Self Driving Cars: Waymo, Tesla

The meeting focused on a presentation and discussion about autonomous vehicles, specifically comparing Waymo's technology to Tesla's Full Self-Driving system. Seth presented a tech talk by Waymo co-CEO Dimitri Dolgov about their AI-powered autonomous driving technology, including their use of LiDAR, cameras, and radars, as well as their safety record and expansion to new markets. The presentation was followed by a detailed demonstration of Tesla's Full Self-Driving system in Boston, where the driver showed how the system uses cameras and machine learning to navigate complex urban environments without traditional coding for specific scenarios. The discussion included technical details about power consumption, weather adaptation, and cultural differences in driving patterns, with participants sharing personal experiences and questions about the technology's current capabilities and future implications.

Seth presented a video discussion on self-driving cars, featuring Waymo co-CEO Dimitri Dolgov. Dimitri explained Waymo's progress in autonomous driving, highlighting their achievement of 10 million fully autonomous paid trips and a safety record showing the Waymo driver is about 5 times safer than humans for serious collisions. Dimitri discussed Waymo's use of sensors, AI, and simulation to develop their technology, and emphasized the potential societal impacts of autonomous vehicles, including increased safety, time savings, and improved accessibility for people with disabilities.

The group discussed autonomous vehicles, focusing on Waymo and Tesla's full self-driving systems. They explored the power consumption of Waymo's LiDAR system, estimating it uses about 1 kilowatt hour per hour. Seth presented a recording of a driver using Tesla's full self-driving capabilities (version 12.5.2) to navigate through Boston, highlighting how the system uses cameras to create a map of the environment and make driving decisions. The discussion included details about Tesla's approach to autonomous driving, which involves training AI systems using vast amounts of video data from Tesla vehicles rather than relying heavily on hand-coded rules.

The video showed Tesla vehicle's self-driving capabilities by sharing screen audio of a test drive through Boston, showing how the car navigated complex urban scenarios including stop signs, crosswalks, and traffic lights. The driver explained that Tesla's approach involves training models using vast amounts of video data from actual human driving rather than hand-coding specific scenarios, allowing the system to learn and adapt to various edge cases. He noted that Tesla's advantage in having millions of vehicles with cameras collecting training data gives them a significant competitive edge over other automakers and robotaxi companies like Waymo.

The group compared Tesla's system favorably to Ford's Blue Cruise and GM's Super Cruise, noting that Tesla's technology can operate in more complex urban environments while the competitors are limited to highways. The group discussed technical aspects of how Tesla's AI system works, with Seth explaining that it uses pre-downloaded software models rather than real-time analysis, similar to how ChatGPT works locally on a computer. Carl shared his positive experience using Tesla's self-driving feature on a highway, while the group explored questions about cultural adaptation of the software and handling of challenging driving conditions like ice on the road.

The group discussed autonomous vehicle technology and intelligent highway systems. They explored how AI and LiDAR could improve road safety, particularly in challenging weather conditions and for detecting motion. The conversation also touched on pedestrian safety challenges, including the difficulty of seeing people wearing dark clothing at night.