-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
How long until a Autopilot accident is reported and the potential public backlash?
- Thread starter meltoots
- Start date
I wonder how people got used to oldschool cruse control... you always need to monitor the speed of the vehicle infront of you and brake when he brakes. Horrorz.
I wonder how people got used to adaptive cruse control... you always need to monitor the vehicle infront of you not to run into a stopped vehicle or traffic lights. Horrorz.
I wonder how people got used to turning the steering wheel ... you must somehow determine how much to turn the wheel without seeing your front wheels and all. Horrorz.
Change is a terrifying thing.
I wonder how people got used to adaptive cruse control... you always need to monitor the vehicle infront of you not to run into a stopped vehicle or traffic lights. Horrorz.
I wonder how people got used to turning the steering wheel ... you must somehow determine how much to turn the wheel without seeing your front wheels and all. Horrorz.
Change is a terrifying thing.
I had a friend once who's father was one of the most senior pilots at American Airlines. He could have flown anything in their fleet, but stayed with 727s because you actually had to fly them. He never used autopilots. He was always willing to let the 1st officer fly, but only hands on. He had pretty much the same opinion that if you weren't flying the plane, you had a tendency to let your mind wander and not pay attention when things start to go pear shaped.
As I understand it you engage it like regular cruise control. In other words it's off until you turn it on.
Humans do have the ability to improvise more than computers do. A computer can only handle situations it's been programmed to handle. If done right, that covers 99.9% of all situations, but the people are there for the 0.1% the software designers didn't think of.
The 0.1% I'm talking about are situations way so outside normal normal driving situations the computer would never learn anything about it until it was too late. Something really out of the box like an airliner landing on the freeway or and earthquake dropping a span of the bridge in front of you. The Bay Bridge in the 1989 Loma Prieta quake had a section of the top deck drop. An autonomous car may not detect it and would drive into the abyss. Programmers in the Bay Area might think of that specific scenario because that particular problem happened in their backyard, but there if you google weird car accidents, you'd probably find some one off situations self driving cars would not have been able to handle it. And the thing about one off situations is that it's difficult to predict what the next one off is going to be.
Humans are generalists and computers are literalists. Computers work much much faster than we do so they can run through far more scenarios end to end faster than we can generalize our way through something most of the time, but in a complex situation with variables that were unanticipated, humans will tend to make the right call more often than a computer which gets confused with undefined input. Computers will have to weigh all the input all the time whereas a human will triage the unimportant details and zoom in on the critical bits.
Tesla makes heavy use of robots in their factory, but they also have humans doing tasks the robots are not well suited for. Computers have strengths and humans have strengths.
Last edited:
Amazing to me that people think pedestrians, which Elon says "It should not hit. Hopefully," drivers of other cars, passengers, and others have somehow relinquished their rights against a car company who puts beta software in the hands of owners which includes the small percentage of idiots (who are now doing things like driving from the back seat, dozing off, reading books, and hands off wheel) and says "have a nice day. "don't be naughty and turn this on when you're not suppose to, you lil devils!" You're wrong. Not paid my $500 an hour as a trial attorney with 35K of trial experience who's never lost a case, ever, to debate on the Tesla board, so instead, simply watch what happens.
Tesla "Beta" AP, doctrine of foreseeable misuse (among others), icey roads, blizzard, around a school getting out with some nitwit owner who turns the system on after told not to, then reads a book. Yup. Juries will NOT blame Tesla for a single buck, after legal discovery reveals the even more insane things Elon has said about AP than what he's been quoted in saying in the media. Yup. After Elon is forced to take the stand and promptly puts his foot in his mouth but keeps talking, because, well, that's what Elon does, Jury will go 'BAD dead evil school children! GOOD Tesla billionare and beta software. Verdict for Tesla! Sure.
If you don't believe that sticking beta software in cars and telling owners not to be naughty and activate it "except where intended" and not have sensors requiring hands on wheel or even disable it when nobody's in the front seats (see youtube) is not going to set AP, autonomous driving, and Tesla back, you're certainly entitled to your belief, but that makes one of us. I want Tesla and autonomous driving to have the best future possible, which I don't find benefited from bad decisions followed by crazy billionaire comments to the press. If google CEO had ever said to the media. "Our google cars shouldn't kill pedestrians. Hopefully." He'd have been forced to resign the next day for being so naive how that comment would be perceived even before the inevitable lawsuits.
To answer the thread, how long until something bad happens after release of 7.0, it might be a while. The numbers of AP OTAs that were delivered simply cannot be that high, so not many of them on the road, and I expect that 7.0 and 7.1 are VERY good, and that a vast majority of owners will indeed use it responsibly just hardly leaves that many AP activated cars in the hands of complete idiots who will turn it on when they're not supposed to, then doze off or read a book.
Tesla "Beta" AP, doctrine of foreseeable misuse (among others), icey roads, blizzard, around a school getting out with some nitwit owner who turns the system on after told not to, then reads a book. Yup. Juries will NOT blame Tesla for a single buck, after legal discovery reveals the even more insane things Elon has said about AP than what he's been quoted in saying in the media. Yup. After Elon is forced to take the stand and promptly puts his foot in his mouth but keeps talking, because, well, that's what Elon does, Jury will go 'BAD dead evil school children! GOOD Tesla billionare and beta software. Verdict for Tesla! Sure.
If you don't believe that sticking beta software in cars and telling owners not to be naughty and activate it "except where intended" and not have sensors requiring hands on wheel or even disable it when nobody's in the front seats (see youtube) is not going to set AP, autonomous driving, and Tesla back, you're certainly entitled to your belief, but that makes one of us. I want Tesla and autonomous driving to have the best future possible, which I don't find benefited from bad decisions followed by crazy billionaire comments to the press. If google CEO had ever said to the media. "Our google cars shouldn't kill pedestrians. Hopefully." He'd have been forced to resign the next day for being so naive how that comment would be perceived even before the inevitable lawsuits.
To answer the thread, how long until something bad happens after release of 7.0, it might be a while. The numbers of AP OTAs that were delivered simply cannot be that high, so not many of them on the road, and I expect that 7.0 and 7.1 are VERY good, and that a vast majority of owners will indeed use it responsibly just hardly leaves that many AP activated cars in the hands of complete idiots who will turn it on when they're not supposed to, then doze off or read a book.
Last edited:
You kinda need to chill and breathe. It's going to be okay. No school kids will die. Tone it down a bit.
This post made me lol a bit.
First, I like how this is at least the second time I've found on this forum that I've been throw in with the "idiots" for the very controlled stunt with autopilot in my video. Not going to re-post for a third time the safety checklist for that video, but suffice it to say you obviously didn't get the memo.
Second, the system was never claimed to avoid pedestrians. In fact it was specifically stated multiple times that the software does not have the ability yet to distinguish things needed for driving outside of a highway environment, like stop signs, lights, and pedestrians. Musk's comment hinted that there is some part of the software that will try to not hit a pedestrian, but that this is not something you should expect the car to do currently. Pretty sure most people got that memo as well.
So, if I'm driving a random Ford vehicle, and I set my cruise control in a school zone, and don't brake when kids are crossing the street, it's the car's fault when I hit them right? I mean, it was controlling the throttle at the time, so, definitely not my fault that I wasn't paying attention. Probably should put the CEO of Ford on the stand. lol.
Sounds ridiculous right? Here's the same sentence except with regard to my P85D to show that it still sounds ridiculous:
So, if I'm driving my P85D, and I set my cruise control or auto steer in a school zone, and don't brake when kids are crossing the street, it's the car's fault when I hit them right? I mean, it was controlling the throttle and steering at the time, so, definitely not my fault that I wasn't paying attention. Probably should put the CEO of Tesla on the stand. lol.
Good luck trying to defend that one.
It is true that people can file lawsuits for anything, but most jurisdictions in the US have a thing called summary judgement which comes into play with frivolous lawsuits. Basically the defense can ask the judge to throw out the case because it's a bogus claim and usually judges do when the claim is baseless. Additionally a lot of states now have laws sanctioning lawyers who bring frivolous lawsuits. They usually get off the first time with a slap on the wrist, but if they bring more they can see their license suspended or revoked. In some places there are fines for bringing frivolous lawsuits.
Someone may do something Tesla told them not to with AP and cause some accident, and they may sue Tesla, but if the instructions said not to and they did it anyway, it will likely get thrown out of court.
Someone may do something Tesla told them not to with AP and cause some accident, and they may sue Tesla, but if the instructions said not to and they did it anyway, it will likely get thrown out of court.
IF there is an incident, I suspect the angle would be Tesla didn't do everything possible to prevent the situation.
Hand on wheel, geo fencing being two obvious solutions, and are fitted to other cars with similar (or even superior) hardware from the same supplier (MobileEye). Certainly here (UK) this could be an issue if Tesla are deemed to have not undertaken all reasonable precautions. Yes I know it's nanny state stuff, but there are precedents based on omitting safety protocol/mechanisms which are readily available, but not implemented without thorough and explicit risk assessments (and those risk assessments may very well include operator error).
All of this is moot here right now as Tesla haven't yet got EU regulatory approval. With approval pending, I have to say, these videos of the system being used outside the design intent certainly won't help convince the decision makers
Hand on wheel, geo fencing being two obvious solutions, and are fitted to other cars with similar (or even superior) hardware from the same supplier (MobileEye). Certainly here (UK) this could be an issue if Tesla are deemed to have not undertaken all reasonable precautions. Yes I know it's nanny state stuff, but there are precedents based on omitting safety protocol/mechanisms which are readily available, but not implemented without thorough and explicit risk assessments (and those risk assessments may very well include operator error).
All of this is moot here right now as Tesla haven't yet got EU regulatory approval. With approval pending, I have to say, these videos of the system being used outside the design intent certainly won't help convince the decision makers
When you name a technology "autopilot", customers have new expectations compared to cruise control: namely that the car will pilot itself automatically. Elon made a mistake by using this word at this early stage of the development of this idealistic technology, especially since Tesla's approach is quite cheap: no highly precise mapping of the environment beforehand, no hardware AI, beta public release (a lot of consumer backlash in the video game industry).
He and other Tesla people made mistakes by talking about this tech unrealistically, and by demonstrating it outside of freeways.
Tesla can lose a lot in court. And in the public eye.
Google are at the same time much more ambitious and much more reasonable, because they understand human psychology, and take into account human weaknesses, and they also face a fundamental fact known by everyone who knows how to write software: computers are stupid. At least today.
He and other Tesla people made mistakes by talking about this tech unrealistically, and by demonstrating it outside of freeways.
Tesla can lose a lot in court. And in the public eye.
Google are at the same time much more ambitious and much more reasonable, because they understand human psychology, and take into account human weaknesses, and they also face a fundamental fact known by everyone who knows how to write software: computers are stupid. At least today.
lolachampcar
Well-Known Member
Tech types need to remember that a lot of Tesla's customers have no clue what a "Beta" release is.
Likewise, the term Auto Pilot means something totally different to a pilot who knows that the responsibility to see and avoid remains with the pilot in command. MS goes further in that it can sense items in its environment and avoid them which is way beyond the capability of most auto pilots. Again, a small group has a different understanding of a particular term then the population at large.
Likewise, the term Auto Pilot means something totally different to a pilot who knows that the responsibility to see and avoid remains with the pilot in command. MS goes further in that it can sense items in its environment and avoid them which is way beyond the capability of most auto pilots. Again, a small group has a different understanding of a particular term then the population at large.
Society is breaking down so fast. My take: The system works if you use it as Tesla instructs. Use it on the highway for long cruises. It works great. Keep the hands on (or by the wheel). Be alert. Put another way: If I drove a Ford Mustang off a cliff, is it Ford's fault?
It's called personal responsibility.
If folks are using it in high risk scenarios, or from the back seat, or in hair-pin turns at 85 Mph, does Tesla own the liability? No more than Ford owns the liability if I drove the Mustang off the cliff.
My 2 cents.
Then why didn't Tesla restrict the system to highways for the time being?
There could be GPS checks etc. disabling this feature off all highway routes.
For example, Nissan implemented a similar feature in their GT-R cars:
"Nissan GT-R Detects When Car is on a Race Track, Disables Speed Limiter via GPS"
http://bit.ly/1jsEvpQ
This was back in 2007 (!).
Why can't/didn't Tesla implement such a security feature in 2015 for the Autopilot?
Don't they supposedly have the best software engineers (as reiterated by Tesla bulls again and again)?
This is a reckless rollout imho. Given the accounts in this and other forums it's only a matter of time before there will be a severe "Autopilot" accident such as a head-on collision.
Elon is certainly pushing the envelope with the AP release. The onscreen statement that the user has to acknowledge before first using AP was certainly vetted by Teslas legal counsel and seems reasonable to me. Eventually someone will have an accident using AP, maybe even on a freeway in reasonable conditions, and AP use may or may not be a contributing factor. Accidents happen. Inflammatory news stories will likely appear and foolish lawsuits may be initiated.
As an owner and shareholder I believe AP is here to stay and that Tesla will weather the storms that may be brewing.
Trial attorney @cswolfe seems quite agitated about AP. I suggest that he not use it.
As an owner and shareholder I believe AP is here to stay and that Tesla will weather the storms that may be brewing.
Trial attorney @cswolfe seems quite agitated about AP. I suggest that he not use it.
steph280
Member
There are going to be stupid people doing stupid things. Like this woman who decided to ditch her car while driving. Maybe she forgot she doesn't have autopilot on that car?
Bizarre Accident - YouTube
Point is, Tesla's auto pilot is not new technology. Other manufacturers have had them for a while now. If there were serious safety issues, we'd see them pulled off the street already.
Bizarre Accident - YouTube
Point is, Tesla's auto pilot is not new technology. Other manufacturers have had them for a while now. If there were serious safety issues, we'd see them pulled off the street already.
I just can't help it...this thread just reminds me of this great movie scene
Anchorman 2- hilarious RV scene - YouTube
Anchorman 2- hilarious RV scene - YouTube
Hey now, the car didn't say cruise control didn't control steering. I think Ron Burgundy has a clear cut case!
Respectfully, you say a lot, while claiming to be holding back - I wonder what you're really thinking (sarcasm...).
I see the argument of culpability you are foreshadowing (I think everyone does); however, this has a flavor of indicating development/advancements should be hindered with the responsibility of being a nanny any potential misuse. Moreover, a developer should consider just how ridiculous people can be in terms of their use of technology, or in general misuse of said objects/consumer offerings and ultimately should be held accountable. This is similar to both sides of the gun debate but so I don't open up a can of controversy pointing to that heated debate (which by the way experienced this same line of accusation to which congress passed a bill protecting them in 2005 called the Lawful Commerce in Arms Act), lets look at the parallels with the concept of suing McDonalds, or worse the supplier of the silverware for making them fat.
Just because someone "can" misuse an item shouldn't mean that the manufacturer should be held accountable for egregious acts of misuse. If there are laws governing the accessibility of the offering then they should abide by them and in general, clear and accessible instructions should be at the ready for all intending to use the offering (including any disclaimers or warnings necessary).
Using the Google car as a point of reference is interesting. On the surface, you are simply making reference to CEO communication. Without saying it though, it appears to be an attempt to show parity with the two technologies yet they are very different. One is expected to operate autonomously; the other, have very limited capability to adopt some of those features but has been introduced and provided as driver assistance.
I realize that law likes to gravitate towards big payout in determining wrong (more like they just name everyone they can in a case like this)...and its clearly more profitable for an attorney to sue Tesla vs. Joe Smith for a decent payday. That said, its been made clear to me as a user of the software that it is 1. BETA; 2. optional in its use; 3. has guidelines for use itself (i.e. - I shouldn't sit in the back seat and expect it to drive myself etc.)
- - - Updated - - -
you didn't read the fine print? you're supposed to put a cork on the fork
Dirty Rotten Scoundrels - cork on fork - YouTube
