Driving in Circles: The autonomous Google car may never actually happen

I absolutely agree that autonomous vehicles are not at all ready for primetime right now, but this article doesn't understand the implications of Moore's Law and the rate that AI is improving as a result of this exponentially increasing computing power.

Specifically, this quote: "To be able to handle the everyday stresses and strains of the real driving world, the Google car will require a computer with a level of intelligence that machines won't have for many years, if ever."

If processing power is doubling every 18 months (which is has consistently for the past 100 years and looks poised to continue), then this statement is nonsense.

This article about computer vision gives a good understanding of progress in AI: http://bits.blogs.nytimes.com/2014/08/18/computer-eyesight-g...

Computer vision is doubling in accuracy almost every years (due in large part to exponentially increasing computing power). Similar trends are occurring in natural language processing, robotics, etc.

I would trust Sergey Brin's ability to forecast how advances in computing power will enable such technologies far more than some non-technical journalist. I think Amara's Law sums this up nicely: "We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run"

Thus, driverless cars may not be ready in the next couple years. But to then conclude they will never be ready is ridiculous.