You bring up a very good point that hits on the most fundamental difference between video tutorials vs text-based tutorials.
Video tutorials can’t be edited, therefore I believe they should not be created for topics that are likely to change in some way. A good example of this is “how to handle collisions in Defold” because the likelihood of changes or additions to collision strategies, the standard API, or custom libraries is very high. If the tutorial is done in text format, then it can simply be edited. If the tutorial is done in video format, then it becomes outdated, misleading, or at worst incorrect.
Something like procedural generation with cellular automata is a good video topic because it’s a technique that’s been out there for a very long time, and the concepts and fundamentals haven’t changed. It works and won’t ever not work.
I think this consideration of being able to edit vs not being able to edit should drive more tutorial-makers to choose the best medium, rather than just defaulting to YouTube.