RE: Ash. We actually have had the total opposite experience. Used it for two years, had a support contract with the library creator’s employer and everything, and the productivity penalty has been so drastic we decided to rip it all out and return to Ecto + standard json views.
We are a team of ten and the codebase is about 200k LoC. Maybe larger teams have a different experience.
Regarding AI, I’ve found that even Opus 4.6 and Codex 5.3 struggle with Ash. For some reason they can’t figure out the “Ash way” of doing standard operations and always overcomplicate things. Stuff like “pull the entire dataset from db and sort in memory” are anti patterns they constantly fall for with Ash. With Ecto they don’t. Maybe because Ecto is so similar to SQL.
