[notakio]

I moved out of SF ~2018, and only recently revisited, and the increased frequency with which I saw robocars was immediately concerning to me, largely based on the potential for things to go wrong, and with just how frequent they were.

My second thought was, after hearing friends and family still there echo nothing for disdain for them, "how long before someone starts setting these on fire?" as some sort of nuisance campaign against the things. Turns out, 1) you don't need fire, and 2) not that long, since the cone thing started happening about a week afterwards.

[standardUser]

I know this is a tired argument, but it's not wrong: There is equal or greater cause for concern with cars driven by people, we're just used to those horrors and reluctant to trade them for a different set of horrors, even if they are conclusively less horrible.

I lived carless in downtown SF for 10 years, usually with a walking commute, and it was not pleasant. Packed streets full of angry, stressed drivers honking and swerving and practically revving their engines as lights change so they can gun it past cyclists and pedestrians down the next narrow street. Anyone trying to replace these agro drivers with marginally-safer robots has my support.

[Eisenstein]

Depending on driverless taxis instead of having emotional people who need parking all the time seems to be a win, as long as they are safe and someone is held accountable when they mess up.

I am very unwilling to change 'people are flawed but (mostly) held accountable' to 'a corporation owned robot car ran someone over because of a software glitch and it is nobody's fault'. The problems with 'oops we got hacked or lost your data or locked you out of a service or deprecated a product because we spend no money on things that are not income generating' can not be passed on to such a system.

[tlb]

Suppose it's the near future and driverless cars are causing 5x fewer injuries per mile than human drivers. (5x fewer != 0, so some injuries and deaths still occur)

What accountability would you like to see in that situation?

Or a table of

  | relative death rate | accountability |

would cover the widest range of future scenarios.

[Eisenstein]

If there is an investigation and if it is found that someone said 'don't bother with those tests just push to market' or 'let's fire the legacy system patching team because they make no money even though the cars are still driving around' and that caused directly or indirectly a car to glitch and run someone over or to get hacked and used for a crime then someone goes to prison and/or the company gets liquidated or something in between that is reasonable to ensure that it is not tempting for another company to do it again.

I am not a lawyer or a legislator and I cannot come with a regulation or law or something that would do this, but I am sure someone can.

[__loam]

It's a lot easier to hold a single corporation accountible than it is for the distributed responsibility of individuals. Uber killed someone and they had to leave the industry. Tesla is being actively investigated for their fly by night approach.

Additionally, when one of these cars fucks up, these companies know exactly what happened because they are collecting data about the cars performance 24/7.

[Eisenstein]

> Additionally, when one of these cars fucks up, these companies know exactly what happened because they are collecting data about the cars performance 24/7.

And if it turns out they had data that showed that they knew someone was gonna get killed, but they would make more money not fixing the bug -- I would want someone to go to jail.

[szundi]

Then no one is going to implement it probably. Or is it just jail time when dead are more than with humans?

[notakio]

I have no dog in this fight. If anything, my final takeaway was that my impressions were entirely anecdotal and/or experiential, and not based on any sort of legitimate analysis with useful data. It should probably be done, but my point was more that it's only getting more complicated as people take action based on their perception, whether that perception is supported rationally or not. So now, on top of a reasonably already complex engineering problem is an additional layer to correct for: irrational human reaction.

[idopmstuff]

I don't think that complicates anything from a policy perspective. If self-driving cars are good (i.e. safer than human drivers), we should be rolling them out. If a small number of people interfere with them or damage them, we should just arrest those people and prosecute them for the relevant crimes. I don't think that vandalism of self-driving cars is going to be a serious issue long term, given that anyone vandalizing them is going to be on camera, plus people will just get used to them and find other things to be mad about.

[notakio]

SFPD won't do anything about property crimes, to include vandalism, other than to take a preliminary report, on a schedule determined by the department (eg: you call them, they show up 8-12 hours later, uninterested in taking your report). This is because, currently, their DA's office isn't going to prosecute any property crimes.

That can all change, but speculating about the nature or timing of any changes like that is well outside my area, so I'll sit and watch from a distance.

And, again, I'm not vested in any outcome here. It'll work out however it works out, and I'm 2000+ miles away from it, so it won't have much effect on my life.

[Barrin92]

>There is equal or greater cause for concern with cars driven by people

No there isn't. Driverless cars drive under a fraction of the conditions and in self selected locations precisely because the tech is still shoddy. Not to mention they only function at all because they're vastly outnumbered by humans who know how to respond to them.

Let's do an actual experiment to compare. 100% driverless cars with current tech, from different companies on all road conditions that humans drive on during busy hours in a city and see how that goes. To even attempt to compare driverless cars to human drivers without seeing the dynamics of a non-trivial amount of them interacting, which humans need to do all the time, is meaningless.

[hadlock]

Given the millions of miles driven since the start and a single pedestrian death in 2018, I would hesitate to call the tech shoddy. SF alone has 20-30 pedestrian deaths per year despite their well funded and almost completely ineffectual "vision zero" program. If there's been a second pedestrian fatality since I can't find it online.

[BeefWellington]

As has been pointed out elsewhere, millions of miles for autonomous vehicles is a miniscule drop in the vast ocean of miles driven per year in just the US. It's essentially where you'd round the number to.

In 2020, a very down year, the Bureau of Transportation stats[1] give 1,934,743 MILLION miles driven. This Verge article[2] shows for 2020 autonomous vehicles drove 1.99 million miles in California. Let's be generous and assume across the rest of the US it's equal to double that, given Cali is the hotbed of testing at the moment. That puts us up to basically 6 million miles driven.

You're attempting to compare:

    1,934,743,000,000 to 
            6,000,000

The data is not in yet on current generation autonomous vehicles.

[1]: https://www.bts.gov/content/us-vehicle-miles

[2]: https://www.theverge.com/2021/2/11/22276851/california-self-...

[hadlock]

Indeed you are correct: we have statistically significant samples and the autonomous vehicles are matching or exceeding fatality rates for miles driven. Meeting or exceeding human error rates would appear to support the hypothesis that it is not "shoddy".

[minwcnt5]

That's not really true. They now do drive the entirety of San Francisco, 24/7, in all weather conditions. I think it's a pretty apples to apples comparison.

The one point where I agree is there aren't as many of them as there are human cars, and it's possible that in large numbers there could be some unintended consequences. Think 1000 of them getting stuck at the same location. There should be provisions limiting how fast they can scale up to make sure that doesn't happen.

[more_corn]

Uber performed self driving car tests in downtown San Francisco for a bit. I personally saw an error where self driving cars turning right on red failed to yield to pedestrians in the crosswalk who had right of way.

It was a bug. The driver looped and tried again while I was standing there and it happened again. It was an incident. It was a safety incident. Getting killed by a robot breaking the rules is no less dead than if it were a human. And there’s no comfort in the thought that it might happen less often with a robot driving.

Your argument that human drivers suck commits the fallacy of whataboutism. Your argument that not all incidents are safety incidents is misleading.

There are safety incidents they’re unacceptable.

[seanmcdirmid]

> turning right on red failed to yield to pedestrians in the crosswalk who had right of way.

> Getting killed by a robot breaking the rules is no less dead than if it were a human.

This "fail to yield to pedistrians" happens way way way too much though! At least the bug in the robo-taxi could, in theory of course, eventually be fixed. The bug in the human drivers will certainly never be fixed, since this has been a frequent problem since forever and will continue to be a problem as long as human drives (or maybe the USA can just give up on right turns on red like much of the rest of the world).

[idopmstuff]

> Your argument that human drivers suck commits the fallacy of whataboutism.

No, it doesn't. Self-driving cars are potential substitutes for human-driven cars. If self-driving cars cause fewer injuries/accidents/etc. per mile driven, then everyone will be safer if we replace human-driven cars with them. If you object to self-driving cars being on the road even if they're safer than human-driven cars, you are implicitly saying that your preference is for more people to be injured.

[jonathankoren]

Wow. This is an argument.

You start with a common, but controversial hypothetical, that in the future self-driving cars may be safer than humans, and then conclude with that opposing self-driving cars today, which of course are not the hypothetical safer than humans car, is advocating for actual humans to be injured or killed.

That’s some real undergrad level bullshit.

[idopmstuff]

No, you just didn't read what I wrote. At no point did I assert anything about the relative safety of self driving cars. I said that if you oppose self-driving cars EVEN IF they're safer...

That's called the conditional. It means that I'm not saying that thing is true; rather, it means that I am saying that if we take that thing to be true, then something follows logically.

The concept of things being conditional is pretty simple, so I guess in that sense it's undergrad level.

[gonehome]

I remember reading somewhere that some large percentage of people can’t understand conditionals - I found it hard to believe, but I’ve been noticing it more and more.

[ameister14]

It doesn't commit the fallacy of whataboutism - it's the same issue and a direct substitute.

Any safety incident is unacceptable is not a realistic standard, and I am not sure why a lower rate of incident is not a convincing argument for you. Yes, getting killed by a robot leaves you no less dead, but if 5 people are killed by robots where 10 people would have been killed by people, isn't that a net positive?

[NoZebra120vClip]

If 10 people were killed by people, then there could be 10 wrongful-death lawsuits and 10 car insurance claims and 10 cases of liability and 10 criminal investigations and 10 driver's licenses sanctioned, where each and every human behind the wheel must accept responsibility and assume liability for the harm caused to other human beings, and/or property.

If 5 people are killed by SDCs, then 5 executors will need to visit our website, create an account, and submit a request for reimbursement for funeral expenses. Please upload your death certificate and all itemized receipts. Our best AI will absolutely make its best efforts to find out which remote human operator caused those cars to begin driving, and then we will launch an internal investigation into whether their pay should be cut, or maybe we'll put them on paid leave instead and connect them with a grief counselor. Thank you for choosing WayMo. Scan this QR code to install our app!

[whats_a_quasar]

You're introducing a new claim, that it will be harder for families to have justice and be compensated when a death is caused by a robotaxi rather than a human. I don't see any reason to assume this. If anything, it ought to be easier to get a rich, large, well known corporate robotaxi company to provide compensation than a random individual who might even be driving uninsured

[CatWChainsaw]

The rich corporation is going to use their riches to exhaust your money supply on lawyers, and in the chance that you survive that test of endurance, offer a measly settlement with an NDA attached.

[NoZebra120vClip]

That's not my only claim. The additional gotcha is that the supposedly responsible human being is remote from the incident, and the regulators may find it more difficult to determine responsibility and assign liability to someone somewhere inside some very large company with a lot of network connectivity and an equal helping of plausible deniability. Compare that with a human at the wheel who hopefully carries a driver's license and proof of insurance?

Anyway, a "rich, large, well-known" company is always going to calculate the cost of a human life taken, vs. the cost of doing business, and run the margin right up to a rounding error. I don't doubt that their actuaries are just as good as GEICO's.

Lest we forget - corporations are people.

[skeaker]

This is not an argument (your situation is non-existent and fallacious) so you've more or less lost this debate here, but I'll give that it was funny.

[greiskul]

So you are saying that you are completely fine with 5 more people being killed, if their family gets money? I don't think thats as ethical a position as you think it is. You are putting human burocracy ahead of human lives.

[NoZebra120vClip]

It's not about "getting money". It's about recourse to the law, it's about humans who take responsibility, and it's about properly assigning liability. A single human life is precious and worth more than gold; you can't put a price on a human life. But in human burocracys, they do that. Perhaps it would be more just if the human drivers were killed in retribution? Death penalty for vehicular manslaughter? I mean, the families don't have to get money. Instead they could just receive front-row tickets in the lethal injection chamber? Is that more just, with less burocracy? We could destroy the cars that kill people, too. How's that? You could crush the car up in a compactor, then extract all the valuable minerals and other material, and award it to the families of the deceased.

Look greiskul, I don't appreciate your attempt to set up a Trolley Problem with my ridiculous and hypothetical scenario that will totally never happen in real life. My comment was intended to highlight the difference of human responsibility, assignment of liability, and recourse to legal means when someone is wronged. Just because the numbers were different in the GP, doesn't change my scenario one iota if you make it 10 and 10, or if you make it 6 minorities and 10 white male landowners, or if you make it 101 Dalmatians and a breeding pair of Tyrannosauri Rex. That wasn't the point.

greiskul, I'll thank you not to make insinuations about my ethical beliefs, especially when those insinuations serve your Trolley Problem agenda. I'm not making an ethical judgement on the state of things today here, I'm simply describing the situation as I see it. The justice system in these United States is set up a certain way, along with insurance adjusters, and the DMV/DOT, and the auto manufacturers and all the regulatory agencies that handle them. Since human dignity is inviolable, and a human life is priceless, I concur that it appears immoral for an insurance company to put a price on that human life, especially a lowball, profitable price.

They say that "money can't buy happiness" but it's a lot more comfortable to do my crying in a Mercedes than on a bicycle.

Unfortunately, SDCs are an excellent method to remove human dignity and living beings from the equation. For better or worse, the roads will be increasingly automated, and there will be fewer humans taking responsibility for their actions on the road. Corporations are people, though, so let's just give them the right to vote and be done with it?