I think that's an intriguing point - and I also feel we generalize "artificial general intelligence" to the point & in such a way where humans have yet to even

I think that's an intriguing point - and I also feel we generalize "artificial general intelligence" to the point & in such a way where humans have yet to even achieve that level of intelligence. How can we build smart systems if we don't know what smart even looks like and human benchmarks turn out inefficient for machines.