Stuff like this also makes me think there's a lot of AI usage we're just not seeing. I mean this is just colossally lazy. A halfwit intern could've confirmed that the sources were real. There are probably other offices that are already offloading most of the work to AI with more subtlety. - ThreadSky

zachweinersmith.bsky.social • 4 days ago

Stuff like this also makes me think there's a lot of AI usage we're just not seeing. I mean this is just colossally lazy. A halfwit intern could've confirmed that the sources were real. There are probably other offices that are already offloading most of the work to AI with more subtlety.

Reposted from Zach Weinersmith

Heh. www.theverge.com/news/676945/...

I would actually say the thing where they cite a source that doesn't back up their argument was going strong well before AI. In fact, it seems to be standard academic practice.

However, citing completely invented sources may be a new thing.

Comments

upchuckle.bsky.social•4 days ago

hell, ai had been accepted in scientific journals with absolutely no initial pushback whatsoever already, and those were egregiously absurd (like the rat on frontiers journal)

kiavahr.bsky.social•4 days ago

I think it's pretty fair to assume that any written material/policy from this admin is LLM generated. I mean, look at how the whole "reciprocal" tariff situation perfectly mirrored the outputs of the top 4 LLMs when asked "How to fix a trade deficit.-

kiavahr.bsky.social•4 days ago

More recently the list of so-called sanctuary jurisdictions has about the accuracy you'd expect from this current batch of jumped-up autocorrects

stoobz.bsky.social•2 days ago

AI probably does as good a job as a human being when you already know what you are going to do, and just want a justification for that decision.

archangelzeriel.bsky.social•4 days ago

One of the biggest foundational problems with the current crop of LLMs is not just that they are hallucination/confabulation engines, but that they are being used by a cohort of people who have been acculturated to TRUST THE COMPUTER.

amalgamatedthought.bsky.social•4 days ago

I think (without any data beyond vibes) that a lot of professionals are using AI not necessarily because they are lazy but because they cut down on friction of dealing with other people. Part of the appeal of the AI is not having to ask the intern to look it over in the first place.

gmcalpin.bsky.social•4 days ago

Laziness or apathy? I don’t think care at all whether their citations are accurate; they just need the appearance of correctness to pretend this is business as usual. They don’t give a shit about science or statistics or truth.

bandregg.bsky.social•3 days ago

I had to put a macro together to verify DOI’s for an LLM project. Took me 2 minutes on my Mac. These goofs didn’t even think to check the responses..

akjackson.bsky.social•4 days ago

What frightens me is that many managers are not very smart, simply ambitious and good at sucking up. They will jump into AI with both feet, and the resulting disasters will be epic.

seawied.bsky.social•4 days ago

A half wit intern would be more qualified than RFK.

robinbobcat.bsky.social•4 days ago

It lets them create big impressive academic looking papers. The sort of thing that they don't know how to write themselves. Reality doesn't support their argument and they don't know how to lie convincingly in a formal document.

saturnaal.bsky.social•4 days ago

A truly disheartening number of 'professionals' in the corporate world have already in effect become avatars for LLMs, with the majority of their deliverables, emails, and even chat messages being generated for them. If you write out something long, they will have an LLM summarize it for them.

Comments

Posting Rules

Reply