Re:OpenAI claims Gold-medal performance at IMO 2025
by falcor84 on Saturday July 19, 2025 @12:57PM
(#44617169)
"Software that sometimes works and very often produces wrong or nonsensical output" can be extremely valuable when coupled with a way to test whether the result is correct.
[ Parent
| on HN ]
There are no comments on this comment yet.