Summary
- William Stein, a technology analyst at Truist Securities, tested Tesla’s Full Self Driving system multiple times and found unsafe maneuvers
- Despite Musk’s claims, Stein doubts Tesla is close to deploying a fleet of autonomous robotaxis
- The latest version of Full Self Driving made illegal and unsafe maneuvers during Stein’s test drive
- Experts warn that Tesla’s system lacks common sense and may not be ready for full autonomy
- NHTSA is evaluating information about a fatal crash involving Tesla’s Full Self Driving system
Article
The Associated Press reported that a technology analyst at Truist Securities, William Stein, tested Tesla’s Full Self Driving system three times in the past four months. Despite Tesla’s claims that the system can operate with little human intervention, Stein reported that the vehicle made unsafe or illegal maneuvers during each test drive. He expressed doubts about Tesla’s readiness to deploy autonomous robotaxis, especially after his most recent test drive left his son terrified. Stein’s experiences raise concerns about Tesla’s goal to present a working prototype of a robotaxi in October.
During his latest test drive, Stein used a rear-wheel-drive Tesla Model 3 equipped with the latest Full Self Driving software. Despite feeling smooth and more human-like compared to past versions, the Tesla made concerning maneuvers such as turning left from a through lane while running a red light. Stein did not intervene as there was minimal traffic, but he had to take control when the car straddled two lanes on a parkway. He concluded that the latest version of Full Self Driving does not fulfill Musk’s predictions of autonomous capabilities or approaching robotaxi capabilities. Stein anticipates significant delays or limitations in the deployment of a fully autonomous robotaxi.
Other experts, such as Alain Kornhauser from Princeton University, have also expressed concerns about Full Self Driving’s readiness for unsupervised operation in all locations. While the system performs well in most situations, there are instances where human intervention is necessary due to unsafe moves made by the Tesla vehicles. Kornhauser suggests that the system may work autonomously in smaller city areas with detailed maps to guide the vehicles, but its current state is not suitable for widespread deployment. The limitations of Tesla’s system highlight the challenges of achieving full autonomy without human supervision in various driving scenarios.
Experts warn that Tesla’s reliance on cameras for its autonomous system may not always be sufficient for real-world situations, especially in adverse weather conditions or darkness. Comparatively, other autonomous robotaxi companies utilize a combination of cameras, radar, and laser sensors for enhanced perception and object detection. The lack of common sense in machine learning raises concerns about the ability of autonomous vehicles solely based on artificial intelligence to handle unforeseen situations, posing a potential risk to road safety. The recent fatal crash involving a Tesla using Full Self Driving underscores the importance of evaluating the system’s safety and effectiveness in real-world conditions.
With over half a million Tesla vehicles equipped with Full Self Driving on US roads, concerns about the system’s reliability continue to be raised. While Tesla enthusiasts showcase their vehicles driving autonomously in controlled environments, skeptics point out the dangers of relying on a system that may not be fully equipped to handle complex driving scenarios. The debate around the feasibility of self-driving cars without additional sensor technologies like lidar and radar persists, with Elon Musk’s persisting vision conflicting with the cautious approach adopted by other autonomous vehicle developers. As regulators evaluate the safety implications of Full Self Driving technology, the general public remains wary about the risks posed by experimental autonomous vehicles on public roads.
Read the full article here
