Musk's arrogant vision has invested nearly 100 billion yuan in autonomous driving, and he resolutely does not use radar, far exceeding Huawei's!
Musk has been a staunch supporter of visual self-driving technology, arguing that autonomous driving can be achieved by relying only on cameras and artificial intelligence algorithms, without the need for other sensors such as lidar. This concept has been questioned because of its obvious divergence from traditional autonomous driving solutions, but Musk has not wavered and has further increased his investment in this area.
Tesla's investment in visual self-driving technology is huge. According to Musk, Tesla's cumulative investment in the field of autonomous driving will exceed $10 billion in 2024, far exceeding the investment level of rivals such as Huawei. This investment is mainly used for computing hardware upgrades, software development, testing and verification, and talent training. Musk firmly believes that as long as there is enough investment, relying on cameras and artificial intelligence will definitely be able to achieve autonomous driving, without the assistance of other sensors such as lidar.
Musk's confidence in visual self-driving technology comes from Tesla's phased progress in this field. At present, Tesla's FSD (Full Self-Driving System) has entered the public testing stage, and although there are still shortcomings, the overall performance has gained a certain degree of recognition. Musk pointed out in an internal email that when delivering new cars, it is necessary to bring customers to experience the FSD function to ensure that customers deeply appreciate the benefits of "full self-driving". He said that almost no one was aware of the excellent performance of FSD, even though it would slow down the delivery process. It can be seen that Musk is confident in the FSD system.
Musk also admitted that it will still take a long time to achieve "full self-driving" in the true sense. He expects it will take at least 10 years for most cars on the road to become autonomous. This is mainly due to the complexity of autonomous driving technology, which requires continuous algorithm optimization, hardware upgrades, and a large number of real-world scenarios for testing and verification. Musk hinted that Tesla's advanced driver assistance system will usher in faster improvements as the bottleneck of its AI training computing power has been broken. But he also noted that most people are ignorant of the potential of Tesla's AI.
In the face of doubts, Musk has repeatedly publicly criticized the use of lidar, saying that "those who rely on lidar are doomed to failure". He believes that although lidar can provide accurate range information, it has blind spots and has limited ability to capture the types and semantic information of recognized objects. The camera can directly perceive the shape, color, motion and other rich information of the object like the human eye, and combined with powerful artificial intelligence algorithms, it can completely replace lidar to achieve autonomous driving.
To prove his philosophy, Musk has gone to great lengths to promote Tesla's visual self-driving technology. He tweeted that Tesla is happy to license Autopilot, FSD and other self-driving technologies to other manufacturers. He emphasized Tesla's eagerness to help other car companies as much as possible to accelerate the development and adoption of visual self-driving technology. This approach will undoubtedly further increase Tesla's influence in this field.
Musk's unwavering support for vision-driven self-driving technology is not unfounded. Tesla's huge investment in this field, as well as phased technological progress, have provided strong support for Musk's concept. Although it is difficult to achieve "full self-driving" in the short term, Tesla is trying to break through the bottleneck and strive to make visual self-driving technology widely recognized and applied as soon as possible. Musk's bold investment and concept dissemination will surely inject new impetus into the development of visual autonomous driving technology.