Adequate quality of software is the only way to go fast. Best practices need understanding not blind following. Same applies to programming in general. Some failures will happen long after you have not failed fast (it all worked okeish, why bother?) when possibly your startup seeks for the exit or becomes a scaleup. At the time the only option may be to rewrite the code, which has been created by incompetent developers.

There is of course the right balance to be found, and the author of the article may have his own understanding of competence, but he has the point. It of course depends whether one wants to "pack up things" immediately or change the situation to the better.