this post was submitted on 06 Mar 2025
107 points (96.5% liked)

Memes

47985 readers
929 users here now

Rules:

  1. Be civil and nice.
  2. Try not to excessively repost, as a rule of thumb, wait at least 2 months to do it if you have to.

founded 5 years ago
MODERATORS
 

take a shot every time someone mentions horses or moveable type

top 6 comments
sorted by: hot top controversial new old
[–] wizardbeard@lemmy.dbzer0.com 9 points 3 days ago* (last edited 3 days ago)

If you want instant liver failure, take a shot every time someone uses an incredibly oversimplified understanding of the human brain and LLMs to assert LLMs have human properties.

Also, shoutouts to !techtakes@awful.systems and !fuck_ai@lemmy.world

Don't go in those if you have any sort of desire to defend modern content generation/content thievery AI.

Also, yes LLM style AI has it's uses. Mainly to generate shit that we shouldn't need in the first place, or stuff that we should pay a person to do well. And the cost is stupid when compared to the minor benefit.

[–] rumschlumpel@feddit.org 10 points 3 days ago (1 children)

"cootie catcher programming" ... ?

[–] Lyra_Lycan@lemmy.blahaj.zone 4 points 2 days ago* (last edited 2 days ago) (1 children)

Apparently some places call a paper fortune-teller (below) a 'cootie catcher'. Still unclear as to how it relates. Closest I got was that it's some kind of LLM script to give users defined answers based on their questions. Maybe that mean OpenAI gave hardcoded responses, therefore lying about intelligent generation?

[–] rumschlumpel@feddit.org 1 points 2 days ago (1 children)
[–] ErsatzCoalButter@beehaw.org 2 points 1 day ago

LLMs are much easier to understand than the people selling them want to pretend. If that's news to anybody, perhaps see if you can spot the similarities between 'paper fortune teller' logic and the models described here.

https://explainextended.com/2023/12/31/happy-new-year-15/

[–] falsemirror@beehaw.org 5 points 3 days ago

Worst part about AI is convincing a generation that we should protect "intellectual property". OpenAI might as well be an RIAA psyop