←back to thread

How large language models will disrupt data management [pdf]

(www.vldb.org)
82 points mfiguiere | 1 comments | 27 Jul 24 01:00 UTC | HN request time: 0s | source
Show context
elahieh ◴[27 Jul 24 02:48 UTC] No.41084133[source]▶
Section header: "Who's data is it?"

Couldn't they have had an LLM proof-read they're paper?

replies(3): >>41084248 #>>41084339 #>>41084378 #
batch12 ◴[27 Jul 24 03:56 UTC] No.41084339[source]▶
I've been seeing this a lot lately. Everywhere from screen printed signs to news tickers. I'm not sure if this is because it's new or if I'm just now seeing it.
replies(1): >>41084445 #
elahieh ◴[27 Jul 24 04:22 UTC] No.41084445[source]▶
Yeah, call me snobby or OCD but I lost confidence in the authors and reviewers when I saw that slipped through.

Then I started wondering if this is going to become the new anti-AI marker. AI-written papers use "delves", "underscores" and "showcasing" too much. Avoid those words, throw in some errors and readers will think your paper was written by humans.

replies(3): >>41084453 #>>41085045 #>>41096457 #
kolinko ◴[27 Jul 24 04:24 UTC] No.41084453[source]▶
As for authors - perhaps english is their second language? Regular spellcheckers don’t check grammar well enough.
replies(1): >>41084467 #
1. rjurney ◴[27 Jul 24 04:30 UTC] No.41084467{3}[source]▶
You fucks, I just learned about this. I feel so stupid!