Isn't that the whole point of Agile MVP development? Sell something that doesn't exist and hope someone buys it? Personally, I don't expect much from it, after working with Airtable's API and hating every second of it.
Theres a lot of reasons why it makes sense to sell something before building it. You want to be up front about this fact but getting commitments to buy something is a good idea before actually spending money and time to build it.
There's some validity to the criticism, but that's more the intersection of short-cycle methods and American "hustle" culture than anything to do with an MVP approach itself. People with more integrity can take an MVP-centered approach and just be honest with customers about where they are.
Indeed, I think that honesty works better; underpromising and overdelivering is a great way to build trust among your initial customer base. Trust that you need to carry people through the inevitable bumps and anticipations of an early-adopter experience.
