Comment by ocdtrekkie | Hacker News Reader

Comment by ocdtrekkie | Hacker News Reader

Parent Full thread

http://fortune.com/2016/01/13/google-self-driving-car-accide... Ten of these would've been the Google car's fault. Given the mileage they've covered in 2015, and the number of accidents, they're terrible drivers statistically... to the point that a human is the safety device.

http://www.opensecrets.org/orgs/toprecips.php?id=D000022008&... Congress, 2014 term. Many of the top recipients actually wrote letters on Congress letterhead begging the EU not to rule against Google, which is a fairly unprecedented action.

http://blogs.wsj.com/digits/2014/09/04/two-googlers-take-top... http://www.govexec.com/management/2014/09/white-houses-roste... Googlers take up some of the top technology positions in the White House, including "Chief Technology Officer". Eric Schmidt and other high ranking Googlers also visit the White House nearly weekly.

Think they behaved themselves? Nope: http://techpresident.com/blog-entry/white-house-deputy-cto-a...

In case you're curious how Google's using their White House positions: It's to get money. http://thehill.com/policy/transportation/265932-obama-pledge...

And some more of our taxpayer dollars. While this article carefully avoids mentioning Chromebooks, they're now the leading purchased portable computers in schools, they're pictured in the header of this blog, and it's written by their "former" Googler, who is the CTO: https://www.whitehouse.gov/blog/2016/01/30/computer-science-...

That's a lot of strong government ties, but are they worried about regulation? Well... for the longest time they controlled the US's regulatory body too:

http://www.buzzfeed.com/williamalden/ftc-commissioner-appear... The FTC was basically run by a paid Google shill, Joshua Wright, for the last couple of years. (The FTC reopened their investigation against Google within a month of him stepping down last year.) His pre-FTC behavior wasn't much on the ethical manner either: He was paid as a university professor to write "academic studies" supporting Google's legal views.

One would wonder by the United States has been one of the slowest and most hesitant countries to regulate and pursue legal action against Google, until one does their research. Then it stops being much of a surprise at all.

> These are encouraging numbers, says Chris Urmson, director of the self-driving project, because most of the incidents occurred early on. Only five took place during the 370,000 miles driven in 11 months of 2015, Urmson wrote in a blog posted on Medium. He expects the rate of these incidents to keep declining.

I mean sure he is biased, but the link doesn't really seem all that negative to be honest. Of course it will have issues in the early stages.

From the report that the fortune.com article links to:

> In 10 of the 13 simulated contact events, the SDC’s predicted behavior would have, in simulation, caused contact (though 2 of these involved simulated contact with traffic cones).

The 13 events happened in the 15 months from September 2014 to November 2015 (both incl). 424,331 miles driven.

Can the average driver drive 30,000 miles without making contact with a traffic cone or another car?

> Can the average driver drive 30,000 miles without making contact with a traffic cone or another car?

The broken pylons at my closest bike lane and highway exit indicate that very few drivers can achieve this.

That's averaged over weeks or months or however long it takes for repairs, and over however many thousands of cars pass by (the US drives trillions of miles each year).

Yes, and so is the 30,000 mile figure.

Detail:

I'm speaking in generalizations here, but it seems that small and even less traveled roads that have pylons often have these damages.

Scarily, on a run today, a "lightly" traveled road in a residential area had multiple "stop for pedestrian" signs that had been knocked down. I find this particularly scary on the (non major side street) that I live - where other than residents (300ppl) and 1 business (< 300 ppl) few people should choose to drive - loses these road pylons in a matter of days.

At least here in Italy no :-)

Yes. According to this pro-Google car post (which is false, when you consider human interventions to prevent accidents), the average mileage between accidents for humans is about 165,000. I've seen claims that humans average closer to 250,000 miles driven between accidents as well. http://mashable.com/2012/08/07/google-driverless-cars-safer-...

I find this statistic highly suspect. I don't know of a single person that had a car for 100,000 miles and didn't have some type of accident.

I suspect those average numbers are being skewed by the rural drivers who rack up lots of miles without interaction with other cars.

San Francisco averages 6.4 years between accidents. https://www.mainstreet.com/slideshow/most-car-accident-prone...

That matches roughly with my experience. And, while the Google car isn't driving San Francisco, the Santa Clara area isn't that much more benign.

Even if you take Mashable's 16,550 miles driven per year figure, and multiply by 6.4, you still get over 100,000 miles. Though San Franciscans may also drive less miles per year as well, since with traffic, you spend a lot more time on the road going a lot less miles.

But then, we can also go into their statistic basing on a single insurance agency, unreported incidents not factoring in, there's definitely a lot of margin of error here.

I'm glad we're talking about it critically though, and not just glossing over the figures.

One of the other potential points is something Tesla said about Geohot's garage-built car. That it's easy to build a car that performs correctly 99% of the time, but it's exponentially harder to build a car that performs correctly 99.9999% of the time. The number of miles may be less material than even I've focused on here.

Google likes to tout the mileage figure, but it ignores the times that it's human drivers take over. The reality is, driving on straight stretches of road is easy. 99% of driving is easy. Figuring out how to evaluate how safe you are in the other 1% of that time is perhaps an even more challenging problem.

You're comparing apples and oranges: reported accident != knocking over a traffic cone.

Obviously the scope of the data for Google's cars are still relatively small. But there's little difference between hitting a traffic cone and hitting a small child. And even if you dismiss the two traffic cones, it's not a rosy record. Certainly nowhere near as rosy as Google paints it each time they say "1.x million miles without an accident!"

I am asking, if anything, that we take a more critical look at the statement being made every time Chris Urmson opens his mouth. Because when he brags about the safety of his cars, he's ignoring or failing to mention these incidents where humans saved him from having a very, very bad day.

> But there's little difference between hitting a traffic cone and hitting a small child.

I agree with some of what you have said, but this is the most bizarre statement I have read today.

There are probably hundreds (or more) traffic cones hit for every small child hit by a car, just because there are way more cars per day passing close to traffic cones than to children.

I personally have had my car hit by others 5 times without reporting it to my insurance company (car was parked, hit and run), as the damage was less than my deductible, and I opted to just live with a few dents. Estimating the number of times a car collides with anything per mile driven is very hard.

Yes, that's true, from a reporting/statistical standpoint. I am okay with dismissing those two incidents when evaluating this record. There's a lot of leeway on human driver statistics. As I said above, I've seen 165,000 miles per accident and 250,000 per accident reported/estimated. I've also seen the claim that human drivers go 300,000 to 500,000 miles between accidents on average, but I heavily assume those are based on underreported figures.

My intention in comparing a traffic cone to a small child is the actual safety issue. If a SDC could hit a cone, it could also hit a small child. They're small road hazards which are not generally going to be mapped by the car in advance. I don't want to dismiss the fact that the SDC would've hit a road obstacle, just because in that case the obstacle was a cheap plastic object.

> I don't want to dismiss the fact that the SDC would've hit a road obstacle, just because in that case the obstacle was a cheap plastic object.

I think the point is that if you counted these kinds of incidents in human driver statistics, your miles per accident numbers would come way, way down.

2015 isn't the first year their Self-Driving cars have been on the road. (Gotta wonder how many accidents they would've caused in 2014!) I picked a quick link while I was trying to find sources for the parent commenter's request. But the report comes from Google, to the California DMV, so you have to look at the numbers and facts, not their PR text around it.

Urmson, of course, whitewashes Google's progress, it's his job. But if you look at the number of miles between safety-related takeovers, or these 'almost accidents', and compare to the estimate of human safety provided by insurance companies, it's hard to believe these cars are yet safe enough to trust on the road. The only reason they haven't been in an accident is because trained, professional humans have intervened.

Self-Driving Cars will probably, eventually, be safer than humans, but Google's PR team has misled people on how far off that point is. Last year Urmson indicated that he wanted to have them ready for consumer use before his kid got his driver's license in four years.

That's pretty unrealistic. Don't believe little old random Internet commenter me? Take it from Robert Scoble, who's been hyping self-driving cars for the last year: https://medium.com/@scobleizer/don-t-worry-uber-lyft-drivers... He's been to see or ride in a variety of self-driving car prototypes from a number of companies working on the problem.

> it's hard to believe these cars are yet safe enough to trust on the road

Well yeah, but is anyone expecting them to be put onto the roads in the near near future?

Yes, which is the scary part. Google's PR seems to keep insinuating we might be able to ride in one as early as next year, or that the law is somehow a burden to the progress of this technology.

There's already people advocating that California's laws requiring self-driving cars be driven by licensed drivers are potentially discriminatory, and Google's already complaining that their statistically-more-dangerous-than-people drivers currently require licensed drivers on the road: http://www.bloomberg.com/news/articles/2016-01-28/self-drivi...

Given where we're at, it's insane to consider California's laws anything but reasonable. We're a long way off from the necessary confidence to allow a driverless car to run without human intervention. But Google's constant indications that this law and others like it are a problem insinuates a much higher confidence in their technology than any reasonable person should have.

Frequently in these sorts of discussions I've often seen people talking about 2018 or 2019 as likely consumer release dates, which seems to be the real incredible claim needing defense.

Another thing:

> Google [...] determined that the driver prevented the self-driving car from making contact with another object 13 times.

This phrasing suggests a notably lower threshold of failure than, say, "prevented a crash" or "prevented injury" or "prevented damage".

Yes, but the phrasing is also Googles and they have all the incentives to spin it that way.

http://www.dailymail.co.uk/news/article-3268421/All-Google-s...

Thanks to the human driver they managed to avoid the accident. /s

But in all seriousness the damn car was waiting at a traffic light and yet according to you it's still a terrible driver.

No, my consideration of the Google SDC's safety record is based on the number of accidents it would have caused (i.e. been at-fault for, from a legal standpoint) had human drivers not interrupted the program and taken manual control.

There's an additional likelihood that the odd/non-human behavior of Google's cars can make other drivers incorrectly predict their behavior, and also cause more accidents, but I'm not really factoring that into my consideration.

Reply on news.ycombinator.com