he was dealing with that situation already in the 01970s—he was using an extensible programming language that people kept extending, so the programs he wrote one day would break the next—which is why he designed τεχ in such a way that you can take any τεχ document from 40 years ago and render it in exactly the same way in current τεχ
one of the drop-dead showstopper tests in the τεχ release process since that time has been the 'τεχ torture test'; it's an enormous random τεχ document which has to produce byte-identical output on each new version of τεχ for it to be released, unless he can justify each difference. so he's not opposed to automated testing, he just tends to do it at a larger granularity