Waymo expands to LA tomorrow (https://lnkd.in/g4AF4H-5), while Automotive News titles “Self-driving industry struggles as DARPA race anniversary nears” (https://hubs.ly/Q02pnGQt0). I disagree with Automotive News. Background ☑️ Inventions take a long time from R&D into safety-critical products: Take ESC (https://hubs.ly/Q02pnD2z0) as an example: Developed in the 1980s, the system had only little market penetration by the 1990s. In the 2000s, regulators had collected data to show its usefulness. In the 2010s, ESP became mandatory for all new vehicles in the US and EU in 2012 and 2014. Now, in the 2020s, it has reached over 80% market penetration. 50 years after invention, it’ll be more than 90 percent. #Autonomousdriving has come a long way already ☑️ Ever since I started to work on autonomous vehicles in 1997, I have been saying that “autonomous driving is always 20 years out.” ☑️ In the 1990s, the EU project Prometheus (https://hubs.ly/Q02pnzsS0) set the foundation for driver assistance systems, which launched around 2000 (https://hubs.ly/Q02pnwDg0) ☑️ In 2000, we managed to automate test driving on Volkswagen’s test track. (I was responsible for data fusion) ☑️ In 2004, the first DARPA Grand Challenge in 2004, the best vehicle completed just 7 of 150 miles (https://hubs.ly/Q02pnGPs0) ☑️ In 2005, 5 vehicles managed to complete a similar course, driving autonomously 132 miles through the Mojave desert. (We, Stanford University won, https://hubs.ly/Q02pnyMX0) ☑️ In 2007, 6 vehicles completed the 2007 DARPA Grand Challenge (Carnegie Mellon University won, https://hubs.ly/Q02pnNlh0) ☑️ Google started Project Chauffeur in 2008, revealed its existence in 2010, by 2015 had driven 1 million and 20 million by 2020. In 2022, Waymo started offering rides in SF (https://hubs.ly/Q02pnGQv0) and just got approval to expand to the Bay Area and LA (https://hubs.ly/Q02pnzVV0). This is their opportunity to scale. Yes, ☑️ productizing self-driving is a monumental task, especially in getting the safety aspect right. ☑️ Yes, we (i.e. domain experts) knew all along that it would take a long time. ☑️ Yes, people in the 2015s overpromised millions of robotaxis within 5 years. ☑️ Yes, some pushback was to be expected. ☑️ And yes, it requires a lot of funding, and some companies run out of funding. But that does not mean the whole industry struggles. Instead, ☑️ a value chain has formed around autonomous driving: Machine learning-based perception, teleoperation, labeling, real-time operating systems (e.g. BlackBerry QNX), SDKs (Apex.AI), vehicle operators (MOIA) are available and speed up development and deployment. ☑️ Today, Waymo is leading autonomous driving, currently in San Francisco and Phoenix and expanding down the Silicon Valley and to LA tomorrow. Summary ☑️ The self-driving industry is making constant progress. But it takes more time than the public expected.
Waymo on LinkedIn: Angelenos and Austinites: we’ve got news for you. Beginning tomorrow… | 28 comments
linkedin.com
Well said! I would add that besides aiming directly for Level 5 (like Waymo is doing), there was also enormous progress on getting from Level 0 and up, so we have already Level 3 vehicles offered comercially to regular customers (e.g. Mercedes Benz DRIVE PILOT). Those are practically autonomous in many driving situations. That approach might be significantly harder, but much more inpactful, in my opinion. I personally cannot find much of the advantage of Robotaxi, compared to regular taxi. But perhaps in US it is different perspective.
For what it’s worth: for years now I’ve disagreed with AN more often than not.
That’s a very concise description and summary, Jan 🙏
Such a reasonable and thoughtful account. A breath of fresh air!
Well said Jan!
Strategic Consultant & Innovator | Foresight Technology Expert | Bridging the gap between Enterprise & Government in China
9moI appreciate the detailed historical and technological analyses, but more emphasis needs to be placed on meeting today's demand for commercializing human and technological advancements. As we move towards a future of autonomous driving, it's important to remember that this technology brings about mandatory social change. To trigger this change, the public must feel "psychologically safe" to accept AD applications. Currently, the city's mass deployment of AV as part of an integrated mobility service fails to scale from a business and societal perspective. However, there is a solution. By operating in a "closed system," where the scale is predictable and profitable, and the customer value is high enough to trigger "psychological safety", we can effectively scale this technology and bring about the change we want to see.