Comment by amoss

The details might bury his point rather than illustrate it. The driving theme throughout seems to be that a tool tuned for correct syntax, with deep understanding of semantics will look like a Dunning-Kruger machine. The specific errors that the author's colleague was oblivious to don't add any weight to that general point, they only explain one specific instance. It's classic omega-consistency.