Points well taken. But still way overhyped/valued (see NVIDIA) IMHO. 
In practicing medicine, there is no prize for being fast/first, one humbly just has to be right about the problem (=diagnoses/treatment).

I believe this to be true of many important fields: energy/nuclear reactors; energy grid; physical engineering problems; computer code solving complex probs…etc. 
The real value is getting it EXACTLY RIGHT. I see the LLMs getting to the answers *fast* and *sounding good*; it is not so clear to me the AI will get it RIGHT compared to a highly motivated, seasoned, thoughtful human who is interacting with the real world.