Faster delivery does not stop silly requests

There is always another request. What happens if you produce better models and reports faster than ever? Nothing different than what happens whenever you deliver too early: you’ll get additional requests.

So the idea for this comic was that early delivery can actually worsen the final result because the stakeholder will ask something anyway, and at that point they can only make silly requests. Perhaps we should instruct the LLMs to include easy-to-detect errors and ask them to be fixed?