I am not familiar with Advent of Code. If all their problems take ~0.1s to solve in C, you are probably right. However, it is a stretch to make a high-level conclusion just based on small exercises, like this:

> There was never a time where rewriting in another language felt like the right way to get better performance.

Similarly, for small exercise problems, dynamic typing rarely hurts, but for large collaborative projects, static typing starts to show its advantages.

> Were I to implement a language […]

So this was not related to AoC but instead to building a language.