Chapter
Tesla's Efforts to Reduce Latency in Self-Driving Cars
1:25:56 - 1:36:47 (10:50)
Tesla's software engineers are working on reducing the latency and jitter caused by the stack-up of processing from the camera to the final actuator on a self-driving car. They are improving the efficiency of compute and using trip accelerators to perform matrix math dot products for better performance.
Clips
Elon Musk, CEO of Tesla, speaks on the intricate neural network and C compiler technology that power their self-driving cars.
1:25:56 - 1:27:11 (01:15)
Summary
Elon Musk, CEO of Tesla, speaks on the intricate neural network and C compiler technology that power their self-driving cars. They have their own C compiler and are continuing to optimize it for maximum efficiency.
ChapterTesla's Efforts to Reduce Latency in Self-Driving Cars
Episode#252 – Elon Musk: SpaceX, Mars, Tesla Autopilot, Self-Driving, Robotics, and AI
PodcastLex Fridman Podcast
Tesla's skilled software engineers are working on improving the efficiency of compute and trip accelerators, without any post-processing of the image.
1:27:11 - 1:30:17 (03:06)
Summary
Tesla's skilled software engineers are working on improving the efficiency of compute and trip accelerators, without any post-processing of the image. The goal is to support eight cameras with as little latency as possible.
ChapterTesla's Efforts to Reduce Latency in Self-Driving Cars
Episode#252 – Elon Musk: SpaceX, Mars, Tesla Autopilot, Self-Driving, Robotics, and AI
PodcastLex Fridman Podcast
Latency and jitter are two major concerns in the development of self-driving cars.
1:30:17 - 1:36:47 (06:29)
Summary
Latency and jitter are two major concerns in the development of self-driving cars. Fixed latency can be anticipated and predicted, but variable latency caused by timing tolerances can be problematic and affect the performance and safety of self-driving cars.