It is a question of versioning and packaging. Basically, v0 of a product is for testing the initial hypotheses that people are interested in buying at all. And it also serves as a tool for finding the initial "fans" that are excited by your product vision and will help you flesh out the product.
"Sell before you build" is just way to communicate this.
btw: The post is very short. Here it is:
> This is such bad advice. It assumes that everyone will nail their product on the first go. Almost no-one does.
> If I went to customers and pitched Songbox to them with what was the initial version of it I had in my head - I would have gotten nothing. And if I'd given up on it at THAT point then I would have missed out on (a couple years later) having a business that's well into X thousand MRR and even more importantly, working with some amazing people and having truly exciting and unique experiences.
> There is no doubt that "sell before you build" will work for some people; but everything and anything will work for SOME people.
> For the general populace of makers it is crushingly bad advice.