That isn't true. The whole point it to quickly pick up statistically significant... | Hacker News

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

		nl 17 days ago \| parent \| context \| favorite \| on: $500 GPU outperforms Claude Sonnet on coding bench... That isn't true. The whole point it to quickly pick up statistically significant variations quickly, and with the volume of tests they are doing there is plenty of data. If you turn on the 95% CI bands you can see there is plenty of statistical significance.

withinboredom 17 days ago [–]

Unless you and I are looking at different web pages… it only goes back to February, not December or January.

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