Google’s all-or-nothing approach to driverless cars was apparently brought about because road tests proved the computer-human tandem incompatible. It was a punch to the gut of Google X at the time, but it forced the emergence of a fully autonomous vehicle. From Alistair Barr at the Wall Street Journal:

Google’s self-driving car project, one of the first to emerge from Google X, also faced major challenges, [Astro] Teller said.

In the fall of 2012, the team thought it had finished because it had built a car capable of driving itself safely on highways. Google gave some of the vehicles to other Google employees to use to commute to and from work and made them promise to continue paying attention to the road, Teller said.

“The cars performed flawlessly. The people did not,” he added. While not providing details, Teller said the employees paid less attention because they assumed the car would take care of any incidents.

“It was not pretty. We stopped doing it. We realized humans cannot be a backup system for the computer,” Teller said.

The team had to re-design a new vehicle capable to driving itself all the way from point A to point B with no help from a human driver. Teller said this was an “existential” blow to the team at the time.•