this post was submitted on 21 Oct 2024
932 points (97.3% liked)

Technology

59596 readers
5104 users here now

This is a most excellent place for technology news and articles.


Our Rules


  1. Follow the lemmy.world rules.
  2. Only tech related content.
  3. Be excellent to each another!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, to ask if your bot can be added please contact us.
  9. Check for duplicates before posting, duplicates may be removed

Approved Bots


founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
[–] EatATaco@lemm.ee 3 points 1 month ago (1 children)

how could they tell it was truly a new thing

Sure, there is a chance the exact question had been asked before, and answered, but we are talking remote possibilities here.

that any description provided for it didn’t map it to another object that would behave similarly when stacking.

If it has to say 'this item is like that other item and thus I can use what I've learned about stacking that other item to stack this item' then I would absolutely argue that it is reasoning and not just "predicting text" (or, again, predicting text might be the equivalent of reasoning).

Stacking things isn’t a novel problem.

Sure, stacking things is not a novel problem, which is why we have the word "stack" because it describes something we do. But stacking that list of things is (almost certainly) a novel problem. It's just you use what you've learned and apply that knowledge to this new problem. A non-novel problem is if I say "2+2 = 4" and then turn around and ask you "what does 2 + 2 equal?" (Assuming you have no data set) If I then ask you "what's 2 + 3?" that is a novel problem, even if it's been answered before.

I mean, I can’t dismiss that it isn’t doing something more complex, but examples like that don’t convince me that it is. It is capable of very impressive things, and even if it needs to regurgitate every answer it gives, few problems we want to solve day to day are truly novel, so regurgitating previous discussions plus a massive set of associations means that it can map a pretty large problem space to a large solution space with high accuracy.

How are you convinced that humans are reasoning creatures? This honestly sounds like you could be describing 99.99% of human thought, meaning we almost never reason (if not actually never). Are we even reasonable?

[–] Buddahriffic@lemmy.world 0 points 1 month ago (1 children)

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.

[–] EatATaco@lemm.ee 2 points 1 month ago (1 children)

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.

[–] Buddahriffic@lemmy.world 2 points 1 month ago (1 children)

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.

[–] EatATaco@lemm.ee 2 points 1 month ago (1 children)

It's a weird world and cool to think about. Thanks for the civil and interesting discussion.

[–] Buddahriffic@lemmy.world 2 points 1 month ago (1 children)

Yeah, it was a good talk. Funny watching the sentiment of the voters throughout, too.

[–] EatATaco@lemm.ee 2 points 1 month ago (1 children)

IMO, one of the best QoL updates for Lemmy is to make the votes invisible.

[–] Buddahriffic@lemmy.world 2 points 1 month ago (1 children)

I suggest the personal firmware update of stopping caring about downvotes. Then they just become interesting data points.

[–] EatATaco@lemm.ee 2 points 1 month ago (1 children)

I dont personally care about downvotes, but those data points sure do shake my faith in humanity sometimes.

[–] Buddahriffic@lemmy.world 0 points 1 month ago (1 children)

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.

[–] EatATaco@lemm.ee 2 points 1 month ago (1 children)

Yeah for me it's less that I rage about getting downvoted, its just when I see a massive number of downvotes for posts that are simply pointing to the facts, or being logical and rational...and they get a massive number of downvotes because it contradicts the circlejerk or what people want to be true.

[–] Buddahriffic@lemmy.world 0 points 1 month ago

Yeah, though on the bright side, it's not nearly as bad here as Reddit was. Based on votes and replies, it seemed like a non-trivial amount of people on there didn't have much skill in the reading comprehension or persuasion department.