On the topic of not enough carrot, I’m curious how impactful the end of support for python2 will be. How many programs stuck in python2 are encountering bugs in the runtime?
I actually find this aspect of the whole thing exciting, perhaps paradoxically. Now that Python 2 is static the runtime can become asymptotically bug-free. Meaning that, if you only change it to fix bugs (as opposed to introducing new syntax|semantics) it's going to approach perfection.
I've considered taking it on. If there were enough money in it, I'd very seriously consider it. It's a very good language.
In ten years, there will be serious money in it as enterprise codebases need to be renewed.
You are assuming that fixing bugs introduces less new bugs.
Python itself won't be as much affected, what will affect you most are your dependencies. Imagine using a library and running in some kind of bug. You check newer version, looks like that bug was fixed, but the library now only works on python 3.6+. What will you do?
Backport the fix from 3.6, or just implement a fix yourself?
Yes, that's one option it will be feasible initially but it will get exponentially harder.