And are your tests abstracted as much as possible? Ie Are you decorating methods to describe the acceptable range and domain and have a hypothesis engine test for that? Are you using snapshot testing to make things easier to trace? I can go on but I don’t know the context nor do I think we should really sync on the specifics of your job. The point is you are at a point where you think the code can’t be made simpler, maybe that is true, but if AI is doing all the boring stuff then I don’t know how one would even care to know.