Can state of the art AI LLM models solve today real world data analysis problems? Well unfortunately it seems like no. DABStep: A new extremely challenging benchmark for model training teams

→ View original post on X — @thom_wolf