Barklikeadog
Active Member
What would that ratio be if the 'autopilot' driver never interfered?It's so, so, so much worse. Talking about "% of the time" is the worst possible metric because it betrays the implicit logarithmic nature of the statement.
Every 1 : 1,000 minutes the system fails: "This works 99.9% of the time."
Every 1: 100,000,000 minutes the system fails: "The system works 99.999999% of the time."
A <0.1% improvement means a 1,000x or 100,000% improvement.
Tesla says their drivers without autopilot are being involved in an accident every 2 million miles. If we assume half of the time they're at fault. That means Tesla humans make a fault every 4 million miles.
If we thought that FSD Beta was 10x less safe than a person that's the difference between FSD being 99.999975% reliable (Human) and 99.99975% reliable.
Likely FSD City Streets is probably at least 10,000x less safe than a human. Even if Tesla the distance between predicted accidents by 50% every 6 months...
log 10,000 / log 150% = 11 years.
(1.5^22.7 = 10,000)
Personally, I won't take the refund because I can't wait to see how Tesla plans to upgrade my car to stay up to date over the next 2-20 years for every necessary update.
Safety warning systems on autos are proven to reduce serious accidents. People willing to pay for more safety systems tend to already be better, more defensive drivers.
I've always looked at it from the perspective that autopilot/fsd has a human backup safety system. That human is on avg a 30-60 yr old six figure income earner which is far and away the safest driving demographic in the road.