I didn't make the claim but I thought to look it up. From NHTSA's report, if I'm reading it correctly, 2021 had 42,915 traffic fatalities, of which 8,174 were related to alcohol. I think those numbers are estimated. I'm not aware if there's better data.
My information appears to be a few years out of date - probably from the last time I had this debate here on HN. 2021 does appear to have been a particularly bad year...
Regardless, the numbers are very good for the US, compared to the rest of the world. iihs.org[1] indicates the US is averaging somewhere around 1.5 fatalities per 100 Million miles traveled.
Compared to the rest of the world... that's pretty decent. Most of the places with significantly high fatality rates are not the places that will be purchasing self driving electric vehicles anytime soon...
You are deluding yourself if you think the US is doing well on traffic fatalities. From the graph at https://www.nytimes.com/2022/11/27/upshot/road-deaths-pedest... we can see that US car deaths/capita have dropped slightly (10-20%?) since 1995. Meanwhile, in France where the rate about matched in 1995 they have dropped by 3x, leaving the US rate at about 2.5x the French rate.
If every driver were drunk at all times, roads would be designed with that in mind. It's much harder to dangerous things in a car with better designed roads
The point is that the design of the vehicle, or anything, needs to be as good as possible regardless of anything else, and there is no overriding excuse not to.
It's not like the poorly designed controls are necessary for some other even more important reason.
If both roads and drivers were all excellent, the vehicle still needs to be as excellent as possible. If all other factors are so good that only 10 people die each year from car accidents, there is still no excuse for making vehicles less safe and make that become 20.
Alcohol, lack of alcohol, good roads, bad roads, speed limits, none of that matters.
