Kudos for tracking down the original study. In my opinion, it should not be acceptable for a Union of Scientists to write "studies have shown ..." without any references to those publications. This is especially the case when journalists are expected to report on those claims.
Like so many other pillars of the progressive nonprofit sector, the UCS has devolved into yet another interchangeable leftist omnicause organization. Their homepage is mostly devoted to begging for donations, internal DEI enforcement is high on their priorities list, they have every vocal opinion one might expect on Middle East conflicts which have no relation to science, etc. I don't think their staff is much concerned with providing objective, credible, evidence-based data.
How embarrassing for The Atlantic, “we should leave people trapped in automateable manual labor jobs so I can talk to them while they work” is a terrible take.
I’ve heard the same types of takes from fellow artists. I saw a mastering engineer and producer tell their online following to never speak with a customer service robot because it was replacing the humanness of the interaction. I told them that as someone who has actually had to work a job like that, their take reeked of privilege, since I’m of the opinion that as few humans as possible should have to work a call center job.
The problem with customer service automation is that, apparently unlike Waymo’s driving robot, the call center AI sucks at its job. It just uses keywords to route people to canned responses and will not understand any circumstances route you to a human.
AI call centers are fine for dealing with the type of callers who don't realize their TV isn't working because it's not plugged in, but are bad for dealing with competent humans who have actually tried to reasonably solve their own problem and only call when the actual company/government office needs to get involved. Even with humans at a call center, I usually have to get them to go off script to actually solve a problem.
Yeah, they do need to improve. The first ones in like 2021-22 were so horrendously unhelpful. Now about 20% of the time I’m able to actually accomplish something. Not good, but progress.
I can't believe that an author who wrote a long piece about how quiet is gentrification which never even alluded to the body of work showing how bad constant exposure to noise is for health would be so shoddy with her evidence.
I don’t know the author so I can’t say this is how she operates, but this article and the one on how noise = good, quiet = bad, between them, make me think of a person who wants “a village” but only wants to take from the village, not put into it.
Service people are paid to do a job, not be your pal.
Those same arguments against Waymo’s (or any AVs) also ignore the benefits that Waymo’s provide for women and people with disabilities.
As someone who doesn’t (can’t drive), the best thing for me is more availability and cheaper access to all forms of transportation. AVs have long been a dream for many and whether or not a privately owned ride share company is the exact vision of the future, it is a clear step towards progress. Of course, yes, I want fast and reliable and affordable public transit. But I also want my own autonomy—one that necessarily doesn’t rely on as many humans to get me to where I want or need to go.
And of course, almost universally the women I know in San Francisco describe the feelings of safety and freedom in Waymo’s. Of course the vast majority of Uber drivers are fine. But we cannot deny that the chance of a very terrible encounter, albeit “small”, has a very meaningful impact on the freedom (or lack thereof) that women and other marginalized people experience with platforms like Uber. Waymo’s are a way to reclaim that freedom.
And I will further add as a pedestrian with low vision but one who still appreciates cars—6 Waymo’s in a row on your street can be frustrating and just look silly. But at the end of the day, I trust and have experienced that they are more likely to stop for me than many of the human drivers in the city we share. 🤷 forgive me for welcoming this change.
You might enjoy Andrew Leland's latest piece on his Substack 'Vile Jelly'. He has a horrendous commute and is so over the forced socialization and the lack of autonomy of Uber dependence. Your local Waymo store can't come fast enough for him. If you haven't read it you might enjoy his book The Country of the Blind; A Memoir at the End of Sight. It's very good.
One thing to emphasize is exactly what's in the study. They take 8 pedestrian detectors, by which they mean computer programs that annotate a still image with where pedestrians are, and test them on a big dataset of images. The detectors are all published and available software, but they are not autonomous driving systems, which have multiple cameras, sensor fusion, and video rather than a single image.
Additionally, the systems they test are not from commercial AV companies such as Waymo. They say that one of the systems is based on something used for the Apollo Chinese AV company, but they are definitely not testing that AV system. The papers are all by authors at universities and companies without self driving systems (such as Facebook).
I found that original Atlantic piece pretty appalling and was hoping you’d write about some of these claims. That bit in the middle was absolutely wild - take one quote by an industry person about potential lives saved and then dismiss it because it’s a large looking but global number, then quote this questionable activist group claiming self driving cars are racist. No serious engagement with the possibility that self driving cars might actually save lives in a domestic context.
I really like the Atlantic overall but it’s pretty frustrating to try to keep a mental list of their contributors who write pieces that are negligent at best, nakedly ideological, deceptive, and bad faith at worst.
I really appreciate Piper's crusade on Waymos and the motivated bullshit that's grown up around them. There's a reason why, when we want to say something is safe, we compare it to driving a car: by far the most dangerous thing we routinely do. When I see people describe Waymos as "antihuman," I get viscerally angry. You don't get to endorse the deaths of tens of thousands every year, hundreds of them kids, and claim to be the prohuman side of the argument.
Thank you for this. The part of her argument that truly made me tear at my hair was the ignoring of base rates. Even if autonomous vehicles were moderately racist--let's be generous to the argument and say 25% more likely to miss a person of color--the 70-95% lower overall death rate would thoroughly wash out that effect. The racist autonomous vehicles would still kill fewer of people of color!
It would also be useful to know if *humans* have differential rates of noticing pedestrians by skin color. I suspect clothing color makes a bigger difference, but it would not be surprising if whatever effects there really are, affect biological neural processing as much as computational neural processing.
My son was hit in his school crosswalk last week by a human driver who ran a red light. After surgery, he's recovering with "only" a snapped leg. But some of his first words were "a driverless car would have stopped." And he's right.
So call it for the lie that it is! Even the technically accurate headline "Waymo has a harder time braking for children than adults" is a lie because of the lingering sense of "Waymo's are bad for children" it imparts on readers not steeped in context and statistics. The correct comparison is "does Waymo or a human driver brake for children in the road more reliably?" and the answer is overwhelmingly Waymo. Any headline that doesn't convey "Waymo is much safer than a human driver" is a lie and journalistic malpractice.
I'm lucky my son is alive. If we quickly adopt this technology, we'll prevent the 600+ annual children pedestrian fatalities, plus countless other parents won't have to go through considering themselves lucky it's "only" his leg.
Gonzalez is a fount of moronic takes, such as advocating for smoking. If only I could get a high paying job that consisted of propagating idiocy every now and then instead of having to do real work for 40h/week.
That said, the solution is a) to have a very high onus on claims of bigotry and b) to also have a reverse onus, basically “prove that the group you’re defending hasn’t brought it upon themselves - do they exhibit elevated rates of antisocial behaviors such as crime, backward thinking, agitation, victim mentality or in group narcissism? Yes? Then improve them before demanding anything”.
Yes she is. She’d probably HATE Sweden, lol. Seriously, I thought that “liking peace and quiet is racist” was a super dumb take. MOST people like peace and quiet at least some of the time. Even hard-partying college kids still want to sleep and study.
You argue that it is not the case that there is a racial disparity with Waymo tech. This is all well and good.
I would like to argue that even if it was the case that we went from a situation where 20 white people and 20 black people died from car accidents to one where 5 white people and 10 black people died from car accidents, that would still be a good thing for everyone. When Waymo's safety is 90% better than the status quo, it's hard to imagine it would make things worse for any particular subgroup.
Perfect equality between every group is not as important as pushing for improvement for everyone.
Good article. I am very pro-Waymo and generally pro technological progress. To that end, I want to push back on a point you made. You say a few times something the effect that, while Waymos may be safer and you think we should have them, it’s a less human world human-connect world with them instead of human drivers.
Is it though? I am on the upper end of friendly, extroverted, gregarious, etc. I take Ubers quite frequently and have had some interesting conversations with drivers. But the median Uber ride for me is not that at all; for a variety of reasons. Sometimes I am working, sometimes I am tired or not feeling talkative, sometimes the driver is in that boat instead, sometimes the driver does not speak English particularly well. So what are we really losing here? Even for me, if I never had a taxi ride with another human in it, it would cut down my experience of human contact minimally in the scale of my life. And I would be willing to be the effect would be smaller for the median (less talkative) person.
But that’s not the only factor here! Let’s think the other way too! If autonomous taxis make going places easier, which they are certainly projected to do, at scale, then people will be able to make trips to see people they actually WANT to spend time with more (Sorry friendly Uber drivers, but this ain’t you)!
Sure, anti-social century and all that. But the people who are choosing to forgo human connection right now, are also not the ones that are going to miss a chat with their Uber driver either.
The fact that people feel the need to justify their pre-existing preferences with (bogus) scientific evidence is a sign of science’s utter and continuing cultural triumph, not its decline. Trafficking in fake evidence is dangerous, and it’s annoying to have to write articles like this one. But in the grand scheme of things, we should take heart that even cranks must argue within the same epistemic paradigm.
For what its worth, with a little baby at home, you would see me way more if I had a self driving car. And you would see the elderly more, and the disabled more. I think a passenger also has a tendency to ignore the driver as a person, which has a lot of class and racial bias to it that I am eager to leave behind.
Ironic about the class and racial bias from someone who wrote about how it’s racist to like peace and quiet. While I’m happy to talk with Lyft drivers if they want to chat, I’m also very happy to STFU and play games or browse on my phone. Or look out the window.
Lyft/Uber/taxi drivers are there to get you from A to B, not be your on-call companionship, jeez. If a Waymo can get me there just as fast and safely, then yay Waymo. Bring on the Waymos.
Kudos for tracking down the original study. In my opinion, it should not be acceptable for a Union of Scientists to write "studies have shown ..." without any references to those publications. This is especially the case when journalists are expected to report on those claims.
Like so many other pillars of the progressive nonprofit sector, the UCS has devolved into yet another interchangeable leftist omnicause organization. Their homepage is mostly devoted to begging for donations, internal DEI enforcement is high on their priorities list, they have every vocal opinion one might expect on Middle East conflicts which have no relation to science, etc. I don't think their staff is much concerned with providing objective, credible, evidence-based data.
How embarrassing for The Atlantic, “we should leave people trapped in automateable manual labor jobs so I can talk to them while they work” is a terrible take.
I’ve heard the same types of takes from fellow artists. I saw a mastering engineer and producer tell their online following to never speak with a customer service robot because it was replacing the humanness of the interaction. I told them that as someone who has actually had to work a job like that, their take reeked of privilege, since I’m of the opinion that as few humans as possible should have to work a call center job.
The problem with customer service automation is that, apparently unlike Waymo’s driving robot, the call center AI sucks at its job. It just uses keywords to route people to canned responses and will not understand any circumstances route you to a human.
AI call centers are fine for dealing with the type of callers who don't realize their TV isn't working because it's not plugged in, but are bad for dealing with competent humans who have actually tried to reasonably solve their own problem and only call when the actual company/government office needs to get involved. Even with humans at a call center, I usually have to get them to go off script to actually solve a problem.
Yeah, they do need to improve. The first ones in like 2021-22 were so horrendously unhelpful. Now about 20% of the time I’m able to actually accomplish something. Not good, but progress.
I think this is just a holdover from the era when calling something racist was a cheat code to win arguments in left-of-center spaces.
Up up down down left right left right Waymo is racist
Only 1980s kids/2010s kids remember.
I can't believe that an author who wrote a long piece about how quiet is gentrification which never even alluded to the body of work showing how bad constant exposure to noise is for health would be so shoddy with her evidence.
IIRC, the same author mad people didn't like her making noise all the time but who would then retreat to a cabin or something to do her writing.
I don’t know the author so I can’t say this is how she operates, but this article and the one on how noise = good, quiet = bad, between them, make me think of a person who wants “a village” but only wants to take from the village, not put into it.
Service people are paid to do a job, not be your pal.
Wait, are you implying the staff at Trader Joe's DON'T care if I have any plans for the weekend???
It's amazing that this is still what I remember her for.
Those same arguments against Waymo’s (or any AVs) also ignore the benefits that Waymo’s provide for women and people with disabilities.
As someone who doesn’t (can’t drive), the best thing for me is more availability and cheaper access to all forms of transportation. AVs have long been a dream for many and whether or not a privately owned ride share company is the exact vision of the future, it is a clear step towards progress. Of course, yes, I want fast and reliable and affordable public transit. But I also want my own autonomy—one that necessarily doesn’t rely on as many humans to get me to where I want or need to go.
And of course, almost universally the women I know in San Francisco describe the feelings of safety and freedom in Waymo’s. Of course the vast majority of Uber drivers are fine. But we cannot deny that the chance of a very terrible encounter, albeit “small”, has a very meaningful impact on the freedom (or lack thereof) that women and other marginalized people experience with platforms like Uber. Waymo’s are a way to reclaim that freedom.
And I will further add as a pedestrian with low vision but one who still appreciates cars—6 Waymo’s in a row on your street can be frustrating and just look silly. But at the end of the day, I trust and have experienced that they are more likely to stop for me than many of the human drivers in the city we share. 🤷 forgive me for welcoming this change.
You might enjoy Andrew Leland's latest piece on his Substack 'Vile Jelly'. He has a horrendous commute and is so over the forced socialization and the lack of autonomy of Uber dependence. Your local Waymo store can't come fast enough for him. If you haven't read it you might enjoy his book The Country of the Blind; A Memoir at the End of Sight. It's very good.
One thing to emphasize is exactly what's in the study. They take 8 pedestrian detectors, by which they mean computer programs that annotate a still image with where pedestrians are, and test them on a big dataset of images. The detectors are all published and available software, but they are not autonomous driving systems, which have multiple cameras, sensor fusion, and video rather than a single image.
Additionally, the systems they test are not from commercial AV companies such as Waymo. They say that one of the systems is based on something used for the Apollo Chinese AV company, but they are definitely not testing that AV system. The papers are all by authors at universities and companies without self driving systems (such as Facebook).
I found that original Atlantic piece pretty appalling and was hoping you’d write about some of these claims. That bit in the middle was absolutely wild - take one quote by an industry person about potential lives saved and then dismiss it because it’s a large looking but global number, then quote this questionable activist group claiming self driving cars are racist. No serious engagement with the possibility that self driving cars might actually save lives in a domestic context.
I really like the Atlantic overall but it’s pretty frustrating to try to keep a mental list of their contributors who write pieces that are negligent at best, nakedly ideological, deceptive, and bad faith at worst.
Maybe all of their articles are bad, but you only detect it some of the time.
I really appreciate Piper's crusade on Waymos and the motivated bullshit that's grown up around them. There's a reason why, when we want to say something is safe, we compare it to driving a car: by far the most dangerous thing we routinely do. When I see people describe Waymos as "antihuman," I get viscerally angry. You don't get to endorse the deaths of tens of thousands every year, hundreds of them kids, and claim to be the prohuman side of the argument.
Thank you for this. The part of her argument that truly made me tear at my hair was the ignoring of base rates. Even if autonomous vehicles were moderately racist--let's be generous to the argument and say 25% more likely to miss a person of color--the 70-95% lower overall death rate would thoroughly wash out that effect. The racist autonomous vehicles would still kill fewer of people of color!
It would also be useful to know if *humans* have differential rates of noticing pedestrians by skin color. I suspect clothing color makes a bigger difference, but it would not be surprising if whatever effects there really are, affect biological neural processing as much as computational neural processing.
Some people just want to watch the world burn (= leveling down)
My son was hit in his school crosswalk last week by a human driver who ran a red light. After surgery, he's recovering with "only" a snapped leg. But some of his first words were "a driverless car would have stopped." And he's right.
So call it for the lie that it is! Even the technically accurate headline "Waymo has a harder time braking for children than adults" is a lie because of the lingering sense of "Waymo's are bad for children" it imparts on readers not steeped in context and statistics. The correct comparison is "does Waymo or a human driver brake for children in the road more reliably?" and the answer is overwhelmingly Waymo. Any headline that doesn't convey "Waymo is much safer than a human driver" is a lie and journalistic malpractice.
I'm lucky my son is alive. If we quickly adopt this technology, we'll prevent the 600+ annual children pedestrian fatalities, plus countless other parents won't have to go through considering themselves lucky it's "only" his leg.
Gonzalez is a fount of moronic takes, such as advocating for smoking. If only I could get a high paying job that consisted of propagating idiocy every now and then instead of having to do real work for 40h/week.
That said, the solution is a) to have a very high onus on claims of bigotry and b) to also have a reverse onus, basically “prove that the group you’re defending hasn’t brought it upon themselves - do they exhibit elevated rates of antisocial behaviors such as crime, backward thinking, agitation, victim mentality or in group narcissism? Yes? Then improve them before demanding anything”.
isn't she the "liking quiet neighborhoods is racist" writer?
Yes she is. She’d probably HATE Sweden, lol. Seriously, I thought that “liking peace and quiet is racist” was a super dumb take. MOST people like peace and quiet at least some of the time. Even hard-partying college kids still want to sleep and study.
You argue that it is not the case that there is a racial disparity with Waymo tech. This is all well and good.
I would like to argue that even if it was the case that we went from a situation where 20 white people and 20 black people died from car accidents to one where 5 white people and 10 black people died from car accidents, that would still be a good thing for everyone. When Waymo's safety is 90% better than the status quo, it's hard to imagine it would make things worse for any particular subgroup.
Perfect equality between every group is not as important as pushing for improvement for everyone.
Good article. I am very pro-Waymo and generally pro technological progress. To that end, I want to push back on a point you made. You say a few times something the effect that, while Waymos may be safer and you think we should have them, it’s a less human world human-connect world with them instead of human drivers.
Is it though? I am on the upper end of friendly, extroverted, gregarious, etc. I take Ubers quite frequently and have had some interesting conversations with drivers. But the median Uber ride for me is not that at all; for a variety of reasons. Sometimes I am working, sometimes I am tired or not feeling talkative, sometimes the driver is in that boat instead, sometimes the driver does not speak English particularly well. So what are we really losing here? Even for me, if I never had a taxi ride with another human in it, it would cut down my experience of human contact minimally in the scale of my life. And I would be willing to be the effect would be smaller for the median (less talkative) person.
But that’s not the only factor here! Let’s think the other way too! If autonomous taxis make going places easier, which they are certainly projected to do, at scale, then people will be able to make trips to see people they actually WANT to spend time with more (Sorry friendly Uber drivers, but this ain’t you)!
Sure, anti-social century and all that. But the people who are choosing to forgo human connection right now, are also not the ones that are going to miss a chat with their Uber driver either.
The fact that people feel the need to justify their pre-existing preferences with (bogus) scientific evidence is a sign of science’s utter and continuing cultural triumph, not its decline. Trafficking in fake evidence is dangerous, and it’s annoying to have to write articles like this one. But in the grand scheme of things, we should take heart that even cranks must argue within the same epistemic paradigm.
For what its worth, with a little baby at home, you would see me way more if I had a self driving car. And you would see the elderly more, and the disabled more. I think a passenger also has a tendency to ignore the driver as a person, which has a lot of class and racial bias to it that I am eager to leave behind.
Ironic about the class and racial bias from someone who wrote about how it’s racist to like peace and quiet. While I’m happy to talk with Lyft drivers if they want to chat, I’m also very happy to STFU and play games or browse on my phone. Or look out the window.
Lyft/Uber/taxi drivers are there to get you from A to B, not be your on-call companionship, jeez. If a Waymo can get me there just as fast and safely, then yay Waymo. Bring on the Waymos.
(Sorry, history of class and racial bias)
Gonzalez has always been sloppy like this.