Building gives you code and screens, which show momentum, while investigating demand mostly gives you negative information, which feels like you're not going anywhere - even when it’s the highest leverage work. For me, I reframe validation as finding out what people already do ... not trying to convince anyone or predict any kind of behavior. It’s much more simple and less satisfying than building, but it helps me make decisions faster.