You could say that LLMs are a general way to fix bad UX in human systems that use language as their interface. Of course, in a better world, this would be fixed via user-friendly online forms instead.

→ View original post on X — @fchollet