alt Hacker News20 minutes planning, 20 minutes coding, 200 minutes review and refactor (includes going for a walk and thinking about the problem deeply).
I know a lot of engineers who skip the last part. They're over confident in their original plan. They're over confident the agent actually fulfilled the plan.
You aren't treating this as a question of ROI. Is it worth spending 5x as much to make sure the plan was OK and implemented well? Or is it actually OK if we discover the bug during testing?
The answer won't be the same for all software, but you're assuming it will be.