Autonomous driving safety supremacy has been achieved

Those goofy Waymo robotaxis that occasionally drag pedestrians under them and get stuck circling roundabouts or honking in parking lots at 4AM have apparently exceeded the safety of the average human-driven car, and not by a small margin either, although they stick to city streets and are still being tested for highway use:

https://www.engadget.com/transportation/waymos-driverless-cars-are-apparently-an-insurance-companys-dream-220746643.html

Insurance savings could make the additional expense of an autonomous car pay for itself within just a few years, and lower risks could help offset the effects for the insurance market in general of newer cars being so much more expensive (not just on average, cheap new cars are rapidly becoming a thing of the past over the last few years!)

The tech will only get better.  As a lover of driving and hooning, I'm not sure how I feel about the implications of that.

In reply to nderwater :

Don't forget that some of the biggest benefits to both safety and traffic efficiency will be achieved on autonomous-only roads.

I can only hope that people who don't want to/shouldn't be driving will adopt this quickly and save the roads for the rest of us. 

Maybe a driver's license can be like a pilot's license requiring ridiculous training and hours logging. Yikes.

The only thing that I have to add is that in San Francisco, which is a challenging environment in all ways except for the lack of snow, they are all over the place and seem to do fine. One eery thing is that for some reason they seem to run in packs. I'll see 2 or 3 in a row. Perhaps that just means that they're splitting a large group of riders. Or they could be looking for.... prey  

iansane said:

I can only hope that people who don't want to/shouldn't be driving will adopt this quickly and save the roads for the rest of us. 

Problem is that every driver thinks they're above average. Which means you need to identify "those who shouldn't be driving". And even taking someone's license away doesn't prevent them from actually driving.

I remember these threads a decade ago. I'm starting to think it's more like fusion reactors. Lots of people hyping the potential to get money, almost none of it ever getting to consumers. 

But this is actual real-world results. Part of the problem is going to be that we as a species are not good at accepting averages. Sure, these may be safer than the average driver by whatever metric, but stories like "dragged a pedestrian" make us think they're not safe. Never mind that the pedestrian was launched at the robotaxi by a car being driven by a human and that human-driven cars do the same thing.

The big advantage they have is that you don't need to teach every autonomous car how to drive. If one learns something, they all learn it. We have to start from scratch with every human driver. They also don't get distracted or bored or impaired or tired. I think there's real potential but we are going to have to accept that 100% infallibility is not in the cards. We're far from it now with humans.

Keith Tanner said:

iansane said:

I can only hope that people who don't want to/shouldn't be driving will adopt this quickly and save the roads for the rest of us. 

Problem is that every driver thinks they're above average. Which means you need to identify "those who shouldn't be driving". And even taking someone's license away doesn't prevent them from actually driving.

Yeah, I was trying to come up with a way to succinctly describe that. Humility is not a burgeoning trait.

In reply to Keith Tanner :

So were the posts a decade ago. 
 

The problem is that people expect perfection out of automation. So being 99% better than humans isn't good enough. 

Keith Tanner said:

But this is actual real-world results. Part of the problem is going to be that we as a species are not good at accepting averages. Sure, these may be safer than the average driver by whatever metric, but stories like "dragged a pedestrian" make us think they're not safe. Never mind that the pedestrian was launched at the robotaxi by a car being driven by a human and that human-driven cars do the same thing.

A jaywalker, hit by a speeding driver who fled the scene. And the robotaxi only pulled over because their original strategy of parking themselves after an incident annoyed people. And the pedestrian survived and I believe fully recovered, by the way.

Yeah, sure, whatever. My printer can't make 3 identical copies in a row without demanding a blood sacrifice, the TV bricks itself for 40 minutes every time an app updates, and when a leaf gets stuck on a sensor my automatic cruise control disables and you expect me to believe that an entire self-driving car is possible outside of VERY strictly controlled circumstances where the infrastructure has been tailored and standardized specifically for the autonomous sensors?

Not happening. 

We just started Waymo here in Los Angeles ,  But its only a small area to start , Santa Monica to  the Downtown LA  area  and a little bit more but not including LAX Airport , 

I signed up for it and would love to try it when it goes from my home to somewhere I need to go ,  hopefully in the next year or so they will expand the area .

The best reason I  can see to use it is to go some place with no parking  or expensive parking , 

And yes I can see it  will be a White Knuckle ride  the first time ,  but they are still not allowed on the Freeways , so we will not be going that fast :)

The Future is "Almost" here !

Kreb (Forum Supporter) said:

The only thing that I have to add is that in San Francisco, which is a challenging environment in all ways except for the lack of snow, they are all over the place and seem to do fine. One eery thing is that for some reason they seem to run in packs. I'll see 2 or 3 in a row. Perhaps that just means that they're splitting a large group of riders. Or they could be looking for.... prey  

When driving around San Francisco a couple of weeks ago for the first time, in the dark, in traffic, with hills and curves, I was thinking "man. This has got to be rough on those robot cars" and yet, they were zooming around without issue, just like you said. Saw more than a few that were in pairs. 

I took a picture of one in daylight and texted my car buddy group text (we live in Virginia). Got the predicted amount of Terminator/robot overlord memes in reply. 

Driven5 said:

In reply to nderwater :

Don't forget that some of the biggest benefits to both safety and traffic efficiency will be achieved on autonomous-only roads.

Tracks.  They're called tracks.  

Efficiency goes way up when you eliminate rubber tires, too!

In reply to Spearfishin :

Yeah, I think that most cities would be cake compared to San Fran. That said, my main concern is what happens when roads get poorly ID'd - construction, snow, et cetera? We've all had plenty of WTF moments when the grid is interrupted. The only thing worse than an unmoving car that doesn't know what to do, is a car that thinks that it knows and is wrong.

Kreb (Forum Supporter) said:

In reply to Spearfishin :

Yeah, I think that most cities would be cake compared to San Fran. That said, my main concern is what happens when roads get poorly ID'd - construction, snow, et cetera? We've all had plenty of WTF moments when the grid is interrupted. The only thing worse than an unmoving car that doesn't know what to do, is a car that thinks that it knows and is wrong.

We call it "snow rules". Normally, if something is blocking your lane, but not the oncoming lane, you wait for the oncoming lane to be clear as you don't have the right of way. But, if the thing blocking your lane is snow? Snow rules! You just go! The other cars will stop. Or swerve. Who cares, you can't see the lines!

At least that's how human drivers act around here during our rare snow events. It's like all understanding of right of way rules, lane integrity, etc just vanish once we have more than a dusting. Hopefully the robots would do better.