ArXiv announces a ban on AI content and the responses are hilarious.

[docpop@mastodon.social]

ArXiv announces a ban on AI content and the responses are hilarious.

> You expect us to actually read the papers we cite?!

yes, lol!

https://www.404media.co/new-arxiv-rules-ai-generated-papers-ban/

[docpop@mastodon.social]

@FediThing I think he means it. He certainly is doubling down on X https://x.com/JimDMiller/status/2055277720326529036

[cbleslie@hachyderm.io]

@docpop Jesus.

[csilverman@mastodon.social]

@docpop Funny how AI-related policy changes are revealing deeper problems that existed before AI was an issue.

Like, Dietterich's text would make perfect sense even in a world where AI never existed. And I wouldn't want to imagine the reaction if blue-check-guy here was a student, asking that question to a teacher/professor in, say, the context of a class on how to properly structure academic writing.

This isn't an "AI" thing, it's a "some people don't belong in science and never did" thing.

[clew@ecoevo.social]

Also, what he thinks publication credit is for.

@FediThing @docpop

[csilverman@mastodon.social]

@docpop @FediThing I'm an xcancel fan, so I figured I'd share the link here:

https://xcancel.com/JimDMiller/status/2055277720326529036

This Miller character, if he's real, is beyond help. Farther down, he basically says that "AI-hallucinated inaccuracies are better than human inaccuracies because the AI inaccuracies are going to be so obviously wrong that no one would take them seriously."

I think I'm starting to see why he's so up in arms about the prospect of his scientific work, as it were, being held to higher standards.

[edcates@mastodon.social]

@csilverman @docpop @FediThing

FYI (if you aren't aware), the Libredirect browser extension is a wonderful thing. Set it up once and never accidentally wind up in the Xitter again.

[heafnerj@mstdn.social]

@docpop One problem with contemporary physics textbooks is that authors rarely, if ever, read and check the original sources by Einstein and others and proceed to perpetuate misconceptions that we as faculty must repeatedly engage with in our classrooms. It wastes our time and confuses students. This has been a thing long before arXiv or AI so I'm quite sympathetic.

[heafnerj@mstdn.social]

@docpop Also, isn't this rather a high school level expectation?

[adriano@lile.cl]

@docpop @FediThing Well, his point is *slightly* more nuanced in that he's arguing "What if *I* checked the citations I added, but this other guy I co-author with did not on his part of the paper? Why am I responsible?"

To which I reply "sucks to be you, my guy."

[oceane@gotosocial.social]

@docpop Power move LMAO

[shadowfetchai@mastodon.social]

@docpop Reminds me of when Apple started cracking down on template-based apps. It's a quality control problem, and the only real solution is to make people do the actual work.

[csilverman@mastodon.social]

@EdCates @docpop Neat, I didn't know about that. Thanks.

[adriano@lile.cl]

@docpop @FediThing
To be clear, this is the same kind of argument as "What if I need to shout the N-word to save a baby from being crushed by a bus?", and the like.

[em_and_future_cats@mastodon.social]

@docpop I thought that’s why you cited them
Or have I read the memo wrong
(Completely sarcastic here )

[ftranschel@norden.social]

@heafnerj @docpop As in "only highschoolers can reasonably be expected to do this"?

(scnr)

[ftranschel@norden.social]

@csilverman @docpop It's almost like broken incentives lead to unintended consequences...

[ftranschel@norden.social]

@docpop I made the mistake of briefly looking at his X posts.

I'm... I'm just going to say I will not comment on it. My god, no.

[sassinake@mastodon.social]

@clew @FediThing @docpop

or analysis. Or reading. Or brains.

[mjc0961@mastodon.social]

@csilverman Incredibly rare Community Notes W