How many incidents have human driven cars had in the same time period? I know the scale is different, but 3 is a pretty low number for the number of hours they’ve racked up.
I just wish we could have less cars overall and more bikes, scooters, trains, buses, etc.
I work directly with the employee who submitted the complaint. He is understating the problem.
The letter also states that Cruise ... intentionally hides from the majority of employees the results of investigations into collisions involving Cruise vehicles and other sensitive, potentially damaging matters.
Cruise has been covering up safety incidents since before the GM acquisition, including an early incident where an engineer suffered a concussion due to violent cabin motion caused by bad motion control code.
I inserted the details into my comment to save you the trouble, and accidentally didn't edit out that sentence. I have to be careful not to reveal my identity by saying too much.
I'd assume that the cars see significantly more utilization than a normal passenger vehicle. An autonomous vehicle doesn't have a driver that gets tired, so in theory it's possible to operate it 24x7, minus the downtime for repairs, cleaning, and charging.
A better metric would be incidents per hour driven, not per vehicle.
There's lot of ways you can measure it to make it sound better.
However it comes across as a sort of centralized future cyberpunk failure mode.
Again we are talking about 3 of the 30 vehicles they have on road failing at the same time.
Imagine if 10% of taxis in your city suddenly froze in place for 20 minutes until a special tow truck came after a pedestrian called the number on the back?
Or if all taxis in your city, when traveling through a particular intersection froze in place, requiring bystanders to call a response team from the taxi dispatch to come out and show the driver what to do?
It would be interesting to see the metrics of number of interventions/calls/responses Cruise is experiencing daily/weekly/montlhy.
That has happened already, although for human driven vehicles. Yandex taxi was hacked earlier this year and all taxis were made to go to a certain location, causing massive traffic jams.
In practice, if companies have these kinds of issues regularly, it'll be similar to how scooters and dock free bikes have been rolled out: they'll be allowed initially, then the city officials will kick them out if they're causing more trouble than they're worth.
Idk if it is bad. The way the article headline is worded it sounds like the cars are crashing into puppies and grandmas. In reality they just caused traffic to get backed up for about 20 minutes because a sensor broke or something. This seems like a misleading headline by kron4.
I think techies make a lot of excuses for tech that just isn't even remotely there yet. Driverless cars is very much that category of tech.
If 10% of the Toyota Camrys on the road in a given city suddenly froze in place and caused traffic jams simultaneously in a major US city, requiring roadside assistance, we'd all be laughing at those incompetent dinosaurs.
Not remotely there? It already drives better than very old and very young drivers, as well as people who are drunk or tired. At least this tech should get our worst drivers off the road.
If only there was some sort of math “journalist” can use to make the article less sensational that takes into account of other evidence like how many hours it drove and how many cars there are on the streets driven by humans. We can even call them priors!
I mean, how many driverless cars are there even? You can just compare absolute numbers. I can't believe I even have to bring it up on here. You mentioned yourself the number of hours. Are we comparing it to the total number of hours for ALL the cars in the same area?
There clearly don't seem to be too many of these cars.
> UNITED STATES—A self-driving vehicle created by Cruise, a self-driving arm of General Motors Co., crashed into another car in San Francisco, leading to the recall of 80 more vehicles. One person who drove the car was arrested for suspicion of DUI. The algorithm that controls Cruise’s cars is reportedly capable of counting pedestrians and cyclists and detecting steel guardrails.[1]
Just found another somewhat amusing link as I was looking for the number of cars. Here's two cruise drivers blocking the street. I guess they deadlocked[2].
Right and I think if driverless taxis ever work (they won't outside small geo fenced localities & narrow use cases) only incentivizes oversupply.
In a taxi or even perma-loss-leading Uber/Lyft, there is a labor component which disincentives between-ride transient driving. Whether they are paid employees costing company money, or independent contractors looking at hours worked vs take home pay..
In a high fixed cost, low operating cost driverless taxi this is not the case. Very easy to imagine fleets of these stupid things clogging the streets even worse than Uber ever did.
Note NYC traffic + TL&C stats already show Uber/Lyft have always had much lower utilization rate per hour/mile than yellow cabs, and thus require more cars on the road to serve the same passenger miles.
I just wish we could have less cars overall and more bikes, scooters, trains, buses, etc.