Last week I shared my 2 cents on the tragic crash of a self-driving Uber that killed a pedestrian.
Since then a lot of new information has come out and it doesn’t looks good for Uber”
- The Volvo safety system was disabled on the car involved in the crash.
- Experts were shocked that the system failed to recognize the pedestrian.
- Uber reduced LIDAR sensors on the Volvo down to 1 from 7 on the Fusion.
- Arizona has suspended Uber’s self-driving cars in the state.
- Uber will not renew it’s autonomous driving license in California, and it will have challenges to get it back.
- Uber is far behind its competitors in the autonomous driving race.
Given all that, it’s no surprise that Uber has reached a quick settlement with the family of the woman killed. The terms are confidential, but no doubt it was a big one to make this story go away as quickly as possible.
It seems like Uber was caught rushing into autonomous driving at a breakneck pace. The rules of “slow and steady win the race” are definitely not practiced at a company that had an internal motto of “Move fast and break things” “Always be hustlin’”
Perhaps it’s time for Uber to license Google’s Waymo superior autonomous driving technology before more damage is done?
Umm, ‘move fast and break things’ was Facebook (and they’ve since changed that to ‘Move Fast With Stable Infra’).
Uber was ‘always be hustlin” – which is taken by many as a symptom of their lame and even dangerous ‘bro’ culture.
Overall, it does look like your point is valid -they apparently moved too quickly here and cut corners on basic protection of human life. Very sad on many levels.
Thanks, mixing up my silicon valley mottoes here 🙂
As you same, the same principle applies though. They managed to kill someone after just 3MM miles on the road, a rate significantly worse than human drivers. They played a dangerous game to get ahead.
Dan,
From this human drivers kill every 500000 miles.
https://en.wikipedia.org/wiki/File:US_traffic_deaths_per_VMT,_VMT,_per_capita,_and_total_annual_deaths.png
1000000000 / 2000
What am I missing?
The US National Safety Council reports a rate of 1.25 deaths per 100 million vehicle miles.
It will be some time until people will accept a lower machine killing rate over a higher human killing rate. You just don’t view a computer killing someone as the same as a human being stupid or making a mistake.
Would make more sense for the NSC to report 1 death per 80 million miles. I mean, you can’t have .25 of a death (it’s like being 25% pregnant).
That would make comparing years a nightmare.
You can’t really judge based on 1 observation (and for sure say “significantly” if there is no statistical significance, which given the numbers above would be my guess). Deaths per miles traveled is not a measure of interval but frequency within a large sample.
Of course it’s not statistically relevant.
But still, that would be remarkably bad luck to happened in just 3MM miles driven if the cars were as good as humans. I’d posit that they are not.
But the woman crossing the street was also somewhat in the wrong. She wasn’t even looking to see if there was any oncoming traffic. I think this event is the culmination of many preventable errors. It’s funny that Uber settled with her family who seems to have been estranged. What a world we live in.
She was most of the way across. Yes, legally she would be in the wrong.
But clearly Uber was enough in the wrong that they quickly paid the family off.
I think that Uber is being held to a much higher standard than a regular driver would’ve been and I’m fine with that; regular humans don’t have lidar sensors. So it is disconcerting that the car didn’t even brake at all (from the articles I read).
It is a shame that everyone is pulling self driving now. Everyday I see dozens of distracted drivers missing traffic signals or hard braking because they can’t keep their eyes off their phones.
Enough in the wrong tends to imply to many that they’re admitting fault when settlements are many times done to simply get the association of negative press away from a company.
In most accidents the victim is partially in the wrong, but in most states that just goes to apportionment of damages. Ie if the injury is worth 1 million and the plaintiff was 20% responsible the defendant would pay $800K
Actually, legally a pedestrian in the crosswalk always has the right of way.
This was not at a crosswalk.
My .02
She saw the car and assumed that the human driver would slow down for her. As much as self-drivers need to adjust to the patterns and habits of human drivers and pedestrians, Pedestrians and human drivers need to adjust to the patterns of self-drivers.
Just curious what inspiring you to write about this topic on your Deal blog seams kind of random?
I’ve been writing about random topics for 14 years in between the deals. Keeps everyone on their toes 🙂
Dan’s been blogging about travel all along! Isn’t this in the ‘travel’ category?? 😛
Anyway Dan, keep ’em coming! We enjoy the variance (…and it broadens your readership…)
Chag kosher v’sameach!
It was declared that any driver would’ve killed that pedestrian. It was not the uber’s Fault.
It doesnt really matter what a human would do. Computers have a much longer vision range and can be more attentive. A correctly function driving AI can see as accurately in the dark as they can in the light.
Uber is right to shut down their self driving AI at this point. They tried to steal their tech from Google, and then they took shortcuts … clearly. This is quite scary for a company which is burning 1 billion dollars a quarter, this tech was meant to bring them into the green, but now that its gone, Uber may die soon, or be acquired.
As Daymond John would say, it seems like a licensing play makes the most sense here.
so for that reason, i’m out (not reading anymore about this issue)
not many facts are accurate here. Uber is not loosing one billion dollars a quarter, they have never had even a 3 billion dollar losing year. Uber has more (net and gross) cash than you apparently could even imagine, so they aren’t going to die for years to come, assuming all they do is make terrible mistakes, like worse than what they’ve done in the last 2 years. Now that they’ve sold off their China and South East Asian businesses, previously their 2 largest losing regions, they’re predictably close to losing a lot less, and if you remove legal costs, autonomous vehicle costs they’re probably net income profitable on a GAAP basis. So unless they mess up much worse, they’ll be a successful company for well than we will be alive. I’m not going to say that the valuation makes sense at reported prices, but that’s a different story.
Is there anyway to find out what the settlement amount was? Maybe free Uber rides for life?
Google’s Waymo is NOT superior for autonomous driving, unless you’re using uber as the baseline standard (which it shouldn’t be). Google’s tech for self-driving would be the slowest to get to the consumer
@Dan, the car in your photo is an innocent autonomous Ford Fusion plugin hybrid. Not the Volvo that killed a pedestrian. For a disclosure, I am a Ford employee.
@Dan: They don’t need to license Google’s technology, they already stole it 🙂
https://www.wired.com/story/uber-waymo-lawsuit-jacobs-letter/
But apparently even after stealing it, they couldn’t use it well enough to avoid this.
They decided to play the same strategy that they used to expand to new cities by ignoring laws and regulations, capturing market share, and then throwing lots of lawyers and “campaign contributions” at the problem when it comes to a head.
That might work when disrupting the taxi industry, but doesn’t go over well when you actually kill someone.
There’s 40,000 people killed in auto accidents every year in the US. If self-driving can save 30,000 lives even we need to get it asap. That’s besides the benefits to fuel economy, street and highway capacity, etc.
Uber may be at fault but this woman illegally walked from a median into traffic going 50mph. No driver could have stopped. If anything sensors can be improved to detect something like this way before a human can.
We need good self-driving asap.