this post was submitted on 28 Jun 2024
1703 points (98.2% liked)

People Twitter

5228 readers
542 users here now

People tweeting stuff. We allow tweets from anyone.

RULES:

  1. Mark NSFW content.
  2. No doxxing people.
  3. Must be a tweet or similar
  4. No bullying or international politcs
  5. Be excellent to each other.

founded 1 year ago
MODERATORS
 
you are viewing a single comment's thread
view the rest of the comments
[–] [email protected] 0 points 4 months ago* (last edited 4 months ago) (3 children)

You'd probably need to verify all submissions

Unless you throw an LLM into the mix

Or maybe there's already some resources giving you all debunked facts with their dates

[–] [email protected] 12 points 4 months ago (2 children)

You believe an LLM can be used to distinguish facts from fiction? I wonder up to which year that misconception was taught in school.

The whole point of LLMs is, to convince their users that the "facts" they generate are actual facts.

[–] [email protected] 2 points 4 months ago (1 children)

They can browse the web, and I never meant it would be 100 accurate just easier. Don't think this is going to be a mission critical website

[–] [email protected] 3 points 4 months ago

That just it, these “facts” won’t be on the web for stuff approximately 2005 and before. No where on the web is the racist and homophobic shit I was taught in the 80’s and 90’s listed on some wiki.

LLM’s are mostly useless anyways at distinguishing real information, they are just shit summary tools and poor search engines.

[–] [email protected] 1 points 4 months ago

LLM is plausible deniability!

[–] [email protected] 4 points 4 months ago

LLMs are not magic, otherwise one just have to request that any submission will have references to reputable sources.

[–] [email protected] 2 points 4 months ago

I would probably start out by proofing or approving them before they post to the site. It say I get a notification read it do a little reading over it and get to a point where I can use a large language model to siphon the submissions.