Musk Called LiDAR Users Fools in 2019
In 2019, Musk stated that those using LiDAR are fools.
He argued that high-priced sensors are expensive and unnecessary, noting that he understands the technology and previously led the development of navigation and docking for SpaceX's Dragon spacecraft, but once visual solutions for cars are resolved, LiDAR becomes useless.
The event-driven pure vision approach has historical validation, with funding flowing towards camera and neural network solutions, while LiDAR faces pressure from automakers.
Source: Public Information
ABAB AI Insight
Musk publicly referred to LiDAR as a "fool's errand" during Tesla's Autonomy Day, emphasizing that human driving relies on vision and neural networks. He had previously applied this technology in SpaceX scenarios but clearly distinguished the needs of automobiles.
Capital trends show Tesla's long-term rejection of LiDAR, focusing resources on visual training data and Dojo, motivated by the goal of achieving low-cost, scalable autonomous driving, specifically through end-to-end learning instead of redundant multi-sensor setups.
Similar to the evolution of biological vision, other automakers are shifting from LiDAR to fusion or abandoning it altogether, with current autonomous driving perception moving from hardware stacking to software-defined convergence.
This is essentially a technological substitution, with the mechanism being that vision combined with neural networks can cover driving scenarios, making expensive sensors a historical burden rather than a necessity.
ABAB News · Cognitive Laws
Scene determines sensor value
Hardware becomes redundant after visual solutions
Early warnings ultimately become reality