Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Ok so add a factor of 20 to match the 8000 questions that OpenAI tested on, it's still crazy efficient compared to the NS result
 help



could be 20, could be 2000000

They have specified a 5% success rate where does 2000000 come from? Even if we ignore their statements there simply aren't enough open questions for them to have a much higher failure rate than this, someone mentioned elsewhere in the thread that they resolve 90 of a slop list of 500 open questions.



Consider applying for YC's Winter 2027 batch! Applications are open till November 2.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: