> I dunno man. I sold a small company back in 2011 and it really was easier back then. Three out of four of our first leads tried to buy us because we were doing "real" machine learning. This was almost a pure API app. We didn't even have documentation we just told people we'd integrate it for them.

You're kind of agreeing with their point. Your example of success was where you were (at least perceived to be) solving a legit need with a real solution.

I agree with the grandparent - people will use a crappy looking app if it solves something they're desperate for (and by implication, isn't solved for already in the market).