logoalt Hacker News

theturtletalkstoday at 12:36 AM1 replyview on HN

Marketing push was there too, everyone was saying Grok had jumped Claude and Codex, yet I never got that when using all 3.


Replies

ascorbictoday at 6:53 AM

Turns out that benchmaxxing doesn't help if it's not very good when people actually try it.