It's a very strange article. It's a strawman, the biggest problem is not that preprints that later get accepted and published are bad. The biggest problem is that some "preprints" are so bad that never get published.