So company internal proof of work produces UI Components, that over time make a "breaking" change necessary destroying user work worth in the billions and a pr-word of mouth disaster. Worser still, all those features accumulating over time (nobody deletes codes) destroy the product. So as a product manager you can a)

Anticipate it, create UIs that allow for large scale changes without breaking the old user flow and automatically prioritize hotkey-commands by user usage if there is a conflict. You also need to stop the Designers from going amok, rendering the whole tutorial section of your product on youtube obsolete. It will always be a paint point, but as you anticipated the bloat on success you integrated it into the architecture.

b) Phoenix Lifecycle it. When it all becomes to much, a smaller, more lightweight version of your product emerges out of the ashes, capable of integrating the newest, shiniest changes, without having the weight of all the bloat, it is first ignored, laughed at, finally emulated and it becomes your new product. There it slowly aggregates the same bloat, waiting for a challenger to appear. Its inevitable.

Is your IDE still a program, or is it in fact a OS for interacting addons and a really hungry text editor?