FYI: Tesla CEO Elon Musk asserts that the new "Full Self-Driving" Beta version 12.4 is a "5-10x improvement" in miles per intervention compared to its predecessor. However, a recent video shows a Tesla with this software driving the wrong way in traffic, shedding light on ongoing challenges.
The Reality of Tesla’s Full Self-Driving in 2024
Elon Musk’s Grand Declarations and Sobering Realities
Elon Musk has continually hyped Tesla’s Full Self-Driving (FSD) software, recently calling the version 12.4 beta a "5-10x improvement" over previous iterations. Yet, paradoxically, a recently positive video from a self-driving enthusiast channel showcases a Tesla with this beta driving on the wrong side of the road. Despite years of development and numerous updates, FSD remains unofficially self-driving, with liability still falling squarely on the human driver.
FSD in 2024: Unfulfilled Promises
Back in 2016, Musk promised full autonomy was on the horizon, with expectations of 1 million robotaxis on the road by 2020. Flash forward to 2024, and the number of robotaxis remains zero. While there have been countless improvements in FSD’s basic driving abilities, glaring faults persist, including a troubling instance of the car navigating the wrong way in traffic.
Highlights and Limitations in AI-Driven Performance
AI DRIVR’s latest video captures some of the more nuanced capabilities of FSD. For example, it shows a Tesla adeptly navigating around a crashed vehicle by slightly deviating off-road, mimicking a human’s instinct to follow the car ahead. However, such solutions are straightforward; the true test lies in how FSD handles being first in line, a scenario still fraught with uncertainties.
In the same breath, the video criticizes the Tesla’s behavior, noting the potentially hazardous tailgating driven by the system’s assumed anticipation of traffic flow. It also echoes Musk’s comments that overtraining on interventions could explain its overly aggressive maneuvers.
The Echo Chamber of Early Access
One fundamental issue in evaluating FSD lies in the sample of testers themselves. Elon Musk tends to grant early access to die-hard fans who seldom question the software’s shortcomings. This self-selecting group often peppers social media with terms like "ChatGPT moment" and "DriveGPT," without critical examination of advanced technology’s capabilities and limitations.
Genuine Impressions and Expert Skepticism
Personally, seeing FSD evolve into a more sophisticated system than originally anticipated is both impressive and concerning. While it’s commendable that it functions reasonably well with modest hardware and basic sensors, the danger lies in testing this technology under real-world conditions with the public.
The progression of the FSD beta version illustrates a cycle of trial and error. AI DRIVR highlights iterations moving from overly cautious to more aggressive behaviors, only to hear promises of future updates rectifying current issues. Yet, new problems emerge with each release, perpetuating a seemingly endless loop of incremental, unfalsifiable improvements.
The Conundrum of Autonomous Driving
While advancing FSD technology is a monumental engineering feat, the simple fact remains that after eight years and high consumer costs, Tesla’s FSD should not be making fundamental errors like driving the wrong way. According to various experts in the field, the current camera-based system will never achieve full autonomy without more advanced infrastructure and sensor integration.
Conclusion
Tesla’s Full Self-Driving Beta has undoubtedly made strides in autonomous vehicle technology, but it still falls short of Musk’s grand visions. As dedicated drivers and technology enthusiasts, the industry must demand more realistic and safer developments.
For all updates and insights on automotive technology, visit Automotive.fyi. Follow us on X @automotivefyi and share your news tips at tips@automotive.fyi.
William Kouch, Editor of Automotive.fyi
