Slashdot Mirror
Volvo Engineer Calls Out Tesla For Dangerous 'Wannabe' Autopilot System (jalopnik.com)
An anonymous reader shares an article on Jalopnik: Tesla's semi-autonomous Autopilot system has been impressing everyone from consumers to journalists, and even other industry experts and executives. But now a Volvo engineer has called Tesla's system out, claiming it's a dangerous "wannabe" autonomous technology. Trent Victor, senior technical leader of crash avoidance for Volvo, had quite a few choice words to say about Tesla's Autopilot system in a recent interview with The Verge, claiming the electric automaker was touting what is essentially a rudimentary semi-autonomous technology as being far more capable than it actually is. Victor fears that Autopilot "gives you the impression that it's doing more than it is." He went on to call Tesla's system an "unsupervised wannabe."
Tesla: "Thank You Volvo for the free publicity!"
Many will see this as market insecurity on the part of traditional car makers, and Tesla's stock will probably go up.
Table-ized A.I.
https://www.youtube.com/watch?...
Just like ULA hating on SpaceX's rocket landing plans or Blackberry hating on the original iPhone, whenever a newcomer comes to market with a disruptive technology, the entrenched players do all they can to trash the newcomer in the media to cast doubt on the viability of the disruptive ideas, rather than pivoting to actually address the market shift that the disruption heralds.
I'll just let autopilot drive for a while.
SJW's don't eliminate discrimination. They just expropriate it for themselves.
Humans do pay less and less to their primary task if their primary task involves little interaction. However, the spread of this technology is pretty narrow. Even if it turns out to be really bad, Tesla is probably not going to deal with more deaths than Toyota had to deal with from not nailing down their floor mats. Not to mention, even in autonomous mode current law puts the person in the driver's seat under responsibility. In the meantime, Tesla is learning all the things it needs to to become a leader in making the driver obsolete and in the end making ground transportation safer. Whoever the leader in this transition is going to have to eat the bad press when things go south the first time.
Considering several of Tesla's Autopilot project leads resigned in protest after complaining that Autopilot was dangerous and not ready for release, and we knew that several months ago, this isn't really news.
If he doesn't have statistical data to back up the assertion, then his assertions are meaningless.
The idea of handling back the controls to the driver whenever the car encounters a situation it can't handle any more is patently absurd. Human reaction time and situational awareness cannot cope with this, and in reality the purpose of this "level 3" concept is to quickly shift the blame back to the human driver rather than the car maker.
Even if the Volvo people are entirely right (and they might be) it doesn't really carry much credibility since they have a clear conflict of interest.
Yes Tesla engages in some puffery. On the other hand they are really doing the most innovative things in the auto industry so it's not entirely without substance. (of course the auto industry is pretty stodgy so being innovative is kind of a low bar) The Model S outsells every model in Volvo's lineup so clearly Tesla is doing something customers appreciate more than Volvo. Makes it sound a lot like sour grapes on Volvo's part even if it really isn't.
"If something goes wrong, the (Volvo) can safely stop itself at the side of the road." 'Cept if the wrong thing is the brakes fail, and I've got a few stretches of high country road I could introduce you to where you would definitely want human judgement involved concerning what you would generously term "the side of the road".
"Win treats sysadmins better than users. Mac treats users better than sysadmins. Linux treats everyone like sysadmins."
Volvo has the tradition of over-engineering. Perhaps they just noticed how hard the problem is to solve perfectly, or at least near to their standards. Or they are making themselves the Nokia in the equation disparaging_comments(Nokia) = success(iPhone).
A company who didn't have a car with anything else than a right angle and looked like a butter box until the end of the 90's.
"I've heard nonsense, compared with which that would be as sensible as a dictionary." Through the looking glass and what
Given that the news two days ago was about the new alliance of Tesla competitors, which includes both Ford and Volvo, I can't imagine why a Volvo engineer might be biased..
All 'autopilot' / 'self-driving' systems will be 'wannabes' for at least the next 10 to 20 years or more. It'll take that long to perfect it -- and even then you'll be nuts to let yourself fall asleep at the wheel with it operating -- no matter what they tell you.
Are YOU using the TOOL, or is the TOOL using YOU? Think about it!
Which is the typical Slashdot way, let me summarize for you guys and have a real discussion.
Volvo consider's Tesla's system a Level 3 system, which they consider dangerous. The reason they consider this is not for technical reasons but for liability and user control protocol reasons. What he's saying Tesla's system does is that in the event of an emergency, the system relinquishes control of the car to the driver, meaning the driver is now responsible (and has associated liability) for any accident or damage caused from the emergency. However Tesla's promotional material about their system is not suggesting that; their promotions are all focused on technical ability but it's ignoring the fact that if you're in an emergency the system basically tosses control to you, so his argument is that Tesla's promotion of the technology gives drivers and users the wrong impression of what it's really doing. So what he's saying is that if you're driving down the road and you fall asleep at the wheel and start to drift, the system will not take over and park you safely, and if in this process you hit a pedestrian or another car, you are liable for the damage caused.
Volvo's system coming out this year is what they call a Level 4. What this means is that the car does not require the driver to be in control at any time, even during extreme conditions. If you fall asleep at the wheel and start to drift, the computer takes over, gets you back in your lane first, then finds a way to pull you to the side of the road and park the car safely. They are so confident in their system that Volvo itself is taking on the liability of their system; if it fails to get you to safety or causes damage, Volvo is responsible, not you. Tesla does not make that claim about their system.
So his argument that the Tesla system which will force control to the driver in extreme driving conditions is actually more dangerous, because the driver is under the mistaken impression that the autonomous system has control and will suddenly find himself thrown into a difficult situation with no autonomous supervision; that surprise will invariably find the driver unable to make rational decisions.
All you Tesla Fundamentalists should stop with your Musk-worship and brand loving and realize that one of the most experienced safety engineers in the world is making a technical and design process argument. Argue his technical merits or the arguments he makes about Tesla's system and it's design decisions; don't just drink the Elon-Aid and dismiss because of the Gee-Whiz factor.
The Model S outsells every model in Volvo's lineup so clearly Tesla is doing something customers appreciate more than Volvo. Makes it sound a lot like sour grapes on Volvo's part even if it really isn't.
Uhh... ok. So I get you're a Tesla fan, but a comment like "the Model S outsells every model in Volvo's lineup" is not an opinion, it's fact that can be checked and in this case is a lie; Volvo had 3 models that outsold the Model S and their total car sales was 10 times Tesla's sales.
Model S sales, 2015 - 50,580
Volvo sales, 2015:
XC60 - 159,617
V40 - 83,357
V60 - 51,333
Total Volvo Sales of all models for 2015: 503,127
It's fine if you're a Tesla fan-boy, but please get your facts right and stop spreading FUD. It makes you sound like you're sour-grapes about a real car brand providing a real critique of your fab-boyishness and doesn't help your argument.
I had no idea Volvo still existed.
Just some basic information about what they mean when they talk about level of vehicular autonomy:
NHTSA defines vehicle automation as having five levels:
No-Automation (Level 0): The driver is in complete and sole control of the primary vehicle controls â" brake, steering, throttle, and motive power â" at all times.
Function-specific Automation (Level 1): Automation at this level involves one or more specific control functions. Examples include electronic stability control or pre-charged brakes, where the vehicle automatically assists with braking to enable the driver to regain control of the vehicle or stop faster than possible by acting alone.
Combined Function Automation (Level 2): This level involves automation of at least two primary control functions designed to work in unison to relieve the driver of control of those functions. An example of combined functions enabling a Level 2 system is adaptive cruise control in combination with lane centering.
Limited Self-Driving Automation (Level 3): Vehicles at this level of automation enable the driver to cede full control of all safety-critical functions under certain traffic or environmental conditions and in those conditions to rely heavily on the vehicle to monitor for changes in those conditions requiring transition back to driver control. The driver is expected to be available for occasional control, but with sufficiently comfortable transition time. The Google car is an example of limited self-driving automation.
Full Self-Driving Automation (Level 4): The vehicle is designed to perform all safety-critical driving functions and monitor roadway conditions for an entire trip. Such a design anticipates that the driver will provide destination or navigation input, but is not expected to be available for control at any time during the trip. This includes both occupied and unoccupied vehicles.
The debate about "autopilot" versus "fully autonomous" is a very real concern, validated by Tesla drivers themselves. You have drivers that stop paying attention to the speed limit, abuse autodrive to violate traffic laws, take their hands off the steering wheel, or just climb into the back seat and let the car drive itself creates not just a danger for the Tesla driver but for every car on the road. This despite Tesla's insistence that people must still stay at the wheel and drive; the technology has advanced enough that people get a false sense of confidence to push the limits even if the technology is not truly ready for it. That's the point that the Volvo engineer is making.
Even if the Volvo people are entirely right (and they might be) it doesn't really carry much credibility since they have a clear conflict of interest.
So you think we should discard information that is actually correct because of conflict of interest? Ridiculous.
Opinions, sure, consider the source. Facts are facts.
I'm not trying to be a Musk fanboy here, and I've looked up the Wiktionary definition of "innovator", but I wonder if the term isn't misused, or being defined incorrectly as compared to its actual popular usage. It seems to simply be a synonym for "inventor".
Obviously, Musk is not an inventor. He didn't invent any of the technologies he pushes.
However, he does popularize these technologies and help make them mature and sellable through both his financing and leadership. I don't know what the exact word is for that, but it's something. It's not that different from what Steve Jobs did; Apple didn't invent the smartphone (or the portable music player before that), but it certainly played a big role in making it as popular as it is now.
let us try on you then, you can stand infront of the car for 1000 tries and see if you survive and hopefully when accident strikes we can make the system safer for next person
The Model S outsells every model in Volvo's lineup so clearly Tesla is doing something customers appreciate more than Volvo
Volvo's customers have fourteen models to chose from.
Volvo sold 52,279 cars in March alone, up 12% from last year. Volvo Car Group Retail Sales By Car Model - March 2016
Volvo is serious about autodrive, with full autonomy coming soon, possibly as early as 2018. But there is no question that Volvo is building some very good cars right now.
Crashworthiness:
Small overlap front G
Moderate overlap front G
Side G
Roof strength G
Head restraints & seats G
Crash avoidance & Mitigation
Front crash prevention Superior [6 Points]
Low speed autobrake 2 Points
High speed autobrake 3 Points
Child Seat Anchors (LATCH) ease of use A
2016 Top Safety Pick - 2016 Volvo XC90
Volvo... yeah, the company that makes wannabe SUVs and no longer leads the industry in ANYTHING but hype. Glass houses...
There seems to be a lot of disagreement lately about what constitutes 'AI' and what constitutes 'automatic'. For me, 'automatic' should mean hands-off, you-don't-even-have-to-be-there automatic. In any weather, on any surface, with zero liability. I have this same standard for self-driving cars. People want to cheer the oncoming of automation and don't seem to see the serious legal issues that lie ahead.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
There are cases where the "upstart" tech just has it so figured out, and the entrenched players are so clueless, that it's an easy bet where things are heading.
It was pretty easy to figure out on day one that the iPhone was going to be fantastically successful, essentially launching a smartphone revolution. The first version wasn't perfect, but all the imperfections were pretty easily fixable (no native third party apps, slow Cell data, no native GPS, no support for push email, no way to access my work email, required connecting to iTunes to load content, etc.) and all the hard stuff was pretty much perfect (battery life, form factor, screen, music, video, web browsing, fit & finish, no locked down crapware, and just generally well-designed and targeted at a broad consumer audience). It was so overwhelming better than nearly every other phone that I would just laugh when people said "this thing's going to fail".
Now it's very hard to predict which competitors will catch up and pass the upstart tech, and a company can always stumble and screw up--but, come on, it was pretty clear the iPhone was going to be conquer the world for at least some period of time.
Fastforward to now, it's clear that electric cars are going to conquer the world. There are going to be winners and losers in that industry which means that some of the entrenched players are going to be tomorrow's Blackberry.
PS: SpaceX also has it figured out, the writing is on the wall for ULA.
An autonomous driving system doesn't have to be perfect, it just has to be safer than human driving, which is not terribly difficult as it turns out.
There are millions and millions of miles being put on Tesla with autopilot as we speak, and statistics works. If that system is actually less safe than human driving then the statistics will bear that out--thus far it is showing the opposite.
Volvo's customers have fourteen models to chose from.
So what? The Model S out sells most of those models and most of them sell at price points far lower than the Model S. The Model 3 gives Tesla 3 models on the market and it seems pretty likely that the Model 3 will be a brisk seller as they have gobs of pre-orders (about 3/5 of the total annual volume of Volvo). Volvo has been around for decades and yet still is a minor player somehow. They've got some fine vehicles but they aren't doing anything that all the other major auto makers aren't working on too in some form or fashion. They've carved out a nice little niche in well made, safe, grocery-getters and they seem to have a hard time thinking much beyond that. Their efforts in automated driving seem aimed primarily at adding bolt on safety features, rather than actually replacing the driver. They are padding their niche rather than trying to develop something new and more advanced.
I was looking at some erroneous data earlier - there are a few Volvo models that worldwide outsell the Model S. The numbers I was looking at were year-to-date numbers rather than full year.
Volvo has the tradition of over-engineering. Perhaps they just noticed how hard the problem is to solve perfectly, or at least near to their standards.
Yes they do. Which is a big part of why they are still a minor player in the auto industry. They make over-engineered, somewhat dowdy looking cars that appeal to a relatively narrow niche. Volvo, very much like Saab, has a really hard time going outside their engineering focused comfort zone. The notion of doing something radically beyond what is currently possible I think is an interference fit with their corporate culture.
Calling Level 3 semi-autonomous control autopilot is dangerous, as drivers can be mislead into thinking they do not need to pay full time and attention just in case the level three returns control to the driver .
If this was an industrial machine control with safety implications, it would be considered unacceptable.
Automation must always fail to a "SAFE" state. Control transfer to the driver can be considered when in a "SAFE" state, or NORMAL not during a attempt to fail into a safe state. The operator workload is too high and the risk of control transfer related failure is just tooooo high.