I didn’t work on this project personally but I feel like “undermining” is a strong word. Did you feel the same way for eg BIG-bench / HELM releases? Do you think it is good that there are more MIT licensed evals on GitHub?

→ View original post on X — @karpathy,