Chapter
Challenges of Scaling Driverless Technology
2:13:47 - 2:22:17 (08:30)
Iteration pace slows down with driverless technology due to fear of regression which requires more rigor. As companies work on multiple products, there are technical and organizational challenges to solving increasingly broad and complex problems, including scaling driverless technology.
Clips
The current hardware and software stack used for driverless vehicles is not well-suited for exponential scaling.
2:13:47 - 2:17:36 (03:48)
Summary
The current hardware and software stack used for driverless vehicles is not well-suited for exponential scaling. Real-world data collection is a valuable asset, but filtering through it at a large scale presents a challenge.
ChapterChallenges of Scaling Driverless Technology
Episode#241 – Boris Sofman: Waymo, Cozmo, Self-Driving Cars, and the Future of Robotics
PodcastLex Fridman Podcast
The iteration pace of driverless technology slows down as the fear of regression forces more rigor, making the already insurmountably hard problem even harder when building a robot-based product that people are going to use.
2:17:36 - 2:22:17 (04:41)
Summary
The iteration pace of driverless technology slows down as the fear of regression forces more rigor, making the already insurmountably hard problem even harder when building a robot-based product that people are going to use. The safety bar gets even higher for full driverless, making it harder to build a robot that successfully perceives and operates in the world.