Model 3 will get the latest technology available at Tesla

Tesla’s Director of Autopilot Programs Sterling Anderson spoke on Tuesday in San Francisco with MIT Technology Review’s editor in chief, Jason Pontin about the future of autonomous driving his team is working on with the main goal to accelerate the world's transition to sustainable transportation and eventually reduce global emissions. Anderson highlighted how Tesla inertly updates its Autopilot software in order to improve the features on its vehicles.

Anderson sat down at the end of the presentation for some Q&A with Pontin who threw a tricky question at him about the Model 3. "Will the Model 3 be the first commercially available autonomous vehicle?" asked Pontin.

"You are throwing me for a loop Jason" avoiding confirmation,  Anderson then explained that the Model 3 will receive the latest technology both in hardware and software as soon as it is tested and implemented on its other models available on the market. What that means is, during the next 18 months of pre-production, the latest hardware and software being implemented on the Model S and X will then trickle down in the production line to the Model 3 come end of 2017. "Every single car coming out of the factory has all the hardware in it to perform these capabilities in it. Every safety feature we develop that we constantly improve upon is included in the car because we don’t think those are optional."

"Our vehicles will receive the latest technologies as soon as we have it, we won’t wait for a new model to release new technology. We fundamentally choose to eschew the concept of model years... we will not hold any technology for Model 3 that we will not have released already when it is ready on our other models." confirmed Anderson.

Anderson admitted that Tesla was far behind the curve - 18 months ago when the hardware first featured in Model S - in terms of advance safety features and autonomy. But they did a very rapid development cycle that leverage uniquely the existence of the hardware on the fleet and their ability to pull data from it and update the software on the cars incrementally.

As new features are conceived in house, the team then run it through several simulations through a large set of data where they simulate each of these new features to insure that - on the net - they improve the safety and the capability of the car. It is then sent for in house validation where the team has access to vehicles to drive, develop and confirm that the software is well suited.

Post in house validation, the team will install - in a logging fashion - an inert feature over the air on all of their vehicles worldwide, today 70,000 vehicles. What that allows them to do is watch how that feature performs over tens of thousands of miles. They can see how the feature has behaved, and within weeks to months they can then activate the feature on the vehicles once they confirm it’s fundamentally safer both in simulation, in house validation and in the field - from the empirical data collected from customers’ cars - if this in fact a feature that is better on the net than what’s in production today. Finally this feature gets tested and approved for over the air update. That lead to rapid roll-out for several features both safety and convenience focused.

You can watch the full presentation here.