That's an extrapolation from
an estimate based on a very small sample. The vets number is from their membership reporting, nationwide. I think I'll stick with that.

Neither number is good for Waymo, as I did a simple linear average. They've had 2 deaths in the last month as they scale.

My source is the American Veterinarian Society. I actually rounded up from 280K because no one really collects these statistics.

I've seen that number asserted, but no sourcing behind it. Can you tell me your source?

Safety is a complex topic. This is one metric. The one you measure means fewer non-fatal accidents measure injuries and property damage.

Which metric you emphasize shows your values.

If Waymos cause less property damage but kill more pets, are they more safe? If they kill more people, are they?

That's what's so astounding: they are less safe than the most unsafe, even when you compare them to that cohort separated out!

Read Cummings 2024.

Humans drive a *lot* of miles. Koopman's attempt at a figure of merit represents that. I adapted it to killing cats and dogs and Waymos failed.

I also refer you to this peer-reviewed paper which puts Waymo in a category with the worst human drivers:

M. L. Cummings and B. Bauchwitz, "Identifying Research Gaps through Self-Driving Car Data Analysis," in IEEE Transactions on Intelligent Vehicles, doi: 10.1109/TIV.2024.3506936.

"Safer than a human driver" means all human drivers, including impaired and student drivers.

This is inspired by Phil Koopman's blog post estimating we need 300M miles without a human fatality for a similar safety estimate.

philkoopman.substack.com/p/perspectiv...

Waymos have killed at least 7 cats and dogs since 2021. Humans kill 1 cat or dog per about 11M vehicle-miles in the USA. (300K dog/cat deaths vs 3.2T vehicle-miles). For us to be 95% confident that Waymo is safer than a human driver, you'd expect it to go 33M miles without a death. It has gone 22M.

Waymo also uses 'made contact with' in their official NHTSA crash reports and sometimes in ways that implies or suggests the other party was responsible, eg:

bsky.app/profile/anic...

Waymos have killed at least 7 cats and dogs since 2021. Humans kill 1 cat or dog per about 11M vehicle-miles in the USA. (300K dog/cat deaths vs 3.2T vehicle-miles). For us to be 95% confident that Waymo is safer than a human driver, you'd expect it to go 33M miles without a death. It has gone 22M.

That looks like a motorcade? Is DC watching?

Waymo is not safer around your pets. It is maybe worse than the worst human drivers.

I know it's hard to understand the scale of human vehicle operation, but it's huge. Americans alone traveled 5,100,000,000,000—five point one TRILLION—passenger-miles in 2023.

Waymo sucks. /FIN

Per Phil Koopman, We'll need on the order of 300 million miles of driving with zero human fatalities to demonstrate they're as good as the worst human drivers, including drunk drivers and unlicensed 12 year olds who stole a car.

/2

philkoopman.substack.com/p/perspectiv...

(His estimate of 20 years for Waymo is probably much too high, but set that aside.)

Let's say we kill twice as many pets as we do humans on the road. Waymo needed to go 150M miles without a single pet fatality to be as good as these worst drivers. It had only gone 100M as of the accident.

/3

LOL a bike ride just rode past a block away