Only few businesses want a well-crafted software. Today's startups are endeavors of treasure hunts. You build a ship just good enough to hit the island. If it doesn't work out(which in all probability it won't) you dump the ship. If it does, you still dump the ship. There is a element of expected failure in the current enterprise of writing software which calls for and pushes people to knowingly write bad fast software. As the author says it will be faster to write slowly if the goal is to write good software. But the goal is rarely that.