I think you've gone from one extreme to the other. Yes, building a product with no distribution sucks. So does doing distribution without having anything to show for it. Take this from a guy who spent 2 years studying distribution channels [1]. There's nothing wrong about spending a week or two developing a MVP before focusing on distribution. I think the key is to start with 1) The minimum thing you can do and call…
I don't plan on spending more than 3-4 weeks building an MVP, which I'll start within 2 weeks.
An advantage that my current approach gives me is a few dozen low-commitment (but qualified) waitlist users that have shown some initial interest.
I obviously have some close friends that will try it out with their teams. But going distribution-first allows me to easily expand beyond that close circle to a more unbiased group of people.