Hacker George Hotz Unveils $999 Self-Driving Add-On (pcmag.com) 80
An anonymous reader quotes a report from PC Magazine: Hacker George Hotz is gearing up to launch his automotive AI start-up's first official product. In December, the 26-year-old -- known for infiltrating Apple's iPhone and Sony's PlayStation 3J -- moved on to bigger things: turning a 2016 Acura ILX into an autonomous vehicle. According to Bloomberg, Hotz outfitted the car with a laser-based radar (lidar) system, a camera, a 21.5-inch screen, a "tangle of electronics," and a joystick attached to a wooden board. Nine months later, the famed hacker this week unveiled the Comma One. As described by TechCrunch, the $999 add-on comes with a $24 monthly subscription fee for software that can pilot a car for miles without a driver touching the wheel, brake, or gas. But unlike systems currently under development by Google, Tesla, and nearly every major vehicle manufacturer, Comma.ai's "shippable" Comma One does not require users to buy a new car. "It's fully functional. It's about on par with Tesla Autopilot," Hotz said during this week's TechCrunch Disrupt in San Francisco.
Did they think of this in the hotel suite? (Score:4, Funny)
Between Eric Bachman's hype and Gilfoyle's latent satanic easter eggs I think I'll wait to see if this actually works. But it's nice to see them beating Hooli.
Uses onboard radar (Score:5, Informative)
The price is possible because it is not a complete system; per TFA they are using (some might say leveraging) onboard radar. That means that unless your car already has a sufficiently useful radar (e.g. it has adaptive cruise control and/or automatic emergency braking) you're not going to be able to retrofit at least the first generation of this system without taking heroic steps.
That's a shame, because it would be really nice to be able to put this into some of the vehicles made in the late nineties, after basic vehicle technology approached the current state of the art but before practically every car sprouted a remotely hackable infotainment system.
Re:Uses onboard radar (Score:5, Informative)
The list of supported cars is currently even more limited: https://commaai.blogspot.de/2016/09/comma-one-supported-cars.html
Re:Uses onboard radar (Score:4, Interesting)
I have an '07 S80 V8 with adaptive cruise and it's my favorite feature. 2007 was kind of early in the adaptive cruise era but it works well, down to 20 mph but won't do a complete stop, though the collision avoidance system will pre-charge the brake pressure.
The problem with adding the hardware, even with a modern Volvo that offered the feature in the same model as an option is that almost nothing in a Volvo can be added without VIDA installing software. My S80 spent 4 days in the shop when I added the iPod connector because they had to load a custom software patch to make the component work.
I think adaptive cruise would be about the sweet spot as a "tack on" -- simple enough to patch it into the existing cruise control system. But lane control and more advanced autonomous stuff would be a lot of mechanical additions, not counting software or electronics even.
Re: (Score:3)
It taps into the vehicle's CAN bus and reads the sensor data from there, then sends commands out to control steering and acceleration/braking.
It will only work with cars that have sensors and steering control (usually for auto parking), and where there is a way to tap in.
Re:Uses onboard radar (Score:5, Interesting)
No, it's not a shame because his system is very dangerous. He built it using reinforcement learning based off his personal driving patterns. Meaning he drove around in training mode and it watched what he did. Then in running mode it copies what he did in similar situations. So that means you can't count on it in any new or lesser seen situations. Any event that didn't occur often enough during training, like a bird flying infront of you, can cause unexpected behavior by the AI. This type of system is inherently unsafe unless you train it against everything. You need to train both what to do and what not to do.
How often has he had someone swerve into his lean or cross in front of him? How many times did someone almost do that? Unless you specifically train for when that happens, this type of learning system could learn to see the natural drifts of people driving between the lines as a hard rule and then freak out when someone suddenly or slowly mergers in front of you. If most pedestrians it sees walk up to the edge of the road and then wait the car passes, it'll never learn to slow down for them because it never saw past that waiting point. Thus when someone does step into the street, you'll run them over before you realize the AI isn't going to stop for them.
There was a research paper who's algorithm made a massive improvement in identifying cars from photographs. Well they thought it did until other people looked closer and figured out all the images with cars were taken later in the day and all the non-car pictures were taken earlier in the day. The AI was only checking the sky color and using that to determine if a car was in the picture or not. Excellent in the training result, horrible real-life performance. The way he trained his system allows this same type of issue to occur.
Re: (Score:2)
No risk, no fun.
Re: (Score:2)
Thinking logically != being a Luddite. Endangering ones life is one thing (but is often selfish), endangering the lives of willing passengers another and endangering anybody near a road yet another. Unless the thingy is 98%+ safe it _is_ 100% crap for normal purposes.
99% safe = death trap (Score:5, Informative)
Unless the thingy is 98%+ safe it _is_ 100% crap for normal purposes.
"98%+ safe" is incredibly unsafe for an automobile presuming you are using any reasonably standard measure of safety like deaths per 100 million miles traveled [caranddriver.com]. That means that it would get into an accident once every 50 miles! For reference current human driven vehicles in the US experience 1.13 deaths per 100 million miles traveled. That is 99.99999998% safe by that measurement. Any automated driving system will have to beat that number and beat it by a lot.
People tend to think that saying something is 99% successful is a good thing but in reality that can be a terrible outcome. A vehicle that was 98% safe under any reasonable measurement would be immediately and rightly labeled a death trap.
Re: (Score:3)
I keep seeing those ATT internet commercials where they brag about 99% reliability as if that's a good thing. I can only think: that means my internet service is down more than 7 hours a month. That's about right, in my experience, so at least you can't accuse them of false advertising. But 7 hours of downtime a month sucks, especially as it usually seems to happen at the most inconv
Re: (Score:1)
I think the infotainment system is less relevant than the need for electronic driven controls like throttle, braking and steering. At the very least is should significantly reduce cost of implementation.
$999 Self Driving Car is really driven remotely (Score:1)
come on .. you know this thing is just being driven by some guy in India working in a call center :o)
Anyone found a link to his "I want a Holodeck"? (Score:2)
At this point i'm almost convinced that it was a fake wiki edit because i've never been able to find any pictures anyone else have any better luck?
forget about it. (Score:5, Funny)
I got a '75 Monte Carlo that I've been restoring and I look forward to making it self-driving so I can send it to get me a fucking gelato, a fifth of Johnny Walker and a carton of Chesterfields.
Re: (Score:1, Funny)
Seeing has how those old ass cars have a hard time going straight with a sober driver, I don't think making it self drive would be better...unless your a drunk with the car.
Re: (Score:1)
How many living rooms do you have that you would need more than one Chesterfield?
Liability? (Score:4, Interesting)
Going cheap can get very, very, VERY expensive, very quickly.
Re: (Score:2)
He looks like he's insane.
Re: (Score:2, Funny)
Does he indemnify the users in case his tech screws up and kills someone?
No because the driver is supposed to remain alert and watching the road and so remains responsible. Besides if it was going to end up spilling lots of blood they would have called it the period, not the comma.
Re: (Score:2)
Product liability (Score:2)
No because the driver is supposed to remain alert and watching the road and so remains responsible.
That is NOT how product liability [wikipedia.org] works. He would be liable for any reasonably foreseeable use of the product and even for negative outcomes he could not foresee. It may not matter at all if he warns the passenger to remain alert. Most likely any case would be held to the strict liability [wikipedia.org] standard which means intent to harm is irrelevant.
Re: (Score:2)
liability waiver does not cover 3rd party victims
Re: (Score:2)
Terrorists must be estatic (Score:5, Insightful)
$999 for a homing missile they can load up with petrol and drive into a civic center.
Re: (Score:1)
Re: (Score:2)
Yep.
Between this and drones, I'm just waiting for the inevitable.
Even if they only run misdirection, you could make people petrified and get them banned overnight.
The same way that because of something that happened 15 years ago, we still can't carry a can of Coke onto a plane.
Re: (Score:2)
Re: (Score:1)
related xkcd [xkcd.com]
Re: (Score:2)
$999 for a homing missile they can load up with petrol and drive into a civic center.
They could already have made a car radio controlled and driven it into a civic center, and they haven't done that yet. It's also not going to retrofit into cheap cars, so they have to steal a car and then get the self-driving system working. It's probably cheaper just to turn someone into a suicide bomber, especially while we're spending the money to bomb them into radicalization.
Re: (Score:2)
$999 for a homing missile they can load up with petrol and drive into a civic center.
Uhhhh, as attacks have shown for quite a long time, terrorists have little issue with committing suicide in order to inflict harm.
Removing the driver is hardly necessary.
Re: (Score:1)
Yes, but this gives them multiple chances: It's a suicide bomber's dream come true!
Re: (Score:2)
Hack (Score:2, Troll)
GeoHotz isn't a hacker, he's a hack. Stay far away.
Manufacturer's Liability Insurance (Score:2)
Sorry (Score:1)
Even if I wanted something like this - "Monthly fee" = "Piss Off"
Re: (Score:2)
Comma? (Score:2)
I think he misspelled coma, as in what you will be in after using this system ;-)
fail (Score:2)
"It's about on par with Tesla Autopilot" (Score:3)
Re: (Score:1)
This boy in a basement have defeated plenty of security measures from large corps who have spent hundreds of millions to protect their technology. Ever heard of a bunch billionaires who started from their basement/garage ?
Re: (Score:1)
It's much easier to find flaws with others stuff than to build something (even with flaws) yourself.
Re: (Score:2)
This boy in a basement have defeated plenty of security measures from large corps who have spent hundreds of millions to protect their technology. Ever heard of a bunch billionaires who started from their basement/garage ?
And that qualifies him to design (ha!), develop and build a mission-critical life endangering system how? Please explain in great detail... how. (I work on mission-critical life endangering systems every day. I look forward to your explanation.)
Also, those basement dwelling billionaires were not building 'high-speed rolling death on wheels.' They built payment-systems and office products. Kinda' big difference there.
Re: (Score:2)
Fuck your subscription model (Score:2)
As Reported By TechCrunch (Score:2)
subscription fee? (Score:1)
People will die. (Score:2)