Comment by springtimesun

I’m at exactly that point where it sounds like you were. I’ve done 3 Access to Rails conversions and I’m hunting for the next one. The one I’m on at the moment is supporting 5 branches over 2 countries and 2 independent machine shops. Even if I can understand what Access is doing under the hood there is no one left to ask why. And I have so many questions. Sit with the users, spec the feature, ground it in whatever data I can find. I don’t think that ever changes for SMEs that take this path (Access or Vibeccess) and need re-writes. I’m also very happy to do them. They are IMO giving me more valuable usage data than any design process ever could.

What is different on this one vs the others is I have Claude to help me data dive and write the boring CRUD parts. I am able to spend so much more time with users testing and getting feedback and just thinking deeply about how to structure things. The quality of what I’m building now has never been higher and I think it’s just because I have more time to spend with it.

My experience with AI has been almost wholly positive and I wonder if Rails is part of the reason. Such well established patterns and structure the agent one shots most things and I spend most of my time wrangling view code based on my preferences.