This.
When working with code that touches money, I don't ever want to hear the reason "because the LLM said so".
Yes, use AI. But let it properly convince you that it's proposal is correct. Just like you would another human, you shouldn't blindly follow what it proposes.
I have been trying to tell people this for a long time now.
As impressed as I am with LLMs, I am equally frustrated.
Just let go of the code…
Nah man, I write business critical production apps that deal with money.
I can’t afford not caring about the quality.
”This a preexisting test failure and not related to our change”
No you dummy, you messed up and broke it.