It appears that Tesla Autopilot is closing in on yet another milestone. Just a year after reaching its 1 billionth real-world mile in Autopilot, the electric car maker’s fleet of vehicles appears to have traveled yet another billion miles, providing Tesla’s Full Self-Driving initiative with a total of 2 billion miles worth of invaluable driving data.
Tesla is yet to issue a formal announcement about its total Autopilot miles to date, though a model by MIT researcher Lex Fridman has predicted that the 2-billion-mile mark will be crossed sometime on November 11, 2019. That being said, Fridman’s model might even be a bit conservative, considering that Tesla’s fleet is growing by the day and more and more vehicles are traveling across the globe on Autopilot thanks to the Model 3 ramp.
Elon Musk, for one, appears to have given his approval of Fridman’s estimate. Last Saturday, Musk liked a link to a blog post predicting that the 2-billion-mile mark will be reached on November 11. This may suggest that the estimate is accurate, or at least close to the actual figures that Tesla is recording from its growing fleet of vehicles.
If Tesla does confirm Autopilot’s 2-billion-mile milestone, it would be a great accomplishment for the electric car maker. After all, it took 4 over four years for Tesla’s fleet of vehicles to reach 1 billion real-world miles on Autopilot. Provided that Lex Fridman’s estimates are accurate, Autopilot’s next billion miles seems to have taken less than a year to accomplish.
And this is with Tesla only finding its stride with the Model 3’s global deliveries. The Tesla Model Y, a vehicle that may very well be the company’s best-seller considering that it’s competing in the crossover segment, is not even in the picture yet. Gigafactory 3, which will be producing affordable, Autopilot-equipped versions of the Model 3 and Model Y for the Chinese market, is yet to begin its full operations either. With Model Y and Gigafactory 3, Autopilot could hit its 3-billion-mile mark in an even shorter period of time.
Real-world data is key to Tesla’s vision of a fully autonomous future. Tesla CEO Elon Musk envisions a future where vehicles could navigate to their destinations on their own without human input, which would require them to traverse inner-city roads on a regular basis. For this to happen, Tesla’s Neural Network must be adequately trained and deployed to the company’s fleet of vehicles, and it must be prepared to handle any corner cases that it may encounter on the road.
As noted by the electric car maker in its Autonomy Day presentation last April, there is no better way to train Neural Networks than to utilize real-world data. The improvements to Full Self-Driving features such as Navigate on Autopilot and Smart Summon since the features’ initial release prove that Tesla is likely correct in its Full Self-Driving approach.