The future failure of Tesla's Self-Driving Car, according to Waymo.
The debate about the most suitable technology for developing a fully autonomous vehicle remains highly relevant. While Tesla continues to defend a strategy based exclusively on cameras and artificial intelligence, Waymo maintains a radically different stance. For Waymo, foregoing sensors like LiDAR means sacrificing a significant safety margin.
That is precisely the message conveyed by Dmitri Dolgov, Waymo's co-CEO, in recent statements. In an interview, he explained why the company continues to invest in a multi-sensor architecture and why he believes Tesla's chosen approach has significant limitations when the goal is to achieve fully autonomous driving.
Waymo believes Tesla's system has a limit that is difficult to overcome.
Dolgov admits that a system based solely on cameras has the potential to achieve great things. In fact, he acknowledges that "if the goal were to approximate the performance of a human driver, using only cameras is a perfectly reasonable option." However, he believes that shouldn't be the industry's true objective.
The head of Waymo maintains that an autonomous vehicle must undoubtedly be safer than a human driver. "If you're looking for driving performance that is significantly superior to that of a human, a system with limited sensors will reach a ceiling far too soon," he explains, arguing that cameras alone eventually face situations where their perceptual capacity proves insufficient.
To illustrate this difference, Dolgov uses a graph comparing the evolution of both philosophies. According to his argument, a system based solely on computer vision can continue to improve thanks to software development. However, there comes a point where this evolution stagnates. On the other hand, the integration of cameras, radar, and LiDAR continues to demonstrate its ability to improve safety by providing more information about the surroundings.
A philosophy completely opposed to that of Elon Musk.
The executive's statements highlight the significant discrepancies between Waymo and Tesla. For years, Elon Musk has defended his position that humans drive using only vision, so, in his opinion, an autonomous vehicle should be able to perform the same function using cameras and artificial intelligence. This philosophy has been fundamental to Tesla's success in first eliminating radar and then ultrasonic sensors from its vehicles, relying entirely on camera-based environmental perception.
Meanwhile, Waymo continues developing its robotaxis by integrating a combination of high-resolution cameras, radar, and LiDAR sensors. The company estimates that this redundancy allows for more accurate object detection, distance measurement, and continued operation with a high level of safety even when one of the sensors loses effectiveness due to adverse conditions such as heavy rain, fog, or glare. Electrek has pointed out that these statements come at a particularly relevant time for the sector. Tesla continues to expand the rollout of its Robotaxi service in the United States, while Waymo operates commercially in several cities with fully autonomous vehicles. While both companies share the same goal, it's clear they're pursuing entirely different technological paths.
At this point, Waymo doesn't question the rapid evolution of artificial intelligence. However, the question arises as to whether using cameras alone will be sufficient to achieve a significantly higher level of safety than a human driver. This is precisely the main point of contention between Waymo and Tesla, and the reason why Waymo continues to consider LiDAR and other sensors essential for the future of autonomous driving.
