Tesla Releases FSD Crash Data That Appears More Honest
8 days ago
- #Autonomous Driving
- #Tesla
- #Safety Data
- Tesla released new crash data for their Supervised FSD system, showing improved safety metrics compared to older Autopilot data.
- The new data indicates FSD-equipped Teslas are 1.5x safer on city streets globally and 4x safer than older Teslas without ADAS.
- Tesla's previous Autopilot data was misleading, comparing freeway crashes (which are inherently safer) to general population crash data.
- The updated methodology now breaks down data by road type (highway vs. non-highway) and region (North America vs. global).
- Tesla differentiates between 'major collisions' (airbag deployment) and 'minor collisions,' though comparisons with non-Tesla data remain imperfect.
- Supervised FSD can improve safety by providing 'two sets of eyes' (human + system), but complacency among drivers remains a risk.
- Ethical questions arise about restricting FSD access to prevent misuse by negligent drivers, even if it benefits diligent users.
- Tesla's data suggests FSD improves safety on average, but skepticism is warranted due to past misleading reporting practices.
- Elon Musk aims for unsupervised FSD by year-end, but current data shows FSD+human only modestly outperforms humans alone on city streets.
- Historical robocar data shows even imperfect systems with human safety drivers can be safe, while good systems prevent not-at-fault crashes.