In late October, Tesla released a software update for a restricted group of Tesla owners in the US enabling the Full Self Driving mode, which allows any Tesla to drive autonomously in all situations on the road and in cities. Named FSD Beta, the update surprised Tesla owners and industry experts for the quantum leap in autonomy. Check it out in the Youtube videos below.
Tesla is the only company in the world to release the technology to final customers. Unlike competitors, which use LIDAR to detect obstacles, vehicles, and pedestrians, Tesla's approach works only with cameras and without the use of high-resolution maps and geofences. It can theoretically work in any situation.
Tesla's secret lies in the software running inside the car which in turn is trained at the company's data centers using data from all their fleet. It uses computer vision and extremely sophisticated neural networks to infer the world the eight cameras see. To understand how the tech works, watch the videos below (a reasonable background in AI is recommended).
Besides the AI software, Tesla has developed proprietary chipsets used in their cars which are many times more powerful than chipsets from Nvidia or Intel for this specific task. However, Tesla's biggest differentiator in the self-driving race is the Dojo inference supercomputer, which will go operational in Q2 of 2021. More information about Dojo in the video below.
Why should I care about this?
Current high school students and undergrads will be introduced to a world where autonomous cars will disrupt entire industries. If you thought about working in the auto industry, please reframe your choices accordingly.
As the majority of cars manufactured by the end of this decade are going to be electric and autonomous, think about how your area of study and expertise will be impacted. For instance, engineers focused on internal combustion engines will likely be out of jobs in the 2030s.
The advent of a world with autonomous cars will disrupt adjacent areas such as urban planning (parking lots will migrate to the suburbs), business models (cars might be summoned like Ubers instead of purchased), computer science (focus on computer vision and prediction models), and entrepreneurship (imagine how many companies will be created to use the autonomous cars’ platform).
Also, the computer vision and inference tech being used at Tesla for their autonomous cars will facilitate enormous advances in other industries such as aerospace, agriculture, and biotech. So, if you thought about studying computer science and artificial intelligence, don't look back! The next decades will be driven by AI as much as the last hundred years were driven by electricity.
To get more articles about tech and entrepreneurship from Innovalab Weekly, please subscribe to our free newsletter.