Good piece, however there is no way any of this is going to happen within the timeframes being advanced by Smith and Shum, and I'd put a modest amount of money on it never being accomplished at all.
At the moment, AI can usually defeat an opponent at the Japanese boardgame "Go." That is the current pinnacle of AI achievement. It's all pie-in-the-sky and speculation.
One day, we will look back on this period of wide-eyed credulity with the same embarrassed irony that we now look back on things like "The Jetsons."
I tend to agree. People are far too optimistic on the tech timeframe and on the positive outcome of it all. People in the 60s thought we would have moon bases by 2000 :P
AGI is long ways away.