Its fair to question just how far AI driving can possibly go at the moment. Most of the AI we have today is being used for classification and recommendation in

Its fair to question just how far AI driving can possibly go at the moment. Most of the AI we have today is being used for classification and recommendation in fairly low-stakes environments (advertising, (search, social)). It _is_ a big jump from that to powering large metal boxes that move at high speed around built up areas. Yes, of course AI driving is coming, but in the next 3 years? Almost certainly not. And then there's the question of where the next advances in AI are likely to emerge from- Google and Amazon? MIT, Oxford or Stanford? A well capitalised startup?