Tesla tests camera-only Cybercab robotaxis in Austin
Tesla has begun public rides for its Cybercab robotaxi service in Austin, Texas, nearly two years after Elon Musk first unveiled the steering-wheel-free vehicle. The launch is a major test of Musk's long-held, contrarian bet that self-driving cars can rely solely on cameras and AI, rejecting the lidar and radar sensors used by rivals such as Waymo, with Tesla now bearing full legal responsibility for any incidents since there is no human driver to fall back on.
The Cybercab runs a modified version of Tesla's Full Self-Driving software, which remains a supervised, Level 2 system in ordinary Tesla vehicles and has been linked to dozens of deaths and thousands of crashes; in the robotaxi it will instead be overseen by remote operators. Musk dismissed lidar as "a fool's errand" at Tesla's 2019 Autonomy Day and removed radar from Tesla's sensor suite in 2022 despite internal engineering concerns, leaving the cars reliant on two-dimensional camera images that critics say are more vulnerable to glare, rain, snow and fog than multi-sensor systems.
- Tesla's driverless Cybercab robotaxi begins public rides in Austin, Texas.
- It uses only cameras, rejecting lidar/radar sensors rivals rely on.
- Tesla now bears full legal liability, with no human driver fallback.
New here? Start with this
Self-driving cars use cameras and software to detect roads, people and other vehicles, then steer and brake without a person controlling every move. Companies take different approaches: Tesla relies on cameras and artificial intelligence, while rivals such as Waymo also use sensors including lidar, which measures distance with laser pulses.
Tesla is led by Elon Musk, who has argued that cameras can provide enough information for automated driving. Its Full Self-Driving software in ordinary cars still requires a human driver to supervise it; a robotaxi has no driver in the vehicle, so remote operators may monitor or assist it.
The approach matters because automated vehicles must work safely in changing conditions such as darkness and bad weather. If a vehicle with no human driver is involved in an incident, questions about safety and legal responsibility become especially important.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
A camera-only system could be cheaper and simpler to build at scale, while relying on AI to interpret the same visual cues that human drivers use. If public rides show it can operate safely and consistently, Tesla could expand access to robotaxis without the cost and complexity of lidar and radar equipment.
The case against
Cameras can be impaired by glare and poor weather, and unlike lidar and radar they provide no independent sensing methods to cross-check what the system sees. Given the serious consequences of errors in a driverless vehicle, critics can reasonably argue that relying on cameras alone is an avoidable safety risk, especially when Tesla’s underlying driving software has so far been used with human supervision.
Art Business Companies Culture Cybersecurity Technology
Read the full article at the source →
Originally published by The Verge as “Elon Musk’s heterodox robotaxi philosophy gets put to the test”.