At least one restaurant opened in your city last week. Of the set of restaurants which opened in your city last week, one of them is last in the alphabet. Without checking either what restaurants opened last week or what the reviews say about them, what did the reviews say about that restaurant?
You should be so lucky as to have everyone know how bad you are.
I don't know any of the restaurants that opened last week, because I get my information about restaurants from restaurant early adopters and reviewers. This is how most people get their information about restaurants. That's why it makes sense for restaurants to follow Geoffrey Moore's advice of using the "bowling alley strategy" from Crossing the Chasm.
If you open a restaurant, you know in advance that you need to have a strategy to get restaurant early adopters in the door. And you also know that restaurant early adopters are the kind of people, by definition, who don't go back to the same place twice if it isn't very good.
Basically, it's only luck if you're successful without knowing this and having a plan to deal with it. And counting on this happening is a poor backup plan at best.
A restaurant that keeps its doors closed for the first two months had better have some great food to reclaim all those overheads when it finally opens. How can you be sure that the food is ever ready if you don't bring in at least a few customers?
Rather than starting with the statement "never launch, just iterate" it would be better to start with the following question:
"Should I start a business that lends itself to iteration?"
That's an interesting question that's worth discussing. But as it stands this post is the wrong answer to the wrong question. (Which also shows how you can often catch sloppy thinking by making explicit the question that a statement is answering, and diffing that with the question you should be asking.)
edit: I think the actual answer is that you should choose to start a business that lends itself to iteration, but that's different than the advice given. It's another case where the given advice is close enough to the truth that it tends to work, but the underlying rule is still wrong.