While the author makes some good points, I think he gets a little too enthusiastic, e.g. near the end:
"Your program will always be better off if you spend a day writing it 6 times, rather than designing it 6 times and writing it once."
Yes, but something like design, write, redesign, rewrite, rewrite... will likely be even better. And the more experienced you get---in the domain, the particular problem and orthogonally, the language, libraries and so on---the more you can do detailed design upfront and less code rewriting.
More times than not, I've found rewriting code leads to more problems than fixing some trivial error. Some small bug in the code is not necessarily a symptom of a horrible design that will destroy your project.
I don't know about the random guy who wrote this article or his experiences, but I rarely find myself with a few weeks to throw away rewriting code that works.
"Your program will always be better off if you spend a day writing it 6 times, rather than designing it 6 times and writing it once."
Yes, but something like design, write, redesign, rewrite, rewrite... will likely be even better. And the more experienced you get---in the domain, the particular problem and orthogonally, the language, libraries and so on---the more you can do detailed design upfront and less code rewriting.