Re:Frontier AI has broken the open CTF format
by recursive on Saturday May 16, 2026 @08:18PM
(#48164989)
I think they are less reliable. For factually verifiable facts LLMs are doing worse than 90% for me. I've been told some incorrect things by educators, but at a much lower rate.
[ Parent
| on HN ]
There are no comments on this comment yet.