I just learned that this exists. The cage match for LLMs.
This is bad.
Now, wait a minute. What could be so bad about crowdsourced, “non-biased” ranking of the LLM models out there? Why is that bad?
The intent isn’t bad. Of course, not. But it’s bad in practice.
The creators have published their methodology. And hence, the model developers have started to game the system. It’s turned into, or will turn into, the college ranking system, corporate productivity KPIs, and rankings of the best companiesto work for.
Basically, useless for it’s stated purpose.
And if you start caring more about where you rank on some made-up list, you have no choice but to stop focusing on how to become actually great.