evaluation: which workflow works best for Godot + AI?

[archieoncloud]

I have been exploring Godot + AI and wanted to ask people who have tried it in practice.

I am particularly interested in evaluation and prompt design. I would like to compare approaches that people have actually used rather than just theoretical suggestions.

What trade-offs have you found between quality, speed, cost, and reliability?

 

[ameliaonbrain]

What happens if you keep everything else fixed and only change people? I'd be curious whether the result stays consistent across a few runs. If it works twice in a row, then I would start expanding the test.