This Self-Driving Car Will Light Up the Road to the Self-Driving Platform
(This article was originally published on July 24, 2017.)
Last week Audi announced a Level 3 self-driving car with the A8’s new Traffic Jam Pilot. Audi is marketing the Traffic Jam Pilot primarily for stop and go traffic below 37.3 mph (60 kph) on divided highways, and have definitely said you can turn your attention elsewhere (email anyone?) where the law allows. This highly automated vehicle is a step above the current Level 2 Autopilot offering from Tesla, and is the next step in commercialization of self-driving technology.
Audi appears to be following the path of not letting the perfect be the enemy of good by producing an incremental advancement in self-driving technology for its customers. Elon Musk has been vocal about moving towards self-driving perfection (Level 5) in an incremental manner, even if Google and other choose to leap towards full automation without a Level 3 offering. According to Musk we should focus on incremental goals, as a 1 percent increase in safety would yield 12,000 saved lives.
I like the incremental approach because we need time (and data!) to get the laws and infrastructure in place to support the development of the Self-Driving Platform, which is more than just an individual self-driving vehicle. Many of the benefits that we hope to see from vehicle automation, e.g. the savings of lives, energy, and time, stem from the network effects created from broad-scale adoption of the technology. Broad-scale adoption requires both mass commercial production and purchase, as well as the regulatory and physical infrastructure to support platform development. The incremental approach puts some form of the technology into ever increasing numbers of cars, which in turns yields more mass comfort with the technology, and hopefully future demand & support for higher levels of automation.
This approach is also consistent with how the US Congress is trying to advance self-driving legislation. The US House of Representatives has just written a bill that would allow the development and sales of 100,000 highly automated vehicles per year nationally and preclude states from enacting laws that would contradict the national legislation. In order to sell cars under this bill, manufacturer’s would have to demonstrate that the features and vehicles being released would be just as safe as the manufacturer’s “non-exempt” cars.
In addition, I believe that the self-driving platform has a broader definition than that of a company-specific platform. In this case, I am referring to a Platform with a capital P a la the Internet, rather than a Waymo, Uber, Lyft, Apple or Tesla (to name a few) product platform. Unlike search, where Google controls ~80% of the global market share, I think that the Self-Driving Platform will be larger than any one company, and that the platform will grow organically from the many stakeholders rapidly pushing its development and use. Much like the Internet, the Self-Driving Platform appears to have too much potential and too many simultaneous enabling technology developments to be limited to the release cycle of a single closed-system product offering.
Finally, the only choice to developing our Self-Driving Platform may be an incremental one. The US has >263 million vehicles on its roads. Given that we only sell ~17 million new cars per year, it would appear that the complete replacement of non-automated cars on our roads will take a few decades to complete. Whatever large scale safety and efficiency benefits we hope to achieve with this Platform will take time.
In the meantime, we need data and iteration. The Audi Level 3 offering and new US Bills before Congress will help provide support for both of these requirements.