You are one of the few writers here on Medium that I end up reading every article of. (What a terrible way to write a sentence).

This article was not an exception. It was complex enough to teach me new things and simple enough for me to understand 80% of it.

I imagine that refinement to the software structure of the thinking machines will make them more efficient, while improvements in the hardware layer will add to that. (Remind me not to write anything first thing in the morning. My brain's word box isn't wording proper).

I imagine we'll get to a point where prompts are analysed for their probably complexity and are then only processed to the point of an answer. So logic problems will get handled differently to nuanced emotional problems (maybe, unless both are equally difficult).

I know a lot of folks are still treating the various bottled intelligences like Search. They ask knowledge questions that they could have easily asked the Google. These type of prompts won't get handed up to the computationally expensive big brain. They'll get handed to an agent I expect.

I'll stop now. My brain isn't braining properly. Great article as usual. Please keep up the good work.

(Are you Italian or Spanish?)