The latter. Ideally you should always talk to potential users/customers first to validate if you you idea truly is one that solves a genuine problem (ie proble

The latter. Ideally you should always talk to potential users/customers first to validate if you you idea truly is one that solves a genuine problem (ie problem/solution fit). Only then go and build an MVP and test it with these same people - who can then also refer other interested people (ie product/market fit). Here you will truly validate the need for the product and/or pivot to something else that can get to product/market fit. Then, and only then should you look to do things like marketing to get the product in front of people. This is the lean startup approach. Following (something like) this significantly raises the odds of success. Building it in the hope of someone finding it/using it is a waste of time/money/energy.