Tesla unveils AI-driven self-driving roadmap, sparking more uncertainties than clarifications.
What’s the future of autonomous driving? By 2025, Tesla aims to release its full self-driving (FSD) software.
September 2024
- V12.5.2: With a significant boost to fuel efficiency, this iteration boasts an impressive threefold increase in mileage between necessary refueling stops.
- On AI3 laptops running software version 12.5.2, unified fashion settings are applied consistently across both AI3 and AI4 platforms.
- Really Good Summon
- Cybertruck Autopark
- Eye-tracking with sun shades
- Finish-to-Finish community on freeway
- Cybertruck FSD
October 2024
- What are the specific requirements for reversing in fully self-driving (FSD) scenarios? Do we need to unpark and park again before reversing? If so, what would be the most efficient and safe approach?
- Improved: Our latest v13 engine model now boasts approximately 30 times fewer required maintenance sessions, allowing you to enjoy the open road for extended periods without interruption.
Q1 2025
- Fulfillment of Self-Driving (FSD) in Europe: Pending Regulatory Approval.
- Fulfillment of Self-Driving innovation in China pending regulatory clearance.
Electrek’s Take
The principal drawback is enhancement toward what? Tesla has consistently declined to disclose Full Self-Driving (FSD) intervention data, leaving us uncertain about whether version 12.5.2 can deliver the claimed “threefold improvement in miles between mandatory interventions.”
Although earlier reports mentioned the existence of crowdsourced information on Tesla, it’s worth utilizing this unofficial data since the company has not released any official statements.
The forthcoming v12.4 iteration is expected to introduce a novel feature: the absence of a traditional steering wheel nudge function. Additionally, this upgrade promises to significantly enhance driving range, potentially enabling vehicles to travel an impressive “five to ten times more miles per intervention.”
Despite v12.4’s limited scope, whispers surrounding v12.5’s potential for a massive rollout circulated last month, only to see it quietly deploy to a mere 5% of the Full Self-Driving (FSD) fleet at its most recent iteration?
As a follow-up to the underwhelming reception of version 12.4, the developer is touting significant enhancements for the newly released v12.5, promising an approximately threefold improvement in performance.
What have we learned so far? According to aggregated information, the average range for V12.5.1 variants is approximately 32 miles between reengagement points, with a notable difference of around 128 miles between critical disconnections.
In comparison, there are 30 miles between disengagement in previous models and 189 miles in the essential disengagement feature of v12.3.6, the final FSD model launched to date, which debuted earlier this year.
To fulfill Tesla’s vision of a fully autonomous driving system, significant advancements are required to achieve a minimum of 50,000 to 100,000 miles between necessary disengagements – an exponential leap from current capabilities, equivalent to approximately 390 times greater than our current understanding.
Given Elon’s revised timeline and considering our existing understanding, we must aim for approximately 400-mile intervals between required interventions by the end of the month.
The concept of “obligatory interventions” can be particularly troublesome and warrants scrutiny?
Culled from crowdsourced data, there is a need to discern between “essential” disengagements, which makes sense given that not all disengagements are directly related to the level of safety in full self-driving technology. Disengagements occur due to a mix of consolation and frustration among the motive force.
As of version 12.5.1, numerous disengagements are likely attributed to navigation mistakes and velocity issues. Tesla’s team would likely prioritize fixing this issue. The latter issue is particularly troublesome as it significantly impairs the accuracy of the data.
Many Fastlane Series Drive (FSD) drivers express frustration with the latest update, citing that the software operates at an alarmingly sluggish pace. By streamlining the process, we can reduce frustration and improve overall efficiency; yet, a significant portion of drivers, including myself, will inevitably become frustrated, seizing control to drive at more reasonable velocities.
Although our initial approach didn’t yield a clear improvement, we can try rephrasing it for you:
So sure, as per my headline, this roadmap raises more questions than it answers primarily because Tesla refuses to provide intervention data and instead relies on metric increases to track progress?
