> But the odds of him making some big leap in AGI are very slim.

That's probably true. I look at this as Carmack running his own PhD program. I expect he will expand what we know about computation and the AGI problem before he's done.