Comment by technoplato

Thank you! My favorite bit (theoretically) would be the forcing function to produce useful open source goods as a result of running these benchmarks and quiet down all the hype about llms. We get it, they’re very good at making 3d models do dumb shit that doesn’t matter