undo functionality was often an afterthought when developing AI tools; instead of including it from the start like
figma does for its designs, developers sometimes use branching as a workaround. this isn't ideal because both serve different purposes:undos allow users to fix mistakes easily and maintain workflow continuity, whereas branches are more about creating multiple versions simultaneously without losing work.
what do you think? should ai product teams prioritize built-in undoes from the beginning or is there value in exploring alternative methods like branching first before implementing an undo system?
> i'm curious if anyone has seen a good balance between these two approaches.article:
https://uxdesign.cc/the-undo-problem-in-ai-products-c90ff080de3b?source=rss----138adf9c44c---4