Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Tesla / Google Partnership, Coming Soon?

This site may earn commission on affiliate links.
Not good enough for driverless cars, no. But I think Google's self-driving cars are over 300,000 miles with only one accident so far, and that was operator-caused to boot. So the technology is ready today to make driving safer for many people, though they'd still be required to sit in the driver's seat and monitor. It's not Utopia, but it's still an amazing feat of engineering IMHO. And it's only going to get better.
That is interesting. I have not thought about it this way... There are definitely potential to make cars more safe, by reducing human drivers to mere operators for most of the time...

But think of those guys who love manual and hate automatic... They would go insane over the idea that car will drive itself 95% of the time....:scared:
 
My thoughts about future transportation:

1. First all cars will be electric
2. Then all cars will be self-driving
3. Then, finally, people will realize that they don't even need their own car. If you just open the taxi app on your phone and you'll have a self-driving car by you in a minute I don't get why people would want their own cars. This way hardly any cars would be left in a parking lot all day. It would make transportation so much cheaper
 
But not if just getting a cab is just as fast and cheaper

Do you live in a city? I think you might, so I can see how that might make sense to you. In a rural area (like I live in), IF I can get a cab, I have to call to the nearest bigger town. That's at least a 15-20 minute wait if they have a car available. Then I have to pay the guy to take me somewhere and also pay to have him wait for me or wait for another car to be called.

contrast that with going down into my garage, getting in my car, and going wherever I want, when I want, and for the cost of my own electrons. No brainer IMO.
 
Do you live in a city? I think you might, so I can see how that might make sense to you. In a rural area (like I live in), IF I can get a cab, I have to call to the nearest bigger town. That's at least a 15-20 minute wait if they have a car available. Then I have to pay the guy to take me somewhere and also pay to have him wait for me or wait for another car to be called.

contrast that with going down into my garage, getting in my car, and going wherever I want, when I want, and for the cost of my own electrons. No brainer IMO.

Yeah, you're right. I'm a student in the middle of a city without a car. I totally see you're point today, but I'm just trying to imagine the future where cars doesn't need drivers. That means that you wouldn't have to pay a cab driver to come and get you. Today, every single person have their own personal car even though they use it for probably about an hour or two every single day. Even in rural areas there could be self-driving cabs driving around just waiting for somebody to need a car, but you wouldn't need a car for every person in the area (i suppose you don't live all by yourself without a neighbor within miles). To pay for the car only when you need it would be really cost saving compared to owning it 24/7. Today this obviously doesn't work because you need to pay a person to drive the car for you, but I'm just trying to imagine where this technology can take us in the distant future
 
Have you seen the video of the Google car racing autocross around the parking garage roof? http://www.motorauthority.com/news/...driving-car-hits-an-autocross-course-at-speed

Sure. I follow driverless cars since DARPA Grand challenge of 2004 and became a big fan of Sebastian Thrun, who's team won the challenge in 2005 and Sebastian later became in charge of Google driverless car program...

But what that video could possibly prove? I mean if we got two situations:
1) one meter peace of colored plastic bag(or styrofoam) blown by the wind on the road.
2) child running crossing the road.

Speed of obstacles are similar. Speed of car is high.
In both situations human or AI could either choose to run over child/plastic, or can choose to crash into different car potentially causing injuries to occupants of driven vehicle and occupants of other car...

Are you claiming that IA is good enough to recognize
a) the situation
b) distinguish between plastic peace and child
c) take best possible actions based on assessment.

I claim that AI for b) is not good/reliable enough to be used in real time. Current state of art in speech recognition, heck, my phone today messed up again when I have said "Iranian" and offered me "Italian" word... You claim that AI in the car is good enough to recognize things in real world?

Read through this thread on driveless cars and get back to us.

You a big guy who get used to tell people what to do, correct? Security guard would be my guess?

And could you collaborate who is "us" you are referring to? Like bunch of you who made a foolish but bold claim that Model S was using NCR18650a cells? And then after it was completely debacled by Chief Technical Officer of Tesla Motors, switched to similar claims without bothering to provide any proof? :biggrin:
 
...
I claim that AI for b) is not good/reliable enough to be used in real time....
I agree that they are not now, but your implication is that they would never be good enough. I was merely pointing out the state of the art of driverless cars is pretty amazing now. Why does speech recognition even matter? Do you want to yell at Kitt?

You a big guy who get used to tell people what to do, correct? Security guard would be my guess?
Yep.

... Like bunch of you who made a foolish but bold claim that Model S was using NCR18650a cells? ...
If you show me where I said such a thing I will apologize for being wrong, hang my head in shame, and turn in my security badge forever.
 
I agree that they are not now, but your implication is that they would never be good enough.

No! That not what I meant. I'm a big fan of Ray Kurzweil and his Singularity theory. So eventually there would be AGIs that are smarter then humans. He predicts that computers would exceed human intelligence at around 2030. I think first systems will be commercialized around 2025.

But ten years earlier or 30 years later, doesn't matter, eventually AGI would be around.

Speaking of the devil, Ray was hired by Google recently too...

I was merely pointing out the state of the art of driverless cars is pretty amazing now.
Yes. I totally agree. It was even more amazing in 2005 to listen to Sebastian of how they did their driverless car that driven in the real world... Without humans inside. And his vision of much safer automotive future.

The very important point that ~30.000 Americans die each year in car crashes. And that driverless cars could dramatically reduce that number, among other benefits. 30k bodybags is 30 times more then America ever lost in Iraq War in a single year. But Iraq was the thing everyone was talking about, on the other hand car crashes were barely on the radar of society.

Why does speech recognition even matter?
Because object recognition is a MUCH harder task then speech recognition. And to drive safely AI would need to recognize as in example I mentioned: is that a piece of plastic or a human child?
ASR while being much simpler task, is much more commercially developed. Millions of people use technologies from Microsoft, Google, Nuance, iFlytek daily. But it still works far from perfect.

Do you want to yell at Kitt?
Sorry, my English prevents me from understanding this probably cultural reference. But I will try to Google it...

If you show me where I said such a thing I will apologize for being wrong, hang my head in shame, and turn in my security badge forever.
Upps, my apologies. I was being wrong.

BTW, Elon's take on driverless cars: they would come into reality as gradual extantion of active safety features. Please check video, not text here: Elon Musk interviewed at USA Today (at around 1:35).
 
I don't get why people would want their own cars.
Open up your favorite mapping site. Find the least "roady" place within 100 miles of where you live. Drive there. Disable your vehicle. Now call a tow truck to get home. I think you'll understand pretty quickly how it sucks to be dependent on others completely.

Bonus points: Do it at 3am when you have no food and it's 30 F outside.
 
Cover article of The Economist this week is one self-driving cars.

@Zzzz..., while I agree that there are some situations where an alert human driver will be better than self-driving cars (at least for a while), there are other situations where self-driving cars will be better. The easiest examples are the blind and certain elderly drivers with impaired visions/reflexes/mobility/etc. Self-driving cars will open up new levels of personal mobility to these and other categories of drivers.
 
Cover article of The Economist this week is one self-driving cars.

@Zzzz..., while I agree that there are some situations where an alert human driver will be better than self-driving cars (at least for a while), there are other situations where self-driving cars will be better. The easiest examples are the blind and certain elderly drivers with impaired visions/reflexes/mobility/etc. Self-driving cars will open up new levels of personal mobility to these and other categories of drivers.

Plus, imagine the efficiency gains on the highway. Trains of cars following at the appropriate distance and speed -- and if cars were aware of others (networked in some fashion), we might avoid traffic jams as well.
 
Plus, imagine the efficiency gains on the highway. Trains of cars following at the appropriate distance and speed -- and if cars were aware of others (networked in some fashion), we might avoid traffic jams as well.
How about they design (and deliver in product) a cruise control that efficiently manages the "freeway hills" on I-5 first? ;) When driving north on I-5 from SeaTac area, setting the cruise to 60 in the Model S is far less efficient than manually maintaining a 55-65 range. At least in my experience.
 
Plus, imagine the efficiency gains on the highway. Trains of cars following at the appropriate distance and speed -- and if cars were aware of others (networked in some fashion), we might avoid traffic jams as well.
Yes, that is another point of driverless cars. Dramatically increasing roads throughput. Effectively preventing society from building more highways and lowering travel time. I do not remember numbers, but those were very impressive and eyeopening.

Other benefit of driverless cars - fuel savings. Car trains are essentially tailgating each other...