I'm no more trusting the benchmarks. other than trying it out myself, what else ... | Hacker News

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

		shnkr on March 27, 2024 \| parent \| context \| favorite \| on: Claude 3 beats GPT-4 on Aider's code editing bench... I'm no more trusting the benchmarks. other than trying it out myself, what else can we do here?

refulgentis on March 27, 2024 [–]

It's already been done (ELO, see LMSYS rankings). I hope we're cresting past the 50% percentile mark of people who haven't heard of it.

shnkr on March 27, 2024 | [–]

I see. thanks for the reference. followed it on x now.

https://twitter.com/lmsysorg/status/1772759835714728217

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