[Retric]

At ~1.2 deaths per 100 million miles in the US you would expect approximately ~10-20 deaths over that distance. Adjusted for type of car and driving conditions.

Autopilot seems to be around those numbers which is promising assuming continued improvements. Though it's not currently dramatically safer their is reasonable expectation it will improve over time.

[jordache]

you forgot that auto pilot likely is only used or only works under certain conditions. So that 1.2B miles is selective. You can't extrapolate that to the 10-20 deaths, which I assume you calculated from ALL miles driven

[Ienuur4i]

But are the parts that are easy for machines also what is easy for humans? Human attention decreases when things get more simpler and more monotonous while an autopilot can deliver constant attention.

[adrianN]

The problem with systems like autopilot is that the human still needs to pay attention. The driver needs to take over in a matter of seconds when the system encounters a situation in can't handle. Autopilot makes things simpler and more monotonous while still requiring perfect attention from the human.

[Retric]

In theory that's a huge problem. In practice we have 1.X billion miles of data and it does not seem to be such a huge deal. Either the systems are already fairly good, or people mostly pay attention.

Granted, I am approaching this from the perspective of a more relaxing driving experience not necessarily from a pure safety standpoint. People spend 20,000+ hours driving in a lifetime making that less stressful is a huge benefit even if they are still stuck in their cars and can't get work done.

[simion314]

We need better number to do this comparisons, similar car models, similar roads conditions, similar driver population, since Tesla keeps the data hidden from the public I incline to believe the data is not in Tesla favor.

[bvm]

Would highway/motorway/autobahn miles be an acceptable substitution?

[Retric]

I specifically mentioned but did not do that adjustment because I don't have good data.

[coldtea]

>At ~1.2 deaths per 100 million miles in the US you would expect approximately ~10-20 deaths over that distance. Adjusted for type of car and driving conditions.

That will be a huge adjustment that they seem to not be doing.

So any numbers of DUI deaths riding an 1990s semi-broken car is compared to an autopilot death in an expensive state of the art car with great engine, speed, and breaks.

How about we get the deaths per miles for non-coke users driving Tesla-like sports cars?

[Retric]

I am not sure how 'huge' that would be. But, be careful when trying to get more accurate estimates as the tendency is to look for things that move Numbers into the directions you think they should go not just to increase the accuracy of each estimate.

[coldtea]

>as the tendency is to look for things that move Numbers into the directions you think they should go

Even if misleading in the too cautious direction, that would be a better step towards safety than the numbers the car makers give.

Besides the same argument could be made for too much drug testing, against e.g. a snake-oil promoter.

[mannykannot]

Tesla insists that drivers pay attention at all times, to catch cases where the system fails. Without numbers on how often that happens, these figures give no quantitative evidence for its effectiveness.

[evfanknitram]

What does "type of car" mean here? Is it brand+price or something else?

[druidcz]

I suppose it should mean something like "full-size luxury AWD limousine", as death rates vary with car size, type, price and driven wheels [1]

[1] http://www.iihs.org/externaldata/srdata/docs/sr5203.pdf