How can it be that during the DARPA Grand Challenge an autonomous car drove 7 miles in 7 hours and 6 years later Google autonomous Vehicle drove 100 000 miles?

Peter Norvig, Chief Scientist, of Google has an explanation: “We don’t have better algorithms. We just have more data.”
The data was the Google street-view vehicle. The data came from humans who drove with the Google street-view cars the roads, equipped with detailed sensors which measured, photographed and collected all the data. The data was stored in the cloud and made available to the Google autonomous Vehicle. This is an example for rethinking human-machine symbiosis.

All this data makes the Google autonomous Vehicle project just possible. It is a fairly hard AI problem to pic a traffic light out of a video stream. It is a trivial AI problem to figure out if it is red or green if you already know it is there.

Read more: e-corner Stanford University’s Entrepreneurship Corner
Stanford Technology Ventures Program (March 6, 2013)