First the good news. It looks like there was no one injured that I could see at the time I passed it.
I live and work in an area where I see dozens of Teslas every day, so I think it's delightful to see more Teslas. The fact that I happen to see some involved in crashes is therefore inevitable.
It looked like the forward car in a three car crash in the left lane of Hwy 1 Northbound after the Park Avenue on ramp, because it was in front and its rear was very smashed and its front was fine. I passed the cleanup stage about 10 minutes ago.
My initial progressive reaction is to hope we aren't waiting for optical recognition driver assist to do more than it can given the current corporate development setup. Tesla might consider swallowing its pride and look into more domestic teams (Stanford, Google, Apple) for superior driver and autonomous abilities. But all that aside, I hope everyone continues to be OK, and hope their insurance options were selected properly and perform well.
The issue we have is driver assisted technology is desperately trying to catch up to driver inattention. According to a recent wired article there has been a significant reduction in claims if you compared cars with driver assisted technology (adaptive cruise control, etc) versus regular cars. So even if occasionally this technology might cause an accident the trade off is it will prevent a lot more accidents from happening.
Driver inattention is rampant with any car. I do think it will get even worse, and it might be REALLY bad with cars with driver assisted technology. Sure I know the limitations of it, but lots of people might be get tricked by Volvo, MB commercials into thinking it can do more than it can. Like what happened with the Volvo where the driver thought they had some package but didn't.
So on the one side you have people who can't pay attention to anything but their phones, and on the other side you have cars with extremely limited driver assisted technology.
So you can try to change peoples behavior while at the same time seducing them with driver assisted technology. Like just what am I supposed to do when an auto-steering enabled AutoPilot Model S is driving me down the freeway? I can make lots of claims that "I'll be paying attention" or "I know better", but the claims won't hold water. I'll be enjoying the scenery or checking google news. Then my car will run into debris in the road that the car didn't see.
Or you can try to fully automate cars.
How do you fully automate cars in short order? You take away peoples privacy.
The biggest problem we have is how people view there cars. They view it from a freedom and entitlement point of view. That they are free to travel the country without anyone knowing where they are going. That's a nice fantasy, but the reality is pretty harsh. Not only are they being tracked (license plate readers are everywhere), but they are tracked on their phones as well.
The reality is we can't really have full automation without lots of tracking. Lots of entering in your destination and the car working with the road system (traffic lights), and other cars to schedule it all out. A database needs to know where cars are at all times. It also needs to know where bikes are and ideally where pedestrians are.
It can't be just optical recognition, or multi-sensory recognition. It's needs to be actual communication identification car to car. Where it's reporting not only what it's currently doing, but what it's intentions are. They also need to report anything on the roadway that shouldn't be there.
It needs to be an absolutely massive deep neural network.
To design a car capable of driving like a good driver is a long, long ways into the future. That's asking it to do too much. You have to simplify it by giving it more information.
- - - Updated - - -
Why are those better than Tesla's? Apple has zero cars, so naturally zero accidents. Google's cars have been in a dozen or so accidents--none were the Google car's fault, but then we don't know if this accident was the fault of the Tesla either.
I don't know if I really trust googles reporting system. I don't have a problem believing that they haven't caused an accident because they're not fully automated. They're driven by people who know exactly what it's capable of.
I'm going to stick with MB/Audi/Tesla until Google/Stanford allow an idiot into their cars.
