Here’s Knuth in an interview:
As to your real question, the idea of immediate compilation and “unit tests” appeals to me only rarely, when I’m feeling my way in a totally unknown environment and need feedback about what works and what doesn’t…
Hmm, people who are “feeling their way in a totally unknown environment”… Like new contributors to an open-source project or a new employee doing maintenance work on a project after the original team has gone on to other companies.
…otherwise, lots of time is wasted on activities that I simply never need to perform or even think about. Nothing needs to be “mocked up.”
Good for him. Here on Planet Earth, developers are often asked to work on projects they didn’t design and implement themselves, and do so in a way that doesn’t horribly break something that already exists. Or work with others because your desired endstate and timeline are not such that you can do it yourself or work things piecemeal and take it back for a redesign. </snark>
James Cape 
Why did you end your post with a “snark” ? I find it mildly offensive!
He is a scientist. Read the rest of the interview about how he approaches problems. If every developer had the time and guts to approach problems this way (and the wage, of course), there’d really be no need for unittests. I use them anyway, as I rarely know my environment as well as he does because nobody pays me for knowing, people pay me for making stuff work, and that’s unfortunately the way our system works.
Other than that, I find unit tests are fun. And I love reading what Mr. Knuth writes, because he’s in my opinion one of the most brilliant men around. If I were as brilliant as him, I’d still use fvwm, too 😉