Slashdot Mirror
Tesla Will Have Self-driving Cars In Just Two Years, Elon Musk Boldly Declares (fortune.com)
An anonymous reader writes: In a new interview with Fortune, outspoken Tesla CEO Elon Musk said the electric automaker is just two years away from developing fully autonomous vehicles that can operate ably and safely in any type of environment. While Musk has long championed an automotive age filled with self-driving cars, this is the most optimistic timeline for their deployment we've seen Musk make yet. In fact, Musk in 2014 said the requisite technology to manufacture a self-driving car was still about five to six years away. "I think we have all the pieces," Musk said, "and it's just about refining those pieces, putting them in place, and making sure they work across a huge number of environments—and then we're done. It's a much easier problem than people think it is."
Self Driving Cars with automated updates, kill switches, remotely accessed cameras and microphones, and any other means of surveillance is not a car I will finance because it will no longer be MY car.
If he means he is willing to accept full liability for any and all problems that's one thing. But I doubt we will have autonomous cars that can replace drivers in anything but freeway and some limited pre mapped city driving situations any time soon. It's difficult for computers to solve visual captchas humans solve easily, likewise humans navigate better than computers today with far inferior sensors. We can essentially duplicate the computational power of a common housefly, but that housefly has superior navigation and SLAM to any man made machine we have today, using far inferior sensors. But hey if it's just a stunt to raise money for research that's not really a bad thing.
He was counting in Mars Leap Years
I have no link, nor Did I read the article, But, Musk has some kind of bet with George Hotz regarding auto driving cars.
Recently G.H. did a preview of the tech with Bloomberg or something. So 2 years? sure.
We don't want another "welcome to the club" tweet like Jeff Bezos just tossed out for landing rockets. ( Apples to Oranges comparison there, but Oh, The Gamesmanship!)
So when will they be able to handle round-abouts and unmarked lanes?
"We can essentially duplicate the computational power of a common housefly"
Common housefly is essentially free.
Self-driving car, $100 billions in research.
Housefly is smarter than car, avoids pesky obstacles.
Obvious answer: teach common housefly to drive car.
When the car gets to ice, broken line markings, rough road, construction code madness, deer coming onto the road, or thousands of other unexpected events, a human's understanding of the situation can sometimes be superior. I guess until you get strong AI, you can drive on a subset of all roads until you fix the harder problems. For 99% of driving self driving cars are good to go. It's the unexpected details that could cause it problems. Maybe going slow and sticking to strict GPS locations of where the road shoukd be will cover most unknown cases...
Elon is apparently still riding the high of Falcon 9 success, a feat which while absolutely amazing, pales in complexity to self-driving cars. I certainly wish him luck, I love his bravado for sure; but me thinks NOT.
Did you ever wake up in the morning, with a Zombie Woof behind your eyes? -- FZ
Time to purchase some Tesla stock.
Every time Musk says something positive about Tesla in a public forum, the stock jumps higher, usually in the $12 to $15 range. Then slowly goes down again over the subsequent 2 weeks.
Look to see a big jump around Monday, after everyone has done with the holidays and logs into their stock accounts over the weekend.
Just 'sayin...
They'll be defending themselves from lawsuits within 3.
So what? They're safer than the average driver. Juries on average may prefer to side with the little guy, but what really convinces juries is showing them exactly what happened. The many, many tort cases about car accidents are all about whose story about how the accident occurred is correct.
Anyone defending a robotic car will be *really* good at showing exactly what happened in an accident. It's not just about witness statements any more.
We're already paying a lot of money because of lawsuits around cars. Insurance will ultimately be cheaper with robotic cars involved.
As I recall, Google has been saying for while that they'd have something ready by 2017. On the one hand it seems like it should be surprising if Tesla manages to make it to market at the same time, since they got a much later start. On the other hand, it probably shouldn't surprise us if multiple companies get there at about the same time, since it's less about the cleverness in building the system than it is about having all of the fundamental technological pieces to do it. In particular, I think deep learning neural networks are the core technology that will make effective fully-autonomous cars feasible (plus the sensors, but those have been available for years), and to a large degree the whole world got access to that theory and practice at about the same time.
What is surprising to me is that we haven't heard more from the likes of Freightliner. IMO, that is the first really major market for self-driving vehicles, and those don't even need to be fully autonomous. If tractor-trailer rigs can just drive themselves on the freeway, freight companies can immediately get rid of 90% of their drivers and massively reduce their costs, by having a human drive the truck to the freeway then hop out and let it travel to the destination city, where another human will drive it through town to its destination. Plus, given the price of a semi tractor, adding $50K or even $100K for an automation suite is a relatively small incremental cost, while it's a rather large chunk of change for a passenger vehicle.
(Disclaimer: I work for Google, but not on self-driving vehicles and I don't know any more about the status of Google's system than what is in the public press.)
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
So what? They're safer than the average driver.
How can you know that about a technology that isn't even built yet?
"First they came for the slanderers and i said nothing."
https://xkcd.com/678/
More context:
“We’re going to end up with complete autonomy, and I think we will have complete autonomy in approximately two years.” That doesn’t mean city streets will be overflowing with driverless Tesla vehicles by 2018 (coincidentally, the company’s Model 3 should be on roads by then). Musk expects regulators will lag behind the technology. He predicts it will take an additional year for regulators to determine that it’s safe and to go through an approval process. In some jurisdictions, it may take five years or more, he says.
Cold fusion, nano-bots, jet packs, flying cars, etc.
BECAUSE THERE'S TROUBLE, RIGHT HERE IN RIVER CITY!
Anyway congrats on landing Falcon rocket. So yeah, he over promises and under delivers. But even his under delivered stuff is way too awesome.
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
We'll having flying Teslas in three years! Finally, our childhood dreams embodied.
Chance favors the prepared mind.
Perfect is the enemy of good.
Self driving cars will only be 2 years away.
"How can you know that about a technology that isn't even built yet?"
Because it is. What's happening now is testing, validating and minor refinements based on existing builds.
Musk said, "and it's just about refining those pieces, putting them in place, and making sure they work across a huge number of environments—and then we're done. It's a much easier problem than people think it is."
On the other hand, perhaps it is at LEAST as difficult as people think it is.
Here's the self-driving technology level list. Which level would you say we are at with current technology?
"First they came for the slanderers and i said nothing."
How can you know that about a technology that isn't even built yet?
Because it is built, and has already been tested over millions of miles of driving. Many of the specific technologies, like adaptive cruise control, and automatic lane following, are already installed in millions of cars.
Again, here are the self-driving technology levels. Where would you say were are currently on that list? What do you think it will take to get to level 4?
"First they came for the slanderers and i said nothing."
One concern I have is if the car drives 99% of the time my driving ability will get worse due to lack of practice. So if I ever need or want to take control my chance of an accident will be higher than if I just drove all the time.
Can someone please give Elon Musk a smack? He's just trolling us now.
You are welcome on my lawn.
Where would you say were are currently on that list?
Legally, we are at level 2. Technologically, we are at level 3.
What do you think it will take to get to level 4?
Two years.
Stop lying you astroturfer.
Wasn't the self driving upate in the November patch !
What do you think it will take to get to level 4?
Two years.
That avoids answering the question. If you don't know what technological advances are still necessary, that's fine.
"First they came for the slanderers and i said nothing."
To ensure such vehicles never see the light of day
Are technological advances actually necessary, or is it just a matter of gathering a lot more data with testing to refine the logic?
This space intentionally left blank
Smack your own self first, he's done more for the world than your bad comments will amount to in a lifetime.
He's done more trolling the world.
see sig....
Get up!
Are technological advances actually necessary, or is it just a matter of gathering a lot more data with testing to refine the logic?
Faster hardware will also help. Most vision and AI algorithms are highly parallel, so more cores helps almost as much as faster cores.
and needs say at least 5-8 years free updates for software / hardware swaps if needed for the auto drive system / safety parts of it.
We do not need after one year want X fixed buy a NEW CAR.
AT least he isn't spewing verbal cancerous diarrhea like you and your fellow cocksucking fanboys.
Will they take on criminal liability as well? and with that you can't hide under some EULA also a EULA will not cover a 3rd party victim.
It depends. If you want an AI that can handle any new road it's never seen before (like a human), then yes, we need to invent that kind of AI.
The Google/Tesla method is to have a very detailed 3D map of the area, and then the car can navigate through the pre-made map. It seems like Tesla is planning on creating a detailed 3D map of the entire United States, probably by collecting data from cars that are already on the road. I am fascinated to see if that is enough. I doubt it, but if I'm proved wrong, then great we have self-driving cars.
Previously reports were that Google car sensors couldn't handle certain weather conditions. Maybe they've solved that by now (they aren't very open at all about what their cars can do), but if not, then we'll need to improve that technology as well.
"First they came for the slanderers and i said nothing."
What about data? Cars may need a lot of local storage can't relay on there being some kind of good cell / wifi network 24/7 even 1 way GPS is iffy in some areas maybe mesh networks but still rural areas and other network black holes.
There's nothing remarkable about it. All one has to do is hit the right keys at the right time and the instrument plays itself.
“He’s not deformed, he’s just drunk!”
"It's a much easier problem than people think it is."
Bwahahahha. When will people figure out that the self driving car isn't about the car, it's about the idiots around the car? I mean, I haven't seen 4 drivers correctly navigate a 4-way stop in years. Tesla will be stuck at the intersection, unable to break into the melee. And it'll only cost you a 911 Turbo for the privilege. ROFLMAO.
"Putting the 'dead' into deadline."
I really really hope they don't rush things to get it done on time.
Don't waste your vote! Vote for whoever you want, unless you live in a swing state it won't matter anyways
"Tesla Will Have Self-driving Cars" is very very very far from "Tesla Will Have Self-driving Cars that are approved to sell to the public"
aaaaaaa
Let them build a reliable and durable human-driven car first.
Can it to tell that the man in front is a cop giving hand signals (and obey those signals) as opposed to some crazy person? Can it know not to stop for a potential carjacker? Can it read road diversion signs? Can it read temporary speed restrictions and roadwork signs? Can it negotiate a crossroads where the lights are out in a way that gives priority and due consideration to other drivers? Can it navigate in a long tunnel, double decker road, multi story carpark or other areas that have no GPS signal? Can it tell the difference between a bus which has stopped to pick up passengers, as opposed to a bus which has broken down and needs to be passed? Can it operate when rain or snow are impeding its sensors? Can it tell the difference between a pothole and a puddle? Can it tell the difference between a plastic bag blowing by and a child running across the street and react appropriately? Can it tell the difference between pedestrians waiting to cross vs those standing with no intention of crossing?
I bet there are a LOT of situation that neither Tesla vehicles or any others can be trusted to operate properly. I expect they'll do fine on motorways and certain predictable lengths of urban road. I expect they'll be so annoyingly bad in cities and towns that they'll be turned off or they'll be the cause of accidents.
Autonomous on controlled access highways. That's all I want and hopefully will be the first thing to market and first to be gov't approved.
Nap time on the interstate.
How about selling 100,000 Tesla's a year for $30k each, and not a pittance volume of luxury cars that don't justify Tesla's valuation or capital expenditures?
Congrats to Musk on SpaceX landing a Falcon 9; he deserves that. But Tesla needs to turn into a consumer product before they roll out a whiz bang gizmo of questionable value or interest like a self-driving car.
That list is clearly put together by some nitwit that has no clue about how autonomous cars works.
I recommend that you throw it away and write your own. You can't do worse than that.
Good. There have been too many glazed-eye douches in Tesla who have nearly run me down in crosswalks. This does not, alas, mean that American car sitters of other vehicle types are really any better at not running down pedestrians, but when you've built a mostly survivable car hell, what can one expect?
The problem with these millions miles is that they are all on the same track, so really means nothing when you go elsewhere or when something changes on the road.
GPS may be hardly usable in some places like South America.
How can you know that about a technology that isn't even built yet?
The technology is partly 20 years old so it is built.
Basically all majour car companies in Europe and Asia have self driving cars since about 10 years. With more or less success, of course.
Cost free eBook I read (by iBook/Kobo/Amazon/ObookO/Gutenberg etc.): "The Green Odyssey" by Philip Jose Farmer.
We are already at level 4.
I never sat in one of those cars, but we have a few at my University more precisely at the "Forschungszentrum Informatik" here in Karlsruhe.
Cost free eBook I read (by iBook/Kobo/Amazon/ObookO/Gutenberg etc.): "The Green Odyssey" by Philip Jose Farmer.
Erm, we are using 4 ARM cores and 2 DSPs and mostly they idle, one DPS does nothing, one ARM is only needed during booting. On something like 500MHz and 16MB RAM. TI board, forgot the name. A pretty simple and cheap one.
We don't need anything faster for simple image processing.
Radar, Lidar and other sensors are processed by different boards.
Cost free eBook I read (by iBook/Kobo/Amazon/ObookO/Gutenberg etc.): "The Green Odyssey" by Philip Jose Farmer.
With more or less success, of course.
lol there's a lot of variance in that phrase.......
"First they came for the slanderers and i said nothing."
"How can you know that about a technology that isn't even built yet?"
Because it is. What's happening now is testing, validating and minor refinements based on existing builds.
Cool, Elon Musk himself is posting on slashdot.
To have a right to do a thing is not at all the same as to be right in doing it
Where would you say were are currently on that list?
Legally, we are at level 2. Technologically, we are at level 3.
What do you think it will take to get to level 4?
Two years.
Do you actually work for Elon Musk, or are you just stupid?
The Google/Tesla method is to have a very detailed 3D map of the area, and then the car can navigate through the pre-made map.
The map is not the territory.
To have a right to do a thing is not at all the same as to be right in doing it
We are already at level 4.
I never sat in one of those cars, but we have a few at my University more precisely at the "Forschungszentrum Informatik" here in Karlsruhe.
So one of the cars at your university can drive me from there to (say) Berlin, while it's snowing, with no human intervention other than my opening the door to get in? Really?
If this is the case, why haven't you licensed the tech to Musk and let him make you all a boatload of cash?
To have a right to do a thing is not at all the same as to be right in doing it
I think what you are saying is that the map will never accurate represent the actual world. In that case, I agree with you, and I will actually be astonished if it works. Pleasantly astonished, but astonished nonetheless.
Also, agile techniques definitely don't work with self-driving cars. Can you imagine an OTA update that isn't quite right, and ends up killing 15000 people?
"First they came for the slanderers and i said nothing."
I guess Elon will license the tech from someone ... what is your point?
At least if he starts from scratch now and does not license anything, or simply buy the required parts (actually no need to license anything if you simply buy the sensors and computers from Bosch and Continental e.g.) he wont make anything in the next 2 years.
After all the other companies are working since 20 years on the required sensors and software.
So one of the cars at your university can drive me from there to (say) Berlin, while it's snowing, with no human intervention other than my opening the door to get in? Really? ... not sure which brand we have right now, I think an BMW and an Audi.
Yes, they are only at the university, more precisely FZI Karlsruhe, to fool around with them. They are mainstream cars
Cost free eBook I read (by iBook/Kobo/Amazon/ObookO/Gutenberg etc.): "The Green Odyssey" by Philip Jose Farmer.
That means that some have 100% autonomous cars, and the others are a bit behind ;D
Cost free eBook I read (by iBook/Kobo/Amazon/ObookO/Gutenberg etc.): "The Green Odyssey" by Philip Jose Farmer.
Well, Google has a car without a steering wheel, and the DARPA challenge was won a while ago, but that doesn't mean we have the technology for a street legal car without a steering wheel.
"First they came for the slanderers and i said nothing."
Legally, we are at level 2. Technologically, we are at level 3.
I'd say that's correct.
But 2 years to get to level 4? No. 2 years we'll be able to expand the list of "specific traffic and environment" situations that are supported, but that's still miles away from level 4.
For level 4, the car has to be able to cope with the car in front of it having a blowout on a bridge, and skidding to a stop on its side. And I'm not talking about just safely braking to avoid a collision; that's the EASY part.
I'm talking about coping with the fact that now there is a traffic jam piling up behind it, and a 2 tons out of commission vehicle blocking the lanes in front of it.
So it needs to get out of the way for the ambulance if necessary. It needs to follow directions to detour around the wreckage, perhaps backup, off the bridge cross over to the oncoming lanes using one of the emergency/service vehicle openings, and then take turns with oncoming traffic to pass the accident as directed by the police.
You think we're two years from a car that can do that? Not me. I don't think we're even close.
And if we don't have that? Then what? We're going to have empty cars, and cars occupied by people who aren't able to drive causing truly epic traffic jams because they simply don't know how to cope with exceptional circumstances like this.
but that doesn't mean we have the technology for a street legal car without a steering wheel.
I would not be so certain about that. And ofc. some questions remains:
a) with or without steering wheel, would an occupant require a driving license (especially if it has a steering wheel)
b) would there be a way, perhaps with an App, to interfere aka giving directions, in case the car is indeed lost, e.g. in a small path through a wood, or if it "sees" tracks of a agricultural machine on a field and likes to divert to follow them from a country road
Hm, I had one more question before, but I lost it.
As far as I see it (checked how far Tesla is, but could not figure how they went there so far, they are basically on par with google I would say) the legal question comes in tough when the majour car companies make pressure. As I mentioned before: all big Germans and a few of the Japanese with which we have cooperations have the technology to sell self driving cars _right now_
I just watched a movie of the Tesla in "autopilot" mode on german roads. The testers where extremely surprised and confident. It seems it only lacks abilities in crossing roads in tows or making turns there, when no traffic lights etc. are available. I guess they have no proper camera/algorithms ready yet.
Cost free eBook I read (by iBook/Kobo/Amazon/ObookO/Gutenberg etc.): "The Green Odyssey" by Philip Jose Farmer.
I would not be so certain about that.
I am lol.
"First they came for the slanderers and i said nothing."