Hacker News Reader

Comment by botanical | Hacker News Reader

I've said it before when a self-driving Uber killed a woman, and I'll say it again:

I don't understand why authorities allow self-driving cars to beta test out in public. I thought Waymo was the best with something like millions of hours on the road but when they started accepting passengers, it failed to proceed past a traffic cone:

https://youtu.be/zdKCQKBvH-A?t=742

I wouldn't trust any of the others like Tesla and Uber that think self-driving is as easy putting it out on the road. And Tesla with their cost-saving no LIDAR nonsense.

These companies should be fined hefty amounts and barred from testing out in public.

View on HN

> [2015] In California alone, 431 people died in accidents involving young drivers[0]

By your logic we should also stop allowing 15-20 year olds to learn how to drive in public.

So far the robots have done less damage.

[0] https://www.inlandempirelawyers.com/blog/young-driver-accide...

In Europe they can't drive normal cars before 18. They will be able to to so after 17 if accompanied by an adult and will be subject to a probation period of 2 years after taking the test. Also speed "limitiers" will be required in all new vehicles after 2024. Most of these would probably be audible nags, like the one that beeps if one drives without the seatbelt.

https://transport.ec.europa.eu/news-events/news/european-com...

Absolutely insane that your society allows children to experiment on unwitting people without consent.

What are you talking about?

You have untrained people using multi-ton motorized equipment on streets where people cross. And you don't even ask them if they'd be okay with this experiment. At least a bit unethical.

They have to take a test after training. Also a two year probation with zero tolerance on drunk driving and other offenses.

Should we give the robots the same test, if they pass they get the drivers license?

Seems like a low bar to me - for both robots and humans. We should instead raise the standards even higher for both!

I don't see why we wouldn't make a robotaxis pass driving exams. Also, I don't see how would this help. Most accidents are caused by licensed drivers

> Most accidents are caused by licensed drivers

Exactly, the point I’ve been trying to emphasize is:

The bar to become a driver in California is really LOW. We should push for the robots to be even better, but given the low bar for humans, I don’t see a good reason to stop them from training on public roads.

So? No need to increase that number by allowing half-developed machine to kill even more people. Having those kids on the road is a necessary evil. Having those dumb machines on the road is not. AI should be trained safely off the roads or not at all. If those companies don't know how then that's their problem, not ours.

Why is it necessary to train kids on the road in real driving conditions but not necessary to train AI on the road in real driving conditions? These AIs have probably hundreds of thousands of hours of incident-free driving in simulation. Kids aren't even required to do a single hour in a simulator before they're allowed on the roads.

> No need to increase that number by allowing half-developed machine to kill even more people

1. To my knowledge there has been only one death by a robot car. It was Uber’s and that program has been shut down for many reasons.

2. Given robot cars already have a higher bar for safety (ie fewer deaths, injuries, etc) than 15 year olds, wouldn’t the more prudent use of time/energy be to invest more into robot cars while changing policy so driving becomes (at minimum) 18+?

> I've said it before when a self-driving Uber killed a woman, and I'll say it again:

But that car had a safety driver, who turned out to be negligent. What's the difference between a company testing a self-driving car with a safety diver killing someone because the safety driver was negligent, and a company doing deliveries with a human driver, and killing someone because the driver was negligent?

The car hit her because Uber purposefully disabled the Volvo SUV's built-in autobraking systems, and because the self-driving system detected and classified her with plenty of time to react...and simply did nothing to avoid hitting her. A complete and total failure of the self-driving control logic.

It's also arguable that she was killed as a direct result of Nevada happily allowing Uber to conduct their self-driving testing after Uber was banned from testing in California for ignoring regulators and their cars routinely being caught violating traffic laws by dashcams, traffic cameras, and cyclist helmet cams.

Even if the safety driver had not been playing with their phone - it's well demonstrated that the whole concept of a 'safety driver' is bullshit; in studies, the drivers are terrible at taking over when they haven't been in control of the vehicle.

Safety drivers were lipstick on a pig, designed to make self-driving cars more palatable to the public.

> Safety drivers were lipstick on a pig, designed to make self-driving cars more palatable to the public.

I go the other way with that information.

Safety drivers being used that way are bullshit.

If they didn't have the safety driver then there's no way in hell the braking systems would have been disabled and need manual intervention.

Therefore it was bad management and the entire idea of using safety drivers that way that was the problem, not the underlying self-driving system.

> The car hit her because Uber purposefully disabled the Volvo SUV's built-in autobraking systems

Are autobraking systems mandatory? If amazon is doing deliveries using vans without autobraking systems, should they also get flak?

>and because the self-driving system detected and classified her with plenty of time to react...and simply did nothing to avoid hitting her. A complete and total failure of the self-driving control logic.

But that's why the human driver is there, right? To catch the system when it fails?

>Even if the safety driver had not been playing with their phone - it's well demonstrated that the whole concept of a 'safety driver' is bullshit; in studies, the drivers are terrible at taking over when they haven't been in control of the vehicle.

No, she was terrible at taking over because she was streaming hulu.

The argument is that self driving vehicles set a false expectation that the driver can relax-- bad design

How do we distinguish if it's "beta" or not? Waymo didn't kill anyone, is it beta? How does non-beta rollout look like?

You're criticizing them for both being dangerous and overly cautious? If anything, the Waymo clip confirms that Waymo's offering is safe.

Those are not mutually exclusive.

I feel like there logically must be a phase where testing in public will have to occur on the path to self driving.

No plan of attack survives contact with the enemy. The same is true for test mocks

> why authorities allow self-driving cars to beta test out in public

because that's the definition of a beta test?

they have gone through a number of test stages over the years, and if the goal is that they drive in public, the penultimate step is test driving in public.

if your criticism is that they didn't do enough prior testing, could be, but you need to make that case.

>I don't understand why authorities allow self-driving cars to beta test out in public

Because self driving cars will save lives in the long run and are potentially the largest QOL improvement of my lifetime.

Not in your lifetime.

> I don't understand why authorities allow self-driving cars to beta test out in public.

Part of Total Information Awareness involves rolling as many cameras around in public as people will tolerate. San Francisco have made their priorities clear: https://reason.com/2022/09/26/san-francisco-police-can-now-h...

I agree 100%. The fact that they let this happen is an affront to the mandate of those safety organizations.

Who is paying whom to clear these hurdles?

Reply on news.ycombinator.com