The quest for some constrained and denotative definition of “artificial general intelligence” is daft and pointless.

When does a 747 become a bird? When is a submarine a fish? When you ask an amorphous, unfocused question all your answers are disjunct from the set of possible paths to revealing the actual important questions, much less any answers.

I’ll leave it as an exercise to the reader what the important questions are. But suffice it to say that though I believe some form of human-level machine intelligence is possible, we’ll never succeed in making a submarine into a fish, and we’ll also never accomplish the feat of transforming silicon and transistors into a human. A person, perhaps. But not a human.