Technology

88803 readers

3716 users here now

This is a most excellent place for technology news and articles.

Our Rules

Follow the lemmy.world rules.
Only tech related news or articles.
Be excellent to each other!
Mod approved content bots can post up to 10 articles per day.
Threads asking for personal tech support may be deleted.
Politics threads may be removed.
No memes allowed as posts, OK to post as comments.
Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
Check for duplicates before posting, duplicates may be removed
Accounts 7 days and younger will have their posts automatically removed.

founded 3 years ago

MODERATORS

submitted 6 days ago by Mustachius_Grumpius@thelemmy.club to c/technology@lemmy.world

41 comments fedilink hide all child comments

you are viewing a single comment's thread
view the rest of the comments

[–] gole@lemmy.zip 11 points 6 days ago* (last edited 6 days ago)

It is sarcastic. But it is also what OpenAI tried.

The reason they just drop these and say nothing is because these are unverified slop.

OpenAI's method of verification is rewrite them as Lean proofs that can be verified by computers. Two things have happened so far:

The Lean proof proved the slop wrong, easy, retract.
The Lean proof, being generated by a LLM, is susceptible to hallucinations. For example the Navier Stokes problem Lean proof turned out to be slightly different than the original natural language proof, because the LLM tried to bend a condition to make the proof compile.

But, in case nobody found any problem, OpenAI gets to claim credit for the discovery until someone can review and prove something's wrong.

You can say these proofs are "Schrodinger's correct"