Hacker News Reader

Comment by redox99 | Hacker News Reader

This article does not adjust for the fact that Teslas drive many more miles on Autopilot + FSD than any other brand. So it's a pretty worthless comparison.

View on HN

Please stop spamming Tesla support over this thread.

"It’s not just that self-driving cars in general are dangerous for motorcycles, either: this problem is unique to Tesla. Not a single other automobile manufacturer or ADAS self-driving technology provider reported a single motorcycle fatality in the same time frame."

Doesn't matter when mileage isn't what's being compared - it's whether or not others have caused the same problem - PERIOD.

Statistically it does matter. If FSD has a 1-in-100M miles driven chance of killing a motorcyclist, and FSD drives 10M miles a day, within 10 days it's likely to see a trend starting. Whereas XYZ other brand may have a 1-in-30M miles driven chance of killing a motorcyclist, but only drives 1M miles a day, so without adjusting, FSD is more dangerous, but in reality, XYZ is.

The distance driven absolutely has something to do with it. If I made my own self driving car that only drove 100 meters, then there would not be enough data to know anything.

There's definitely an issue with Tesla's approach, though.

That's just probability. In a binomial distribution with 0 < p < 1, the probability of 0 successes is nonzero P(x = 0) > 0

> Doesn't matter when mileage isn't what's being compared

Maybe the mileage on other brands is lower, because Tesla is unique in hyping as "full self-driving" a product that's not ready? It would still be a problem unique to Tesla then.

so if the other cars have 5x less miles on autopilot compared to tesla you expect to see 0 crashes even though the probability of a crash is the same

Why are we moving the goal posts to « other manufacturers drive less miles so statistically it’s still OK to kill people »?

I would like the goal post somewhere around “Is safer in every situation than a skilled and concentrated human driver”

Right, and so by that standard, self-driving companies should have literally zero death attributed to them per year given how few miles they drive compared to human drivers.

That is: until self-driving cars ride much more, obviously.

Because currently 40,000 people die every year in the US from car crashes, so it has been decided that it's ok to kill a certain amount of people with cars for a long time.

Reply on news.ycombinator.com