News for hackers. Stuff that matters.

A comment on OpenAI's o1 correctly diagnosed 67% of ER patients vs. 50-55% by triage doctors. Parent comment

Re:OpenAI's o1 correctly diagnosed 67% of ER patients vs. 50... by SirHumphrey on Monday May 04, 2026 @01:56AM (#48005114)
And with the code, the closer you come to the physical world the worse LLMs fair.

Claude can’t really write Openscad and when I was debugging some map projections code last week it struggled a lot more than usual.

[ Parent | on HN ]
Re:OpenAI's o1 correctly diagnosed 67% of ER patients vs. 50... by prplxd_nihilist on Monday May 04, 2026 @03:05AM (#48005551)
Until anthropic hire or steal code from acquired companies and train with it.
[ Parent | on HN ]
Re:OpenAI's o1 correctly diagnosed 67% of ER patients vs. 50... by DrewADesign on Monday May 04, 2026 @12:29PM (#48010923)
I think that might help a little, but is not a solution. When you’re figuring out some new way to combine code instructions to perform novel coding tasks, you’re just finding new configurations for existing patterns to get results you can easily test. The world outside of computers is infinitely more complex, random, and novel.
[ Parent | on HN ]