telling all the big tech companies to stop blindly competing on video models. Feels a bit against the current, since everyone’s diving in right now. But if you think about it differently, all this homogeneous model stacking and parameter comparing might just end up with a bunch of products that look the same, with no one pulling ahead.
What’s really scarce is the actual use cases and workflows that make good use of the models, not just another engine that can generate videos. By “blindly competing,” they mean pouring in huge resources without finding a real landing spot. This advice applies to startups too. So, what do you all think—should we focus on the models or the applications?