Doesn't matter how many lives the system saves, its the lives that it takes and resulting lawsuits that will eliminate this stuff.
Texas Instruments used to have a control division that made switches for autos. Turns out brake fluid or something dripped on them in ford cars and caused garage/car fires. They quietly and quickly sold off that division to Bain Capital after the lawsuits, never saying why. The Texas Instruments model 9F924 speed control deactivation switch. Same thing applys to Tesla/Mobile Eye, as Ford/Texas Instruments. after the lawsuits things usually change. Who cares about improving the technology.... your still liable. Mobileye should concentrate on something that doesn't kill people when it fails.
Maybe MobilEye should concentrate on selling to the military so that there are no lawsuits when their stuff fails.
Point is, we sure don't need to "wait" for Tesla or Mobileye, to witness multiple millions of accidents. Human drivers are more than capable of offering those up, every single day.
The only credible issue with autonomous driving is the tangential legal one. With human drivers, pockets are only so deep. So even though incompetence is rampant, there is only so much suing that can be done.
With autonomous driving, the deep pockets of automakers will become the lawyers' obsession.
Aside from that, there is no believable downside to autonomous driving being introduced in the coming years. It's not like autonomous driving has a hard act to follow. It's not like people are so expert at driving that machines can't easily do it better. Just the legal issues.
Even these problems you repeat, with ignition keys, not only pale with comparison to human incompetence, but are also convincingly fixable. Humans are not. That's the bottom line.
They have opened themselves up to liability from the end-customer, the drivers who they will have killed.
Airbags, Ignition switches etc can't even be designed right. Look how much they have cost their respective companies. Those are just passive devices meant to do one simple thing. And now we have a company that takes over the entirety of driving. Somehow I think more people will die and more lawsuits will be filed.
Mobileeye must beleave that Tesla has put them at higher risk, especially the fact that you can still drive for minutes with no hand on the wheel despite the car giving warnings.
Waiting for more deaths and lawsuits.....just a matter of time. The more Teslas with autopilot on the road the more likely....
I would think in cases like this one, if Mobileye clearly described the limits of their vision system, e.g. in terms of vertical resolution, it would be to their advantage to only be seen as an equipment vendor for Tesla. Otherwise, if they had been a more equal participant in the design of the autopilot system, they would risk being sued as much as Tesla.
I dunno, this all seems like to-be-expected posturing. The facts in this case seem quite straightforward to me. I'm sure that sooner or later, we'll witness something baffling, in the autonomous driving arena, but this accident ain't it.
Tesla may well be like Apple or Microsoft, who want to own everything needed for their products. If that's the case, though, this accident would at best have been the catalyst to accelerate the process.
I think we will see more such conflicts as semiconductor companies attempt to move up into vertical markets. MobileEye started in chips, but quickly evolved into a computer-vision system company, even though they also sell chips to key customers. Since Tesla is the leader in the adoption of ADAS, the company is probably not thrilled about helping a supplier who can quickly enable competing car companies to catch up. As chip vendors try to compete in these automotive vertical markets, the market power will likely remain with the companies that actually build the complete automotive platform. Soon that club will include Apple and Google, but it remains a short list of potential customers for the "chip" companies that are dealing with the continuing trend toward vertical integration.
To be more specific, it's not entirely new to see OEM's developing their own silicon. Apple is doing it and before them Cisco has been doing it for a long time. What's new is that OEMs are now developing their own processor architectures.
Really interesting article and comments. Looking at the list of people who joined Tesla recently, Tesla is obviously developing its own processor for autonomous driving. This follows Google's announcement that they have developed their own processor for big data analytics. It is very interesting to see that the OEMs are now doing their own core components in order to differentiate. We went from IDM to fabless semiconductor companies and now we are entering a new phase in which Semiconductor companies are not needed anymore. OEMs will go straight to foundries. Interestingly, ARM or Intel may not be needed either as these OEMs develop their own architectures. Because such processors are specialised for a certain domain, they don't have to carry a huge software legacy so they don't need to keep X86 or ARM compatibility. I am less and less sure SoftBank made the right bet.
Sounds like tesla went too far and opened up liability to mobileeye.
Its all about the lawyers and liability now because the driver can't be blamed if the machine is in control.
I don't understand how their autopilot keeps driving without hands on the wheel and just gives warning beeps and large time delay before it shuts off. Why not immediatly shut it off if no hands on wheel? I guess because no one would use it except for bumper to bumper traffic.