The problem is that user research and competitive research are also not the truth. I prefer to ship something I know I like than what someone else thinks a third abstract person might like.
Congratulations! Looks awesome.
1. I found it very intuitive.
2. If I could have smart filtering using llm classification, that would be very powerful. Any plans on doing that?
TL;DR - the benchmark depends on its specific dataset, and it isn't a perfect representation to evaluate AI progress.
That doesn't mean it doesn't make sense, or doesn't have value.