When I was a manager, I classified mistakes in three categories:
human error (To err is human). For example, on average I hit a wrong key on a keyboard every so many keystrokes. As long as the human error rate is within the acceptable range, there is no problem. We design systems with this taken into account.
BAD mistakes: doing again the same thing that didn't work before. Also know as stubbornly stupid mistakes. Work to eliminate such mistakes.
GOOD mistakes: these are a result of a reasonable (but erroneous) decision near the limits of a person's knowledge and capability. They come from pushing the limit and result in learning. I sometimes had to actively encourage such mistakes (otherwise my people would perform below their capability and not learn). Too many such mistakes can be an overload. I generally budgeted 15% of my time for dealing with GOOD mistakes my people made.
"I do not like to err" was my answer, because it is much more concise than "I don't like to make mistakes", which aside from not being at all poetic, clumsily has to build a single verb action from a generic verb and a noun. So it is efficient and eligant, just like the Esperanto.
To herr is uman! It's not so funny in English or Esperanto, but we often write "Herrar é umano" (kidding, obviously) in Portuguese because "h" is a diacritic letter (it has no sound by itself). Correct sentence: "Errar é humano. Ficar no erro é burrice" (To err is human. To stay in error is silliness)