I am suspicious.
The first ever fatal crash involving a fully driverless vehicle occurred in San Francisco on January 19.
Elaine Herzberg was killed by a self-driving Uber with a safety driver in 2018. She was crossing in an unsafe location, but both the software and human had time to react but they both fucked up and did not, and it hit her at 45 mph and killed her.
I suspect that “fully driverless” is added for no other reason than to avoid having to talk about this.
Most Waymo crashes involve a Waymo vehicle scrupulously following the rules while a human driver flouts them: speeding, running red lights, careening out of their lanes, and so forth.
This is impossible to verify, because Waymo successfully sued the California DMV in 2022, arguing that detailed safety records are a trade secret and they shouldn’t have to release them to the public. Presumably, they are so secret, and their competitors would get a competitive edge so strong were they to be aware of Waymo’s safety record, that they don’t want to release them even though all they would do is confirm how safe Waymo cars are.
Also, framing this in this loaded fashion, painting a picture of human drivers “careening” around while the Waymo sits peacefully stopped at its red light, is a textbook propaganda framing designed to paint a vivid and lasting picture in the reader’s mind. Since Waymo has forbidden us to look at the data, such pictures are all we have, so I guess we might as well make vivid ones.
I am suspicious.
In fairness to Waymo, Swiss Re (who are unlikely to be easily fooled) also back up these claims.
https://futuretransport-news.com/waymo-and-swiss-re-analyse-safety-benefits-of-autonomous-vehicles/
That doesn’t look like Swiss Re compared anything. That looks like Waymo compared their own data (which they’re refusing to show us) against Swiss Re’s data. They also didn’t correct for the data being across very different driving conditions (surface streets versus freeway), but regardless of that, the biggest takeaway to me is:
Reduction in Claims: Waymo vehicles demonstrated an 88% reduction in property damage claims and a 92% reduction in bodily injury claims compared to human drivers.
That’s incredible! As in, I don’t think it’s credible. Is there any insurance company that’s willing to insure Waymo’s cars for 20% of the cost of an individual human policy? If so, they should be able to just pocket the other 8%/12% and make out like gangbusters. Want me to call up Swiss Re and see if they’re willing to do it?
Yes, their interpretation of Swiss Re is open to debate. That said, I suspect self-driving cars are already safer per kilometer than human driven cars.
Also, they’ll only keep getting better, while human abilities plateau.
“Waymo cars are safer!”
We don’t actually know that, specifically because Waymo wants to keep the data away from us that we might use to verify it.
“Swiss Re is trustworthy and they say they’re safer!”
They don’t actually say that.
“I suspect it’s true though!”
Humans are terrible, it’s almost like we’d be better off with just using mass public transport
