I don’t mean it needs to be the exact question, just something with equivalence. If someone talked about stacking boards and other things, the board could go on the bottom and then maybe someone else talked about stacking balls and books that way, so it used that because “eggs” were associated with “round”. Follow up with the nail thing from another conversation.
It’s definitely a form of intelligence, but I don’t think it’s anywhere close to 99.9% of human thought. I think it’s missing entire dimensions of thought.
I’m not saying it’s 99.9% of human intelligence, I’m saying you’re describing 99.9% of human thought.
This is what humans do, we hear about something thing and then we learn how to apply it to another. You even mention here “stacking balls” and then making the connection that eggs are also round and would need to be stacked in the same way to prevent rolling. This is reasoning, using what you’ve learned and applying it to a novel problem.
What you are describing as novel problems are really just doing the same thing at a completely different level. Like I play soccer, but no matter how much I trained, there is no way I would ever reach Messi’s skill, because he was just born with special skill in that area, but still just human like the rest of us.
And remember I’m mostly just pointing to the “text predictor” claim. I’m not convinced it’s not, and I think that appeared true for early models, but not so easy to apply to current models.
Yeah, it is hard to say if the “glorified text predictor” is completely accurate, since the sheer size of the model allows for some pretty deep connections.
And, thinking about it since making that post, it’s hard to say for sure that even Einstein or Newton were doing anything differently or were just the first/most famous to put those particular things together.
I don’t mean it needs to be the exact question, just something with equivalence. If someone talked about stacking boards and other things, the board could go on the bottom and then maybe someone else talked about stacking balls and books that way, so it used that because “eggs” were associated with “round”. Follow up with the nail thing from another conversation.
It’s definitely a form of intelligence, but I don’t think it’s anywhere close to 99.9% of human thought. I think it’s missing entire dimensions of thought.
I’m not saying it’s 99.9% of human intelligence, I’m saying you’re describing 99.9% of human thought.
This is what humans do, we hear about something thing and then we learn how to apply it to another. You even mention here “stacking balls” and then making the connection that eggs are also round and would need to be stacked in the same way to prevent rolling. This is reasoning, using what you’ve learned and applying it to a novel problem.
What you are describing as novel problems are really just doing the same thing at a completely different level. Like I play soccer, but no matter how much I trained, there is no way I would ever reach Messi’s skill, because he was just born with special skill in that area, but still just human like the rest of us.
And remember I’m mostly just pointing to the “text predictor” claim. I’m not convinced it’s not, and I think that appeared true for early models, but not so easy to apply to current models.
Yeah, it is hard to say if the “glorified text predictor” is completely accurate, since the sheer size of the model allows for some pretty deep connections.
And, thinking about it since making that post, it’s hard to say for sure that even Einstein or Newton were doing anything differently or were just the first/most famous to put those particular things together.
It’s a weird world and cool to think about. Thanks for the civil and interesting discussion.
Yeah, it was a good talk. Funny watching the sentiment of the voters throughout, too.
IMO, one of the best QoL updates for Lemmy is to make the votes invisible.
I suggest the personal firmware update of stopping caring about downvotes. Then they just become interesting data points.
I dont personally care about downvotes, but those data points sure do shake my faith in humanity sometimes.
If I get downvotes without any reply, I generally assume I’ve angered some trogdolytes, similar to getting replies full of vitriol.
Which has helped with my own urges to rage reply.