Yeah, anyone that invokes Turing machines to claim unittests don't work has completely missed the point of testing their code. Perhaps if the article was talking about "formal verifications of code" it would bear some resemblance to good advice.