Google's new parent company, Alphabet, just posted their first quarterly earnings, and they've just become the most valuable planet on earth, overtaking Apple. Google now stands for 'G' and is part of a myriad of different companies that form the new corporation that is Alphabet. The idea being that these esoteric firms, such as X, can better function as part of Alphabet, rather than part of a company born to make the most of the Internet. Google still has a handful of questionable endeavours, compared to their traditional Internet antics at least, and the self-driving car is just one of those. As more data on these self-driving cars emerge, it's pretty fascinating to see what lengths Google are going to.
In recent years, Google has been known for answering our questions outright, rather than simply serving up relevant search results, and how the Internet giant does this is through machine learning hosted throughout their many data centers. These data centers are being put to work by their self-driving cars too, or at least their brains, anyway. It's been reported that the software inside of these cars is ran through simulations of 3 Million miles using data centers all over the world. That's a pretty massive number, and is so large you could spend hours and hours thinking of equivalent measurements, but the real takeaway here is that Google is dead serious about their testing. These simulations give Google the opportunity to ensure that new code or changes are safe before they hit their mix of "Panda" cars and Lexus SUVs currently cruising Mountain View and Austin.
As it stands, the fleet of autonomous cars already do 10,000+ miles on the roads testing in the real world, but these simulations take things to a new level. Simulators have become incredibly sophisticated in the automative world, and much of the off-season testing during Formula 1 is done via a simulator. For Google's self-driving cars the question seems to be less of whether Android dream of Electric Sheep and more of "when do I get a day off?"