Why Tesla will win the Autonomous Driving race

[Turing]

The nice thing about Tesla's position is that with their incremental approach and OTA updates, they can essentially release a production self driving car way before others under the rules of a driver assist system while the other competitors have to reach the much higher goal of full on self driving car and wait around for the possibly delayed regulatory approval. With tesla it will just get to the point where tesla's a/p system is so good that drivers almost never have to take over and the data clearly verifies that A/P mode is safer than manual mode. The downside of this, of course, is the system being good enough that drivers get complacent and don't pay attention at all, even though they sign off on it being a driver assist system.

 

[CarlK]

Why are you pretty sure of this?

From Elon:

“What we’ve got will blow people’s minds, it blows my mind …it’ll come sooner than people think.”

“It blows me away the progress we are making. And if it blows me away, it’s really going to blow away other people too when they see it for the first time.”

There is no real winner in autonomous driving.

Puts millions of taxi drivers out of work. Puts over the road truckers out of work as well. Already putting fork lift workers out inside of automated warehouses. But good for profit margins of companies that possibly some who chat here may own and are looking to cut those pesky and costly workers.

People said the same thing since the industrial revolution but we should also know increased productivity is always good for everyone's living standard. We need to adjust our socioeconomic structure to fit the change of course but there is nothing wrong with technology that helps increase productivity. I don't think back when more than half of population work in the field and everyone works 80 hrs a week is the kind of full employment you'd like to have. Even today there are probably still people in part of the world working under that kind of environment. We are blessed with technologies we have please do not complain about what has provided the better life you are enjoying now.

 

[3Victoria]

Consider, if TM beats Uber to autonomous cars, they can duplicate Uber's model quite easily. Uber needs to have skin in the game.

 

[CarlK]

Consider, if TM beats Uber to autonomous cars, they can duplicate Uber's model quite easily. Uber needs to have skin in the game.

Elon's in his second secret plan mentioned share mobility and says Tesla will have its own fleet in cities where there are larger demands than supplies from owners. I'm sure Uber felt the great urgency already. Kalanick has said if Tesla can make 500,000 autonomous cars by 2020 he will buy them all. He might even have approached Tesla but Tesla probably has little reason to work with him. If Tesla can make the car, and it's pretty simple to create the car sharing app, why it needs to go through the middleman?

 

[green48]

From the responses so far, this "battle" for autonomous driving seems to be far from decided. That was an impressive video from Carnegie Mellon and I think that some of the Tesla shareholders on these boards are getting a little ahead of themselves. Until AP 2.0 comes out or we see some hard evidence that Tesla is truly better than Uber, Google, etc I think this is Tesla's least certain "win". Tesla has created and scaled cars that crush their competition, it's by far the leader in creating and scaling solar and storage, but the autonomous driving front still seems undecided.

 

[CarlK]

From the responses so far, this "battle" for autonomous driving seems to be far from decided. That was an impressive video from Carnegie Mellon and I think that some of the Tesla shareholders on these boards are getting a little ahead of themselves. Until AP 2.0 comes out or we see some hard evidence that Tesla is truly better than Uber, Google, etc I think this is Tesla's least certain "win". Tesla has created and scaled cars that crush their competition, it's by far the leader in creating and scaling solar and storage, but the autonomous driving front still seems undecided.

Tesla's approach is opposite to Google/Ford/Uber's but what that Uber car could do Google has had it five years ago. The latest Google car is much more advanced than that. It does not even have a steering wheel or pedal. Just that you know there is really nothing worth to mention other than a PR stunt of that "driverless" Uber car.

From Wiki:
"On March 28, 2012, Google posted a YouTube video showing Steve Mahan, a resident of Morgan Hill, California, being taken on a ride in Google's self-driving Toyota Prius. In the video, Mahan states "Ninety-five percent of my vision is gone, I'm well past legally blind". In the description of the YouTube video, it is noted that the carefully programmed route takes him from his home to a drive-through restaurant, then to the dry cleaning shop, and finally back home"

 

[Alketi]

Tesla's approach is opposite to Google/Ford/Uber's but what that Uber car could do Google has had it five years ago. The latest Google car is much more advanced than that. It does not even have a steering wheel or pedal.

As far as I'm aware, Google's self-driving "prototype" may not have a steering wheel or brake but this is not the vehicle that Google is testing on public roads.

Google is road testing vehicles with steering wheels and brakes and with a safety driver, exactly as you'd expect when you're aiming toward L4 but not quite there yet. In fact, Google rates their progress by the number of times the safety driver has to intervene.

 

[Garlan Garner]

No real winners, are you kidding? How about the 10,000's of people who needlessly die every year in human error caused auto accidents? How about the millions of hours wasted annually in mind numbing traffic related mostly to human driver error? Not to mention there is quite a lot of research showing how much better off lower income people are with good and cheap transportion, something that is severely lacking in the US but which could be mostly solved with the application of self driving technology bringing down the cost of transportation, making it not a significant financial barrier for those stuck in poverty in many places wrt to owning and maintaining a car.

Thanks. This is a very nicely written post. You covered many things in this post and you are correct in your assessment concerning poverty related issues. One thing I wish to make sure we emphasize is that many poverty challenged areas don't have a problem with the purchasing of cheap cars. Its the maintenance that is HUGE. Actually I wouldn't say that the folks who can afford a M3 are anywhere near poverty challenged. I understand what you are saying though.
One of the reasons I'm not riding in a new MS (even the 60) isn't because I can't afford the car payments.... its because I would have to also eat every now and then.

 

[Garlan Garner]

As far as I'm aware, Google's self-driving "prototype" may not have a steering wheel or brake but this is not the vehicle that Google is testing on public roads.

Google is road testing vehicles with steering wheels and brakes and with a safety driver, exactly as you'd expect when you're aiming toward L4 but not quite there yet. In fact, Google rates their progress by the number of times the safety driver has to intervene.

I would agree that Googles car is close to self driving while Tesla is currently a Driver Assistance type application. Am I criticizing Tesla - Absolutely not. Teslas application is far more trustworthy as it is booking a million miles every 10 hours.

I would much rather have quality in this area than an application from a company that is simply racing to be first.

 

[Garlan Garner]

The nice thing about Tesla's position is that with their incremental approach and OTA updates, they can essentially release a production self driving car way before others under the rules of a driver assist system while the other competitors have to reach the much higher goal of full on self driving car and wait around for the possibly delayed regulatory approval. With tesla it will just get to the point where tesla's a/p system is so good that drivers almost never have to take over and the data clearly verifies that A/P mode is safer than manual mode. The downside of this, of course, is the system being good enough that drivers get complacent and don't pay attention at all, even though they sign off on it being a driver assist system.

I don't ever believe that the "car" will ever be responsible for an accident. It will always be the driver or owners fault. The car won't pay the insurance bill every month. A human will.

 

[Adrien]

Hmmm. Lots of misinformation here with sprinkles of unrealistic expectations AND realities at the moment.

Let's define a few things:

Pulled from the internet
"
Level 0: This one is pretty basic. The driver (human) controls it all: steering, brakes, throttle, power. It's what you've been doing all along.

Level 1: This semi-autonomous level means that most functions are still controlled by a driver, but some (like braking) can be done automatically by the car.

Level 2: In level 2, at least 2 functions are automated, like cruise control and lane-centering. It means that the "driver is disengaged from physically operating the vehicle by having his or her hands off the steering wheel AND foot off pedal at the same time." The driver must be still always be ready to take control of the vehicle, however. So, level 2 means two functions automated. Easy to remember, right?

Level 3: Drivers are still necessary in level 3 cars, but are able to completely shift "safety-critical functions" to the vehicle, under certain traffic or environmental conditions. It means that the driver is still present, but is not required to monitor the situation in the same way it does for previous levels. Jim McBride, autonomous vehicles expert at Ford, said, "the biggest demarcation is between Levels 3 and 4." He's focused on getting Ford straight to Level 4, since Level 3, which involves transferring control from car to human, can often pose difficulties. "We're not going to ask the driver to instantaneously intervene—that's not a fair proposition."

Level 4: This is what is meant by "fully autonomous." According to the DOT, level 4 vehicles are "designed to perform all safety-critical driving functions and monitor roadway conditions for an entire trip." It's what Tesla says will beavailable by 2018.

*Level 5: It should be noted that some organizations, like the Society of Automotive Engineers (SAE), have their own charts that refer to "Level 5" vehicles. This refers to a fully-autonomous vehicle that does not have any option for human driving—no steering wheel or controls.
"


So we are at Level 2 legally ATM. Level 3 will be the next case where you must be in the front seat, but you can nod off or play on your phone legally. If there is a situation the car can't deal with it will give you ample time to take control otherwise it will likely safely pull over and wait for input. To get to level 3 and 4 will require approval and to get approval is just a data game and we KNOW Tesla is ahead with that. Let's not mistake ALL players in the game have full level 4 even level 5 prototype cars out there. Google, Apple, Universities, Mitsubishi, BMW, and yes, Tesla does too. I gather from Elon's "blowing people's minds" comments that he's ridden in one. As someone mentioned though, some prototype is not suitable for the masses. It must be proven and then approved by the DOT. So far everything I've said is accurate, what comes next is speculation. Getting the right hardware (integrated beautifully) and redundant control systems, I think Tesla is ahead of the game by a year or so. Software is up in the air - So many management and talent personnel being traded right now between companies, it's crazy haha. Again, data should win the game here and we know Tesla is winning at this. It'll also be the first to mass market the AP 2.0 hardware and if the speculation that AP 1.0 data is no good, they will still be the leaders as there are no other mass market cars delivering this.

 

[CarlK]

As far as I'm aware, Google's self-driving "prototype" may not have a steering wheel or brake but this is not the vehicle that Google is testing on public roads.

Google is road testing vehicles with steering wheels and brakes and with a safety driver, exactly as you'd expect when you're aiming toward L4 but not quite there yet. In fact, Google rates their progress by the number of times the safety driver has to intervene.

Google is now road testing an in house designed self driving (electric) car, not regular production cars fitted with sensors and computers. That car is easily recognizable you can see them everywhere on the streets near its headquarters here. Not sure how extensive the car is tested elsewhere though. It has a "driver" for safety reasons but no steering wheel and pedal.

 

[Alketi]

Google is now road testing an in house designed self driving (electric) car, not regular production cars fitted with sensors and computers. That car is easily recognizable you can see them everywhere on the streets near its headquarters here. Not sure how extensive the car is tested elsewhere though. It has a "driver" for safety reasons but no steering wheel and pedal.

Carl, I'd be seriously surprised by this. According to Scientific American, California requires steering wheels and brakes during testing, as one would expect: Google's Self-Driving Cars to Hit Roads, with Steering Wheels

Have you personally looked into the car and seen no steering wheel or brake? What does the driver do if there's a problem -- just hit a kill switch and sit there? It would be shocking to me if that's how they're being publicly tested. There's no reason for them to do that; it's a safety issue. /shrug

 

[Adrien]

Carl, I'd be seriously surprised by this. According to Scientific American, California requires steering wheels and brakes during testing, as one would expect: Google's Self-Driving Cars to Hit Roads, with Steering Wheels

Have you personally looked into the car and seen no steering wheel or brake? What does the driver do if there's a problem -- just hit a kill switch and sit there? It would be shocking to me if that's how they're being publicly tested. There's no reason for them to do that; it's a safety issue. /shrug

San Francisco has special laws:

 

[Intl Professor]

While I normally enjoy and respect Cringely's comments on computers and the business, especially his observations on why IBM is destroying itself, today he surprises with a blog entry on autonomous driving. Not new info for this thread, except perhaps for history, but confirmation of the bias I have along with the inevitability of this technology.

The self-driving car is old enough to drink and drive - I, Cringely

 

[neroden]

Full "Level 4" is impossible. The driving environment is simply too varied and heterogeneous. I drive on grass more than once a month. Good luck automating that.

We'll see Level 4 for specific environments -- popular ones -- not for the general case which has all kinds of weird corner cases. Given this, I think Tesla is taking the right tactic.

 

[Intl Professor]

I drive on grass more than once a month. Good luck automating that.

I'm not a golfer. But you're probably right.

Incidentally, Consumer Reports' most recent issue predicts all cars produced by 2040 will have autonomous driving, and a few other interesting possibilities. I'm mostly concerned, as many TMCers are about how people will be paid as machine learning progresses so fast. Already machines can best cancer specialists at diagnosing skin cancers.

I started, but haven't finished reading a book by a successful game developer type who says we will have to be paid by producers for the information about our consumption. I should take another look at it.

 

[Fallenone]

The single most deciding factor is data. With the ever growing fleet of Tesla, all other potential competitors combined can't rival the amount of Tesla has and the gap is not decreasing. Everyday, Tesla increases the size of the database equals to what Google collected over 7 years.

MBLY is partnering with other OEMs and collecting data to build a high precision map too. But the richness of this data is still not comparable to Tesla's. So in all likelihood, MBLY is more likely to create a high precision map first. But that's one of the many pieces of autonomous driving. In the domain of richness and amount of real-world data, no one comes close to Tesla at all. Autonomous driving development is also more and more relying on machine learning. And the key to machine learning is data. The one who doesn't have the best data cannot possibly become the best player. This is a component that neither talent nor money can grant you overnight.

8.0 shows exactly why data is crucial in establishing Tesla's lead in autonomous driving. And why it is the deepest moat Tesla has.

 

[Saghost]

8.0 shows exactly why data is crucial in establishing Tesla's lead in autonomous driving. And why it is the deepest moat Tesla has.

Actually, I think the fleet of sensor and actuator equipped OTA updateable cars is the biggest factor - as 8.0 shows. Without a hundred thousand cars they can update next week and send out to gather the data, the approach taken in 8.0 wouldn't be very practical - it would take years to build up a list if they had to send dedicated vehicles out like google streetview, and it'd only help you with you commute if they had only a few owner cars on the road.

 

[Fallenone]

Actually, I think the fleet of sensor and actuator equipped OTA updateable cars is the biggest factor - as 8.0 shows. Without a hundred thousand cars they can update next week and send out to gather the data, the approach taken in 8.0 wouldn't be very practical - it would take years to build up a list if they had to send dedicated vehicles out like google streetview, and it'd only help you with you commute if they had only a few owner cars on the road.

Yes we're talking about the same thing. Tesla won't be able to collect the data without the connectivity they've built into their cars.