The Safety Threshold: Tesla Promises Human-Surpassing Reliability in FSD v15
Elon Musk claims Tesla's upcoming FSD v15 will 'far exceed' human safety levels, though skepticism remains as previous versions v12 and v14 faced similar promises while still requiring active supervision.
Tesla is once again making bold claims regarding its Full Self-Driving (FSD) software. According to CEO Elon Musk, the upcoming version 15 is expected to 'far exceed' human levels of safety. This announcement comes as Tesla continues to iterate on its end-to-end neural network-based approach, which moved from modular code to a more unified 'vision-to-action' deep learning model in version 12.
While the hardware in Tesla's fleet remains largely unchanged, the software refinement aims to reduce 'disengagements'—the moments where a human driver must intervene. Musk’s latest projection targets a level of reliability that would theoretically allow for unsupervised autonomy. However, the industry remains cautious, noting that similar milestones were predicted for versions 12 and 14, which, while improved, still require the driver to remain fully attentive and ready to take control.
The safety of ADAS (Advanced Driver Assistance Systems) has become a focal point for regulators and consumers alike. As Tesla pushes the limits of vision-only systems, the goal for v15 is to demonstrate superior handling of complex urban scenarios and unpredictable human behavior. Whether this version finally bridges the gap between 'assistance' and true 'autonomy' remains to be seen, but the pressure to deliver a demonstrably safer-than-human system has never been higher.
Source: Electrek
