VibeBench: Measuring 1k Engineers' Opinions of New Models
12 points
1/21/1970
2 days ago
by jpschroeder
Comments
mhi3
2 days ago
jpschroeder
Oh very true. Benchmaxxing itself is basically gaming them.
2 days ago
ramon156
Love the idea!
Page is incredibly slow on mobile, probably the avatars
a day ago
memoryleakgame
800 commits in a year...
a day ago
"Published benchmarks are gamed, optimized, and overfit, and no longer yield a useful signal."
Is this true?
But I love this concept!