Waymo drove a total of 3M miles in 2022, for reference (driverless and non-driverless.) For fatalities, there's a national average of about 1.3 deaths per 100 m

Waymo drove a total of 3M miles in 2022, for reference (driverless and non-driverless.) For fatalities, there's a national average of about 1.3 deaths per 100 million miles driven, so you wouldn't expect a Waymo fatality for a while yet if it was a standard human driver. Obviously this kind of thing lives and dies by the edge cases and in my opinion we're going to need a massive expansion in the fleets to really make them safe - by definition you will encounter very few edge cases if you do far less driving. Total miles driven in the US is over 3 trillion per year, so Waymo drives less than 1 millionth of the total miles in the US. For an intuitive way to visualize this, you can think of Waymo as a fleet of 30 truck drivers, each driving 100k miles per year (which is a typical number for truck drivers.) Thirty truck drivers having good insurance rates is a great sign, but obviously if you hire a 31st guy and he causes a ten-car pileup then your statistics will be in the toilet. Thirty truck drivers is a tiny company. The Stevie Wonder Institute For Trucking has some 15-20k drivers and recently went nine days without an accident [0]. That's not meant to belittle Waymo, of course, their achievements are very impressive and they are clearly leading the pack. One argument you could make in favor of Waymo is that highway driving is simpler, less risky statistically, and makes up most US miles driven - and they do very little of that. I'm looking forward to an expansion there, that'll really get their mileage up. [0] https://www.youtube.com/watch?v=dvkrnWS1ea8 https://www.youtube.com/watch?v=dvkrnWS1ea8 [1] [1] yes I did just link to "Mutha Trucka News" on HN.