I actually think that could be an amazing resource to benchmark the LLM - essentially incentivizing people with internet points to find particularly difficult o

I actually think that could be an amazing resource to benchmark the LLM - essentially incentivizing people with internet points to find particularly difficult or edge-case questions, that would help improve the model.