Beehaw
  • Communities
  • Create Post
  • search
    Search
  • Login
  • Sign Up
Mazdak ( Mazdak@lemmy.org ) M to Technology - Lemmy.org@lemmy.org · 1 year ago

OpenAI admits AI hallucinations are mathematically inevitable, not just engineering flaws

www.computerworld.com

external-link
message-square
1
link
fedilink
  • cross-posted to:
  • futurology@futurology.today
  • technology@lemmy.zip
2
external-link

OpenAI admits AI hallucinations are mathematically inevitable, not just engineering flaws

www.computerworld.com

Mazdak ( Mazdak@lemmy.org ) M to Technology - Lemmy.org@lemmy.org · 1 year ago
message-square
1
link
fedilink
  • cross-posted to:
  • futurology@futurology.today
  • technology@lemmy.zip
In a landmark study, OpenAI researchers reveal that large language models will always produce plausible but false outputs, even with perfect data, due to fundamental statistical and computational limits.
alert-triangle
You must log in or register to comment.
  • lemmysmash ( lemmysmash@beehaw.org ) 
    link
    fedilink
    arrow-up
    1
    ·
    1 year ago

    What do you mean “admits”? It has been known from the very beginning to everyone who has read about LLMs at least from the output of those very LLMs (the irony, huh?).

Technology - Lemmy.org@lemmy.org

technology@lemmy.org

Subscribe from Remote Instance

Create a post
You are not logged in. However you can subscribe from another Fediverse account, for example Lemmy or Mastodon. To do this, paste the following into the search field of your instance: !technology@lemmy.org
Visibility: Public
globe

This community can be federated to other instances and be posted/commented in by their users.

  • 1 user / day
  • 2 users / week
  • 2 users / month
  • 10 users / 6 months
  • 2 local subscribers
  • 411 subscribers
  • 575 Posts
  • 30 Comments
  • Modlog
  • mods:
  • Mazdak ( Mazdak@lemmy.org ) 
  • BE: 0.19.21
  • Modlog
  • Legal
  • Instances
  • Docs
  • Code