Yeah most of my team is using an LLM for "make this function better", or learning, or just somewhat smaller bites or code that an LLM will work well with. So we don't see the "hey rewrite this whole 20 year old complicated application, omg it didn't work" kind of sitiatons.