• UberPeople.NET - Independent community of rideshare drivers. It's FREE to be a person and enjoy all the benefits of membership. JOIN US! CLICK HERE

To understand Uber's predicament, you have to understand self-driving cars.

tomatopaste

Well-Known Member
When Uber launched in 2009, self-driving cars were still stuff of science fiction. In fact self-driving cars caught Uber off guard. About two years ago smart people around the world began to realize - hey this thing is really going to work.

Two years ago is also when Uber began their doomed entry into the world of self-driving cars. What else happened two years ago? Uber slashed fares.
Self-driving cars upended Uber's entire game plan, now they find themselves between a rock and a hard place. They know they can't play, and definitely won't win the the costly self-driving game, so they're hoping to create value with the largest customer base. In the end it's not gong to work. At this point they're just trying to buy time while they come up with a new game plan. Thus the 180 days of window dressing.

First commercially available self-driving cars will be on the road in less than five years. Within 20 years, everything is self-driving.

Here are the numbers: (and this is just for the U.S.)
1. Three trillion miles driven in the U.S. annually.
2. Those creating the technology think they can get the cost down to ten cents per mile.
3. That's 300 billion dollars annually. Over a trillion dollars worldwide. ANNUALLY!

The short version:

The long version:
 

Atom guy

Well-Known Member
Self driving cars solve no actual transportation problem. Neither does Uber. Ride share cars actually cause MORE traffic congestion, not less, because unlike a solo driver in their own car who drives to a destination then gets off the road for 8 hours, the ride share car remains on the road all day, and all those empty miles between fares is excess congestion that otherwise wouldn't exist.
 

Lissetti

Rebel Honey Badger
Article Manager
Moderator
Yep! I know that's the future. I'm in school right now getting my Bachelor's degree in computer science. Formerly I've been a semi truck driver, and I've spent most my teens restoring and rebuilding VW's, such and the one in my profile pic, my 1968 Bug. I'm going to take all this former experience, and my computer science degree, and be an Antonomous Vehicle software engineer. It's the future. It will become a norm. Also I'll do side work in robotics, such as Amazon's Kiva 800, Robo-Stow, and the drones.
 

tomatopaste

Well-Known Member
  • Thread Starter Thread Starter
  • #5
Self driving cars solve no actual transportation problem. Neither does Uber. Ride share cars actually cause MORE traffic congestion, not less, because unlike a solo driver in their own car who drives to a destination then gets off the road for 8 hours, the ride share car remains on the road all day, and all those empty miles between fares is excess congestion that otherwise wouldn't exist.
Self-driving cars solve almost every transportation problem:

1. parking
2. traffic
3. cost
4. road rage
5. drunk driving
6. accidents/fatalities
7. Mobility for the disabled/elderly/kids
 

Atom guy

Well-Known Member
Self-driving cars solve almost every transportation problem:

1. parking
2. traffic
3. cost
4. road rage
5. drunk driving
6. accidents/fatalities
7. Mobility for the disabled/elderly/kids
Parking - yes
Traffic - definitely not - these cars will be in continuous motion on the streets instead of parked most of the day
Cost - highly unlikely - all the autonomous equipment will not be cheap
Road rage - for the people in these cars yes, for the human drivers, no
Drunk driving - yes, but drunks may have a hard time communicating with the car
Accidents - only when all cars are autonomous
Mobility for disabled/elderly/kids - doubtful. These groups are most in need of autonomous cars, but the least able to use the technology.
 

koyotemohn

Well-Known Member
Self driving cars will cause lots of injuries/accidental death for pedestrians and bicyclists.

Also you need a lot more than waze and some cameras in a hi power CPU. You need a massive server and a whole lot of miles on the road as the cars and server "learn" each city and highway zone. Add in road maintenance and construction and rush hour congestion and general intengibles.

Not just that...but hackers can disrupt the server or take over a drone.

5 years?

Minor tests are happening now. It's more like 25-50 years....and even then human companions should be a mandate.

Had Uber been smart...they could have partnered with a company and syphoned some of those learning/algorithms of a vehicle...maybe at a discount to the car owner/operator.

It's a great fanciful concept. How many people will have to be injured or hurt by autonomy before humans realize we still need a whole lot of humanity in the equation of being a driver?
 

Uberingdude

Well-Known Member
+1.

Liability and navigation are the achilles heel....no solution for either of those presently.
My car has sensors for cruise control and other things. When it is raining hard the sensors stopped working. I don't know how they would overcome this with self-driving cars.
 

Atom guy

Well-Known Member
My car has sensors for cruise control and other things. When it is raining hard the sensors stopped working. I don't know how they would overcome this with self-driving cars.
Today I came across a single traffic cone with caution tape around it tied to a fence on the side of the road, blocking one lane of a 2 lane road. IDK how an autonomous car would even notice that, let alone understand what it means. There are millions of variables these systems would have to understand. As advanced as they seem, these systems are still very primitive.

+1.

Liability and navigation are the achilles heel....no solution for either of those presently.
For all the effort in getting autonomous cars to be able to see and understand the road ahead, current maps are woefully incomplete and inaccurate. I would see ride are passengers having to hunt for their cars on a regular basis with current maps.

Today I did a pick up on a road that was on the map, but not part of the gps directions. Luckily the passenger had dealt with this before and called me while I was headed towards him. The gps would had me stop on the side of the road about a mile away from the actual destination, and when I finally got there his house was set way back from the road on a long driveway. Good luck AI. Ubers navigation also had me go down a dead end road that it believed took me to my destination. The road was so dead end that you couldn't even see where the road used to be.
 

tomatopaste

Well-Known Member
  • Thread Starter Thread Starter
  • #15
Parking - yes
Traffic - definitely not - these cars will be in continuous motion on the streets instead of parked most of the day
Cost - highly unlikely - all the autonomous equipment will not be cheap
Road rage - for the people in these cars yes, for the human drivers, no
Drunk driving - yes, but drunks may have a hard time communicating with the car
Accidents - only when all cars are autonomous
Mobility for disabled/elderly/kids - doubtful. These groups are most in need of autonomous cars, but the least able to use the technology.
Traffic definitely yes. Freeway traffic is caused by someone hitting his brakes, then a chain reaction happens. Eventually all cars will be communicating with each other. In the morning all lanes could be going toward downtown. Reverse in the evening.

Cost yes. Avg person uses their car 4 percent of the time.

The car will have a speaker and the human at the call center will talk to the drunk. If still having issues after 30 seconds, the car will just run over the drunk. Next problem.

Accidents. 95 percent of all accidents are human error. The more human you take out of the equation the fewer accidents.

Mobility. What? Please show your work.

Self driving cars will cause lots of injuries/accidental death for pedestrians and bicyclists.

Also you need a lot more than waze and some cameras in a hi power CPU. You need a massive server and a whole lot of miles on the road as the cars and server "learn" each city and highway zone. Add in road maintenance and construction and rush hour congestion and general intengibles.

Not just that...but hackers can disrupt the server or take over a drone.

5 years?

Minor tests are happening now. It's more like 25-50 years....and even then human companions should be a mandate.

Had Uber been smart...they could have partnered with a company and syphoned some of those learning/algorithms of a vehicle...maybe at a discount to the car owner/operator.

It's a great fanciful concept. How many people will have to be injured or hurt by autonomy before humans realize we still need a whole lot of humanity in the equation of being a driver?
Watch the videos. Self-driving cars are already better and safer drivers than humans. Minor test? Google runs 3 million miles of simulations per day.

Today I came across a single traffic cone with caution tape around it tied to a fence on the side of the road, blocking one lane of a 2 lane road. IDK how an autonomous car would even notice that, let alone understand what it means. There are millions of variables these systems would have to understand. As advanced as they seem, these systems are still very primitive.



For all the effort in getting autonomous cars to be able to see and understand the road ahead, current maps are woefully incomplete and inaccurate. I would see ride are passengers having to hunt for their cars on a regular basis with current maps.

Today I did a pick up on a road that was on the map, but not part of the gps directions. Luckily the passenger had dealt with this before and called me while I was headed towards him. The gps would had me stop on the side of the road about a mile away from the actual destination, and when I finally got there his house was set way back from the road on a long driveway. Good luck AI. Ubers navigation also had me go down a dead end road that it believed took me to my destination. The road was so dead end that you couldn't even see where the road used to be.
The maps they're using for self-driving cars make make waze and google maps look primitive

Today I came across a single traffic cone with caution tape around it tied to a fence on the side of the road, blocking one lane of a 2 lane road. IDK how an autonomous car would even notice that, let alone understand what it means. There are millions of variables these systems would have to understand. As advanced as they seem, these systems are still very primitive.



For all the effort in getting autonomous cars to be able to see and understand the road ahead, current maps are woefully incomplete and inaccurate. I would see ride are passengers having to hunt for their cars on a regular basis with current maps.

Today I did a pick up on a road that was on the map, but not part of the gps directions. Luckily the passenger had dealt with this before and called me while I was headed towards him. The gps would had me stop on the side of the road about a mile away from the actual destination, and when I finally got there his house was set way back from the road on a long driveway. Good luck AI. Ubers navigation also had me go down a dead end road that it believed took me to my destination. The road was so dead end that you couldn't even see where the road used to be.
You obviously didn't watch the videos. I would suggest you watch the videos.

My car has sensors for cruise control and other things. When it is raining hard the sensors stopped working. I don't know how they would overcome this with self-driving cars.
That's ok. The engineers do.
 

koyotemohn

Well-Known Member
You can't simulate a bunch of kids getting crippled or killed on a school bus due to a malfunctioning a.i. with 3 million simulations per day. I know that deep learning on high end mainframes is fascinating and makes you want to run back and start your python classes in earnest.

A.i. Can get hacked, corrupted or make a good math on paper ....really bad results in regards to civic policy.

Just like your assertion that human error is the primary cause of accidents, accidents will still happen with as a.i. Proliferate into the transportation sector.

Uber cannot use google's technology anyhow...due to their dubious acquisition of the intellectual brain trust property.

I know tha everyone is super certain that this will happen in the near future.

I don't buy it.

I'm not saying that it will never happen. I just don't think Uber is run responsibly enough to implement that kind of technology.

I'm not against automation. I just think that people get infatuated real easy with what is on the horizon instead of dealing with what is happening now.
 

Cableguynoe

Well-Known Member
I'm not saying that it will never happen. I just don't think Uber is run responsibly enough to implement that kind of technology.
Don't forget it's not just UBER that thinks this is close to happening. Google is all in also

I don't believe a company like Google would be working on this if they didn't believe it was possible.
I understand Google has money to lose, and if they completely fail at this they'll still be ok.
But I just don't see them getting behind a 50 year plan. Not even a 20 year plan.
They obviously know a lot more than us and they believe it's possible.
So that leads me to believe that this is closer than we think.
 

swingset

Well-Known Member
The maps they're using for self-driving cars make make waze and google maps look primitive
Lol. Cool aid. No, they're not more accurate, they're just more detailed. Too much involving mapping is predicated on voluntary reporting that never gets done. Just because a map is accurate to identify a street/building is not synonymous with the ADDRESS and occupancy being correct. We all know this as drivers....how many times does the app guide us to something that isn't there, or is closed, or abandoned? That's not fixable with more $$ and tech....if people don't report it, it doesn't get updated. It's a two way system, and NO ONE has a fix for this. Johnny Robot showing up to the wrong place, or guided to the middle of a bridge that's supposed to be his building is going to happen. A lot. There's a solution to this (guidance to pin/physical location of the phone), but that requires an incredibly adaptive guidance tech, and pax who will facilitate the robot's needs....likely? Lol.

So that leads me to believe that this is closer than we think.
Remember the optimism built around Segways? "They'll build cities around them!". Well, they're stupid gadgets now and a frivolity, nothing more.

Just because a lot of people are pushing a tech doesn't mean society will embrace it tomorrow or that the legal & government road blocks won't be substantial.

Look at it in the simplest terms possible. If you screw up behind the wheel, who gets sued? Raytheon who made your ECU? Chevy? Nope, you.

Who gets sued EVERY. SINGLE. TIME. an autonomous car makes a bad ethical decision and someone gets hurt? Or, the sensors are fouled or dirty and it mows over an Amish family? Or it can't detect a crowd of people wearing black in the rain and blows through them? That's Google or the maker, every time. How many lawsuits can they withstand? Remember, zero liability from owners or operators, 100% on the maker. That stuff is not going to be commonplace, but it's going to happen as these things go prime time and teething issues are worked out.

I think autonomous vehicles will happen, absolutely, but it's not just around corner. Point to point maybe...across a desert to a distributor? Sure. Not doing what we do every night in a crowded bar district where that alley that used to be a drop off point is now off-limits. There's no way they're even close to that kind of decision making and accurate mapping detail to make that work. Our infrastructure just won't allow it. Not yet.
 

tomatopaste

Well-Known Member
  • Thread Starter Thread Starter
  • #20
Don't forget it's not just UBER that thinks this is close to happening. Google is all in also

I don't believe a company like Google would be working on this if they didn't believe it was possible.
I understand Google has money to lose, and if they completely fail at this they'll still be ok.
But I just don't see them getting behind a 50 year plan. Not even a 20 year plan.
They obviously know a lot more than us and they believe it's possible.
So that leads me to believe that this is closer than we think.
The guy is these videos, Chris Urmson, says within five yrs. He also says they are very close. He is one of the top engineers, if not THE top engineer, in self-driving cars.
 
Top