It seems like it's fundamentally different. I've seen 4 different senior engineers use the AI drastically differently to improve the way they write code and think about problems.