Hacker News Reader

The bar is low. I don’t want comparison with an alcoholic with multiple DUIs who still drives and crashes.

The benchmark should be the top decile of drivers.

The number of people who think they're top tier drivers never ceases to amaze me.

Yes statistics are full of drivers with more than one accidents at fault. 15% of the drivers are causing more than 70% of the accidents.

I don’t care about the average driver. I care about the median.

It is not a high bar to expect an autonomous system to be better than 90% of the American drivers.

Median and average driver are close to identical. This isn't income.

(Edit: I now think my comment above is wrong)

Driver risk is a Pareto distribution.

I mean, if some subset of those 15% used Waymo, we'd be better off, no?