>We now have a tool that can be useful in some narrow domains in some narrow cases.

I get being reserved about where this goes, but saying something like this is quite insane at this point.

I'm not an enthusiast. I'm a Butlerian.

Imagine hearing pre-attention-is-all-you-need that "AI" could do something that Donald Knuth could not (quickly solve the stated problem in collaboration with his friend).

The idea that this (Putnam perfect, IMO gold, etc) is all just "statistical parrot" stuff is wearing a little thin.

You must have forgotten the /s at the end of your comment?