@lalleal What is needed is products that solve actual use cases and provide a real value add beyond training ML models. Not sure if it is hard to generalize or we just haven‘t found ways to do so yet.