This is my eternal argument with mostly everyone I work with. MVP’s aren’t supposed to be products but hypothesis of what a product can be. From a development perspective, most MVP’s are quick iterations that are mostly thrown away whenever that original hypothesis is proven. Once that happens, you start from scratch having enough insight from the previous iterations to build a proper product. There’s no shame in “throwing away” to start from scratch as long as you do it having learnt enough to move forward.