether the current approaches can be scaled 100x or 1000x.
[ Parent
| on HN ]
That's a hardware question, not a software question.
I don't know if the hardware can be scaled up.
Re:Does current AI represent a dead end?
by bee_rider on Friday December 27, 2024 @12:21PM
(#42524008)
It is probably a both question. If 100x is the goal, they’ll have to double up the efficiency 7 times, which seems basically plausible given how early-days it still is (I mean they have been training on GPUs this whole time, not ASICs… bitcoins are more developed and they are a dumb scam machine). Probably some of the doubling will be software, some will be hardware.
Re:Does current AI represent a dead end?
by sanderjd on Friday December 27, 2024 @02:34PM
(#42525303)
Yep, agreed.
I'm pretty skeptical of the scaling hypothesis, but I also think there is a huge amount of efficiency improvement runway left to go.
I think it's more likely that the return to further scaling will become net negative at some point, and then the efficiency gains will no longer be focused on doing more with more but rather doing the same amount with less.
But it's definitely an unknown at this point, from my perspective. I may be very wrong about that.
Re:Does current AI represent a dead end?
by sanderjd on Friday December 27, 2024 @02:29PM
(#42525257)
The question is essentially: Can the current approaches we've developed get to or beyond human level intelligence?
Whether those approaches can scale enough to achieve that is relevant to the question, whether the bottleneck is in hardware or software.
Re:Does current AI represent a dead end?
by s1artibartfast on Friday December 27, 2024 @12:58PM
(#42524378)
That depends on if efficiency is part of the scaling process