Slashdot Mirror
Selling Full Autonomy Before It's Ready Could Backfire For Tesla (arstechnica.com)
An anonymous reader quotes a report from Ars Technica: Tesla has an Autopilot problem, and it goes far beyond the fallout from last month's deadly crash in Mountain View, California. Tesla charges $5,000 for Autopilot's lane-keeping and advanced cruise control features. On top of that, customers can pay $3,000 for what Tesla describes as "Full Self-Driving Capability." "All you will need to do is get in and tell your car where to go," Tesla's ordering page says. "Your Tesla will figure out the optimal route, navigate urban streets (even without lane markings), manage complex intersections with traffic lights, stop signs and roundabouts, and handle densely packed freeways with cars moving at high speed." None of these "full self-driving" capabilities are available yet. "Self-Driving functionality is dependent upon extensive software validation and regulatory approval, which may vary widely by jurisdiction," the page says. "It is not possible to know exactly when each element of the functionality described above will be available, as this is highly dependent on local regulatory approval."
But the big reason full self-driving isn't available yet has nothing to do with "regulatory approval." The problem is that Tesla hasn't created the technology yet. Indeed, the company could be years away from completing work on it, and some experts doubt it will ever be possible to achieve full self-driving capabilities with the hardware installed on today's Tesla vehicles. "It's a vastly more difficult problem than most people realize," said Sam Abuelsamid, an analyst at Navigant Research and a former auto industry engineer. Tesla has a history of pre-selling products based on optimistic delivery schedules. This approach has served the company pretty well in the past, as customers ultimately loved their cars once they ultimately showed up. But that strategy could backfire hugely when it comes to Autopilot.
But the big reason full self-driving isn't available yet has nothing to do with "regulatory approval." The problem is that Tesla hasn't created the technology yet. Indeed, the company could be years away from completing work on it, and some experts doubt it will ever be possible to achieve full self-driving capabilities with the hardware installed on today's Tesla vehicles. "It's a vastly more difficult problem than most people realize," said Sam Abuelsamid, an analyst at Navigant Research and a former auto industry engineer. Tesla has a history of pre-selling products based on optimistic delivery schedules. This approach has served the company pretty well in the past, as customers ultimately loved their cars once they ultimately showed up. But that strategy could backfire hugely when it comes to Autopilot.
Considering that they are topping the table for car-ownership satisfaction rates, I don't think it's such a big issue. https://www.consumerreports.or...
Selling Anything Before It's Ready Could Backfire For Anyone
Of course, now it looks more like an Onion's headline, but that is, in itself, a hint...
In Soviet Washington the swamp drains you.
I also thought about how hard it would be to make a real autonomous vehicle that worked under all conditions
Getting to 90% has been done. Good weather, good visibility, few unexpected hazards
Getting to 95% will be harder, and it gets exponentially harder as you asymptotically approach 100%
The billion dollar question is...How close is close enough?
No matter how good it gets, someone will always sue, claiming it isn't perfect
The law needs to be adjusted to accept the reality that nothing is perfect
The people who think it takes hard AI to achieve full autonomy in a self driving car vastly overestimate the cognitive abilities of human drivers. The computer does not need a complete and entirely correct model of the environment to be a better driver than a person. People are very easily overwhelmed by complex traffic situations and make tons of mistakes. Roads are designed to enable safe traffic regardless of these cognitive deficiencies. Computers can take advantage of that too.
Likely they cannot afford the payments and operating costs.
Tesla's are nice cars, but they come at a premium price. I'll never afford one. My guess is as folks start to realize they are mileage limited with long recharge times (range anxiety), are paying a LOT for that 4 door sedan, AND they could get a REALLY NICE fossil fueled option for less money so they put it up for sale.
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
No internal combustion engine and we are discussing a backfiring Tesla? How's that possible?
(To you literalist... I'm making a joke.. )
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
The people who think it takes hard AI to achieve full autonomy in a self driving car vastly overestimate the cognitive abilities of human drivers..
And others vastly underestimate the challenges of designing a system that can perform better than humans without human oversight. Particularly in an environment that was designed specifically for human drivers interacting with other human drivers. We have a very long way to go.
Why would anyone purchase non-existent driving software with their Tesla when, presumably, that feature could be purchased separately later, at a time when it really exists? You are only giving Tesla a free $3,000.00 loan for an indefinite period by ordering it now with the car.
Ceci n'est pas une signature.
The Tesla Theranos Edition.
I own a Tesla and the operating costs are substantially less than any fossil fueled car I have ever owned. Charging is much less expensive than gas and the only real maintenance is replacing the tires. I normally don't charge the battery past 200 miles unless I'm going on a long trip and my typical home charging time is less than an hour. The customer service absolutely blows away any other car brand I've owned and the car just keeps getting better with new software updates. I have never heard of any of the major auto manufacturers doing updates at all for anything that wasn't a serious problem and certainly not free over-the-air updates. I paid around $70K after the tax rebate which is not even close to the high-end for a luxury car.
The people who think it takes hard AI to achieve full autonomy in a self driving car vastly overestimate the cognitive abilities of human drivers. The computer does not need a complete and entirely correct model of the environment to be a better driver than a person. People are very easily overwhelmed by complex traffic situations and make tons of mistakes. Roads are designed to enable safe traffic regardless of these cognitive deficiencies. Computers can take advantage of that too.
And people who make claims such as yours tend to forget that roads were designed for humans, not computers, and that some things that humans do very easily are very difficult for computers to do (and of course, vice-versa).
If we had roads which were designed for computers, I am sure computers would very quickly outperform human drivers. The problem is - we don't.
I believe that self-driving efforts are focused on the wrong thing: trying to reproduce a human driver, and claiming it's better if it, on average, messes up less than a human. Instead, it should be focused on trying to create an infrastructure that supports self-driving vehicles. Does that mean they will be able to drive on every imaginable road? Probably not...but likely on 95% of roads (such e.g. all roads and streets in cities), once the roll-out is complete. In this case, it probably will be safer...but if it satisfies a bunch of conditions first.
The other problem that people who claim "humans are not that smart, computers are better than them on average" miss is the variability among humans. No human driver is identical; almost all make mistakes, but the types of mistakes that are made (and the situations which they are made in) can differ widely. However, software flaws are replicated identically across many units...potentially millions of them. Joe in Chicago being a bad driver does not affect Bob in Cleveland. An undetected bug in Tesla's Autopilot will affect all Tesla owners in the world potentially, and probably under the same (or very similar) conditions. You might be in a situation where you have a 100% probability of a screw-up under certain conditions. That is simply not the case with any human driver. Now imagine if say, 20% of the cars on the road are self-driving...what you are potentially setting things up for is a black swan type of event: software-driven cars may be much safer 99% of the time, but could be prone to major screw-ups 1% of the time that will dwarf the combined effects of bad human driving.
Variability among human drivers also allows for evolutionary selection: reckless drivers will typically die, or have their licenses taken away from them. This does not remove all of the bad drivers - but does remove a great deal of them over the long run. How do we do that with self-driving cars? A destruction of one self-driving car with a flaw will not end it, because then likely all other cars of the same model and series have the same flaw. Removing just that particular car won't do it, you'd have to do a recall...now recalls can already get quite bad, imagine all of the recalls we're gonna have with self-driving vehicles (where the authorities are bound to more paranoid)...and it won't all be software problems you can patch, there will be hardware problems too.
I'm not saying these are all insurmountable problems. They can be addressed. I'm just saying there's a lot more to autonomous vehicles than the techno-optimistic "self-driving cars just need to cause less accidents on average than humans" stance. A lot more. It cannot be reduced just to a single metric. Or just a bunch of numeric metrics.
I happen to have a Tesla S and I do love the car.
They are a joy to drive. The acceleration is second to none. The handling is great, almost on par with the ( now deprecated ) hydraulic steering on BMWs's I used to drive. The cabin technology is great. I also really like the fact that the car gets free updates that actually improve it on a continuous basis!
I also feel extra pleased with my purchase because I get free EV charging in my office building, premium EV parking at most malls, and I just learned that Tesla will install a supercharger ( which I also get for free ) at my local mall. Heck I even get premium parking and free power at Ikea! ( I would get these with a lower cost EV but their really just a 'feel good' bonus. )
The autpilot is def misnamed. It's nowhere near autonomous. Just now it's more like a junior co-pilot. Would I like it to be better? Yes I sure would. But I will say that once you know it's limitations it is very helpful and - contrary to popular opinion - it does an excellent job in all weather conditions except heavy snow which covers the road. I'd say 70% of my daily commute is now handled by autopilot.
It does take hard AI. The dumbest person is infinitely smarter than the best computer.
In fact, I think Tesla is probably the least of the offenders in this case; all (so-called) 'self-driving car' developers are rushing their 'product' to market, because they've spent so much more money 'developing' it than they ever thought they'd need to, only to find that it's way, way more complex than they ever thought it was, and they need to start showing a profit or heads will roll. Therefore they expect us, the general public, to be their 'alpha-testers' (not even BETA-testers, the damned things aren't even that good).
The billion dollar question is...How close is close enough?
You only have to be SLIGHTLY better than the average human driver, and you start saving lives by getting marginal drivers off the road.
Judging from how I've seen people drive around the world, 80% is PLENTY HIGH.
I am serious; if you took some of the worst drivers today and gave them self driving cars with existing tech, you would be saving lives and reducing accidents.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
My guess is as folks start to realize they are mileage limited with long recharge times
I doubt if any Tesla owner was unaware of these issues before they bought the car.
they could get a REALLY NICE fossil fueled option for less money so they put it up for sale.
FF cars come with zero nerd-cred. Nobody cares. You have to park at the back of the lot, while the EVs get the premium parking spaces near the entrance.
Disclaimer: I am a very satisfied Tesla owner. Well, technically my wife owns it, but she lets me drive it if I wash the dishes everyday.
You paid $70k for a car. No one said anything about a high-end luxury car. Tesla's certainly aren't that. They are the equivalent of a $30k Ford in terms of luxury. But at least you get free updates!
Both for their cars and their factory. They're getting sued for their FSD fraud, and they've wasted time and money over auto mating their Fremont factory for the Model 3 line. Now Musk is going to have a 24/7 shift for the Model 3 line. How ever close they were to positive margins on the $60K Model 3 have now become that much more elusive. And the $35K Model 3? Phffft.
What happens when they start selling them and the courts find that all liability is with the software/hardware manufacturer?
Then the manufacturer will pay for insurance, rather than each individual paying for their own. It will just be built into the price of the car, but will be less expensive and more efficient than the current system because of better transparency and lower transaction costs.
Consumers will win, since they will save money and hassle. Car manufacturers will win since "no-insurance-needed" cars will sell better. Insurance companies will lose, since they will be selling to informed manufacturers (who may opt to self-insure) rather than to confused consumers.
And others vastly underestimate the challenges of designing a system that can perform better than humans without human oversight
And still OTHERS appear to be utterly ignorant as to the state of the art in self-driving car research and delivery.
Kind of strange for a place like Slashdot to have some many people so very, very ignorant of technology.
Please explain the "state of the art" for all those ignorant people. You could be one of them as far as I know.
The law needs to be adjusted to accept the reality that nothing is perfect
The problem (traffic, roads, laws, standards) will get redefined to fit the new AI solution, same as when we transitioned from horses to cars. When you get enough autocars out there, the roads will begin to get engineered to mitigate the weaknesses of AIs, bit by bit.. This is exactly what happened to cars: we created and adapted roads, laws, enforcement, etc to match car's needs and continue to do so. The Model T was high off the ground to deal with the rutted, muddy dirt roads (or no roads at all) they were likely to encounter. Today, we have aerodynamic skirts a few inches off the pavement for efficiency. Pavement--smooth pavement--is simply assumed.
That's what always happens with any disruptive technology: we end up adapting everything, including ourselves, to meet it part way.
An example of things that will probably be changed soon than later: road construction zones will be required to implement certain protocols (signage, markers, notifying some central database, whatever) to make them easier for AIs to traverse. Failure to do so will entail liability for accidents.
My guess is that true full autonomy will first roll out in a big way on certain long-haul trucking routes. Many freeways are a fairly clean, well-defined situation and the prize for trucking companies is too big to ignore. Those parts of the chosen freeways that are problematic for the AIs will be upgraded, either due to lobbying by large trucking firms, and/or those firms'll kick in some of their own $ to make those changes happen sooner.
- everybody who is tired of waiting on the feature can ask their money back from Tesla.
- not a lot of people who have enough money to buy these cars are stupid enough to 'get confused'
- it's a duplicate of a similar story which ran 1 year ago. Tesla Autopilot has gotten better and will get better
- Tesla has already promised that if a hw upgrade is necessary (which is likely), they will upgrade free of charge.
Aside from that, we love the AP capabilities of our Tesla's.
What happens when an accident happens because the owner did not do proper maintenence
Very few accidents are a result of deferred maintenance. Nearly all are from impaired driving and human error.
bad brakes, bald tires, mechanical failure, didn't take the vehicle in for updates, etc)?
Tesla brakes don't wear out. The automatic regen absorbs 95% of braking energy. The brake pads last the life of the car.
Updates are OTA.
There isn't really much mechanics to fail. The only moving part in an electric engine are the bearings.
Tires wear out faster, because of the fast acceleration, so that is an issue. But existing liability law already covers out-of-warranty abusive use.
What happens when someone damages your car? The manufacturer isn't going to cover that.
Why not? If they are selling "no insurance needed" as a feature, then of course they would cover it. Otherwise, you would buy something else.
You could expand this to most of the current AI goldrush. It wouldn't be the first time that capabilites of AI have been oversold.
Also, concerning this article: The terms for AP and FSD haven't changed in a year. There is no "news" behind this article. By selling FSD separate from AP, it makes it explicit what you're buying. You cannot choose just AP and think "I'm getting full self driving". Meanwhile, the FSD option is plastered with all sorts of weasel words like "in the future" and "eventually", with no hard dates. You cannot choose FSD and think "I'm getting this immediately, or at least the day after tomorrow!"
Why this article? Why now? Again, literally nothing has changed about the terms or wording in the past year. It's not like there's a lack of news. Just yesterday, we learned that not only was the 2k+/wk Model 3 production rate not a burst rate, not only have they maintained it for three weeks, but that when they're done with the upgrades this week, the line should come back at 3-4k/wk; and Tesla is now targeting not just 5k/wk at the end of this quarter, but 6k/wk (with an expectation that at least one supplier or process won't get it all the way, in order to ensure that they get at least 5k, with the intent to get it up to the full 6k in Q3).
That's actual news. This is concern trolling. And it's full of statements that are just plain wrong, such as that camera-based AP systems "don't work well in low light conditions". Nonsense; AP often works even better in low light conditions than it does in bright conditions. There's no "glare" at night, strong contrast between headlight-illuminated markings and obstacles vs. the road, and the cameras have good low-light sensitivity (better than the human eye). The article also makes it sound like they've not taken any time at all to research how Autopilot works, writing things like "If a sensor fails, Tesla will have to choose between disabling self-driving capability until the customer repairs it or allowing the car to continue operating with a higher risk of a crash." - seriously, virtually any Tesla owner can tell you that if you have a sensor failure, AP is disabled until you get it fixed. Just bloody ask. They then go on to imply that Tesla doesn't have redundant control systems. That's simply not true - go see them with your own eyes. The mechanical systems to control the brakes, steering, etc are all redundant. Concerning the computer, they link to a teardown of a Model S (older model) from last year. Why? Do they think that nothing's changed in the past year? And yes, the computer has on-board redundancy.
What the heck is up with all of these hit pieces of late?
I also could not disagree more with one of their main conclusions - that going straight for "full autonomy" is the safer option. Tesla has had three deaths around a billion miles of AP driving - the exact amount of mileage on AP today is not known, but was 300 million in November 2016, so probably around a billion today. The normal rate of deaths per mile is 1 per 80 million. Meanwhile, Uber had a pedestrian fatality in its first week operating in Arizona. If you try to go straight for full autonomy, you guarantee driver inattentiveness. Vehicles should be locked at no more than Level 2 autonomy (and level 3 should be illegal) until a high degree of safety can be guaranteed without driver involvement in the real world. Tesla and most of their emerging competitors (like the laughably bad Mercedes Drive Pilot - which actually does claim to be self-driving, unlike Autopilot) make you keep your hands on the wheel to try to ensure you're paying attention. With AP, it's not enough to simply touch it - you have to actually apply torque. But honestly,
I will pull over this spaceship right now!