What's terrifying is it could've just as easily been a child that was playing and trying to retrieve a ball, thinking the car's parked because there's no driver.
If you watch the video, the cat is unfortunate, but not particularly concerning. It was night time, and the cat ran underneath the car from the sidewalk. It's very unlikely that a human would do any better (though, I'd welcome Waymo adding more sensors to the undercarriage to prevent this in the future).
What is concerning is that a woman went up to the Waymo to try to get the cat, backed away a foot or two, and then the Waymo drove off. That seems dangerous, both for the fact that the woman was so close and also because her crouching down in front of the Waymo indicates that there's something exceptional about the environment so the Waymo should proceed with more caution.
>If you watch the video, the cat is unfortunate, but not particularly concerning.
I suspect you'd feel different if it was your cat.
>It's very unlikely that a human would do any better (though, I'd welcome Waymo adding more sensors to the undercarriage to prevent this in the future).
Except the human driver would've almost certainly communicated with the woman trying to collect the cat. I'm surprised they don't have a robust undercarriage array given the hypothetical risk to children, even adults.
It’s still a red herring. I lived in the country and had many “pets” (well… stray cats) get killed by cars growing up. Humans kill soooo many pets with cars every year.
The only question is whether Waymo is less dangerous than humans on average, not whether Waymo can achieve a flawless track record.
It’ll be great if Waymo models can learn from this and become even safer. Human drivers absolutely don’t do that.
I don't think so. Not if it's easily within their engineering capabilities to make scenarios like this much safer or even completely safe.
Historically there's always been frequent accidents involving low speed fatalities of pets and children in similar fashion to this that tend to happen in driveways, so the fact the scenario hasn't been accounted for adequately is a little shocking.
>The only question is whether Waymo is less dangerous than humans on average, not whether Waymo can achieve a flawless track record.
Even if it's vastly safer than humans on average already, if this particular area is less safe than humans, then that warrants attention.
Won't someone please think of all the hypothetical deaths of cat sized children crawling under cars in the middle of the night?
Come on. If we're going to go into hypothetical deaths of unrealistically small children, I'm not sure why we're singling out Waymo given all the cats killed by buses, trains, trucks, fire engines, ambulances, motorcycles, tractors, etc. Even cyclists have run over and killed cats.
Unrealistically small? The example given was a child retrieving a ball. Even if they aren't completely under the vehicle as in the cat's case, the risk remains real nonetheless. It could've been a typical 7-year-old's head or arms under that rear tire instead.
Sorry Ms. Lovejoy, but I fail to see the harm in singling out Waymo here. It's an automated vehicle that purports to be intelligent, so when it kills a pet in a fashion that's plainly preventable, and elements of that scenario can translate to human children, it seems worthy of discussion.
The point in this thread is that Waymo's control system evidently has zero ability to interact with or appropriately consider humans, only the physical objects it treats as inanimate. Beyond ignoring school bus [STOP] signs, cited here are crowding other drivers parallel parking, ignoring people coming out to check in/around/under a Waymo car, and ignoring small or low living things.
Yes, accidents happen and kill pets and children with all those other vehicle types you cited. That is NOT the point.
When those accidents happen, it is not usually because the human simply ignored the other human, and if it was, it is considered a crime. Child or pet runs out and trips, adult runs out after and starts frantically looking under your car — you just ignore the human and proceed ahead and injure or kill a child or pet under the car, it isn't considered an accident, but an act of willful negligence.
Crowd a person trying to parallel park? Maybe not life-threatening, but are certainly an asshat if you do it, and also stupid since it will take longer for them to park and clear your way.
Not stopping for a school bus dropping off children? That will get children injured and killed, and is flat-out illegal.
In this case, Waymo is ignoring all those types of situations, likely not maliciously, but because those cases are inconvenient to fully deal with in the rush to achieve profitability.
Risking and threatening the lives of people and pets in a public arena where they have no ability to opt-in to the hazard for mere profit, is flat-out irresponsible. Get it sorted on your own testing grounds first. Oh, that's not profitable? Well then, you do not have a working business model.
"Fixing" your business model at the expense of risking the lives of innocent people is about as antisocial as it gets; "Some of you may die but that's a risk I'm willing to take" — is pathological capitalism.
I'd give them a couple of chances to fix it, but if they say they did, and then it happens again, boot them off the streets until they prove they did.
What's terrifying is it could've just as easily been a child that was playing and trying to retrieve a ball, thinking the car's parked because there's no driver.