Kurz is an exceedingly clever and well informed polymath. I was once offered to run one of his R&D centers in MA. Read his books. But the "fruit fly" goal posts cited by Mead & I have been 'day one goalposts' from both proponents and skeptics. If you read what Mead says in detail he and others like him are well aware of what I call computational/algorithmic AI, stuff like playing games, analyzing voice FFTs, whatever. Recently they've accomplished great stuff like the DARPA car challenge prize.

But Mead's point is that with all the algorithms and computational hardware we STILL can't do what the AI visionaries of the 70s touted as iminent with increased computational power, e.g. Minsky, Moravec, et. al. Reading Moravec's "ROBOT" you can see how he claims we've vastly exceeded inscetile AI, yet truth be told all the AI in the world can't navigate thru your house and avoid being swated as well or as quickly as a housefly. Digital computational systems are nothing like biological systems, which despite low power. tiny size and really slow 'clock rates' outperform the petaflop machines, embarrasingly so. We do not have a scientific or a pragmatic engineering clue how to build a 3-D real time predator and avoidance system like a fly embodies.

AI rightfully claims credit for 'expert systems' that improve our daily lives, but hard AI as touted by the True Believers is a different animal. Like a fruit fly. :)

Ain't happenin soon despite all the touted exponential curves.