Comment by mathisfun123
2 months ago
The funniest thing about LLMs is the cognitive dissonance they cause people. People clearly recognize (and bemoan) the fact that LLMs produce derivative breathless prose ie they fundamentally fail at "unstructured creativity" (something the might accurately labeled intelligence) but are then shocked that the same LLMs can do math.
It's reasoning from a flawed premise that math universally requires intelligence and creativity. It does not. Anyone that's proved things via "diagram chasing" can affirm that. The conclusion you should draw is that math (at least the kind they excel at) isn't actually a creative endeavor.
I would think it's more that some aspects of math (like anything really, including programming) don't require creativity and can be solved through "brute force" or whatever you want to call what LLMs do. But it's pretty obvious that LLMs are not capable of solving the vast majority of problems in math (or programming) at this time. Eg. Google went 9/353 on Erdos problems. If a more powerful LLM is capable of solving those or if they require a certain je ne sais quoi of the human variety is still up in the air at this point. In either case it seems like they require a long and detailed prompt from a domain expert (ie. human) regardless.
Didn't humanity also score very low on 353, namely 0 since they were open? Probably collectively we could have gotten a slightly better score if all of Math started trying to solve those, but not by that much, I think, since they are precisely still open.
I dunno, this screams of goal post moving. Even if LLMs lack whatever nebulous definition of "creativity" that someone favours, there's no inherent reason for "creativity" to be required to solve any problems at all, "creativity" could just be a human method for solving problems that evolved because of it's broad applicability but is suboptimal at any given task.
> could just be a human method for solving problems that evolved because of it's broad applicability but is suboptimal at any given task.
Ya sure let's just posit another random hypothesis about evolutionary biology in order to substantiate the claim that LLMs are intelligent.
Or (bear with me) you can recall your (likely) experience proving stuff like SAS triangle identities and reflect on whether that required intelligence or just computation.
What's more reasonable?
Why is your example SAS triangles and not something like the modularity theorem?
4 replies →
Hilariously absurd cope.
They're still not very good at writing, but you've flipped the takeaway. The correct conclusion is writing style isn't very relevant to intelligence.
I think the real tragedy is that they are good at writing, just by default are tuned to have a kind of bland corporate tone. If you give the LLM a few pages of writing you like and tell it "Continue, but using this style" it will do a pretty good job of it. Most people just .... don't bother to do that.
Yeah. And really their default tone isn't that bad objectively, it's more that it becomes annoying when it's everywhere. There are a bunch of writers with distinctive styles that I enjoy, but would quickly get sick of if half of everything I read was like that.
It's just Moravec's paradox all over again