logoalt Hacker News

root_axis • today at 7:11 PM • 5 replies • view on HN

What value is there in just dumping something into an LLM and saying it produced errors?


Replies

arcastroe • today at 7:18 PM

It's being hurriedly announced as fully human work. I wonder if the author had carefully avoided ever running the full argument through frontier LLMs (OpenAI/Anthropic) for fear of it being stolen or incorporated into AI training.

This may be the first time this proof is being proof-read by frontier models

(this is pure speculation, but provides an answer to your question)

➕ show 1 reply
markasoftware • today at 7:16 PM

Have you tried doing this recently? For my work on software at least, the latest LLMs have a very low false positive rate for bugs. They almost never hallucinate issues. My understanding is the same applies to math.

➕ show 7 replies
zone411 • today at 8:47 PM

Very high value. The LLMs are excellent at auditing math proofs.

mtlmtlmtlmtl • today at 8:30 PM

Seriously. At least be constructive. Share what the errors were. Make an effort to understand them, anything. That tweet just amounts to low effort FUD.

augment_me • today at 7:34 PM

[flagged]

➕ show 1 reply