Summary of Chris Urmson: Self-Driving Cars at Aurora, Google, CMU, and DARPA | Lex Fridman Podcast #28

00:00:00 - 00:40:00

Chris Urmson discusses the challenges and opportunities of developing self-driving cars. He notes that costs must be reduced to make self-driving cars economically viable, and that human factors such as vigilance and complacency must be taken into account. He also discusses the importance of active safety systems.

• 00:00:00 Chris Urmson was a CTO of the Google self-driving car team and a key engineer and leader behind the Carnegie Mellon University autonomous vehicle entries in the DARPA Grand Challenges and the winner of the DARPA urban challenge. He's the CEO of Aurora innovation, a autonomous vehicle software company, and the former lead autonomy and perception lead at Uber. He is one of the top roboticists and autonomous vehicle experts in the world and a longtime voice of reason in a space that is shrouded in both mystery and hype. He both acknowledges the incredible challenges involved in solving the problem of autonomous driving and is working hard to solve it.
• 00:05:00 Chris Urmson discusses the history and development of autonomous vehicle systems, highlighting key technological innovations. He talks about the Grand Challenges of autonomous vehicle technology, and how advances in mapping and Bayesian estimation have helped to make these challenges more manageable. He discusses the current state of autonomous vehicle technology, and how it is still evolving.
• 00:10:00 Chris Urmson, the creator of the self-driving car project at Google, CMU, and DARPA, discusses how lidar technology is essential for autonomous vehicles, but cameras and radars are also important. He also discusses how the urban challenge helped develop perception tasks for autonomous vehicles.
• 00:15:00 Chris Urmson, the head of Google's self-driving car project, discusses the challenges and opportunities of developing self-driving cars. He notes that costs must be reduced to make self-driving cars economically viable, and that human factors such as vigilance and complacency must be taken into account. He also discusses the importance of active safety systems.
• 00:20:00 Chris Urmson discusses the potential benefits of self-driving cars, and the need for careful communication of the technology's capabilities. He believes that the technology will diverge from the path needed to achieve truly self-driving cars, and that the economics of gradual increasing from level 2 will not be cost-effective in the long term.
• 00:25:00 Chris Urmson discusses the challenges of demonstrating safety for self-driving cars, and explains how Aurora's safety process works. He also mentions a metric he doesn't like, number of disengagements.
• 00:30:00 Chris Urmson discusses the various efforts being made to create self-driving cars, including at Aurora, Google, CMU, and DARPA. He argues that there will be individual metrics to measure safety outside of fatalities and that ultimately the goal is to make roads safer while reducing the cost of mobility.
• 00:35:00 Chris Urmson, head of Google's Self-Driving Car project, discusses the challenges of developing self-driving cars and how they will become more advanced over time. He also discusses the importance of having a good perception forecasting capability to anticipate future traffic patterns.
• 00:40:00 Chris Urmson, the CEO of self-driving car company Arora, discusses the company's competition and how they plan to outcompete them in the long run.