Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

How we measured AI writing across arXiv, and where the measurement breaks

unslop.run

Post reply on HN
Print view View original

101–110 of 185 posts

Re: How we measured AI writing across arXiv, and where the measurement breaks

#101

Post by alexpotato »

I work with several researchers (primarily working on consensus protocols etc).

We were discussing research in general and I asked them: "Do you prefer the writing of the papers or the research?"

They, almost unanimously, agreed that they preferred the research. This makes sense as if they preferred writing they probably would have chosen another profession.

I say this b/c having LLMs available to turn research diagrams, code etc into a paper (or at least the starting point of a paper) will probably lead to MORE quality research papers. This is b/c I'm sure there was some friction in a researcher's mind of "I would love to do the research on this but don't want the trouble of writing the paper".

Put another way: on a 2D plot with one axis being the skills as a researcher and the other being hatred of writing, LLMs may "unlock" the people high on both axes to get more papers out.

Post Script: I agree that this could also lead to more BAD papers but the net may turn out to be positive in the long run.

Re: How we measured AI writing across arXiv, and where the measurement breaks

#102

Post by ffsm8 »

ryandvm wrote: post #85

There some real game theory mechanics at play when it comes to LLM usage in corporations. Devs are cranking out superficially superior code and documentation by just aiming the Claude Code fire hose at everything they can. Leadership encourages this because from what they can tell, there is no downside. It's hard to say if this code is structurally better or worse than before, but it's certainly voluminous and as far…

> It's hard to say if this code is structurally better or worse than before,

Huh?

But that's easy to tell?

It's better then incompetents code, worse then a motivated average dev... But good enough hence the real question is value aka time& money invested/quality.

That's much harder to tell, and I currently think opus/fable generated code is decent enough to be safely in favor - at least on subscription

Re: How we measured AI writing across arXiv, and where the measurement breaks

#103

Post by probably_wrong »

guywithahat wrote: post #78
probably_wrong wrote: post #23

Earlier quoted context omitted.

I recently desk-rejected a paper where every single citation in its Introduction was hallucinated. That means that the entire connection between what the author(s) did and how it relates to existing research was simply made up. I've never seen this happening before AI but now there's at least one paper in every cycle pulling something similar. My problem therefore is: we are seeing more and more papers written with t…

Yeah but that's an issue with the researcher putting out a bad paper, and it suggests you'll have to reject more papers. We wouldn't ban email because many of the emails are spam, it just means we need new tools to filter out junk. AI will allow researchers to be more productive all together and take less time to publish a paper, which is good.

The problem with spam is that it's not a technology like AI is. So I suggest taking cars instead.

Cars have plenty of advantages, and yet no one would say "the number of pedestrians killed by cars is rising, but that's an issue with the drivers". In fact, the opposite is true: from fines and school zones to speed bumps and bollards, we have accepted that cars bring structural problems with them that cannot be solved at the driver level alone.

> we need new tools to filter out junk

Agreed, but if my office suddenly was flooded with garbage my first thought wouldn't be "I need more, bigger trash cans" but rather "who brought all this junk here and why?". To simply assume that the garbage is a sudden natural phenomena that I have to live with seems, at the very least, unfair.

Re: How we measured AI writing across arXiv, and where the measurement breaks

#105

Post by OleksandrC »

jknoepfler wrote: post #97
ryandvm wrote: post #85

There some real game theory mechanics at play when it comes to LLM usage in corporations. Devs are cranking out superficially superior code and documentation by just aiming the Claude Code fire hose at everything they can. Leadership encourages this because from what they can tell, there is no downside. It's hard to say if this code is structurally better or worse than before, but it's certainly voluminous and as far…

> "superficially superior code" Can you unpack that a bit? It produces measurably, meaningfully inferior code everywhere I see it in use. > "leadership encourages this because from what they can tell, there is no downside" As said 'leadership' I find this a bit puzzling. I'm seeing strong, quantifiable evidence of increasing churn, increasing incident count, and length of downtime from the date of our biggest push in…

The code written by agents generally seems to reflect what you (the human driver) asked for. If you discuss the approach and architecture first, asking the right questions in the process, and then let it implement - the result is quite on point with the frontier models. Might need some minor touch-ups if agent missed some common conventions or guessed the expectations wrong (but again, salvageable with follow-up prompts). And when it comes to line-by-line logic within functions, I would argue that today's models are LESS likely to make a mistake in there than humans - especially if you cross-review with another model (e.g. "write with Claude, review with GPT").

Humans write slop too, you know. Just saying.

Re: How we measured AI writing across arXiv, and where the measurement breaks

#106

Post by addaon »

ryandvm wrote: post #85

There some real game theory mechanics at play when it comes to LLM usage in corporations. Devs are cranking out superficially superior code and documentation by just aiming the Claude Code fire hose at everything they can. Leadership encourages this because from what they can tell, there is no downside. It's hard to say if this code is structurally better or worse than before, but it's certainly voluminous and as far…

> It's hard to say if this code is structurally better or worse than before, but it's certainly voluminous and as far as leadership could ever tell, with their flawed metrics, that is all that matters.

Hah. I've been working with ChatGPT-5.6 Sol and a lot of guidance to port a particular geometry from 2800 lines of Python-scripted Rhino3D (basically a custom parametric CAD kernel in there, plus use of it) to Python-scripted OCCT/FreeCAD (an existing CAD kernel) and it's up to 36,000 lines and only about half way there. And I've been setting goals and targets for duplication minimization, code size reduction, etc. The results are fine for my purposes, but if I put a positive value on "voluminous" I'd be super impressed; with my actual negative value assignment on lines of code, this is more "absolute crap but still useful to me."

I think "absolute crap but still useful to me" is a pretty high value and not worth neglecting, but I'm definitely scared by leadership who sees the toilet overflowng and assumes that means that value has been maximized.

Re: How we measured AI writing across arXiv, and where the measurement breaks

#107

Post by Der_Einzige »

jsrozner wrote: post #32
dopamine_daddy wrote: post #6

Earlier quoted context omitted.

Surprisingly I agree with you. My opinion is: if it makes communicating research more effective, while not reducing the quality of the output substantially, I see no issue. A possible conclusion for this could be: If the majority of CS papers is AI written, let's just accept this reality universally and stop worrying about it altogether.

I reviewed for ACL (big NLP conf) and another similar conference recently. Lots of crap (though there always has been). One superficially well-written paper was likely AI-plagiarized (i.e. the AI sampled an idea from prior work and rewrote it) and got desk rejected for it. Reviewers were also totally unengaged. Of 20 reviews I read (from my reviewers or from reviewers on the same papers), maybe 2 were mediocre, and t…

And yet the quality of work at ACL is still higher than NeurIPS, and that's with ACL this year having over 40% of all posters looking nearly identical due to everyone claude coding their posters...

Re: How we measured AI writing across arXiv, and where the measurement breaks

#108

Post by sandeepkd »

jknoepfler wrote: post #97
ryandvm wrote: post #85

There some real game theory mechanics at play when it comes to LLM usage in corporations. Devs are cranking out superficially superior code and documentation by just aiming the Claude Code fire hose at everything they can. Leadership encourages this because from what they can tell, there is no downside. It's hard to say if this code is structurally better or worse than before, but it's certainly voluminous and as far…

> "superficially superior code" Can you unpack that a bit? It produces measurably, meaningfully inferior code everywhere I see it in use. > "leadership encourages this because from what they can tell, there is no downside" As said 'leadership' I find this a bit puzzling. I'm seeing strong, quantifiable evidence of increasing churn, increasing incident count, and length of downtime from the date of our biggest push in…

> leadership

Not sure if its just me, lately I have started feeling pretty offensive about the increased usage of the word. Its management not leadership by any means.

> "leadership encourages this because from what they can tell, there is no downside"

For most people in management its easier to pick the current set of slangs/abbreviation's, general trend and go with it. Understanding the details would take time, raise questions and no one in management has time or political capital to spend on it.

Re: How we measured AI writing across arXiv, and where the measurement breaks

#109

Post by mrguyorama »

hereme888 wrote: post #82

I run most professional statements and articles through LLMs before submission. This helps correct grammar and improves accessibility through better syntax (because I write exactly how I think). The article doesn't seem to mention consideration of AI for polishing human work.

What was ever wrong with just producing a second draft of your work?

Re: How we measured AI writing across arXiv, and where the measurement breaks

#110

Post by akomtu »

ryandvm wrote: post #85

There some real game theory mechanics at play when it comes to LLM usage in corporations. Devs are cranking out superficially superior code and documentation by just aiming the Claude Code fire hose at everything they can. Leadership encourages this because from what they can tell, there is no downside. It's hard to say if this code is structurally better or worse than before, but it's certainly voluminous and as far…

IMO, cognitive athrophy is the point. If humans become dependent on AI, those corporations will get enough leverage to change laws in their favor. For example, people don't want a huge data center in the middle of their city, but the corporations will say that the same people can't do anything without AI powered by this data center. That's why I think that making humans deficient and dependent on AI is the entire point.
Post reply on HN