I totally get your point and agree to an extends, though I have not yet been able to create that trust with the LLM. With human teams, yes, with LLMs, feels like I still have to verify too much.