Smoke And Mirrors? Tesla Claims Accident Risk Drops By 50% Using Autopilot Feature

How would this be smoke and mirrors? I'm surprised the number isn't even higher than a 50% reduction.

This is smoke and mirrors because it is not apparent that a verifiable experiment/study was made that can lead to such a claim. Elon Musk is just shooting from the hip and citing some random airbag deployment statistics. Its called the "association-causation" fallacy; its heresay.

If Autpilot actually DOES lead to safer driving, that’s a good thing because there are lots of accident photos on the internet with Tesla’s wheels falling off.

https://www.flickr.com/photos/136377865@N05/sets/72157658490111523/

In every Tesla accident photo shown in this particular compilation, no matter how minimal or how severe, the Model S' lost at least one wheel - either snapped off at the half-shaft, separated at the hub, or with cracked rims.

Maybe the Tesla Model S, at a curb weight of 5,000 pounds is too heavy for its wheels?

Not so crazy, the first generation (2005-2012) Bentley Continentals Flying Spurs which weighed 5,100 to 5,500 pounds curb weight were commonly know to shatter their rims and/or destroy their tires because the rims/tires were not designed to handle the weight. This issue went away with the "refresh"/new-gen in 2012/2014 when they all got 20" rims as standard and more robust tires.

Maybe this is the next problem that Tesla has to look forward too?

FYI - the curb weight of a Model X is 5,450 pounds. Wonder when those wheel issues will start appearing?

What's different about Tesla? A number of vehicles with auto cruise control braking, lane warning etc. etc. can make that claim.

Driving in Boston the other day and it was clear that there is no way an autonomous vehicle could make it there. Roads that are old cow paths paved over, intersections with six roads coming together, drivers that follow the rule "never make eye contact", yellow light means step on it, etc. etc. An autonomous car's computer would just freeze up.

You would have to drive one, or let Autopilot drive you to really understand what's different. The implementation is much smother, very rarely it gives up and when it does it gives the driver plenty of warning. Also the system constantly improves via over the air updates comprised of crowd-sourced data from the whole fleet.

All that said it really depends on how safe the driver is that is compared to Autopilot and how wisely the use Autopilot as it is not appropriate to use in some circumstances. If Tesla is comparing it to an average driver I'd say it is much safer. It also doesn't get tired and it doesn't get distracted, so for a long trip it really makes a lot of difference.