Introducing Navigate on Autopilot
tesla.comI'd be more impressed by this if they had fixed the problem where the car crashes into barriers at lane divergences. Kind of a turn-off to the whole thing until that's fixed, to be honest. Frankly I don't understand how the company hasn't been buried in lawsuits for this.
https://www.reddit.com/r/teslamotors/comments/9pkvy0/psa_v9_...
That actually is a part of v9 autopilot but Tesla is never going to put that in writing.
The fixing of bugs like this without admission of guilt isn’t unique to Tesla. Apple, Google, Samsung, and many others do the same.
The reason they haven’t been buried in a lawsuit is because these are driver assistance features, not self driving features. It’s the responsibility of the driver to keep control of their car at all times.
Arguing that “people think they can drive with their hands off the wheel” isn’t a real legal argument. This is why Tesla warns drivers in documentation that they need to keep control of the car at all times.
This is the same legal argument that has caused warnings in microwave manuals about not putting animals inside them to dry them out. Just because someone misinterprets the capabilities of technology they purchase doesn’t mean the technology isn’t doing what it’s designed to do.
All self driving cars and driver assistance tools will be imperfect. The question is whether or not they are statistically safer than the average human. If you are an above average driver (by your record, not your impression of yourself), these tools may be unnecessary or may actually be worse for you. It’s the average and below average people that they do wonders for.
Please stop using one or two data points to form an opinion about something that has millions and millions of data points.
> All self driving cars and driver assistance tools will be imperfect. The question is whether or not they are statistically safer than the average human
I disagree. If you're arguiung for driverless cars, the comparison should be with the average driver aided by a driver assistance system.
I don’t understand what you disagree with. I’m not arguing for or against driverless cars.
I’m saying that our basis right now is whether or not driver assistance systems are making driving safer or less safe.
Driverless cars are currently only used in specific areas under tight regulation. They are not generally available and this article is about Teslas which are not self-driving, they have driver assistance features.
The original commenter above was making a remark about how the driver assistance features in a Tesla are “unsafe” due to a highway barrier accident.
I was making the counter argument that the driver was acting in an unsafe fashion by using driver assistance features as “self-driving”
I agree that if driver assistance features with a human driving are safer than driverless cars, driverless cars are not ready.
The point I was trying to make is that rather than talking about a single traffic accident, we have to look at the data. This is a law or large numbers thing, not penny press headlines.
I don’t think this a choice between driverless cars or not. We should be looking for the solutions which reduce traffic fatalities and injuries and also those which reduce the amount of hours wasted in traffic. Driver assistance features or self driving is only one angle for that. Improving road design patterns, building roads that actually follow the principles of psychology and fluid dynamics related to driving, properly maintaining roads, and other non technology related solutions are also potential factors for improvement.
Exactly, the worst driver is you, when frustrated, distracted, or bored. Driver assistance means that stop&go traffic or long stretches of interstate can be made more tolerable and safe. It means rubber necking as you pass that accident on the other side of a barrier won't cause another accident on your side.
Of course reliance on this breeds complacency and allows longer commutes, but so do automatic transmissions, straight tracking wheel alignment, or cruise control. Maybe we would all drive more safely if there were giant spikes that killed us instantly rather than airbags, but we aren't going back... and hopefully one day (before I need it) we'll have true self driving.
Yea, it is frustrating that these driver assistance technologies aren’t self driving. It’s also frustrating that they are over hyped.
At the moment, they are really good in certain situations (stop and go traffic, highways that are maintained properly and designed really well, roads without a lot of high speed curves, etc)
They won’t add value to many commutes but they add tremendous value to others.
Like all products, they are definitely not one size fits all
Sure then I agree with you.
I want to make clear the difference between driverless and autonomous/self-driving/assistance. The difference is frequently muddled by Tesla and Elon musk with false marketing.
I totally agree that the marketing hype is ramped to the maximum. This is a result of the abysmal lack of consumer protection in US advertising laws. A large portion of advertising in the states is making overblown promises and unfortunately that is the standard. Tesla is not an outlier when it comes to over hyping their products.
Unfortunately so. I really wish advertisers were held to higher standards
I'm surprised regulators don't force Tesla to rebrand the driver assistance package to a name that doesn't deceive drivers.
Would love to see that kind of consumer protection across the board. Almost every US ad looks like this these days:
100% effective, works wonders, fully natural!
(These statements haven’t been validated by any organization; our product may not be effective at all. Natural has no legal meaning but it seems good for you! Actually, our product is a placebo but thanks for throwing away your money!)
I watched the video. My Nissan has lane-keeping and it does exactly the same thing as in this video. My impression is it's because it follows the left lane line when it's not sure.
My Honda would interpret it as the lane getting wider and try to stay in the middle of that new wide ‘lane’.
I’ve noticed that behavior when there is a turn off on the highway that isn’t marked off. The car will start to drift to the right (to stay in the “middle”) if you don’t kee i’ve noticed that behavior win there is a turn off on Highway that isn’t marked off. The car will start to drift to the right (to stay in the “middle“) if you don’t intervene. I imagine if I kept going it would then decided it was outside the lane (after turnoff was passed) and yell at me.
Same in VWs that have lane assist. Tesla just figures out how to rebrand lane assist. No reason the other auto companies couldn’t remove the distance limitation on lane assist and the it up to their nav system.
Maybe the state should take responsibility and realize that poorly maintained and marked roads lead to accidents that cause injury, including death.
The Model X that hit the barriers in the Bay Area would not have happened if A. Lane marker painting had been kept under maintenance. They were mostly missing.
B. The crash barrier had been in place, instead it was missing from an earlier crash. The missing arresting system caused the car to strike the Jersey barriers with no safety controls. This is the equivalent of driving into a 6" wide concrete wall at freeway speed.
Sure, there is some amount of fault for the system needing more maintenance. But road conditions are often going to be far from perfect, and that crash is certainly not atypical conditions.
Either California can shut down the entire highway for a week until they replace the crash barrier, or Tesla can build their self-driving cars to be able to recognize only moderately faded white lines.
That's my point. A human driver had hit those barriers just ten days prior. If there is a part of the road people are driving into constantly maybe we should look at road design over drivers.
This is definitely a major challenge for driver assistance features. Poorly maintained roads are less of a burden to humans who can more easily account for flaws in road design or maintenance.
What is also a major issue is lack of design patterns on roads. Rather than using a standard set of merging lanes, circles, right and left turn lanes, etc, there are hundreds of variations across municipalities. What makes it worse is that even the patterns in a single municipality aren’t often followed. We notice this when driving on 290 from Austin to Houston where the turn lanes are sometimes marked with a white line on the left and other times marked with a yellow line.
It’s going to take a concerted effort by legislature to help driver assistance and self driving cars by adjusting roads to make them safer rather than the free-for-all we have now.
It also would not have happened if a human was driving.
Do you mean except for the crash barrier that the earlier crash destroyed?
> Navigate on Autopilot can be customized to a driver’s preferences, including four settings for speed-based lane changes (Disabled, Mild, Average, or Mad Max).
Mad Max is definitely a feature I want self-driving cars to have. </sarcasm>
No kidding. This does not inspire confidence. v9 already does weird things like show stopped cars bouncing around randomly. This kind of thing along with crazy branding really makes it feel unsafe.
It’s one thing that the stereo goes to 11, that’s funny. But I’m not going to die if the stereo has a bug.
The stereo only goes to 11? The stereo in my previous car went to 60.
What's with all the pop culture references in this car? Is this the Ready Tesla One trim package?
There isn’t any relationship between your car’s 60 and Tesla’s 11. Tesla’s volume has more than 11 unique volume settings.
https://forums.tesla.com/forum/forums/reason-why-radio-goes-...
The linked forum post doesn't have any more information than your post, other than confirming the pop culture reference.
I agree, but to give Elon the benefit of the doubt for a moment, labeling it "Mad Max" may actually discourage people from setting their car to it (as people often perceive themselves to not be reckless), whereas labeling that setting something like "Often" may lead to it being selected more frequently. Of course, this can also backfire if people select "Mad Max" because it's the only one with a funny label.
Honestly I'm baffled by how Musk/Tesla get away with all their branded self-driving-but-you-must-always-be-in control messaging and jokey terminology like Mad Max mode etc.
Given how aggressively US DoT enforces safety regulations for vehicles you would have thought there would have been far more pushback, not just on truth in advertising but more importantly road safety for both SD vehicles and the people around them.
Presumably the bureaucrats are asleep and not paying attention, so there is no formal adherence to the levels of self driving that everyone except regulators seem to regularly quote (Toyota et al) https://www.economist.com/science-and-technology/2017/05/25/...
It's a joke, just like the volume control going to 11. Fun is okay. Whimsy is okay. I've never understand why so many people disapprove of little bits of joy scattered throughout the world. Little moments of fun make life a more tolerable experience.
Thoughtless criticism like this, even as a joke, is what will prevent us from getting any self driving cars.
There are three modes for how aggressive to be with lane changes. A least aggressive mode, a most aggressive mode, and a middle setting. You can also turn it off completely.
Naming the highest setting Mad Max mode doesn't make that setting more dangerous -- it makes the average car less aggressive by discouraging people from selecting that option. Contrast that with if they called the options "drive super slow", "drive slow" and "normal" without changing the behavior of each setting.
Very impressive. However these autopilot features are getting subjectively less impressive as time goes on. I want a bit more 'wow'.
Autopilot is not so compelling if you live in Europe with 'analog' roads rather than the mix of well designed, nice and wide 'residential', 'street level' and 'interstate highway' roads that there are in the U.S.A.
Reminds me of Space Shuttle launches, at the start of the programme people would be transfixed for hours waiting to see what would happen, a couple of years later it was 'meh', no need to put the TV on. Even if there was a cool spacewalk some of the magic was lost, awe levels dropped from 10/10 to 7/10. Autopilot is a bit like that, kind of need to see it work going through the streets of Paris, Cairo or Bombay for it to be back in the realm of stunningly amazing.
Living in Paris center, I'm convinced it will be the last place where autonomous cars will be allowed. Roads are smalls, the city plan is old, there are experiments with new e-scooters, people don't respect priorities, etc
It could only work if ALL vehicles were automated and no human driving was allowed. But even then, walking people will exploit the overly-safe vehicles' decisions and just force the priority (which is probably a good thing to give the city back to walking people).
In my opinion, the capital cities won't be the same once autonomous cars are fully developed, there will be probably more public transport, and more car-pool sharing drive services, more escooters, etc. The most interesting long term usage of autonomous driving is on highway and smaller cities than capitals, with less traffic
Interesting that you consider US highways and streets to be well designed, given that they are objectively substantially more lethal to their users than European roads.
There might be additional factors.
a) training and education - it is way more difficult and takes longer to get a driver's licence in Western Europe than in the US
b) mandatory vehicle inspections - I think this varies by state in the US, but even the strictest state legislation is still less strict than Western European countries when it comes to allowing dangerous vehicles on the roads
c) less regulation for truck drivers' work hours - I think in the past decade, sleep deprivation amongst transportation workers and the results thereof has received increasing exposure in the press, which is good, but awareness only goes so far
Sure. But highway design, car centric cities, and high-speed suburban arterials are a factor as well.
The deaths per VMT measure is EU-wide, not just Western Europe.
Yup, sure looks like Europe is substantially safer than the US in that map.
In what way are they “objectively substantially more lethal?”
Double the deaths per VMT.
https://www.newsweek.com/2014/04/04/you-drive-american-24806...
I think attributing that to the design of the roads is kinda silly. We have very low standards when it comes to acquiring a license to operate a vehicle.
I think attributing it to any one cause is kind of silly. It's a fact that American cities are much more car centric with higher speeds and fewer areas with traffic calming measures, all of which make the roads much safer.
Daily highway driving is much more common for commutes and errands, suburbs have dangerous arterial street designs.
I'm sure there's more to it, such as licensing and vehicle safety measures, but American highway design (and making local roads into mini highways) is a part of it.
Also the EU measure is EU-wide, and not all countries in the EU are as strict as say Germany for licensing.
You’re going to be disappointed for quite some time then. Most of the improvements to autopilot in the near future will be refinement of the cars reactions to sensor data and not new public facing features. They are closing in on the line where regulation does not allow for further “self driving” festures. Having the car change lanes itself is “self driving” and that’s why the driver must initiate the lane change. The same argument can be made about stopping at stop signs and red lights which is the next killer iteration of driver assist. Hopefully there will be enough data collected and good decisions made to allow for iterative improvements to driver assist features which would improve overall traffic safety.
My 2018 Mercedes E class can do lane changes - I just have to put the signal on and it does the rest. They call it "Drive Pilot". It seems to be just as advanced as Tesla but much less "marketed".
Yup, Audi offers similar driver assists with lane changes(https://www.audi-mediacenter.com/en/technology-lexicon-7180/...), Land Rover has lane keeping (https://www.landrover.com/ownership/incontrol/driver-assista...), so do Honda/Acura (https://www.acura.com/rlx/modals/lane-keeping-assist-system-...) and many more.
I agree that Tesla overhypes their system the most and that contributes to consumer confusion. They use the classic Silicon Valley tactic of making their features which are comparable to competitors seem light years ahead through marketing. On one hand, it sucks for consumers, on the other hand, this type of advertising is rampant and they are playing the game.
Tesla’s over promising does erode trust. There’s only so many times you can promise stuff that’s obviously overhyped or years out (FSD) before consumers will stop believing the hype.
What’s interesting about Tesla is that almost all of their controversy is self inflicted by Elon’s comments.
What’s not clear to me is if this is intentional as he wants to stay in the news cycle or if it’s driven by his ego.
> While initially the feature will require drivers to confirm lane changes [...], future versions of Navigate on Autopilot will allow customers to waive the confirmation requirement if they choose to.
So at least technology wise my dream becomes true: a car that takes care of the biggest and most boring part of long drives, the highways. Hopefully legislation will follow soon...
You don't think your drive will become even more boring when your car is "doing everything" but you are required to "pay attention and be ready to take over at all times?"
I'd agree if you could trust it and didn't have to be actively involved in watching your car drive for you. Sounds pretty boring as is though. Sleep inducing.
It'd be like having a normally very reliable chauffeur that would occasionally get horribly confused and do something very wrong, and you had to watch their every move, most of which are perfect, for the remote chance of terrible confusion and be ready to grab the wheel.
No, it doesn't get more boring. It is very relaxing and leads to less stress. I use it almost every day, including driving from LA to San Francisco on Friday.
I think this is still a super under appreciated point. The focus is often on full autonomy, but we don't need full autonomy in complex city driving environments to unlock a potential sea change in behavior. Imagine the impact of simply being able to trust your vehicle to do the heavy lifting for an hour to get you from, say, Sacramento to San Francisco, only needing to take control at the highway end points for 30 minutes on each end.
Waiving the confirmation still doesn't get you there; you still will have to pay attention the whole time and have hands on the wheel at this point.
And then people argue that Tesla does not do deceptive advertisement and everybody knows the legal and technical limitations of autopilot.
> In both of these scenarios, until truly driverless cars are validated and approved by regulators, drivers are responsible for and must remain in control of their car at all times.
This makes it sound like it's just a problem with evil legislators not making it legal for Tesla drivers to drive without paying attention.
Is this the case?
It doesn't sound like that to me. It sounds like they are saying that this is NOT a driverless car and you need to pay attention. However, it is a little bit misleading in that this is all Tesla's software can do right now. Even if the regulatory framework was in place, they couldn't just enable it instantly. They have a lot of dev / testing work to do before that happens. But they do have stuff working in the lab that is more capable than what they've released publicly.
No, it's not. This can be read either as Tesla wants you to read it, that the only reason they're setting it this way is because legislators demand it, or the way you should read it, which is that Tesla will find any possible excuse to blame you for not paying enough attention when the car makes a mistake and kills you.