it’s getting to the point where I notice people say it a lot, especially IRL now for whatever reason recently.
And for clarity I’m not in research or anything, so these people just mean ‘LLM/image gen’, not utilities like OCR or (usually not) transcription.
Some have argued it’s just more efficient (which I can kind of get), while others think you’re actively hindering your intelligence somehow.
On the first point:
I’ve tried it occasionally to see how it compares to my own skill, and while it produces a functional result, it’s always very derivative work to the point where you can find things with the exact same names of other ‘public’ (but not libre) works, and often isn’t the ideal solution to what it targets. So I can see how you can get things out of it, but it never felt really that profound to me.
But for the second… isn’t this supposed to be the tool for people to do things they aren’t experienced in? If anything, you probably need to be able to understand how to write pertaining to the task so the token probabilities are biased toward writing from that area.
And even then, if all you end up doing is prompting AI, then wouldn’t you ultimately serve no purpose outside of being glorified QA?
I guess I’m trying to figure out what exactly non-users would be ‘falling behind’ in that affects them more than those who use AI?
llms are able to do a lot of things many people simply can’t (at least without spending a long time learning), although the result can be mediocre. it’s being misused widely, but doesn’t mean it’s entirely incapable of anything useful.
recently i’ve seen it being used for reverse engineering, seemed like a pretty good usecase.
deleted by creator
It really cannot. But people who have no knowledge in a field absolutely think LLMs are proficient in that field. As soon as an expert in that field look at what the LLM is doing in their field, that illusion quickly vanishes.
Unfortunately, that’s not true. LLMs have proven new results in mathematics, at extremely high levels, conjectures open since decades.
I am personally very against this use of AI, but let’s be clear on its capacities.
About some of those proofs…
https://cims.nyu.edu/~tristanb/statement.pdf
Yeah, I had read earlier today. But even the author said that his own draft was LLM-based.
Assisted, not based. Read through to see the difference
They write “with a great deal of help”, so I am usure where the boundary between the two lies and where they would find themselves.
names Fremats last theorem proved by Anthropic
Uhm? The one proven in the 90s?
https://www.anthropic.com/research/formalizing-fermats-last-theorem
Every one of those mathematical milestones made by AI is followed by an article from a mathematician pointing out the ‘discovery’ was a flawed rehash or not really applicable to the problem. Same for all the claims about finding bugs, they are old and known bugs and occasionally it can identify already known types of bugs faster as long as a human can filter out the wrong results.
AI does do a fine job of filtering large data sets in controlled environments, like how it is used for finding exoplanets, but just for speeding up an existing process.
It doesn’t do anything new.
A reference with a reasonable overview: https://arxiv.org/pdf/2608.02859