Uber's Hiring Plans Show Outlines of Self-Driving Car Project

itwbennett writes The most interesting people that Uber is now hiring aren't drivers: they're engineers. The mobile ride-hailing app has listed a slew of jobs at its new Advanced Technologies Center in Pittsburgh. In particular, Uber is looking for engineers in the areas of robotics, machine learning, communications, traffic simulation, vehicle testing, and software and hardware development.

Comments

Great jobs except...

[Anonymous Coward]

Why on earth are they running it out of Pittsburgh of all places? I could see a tech area like Silicon Valley, RTP, etc. or near where other automakers work like Detroit. But Pittsburgh?

Re:Great jobs except...

[Stormy Dragon]

Yeah, why would a robotics company want to be near Carnegie Mellon University?

Re:Great jobs except...

[Anonymous Coward]

Yeah, because that's the only place doing research on robots and autonomous cars...

Re:Great jobs except...

[ShanghaiBill]

Yeah, because that's the only place doing research on robots and autonomous cars...

It isn't the only place, but it is easily one of the top three. Pittsburgh is not a tech backwater.

Re:Great jobs except...

[plopez]

1) Good wages. I mean really good wages.
2) Low housing costs
3) Some really good unis in the area
4) Brew pubs
5) Go Steelers!
6) For all you straight single young male professionals out there, a MUCH better male to female ratio.
and best of all......

No Californians! (or Texans)

But seriously I interviewed for a job there and the more I investigated it, the more I liked it.

Re:Great jobs except...

[Anonymous Coward]

If 7) is no hipsters, I might apply!

Hello!

[Stormy Dragon]

Welcome to jonny cab. Please state your destination?

Re:Hello!

[BarbaraHudson]

Once you get rid of the drivers, then you're not providing a taxi service, but a car rental service.

Re:Hello!

[Anonymous Coward]

Once you get rid of the drivers, then you're not providing a taxi service, but a car rental service.

If you are splitting hairs, both terms have baggage that does not convey the intended nuance. "Johnny Cab" is perfectly clear to anyone familiar with the cultural reference. "Robotaxi/robocab" would also work.

Re:Hello!

[Livius]

Once you get rid of the drivers, then you're not providing a taxi service, but a car rental service.

There's a difference between being a driver and being a passenger.

Re:Hello!

[plopez]

robots are people too you know. If a corporation can be a person so can a robot.

Re:Hello!

[BarbaraHudson]

Once you get rid of the drivers, then you're not providing a taxi service, but a car rental service.

If you are splitting hairs, both terms have baggage that does not convey the intended nuance. "Johnny Cab" is perfectly clear to anyone familiar with the cultural reference. "Robotaxi/robocab" would also work.

You missed my point - car rentals are not regulated like taxi rentals, so a huge legal problem for Uber disappears.

Too bad (?)

[Anonymous Coward]

That's a shame actually. One of the nicest aspects of Uber in my opinion is that they provide very decent employment for people who would be out of a job otherwise or need the extra income. I've never once taken Uber and thought gee, I wish I didn't have to deal with a person.

Re:Too bad (?)

[Fire_Wraith]

Unfortunately it's the way things are going to go, sooner or later. It's going to put a lot of people out of work, because driving is a key occupation for a lot of people, not just Taxis. It's also not just college students that deliver pizza, but all the delivery drivers, long-haul truck drivers, and it's these we should be worried about, because there's a lot more of them.

Even just going by the 2012 numbers from the Bureau of Labor and Statistics (data.bls.gov), there's something around 1.25 local/regional delivery drivers, and 1.7 million heavy and tractor trailer drivers, plus about 250k taxi drivers/chauffeurs. Roughly speaking, we're talking about something like 1% of the population here that will likely be replaced almost entirely by robots in the next 20 to 30 years, if not sooner. The number of jobs that take their place will be minimal.

Now, this will make the economy vastly more efficient in any number of ways. Robots don't doze off at the wheel because they tried to drive too long in a day, they try to attack their taxi customers, and they can work far, far longer than any human can. At the same time, these are not high skill jobs that are being eliminated, and many of these people will not be able to easily transition to other work, if at all. What are we going to do with that? Sure, eventually people will stop having the expectation that they can simply go into truck/taxi driving as a career... but I also don't think many were directly planning on that when in school, to begin with.

Instead, we're facing a situation where the amount of viable work for no or low skill workers is becoming smaller and smaller, and we're going to have to figure out what to do about that as a society where increasing amounts of people are simply unable to earn a reasonable living no matter how hard they're willing to work.

Re:Too bad (?)

[ShanghaiBill]

One of the nicest aspects of Uber in my opinion is that they provide very decent employment for people who would be out of a job otherwise

Maybe those people can spend their spare time reading about economic fallacies. The purpose of employment is the creation of goods and services, not "keeping people busy".

Re:Too bad (?)

[Fire_Wraith]

It is. Providing occupations is a side benefit, as is allowing people to earn enough to make a living. They've all gone together for so long that a lot of people take it for granted. There was a time, too, when economic activity was pretty much entirely comprised of human labor, plus a bit of work from the draft animals we had. That's not true anymore, and while you still need human involvement, it's a much smaller fraction than it used to be.

Moreover, we've gone from "making people more efficient in what they do", such as having a worker drive a horse-drawn cart, or later a truck, rather than hauling goods on their back to not even needing the worker at all, and instead having one or two people in an office directing all the various trucks their company runs remotely (the same people who provide instructions to the existing drivers, doing so for the robots instead). We're eliminating the jobs entirely, not just making them able to do more with less workers.

The question then becomes, what do we as a society do with those people? How are they going to earn a living, as society expects them to do? Some will be able to retrain into higher skilled occupations, but those people are the exception and not the rule. Minimum wage work isn't enough to survive on without significant government assistance, and even those jobs are going to become increasingly scarce.

My thinking is that we're eventually going to have to divorce economic survival from employment. Put in a guaranteed basic income, such that everyone gets enough to get by, and any wages earned go on top of that. It can take the place of pretty much all the current benefits (SS, WIC, Food Stamps, etc). You can get rid of the minimum wage, because no one needs to worry about earning enough to live on - all they're earning is for luxuries (however small), so markets can freely set the value of labor (because people would be free to quit without worrying about unemployment threatening their survival). Now, you'd probably also have to radically alter the tax base, and tax the output of capital (namely, robots and other highly automated or independent machinery) - partly since that's where more and more of the money will be. There's some precedent for this, as at the time the Constitution was signed, tariffs and excise taxes were the primary funding source for the government. Income tax didn't come until much later.

Re:Too bad (?)

[Anonymous Coward]

The number of jobs that take their place will be minimal.

There will still be people needed to load and unload the cargo from automated delivery vehicles. Humans are great at manipulating and sorting small and medium sized objects--still better than robots. Instead of the fish company employee unloading the truck, a restaurant employee will instead. If the soda truck is driverless, who stocks the vending machines?

It's going to be disruptive, and people will lose their jobs, but delivery encompasses more labor than just driving the truck. Self-driving trucks won't replace all of that.

Re:Too bad (?)

[Gr8Apes]

The number of jobs that take their place will be minimal.

There will still be people needed to load and unload the cargo from automated delivery vehicles. Humans are great at manipulating and sorting small and medium sized objects--still better than robots. Instead of the fish company employee unloading the truck, a restaurant employee will instead. If the soda truck is driverless, who stocks the vending machines?

It's going to be disruptive, and people will lose their jobs, but delivery encompasses more labor than just driving the truck. Self-driving trucks won't replace all of that.

That is extremely short term as robots will also be doing that work within 5-10 years. Personally, the self-driving car bits will be 10 years or less, probably closer to 5 before it starts making large inroads.

Re:Too bad (?)

[Gr8Apes]

The "utopian" future is coming whether we like it or not. The question is whether it will be Star Trek or Blade Runner.

Re:Too bad (?)

[CanadianMacFan]

Don't forget about all of the bus drivers. There probably will still be a need for someone to supervise on school buses but on city and between cities I can see them becoming automated.

Re:Too bad (?)

[alvinrod]

I suspect that we could stand to devout more resources to education. Even people who aren't exceptionally bright can likely teach at a primary school level, especially if they special in a topic area that they're good at. If the future is going to have even fewer low skill jobs, it's more important than ever to improve educating the next generation to fill the jobs that haven't been automated.

If we ever do reach a point where everything can be automated, or at least everything necessary to sustain human life, I suspect we'll have to move on to some other economic system that fits with the times.

Re:Too bad (?)

[plopez]

Or Charles Dickens

Re:Too bad (?)

[Hognoxious]

If you're loading from the same place you could do it with a static robot/smart crane. These have been in use in warehouses for decades.

For loading/unloading where the layout isn't standardised, perhaps a smaller version of this would work. http://www.bbc.com/news/techno...

Re:Too bad (?)

[Anonymous Coward]

A negative income tax or universal basic income would suffice in that situation.

Even if we do automate driving to the point that vehicle can handle any terrain and obstacle, things can still go wrong. How sure are we that loads will be 100% secure? Someone will need to monitor things. Maybe a tire blows. Who will change it? Etc. Maybe truck drivers have job security, but pay is a different issue. As for truck trains, I guess that would lower the pay even more.

Regardless, see The Simpsons episode Maximum Homerdrive.

Also, someone has to carry the pizza to the door. Although, receiving a text when it arrives would help. But no, I wouldn't want to do extra work. And when/if robots take the place of cashiers in fast food places, I'll most likely request to have a human take my order. (Not fond of the self-checkout machines in stores. Seems too impersonal for no financial benefit to me.)

great example of forward-thinking

[ks9208661]

I think Uber is doing it right. Kudos to them for thinking ahead. Meanwhile, they should hire more lawyers too, to ensure they company will still be around to make use of the newly-hired engineers.

Re:great example of forward-thinking

[ShanghaiBill]

I think Uber is doing it right.

Are they? Wouldn't it make more sense for them to just buy self-driving cars, or team up with someone like Google, or an automaker, rather than doing their own R&D?

Re:great example of forward-thinking

[Anonymous Coward]

Logistics of a taxi company any associated risks is somewhat different from a "self driving car".
 
Different research is required here for efficiencies for the job at hand, but as you said teaming up with another party would make a lot of sense.
 
However, if you are already a billionaire in cash, why not take the risk and potentially make yourself the richest person on the planet by orders of magnitude if you get it right?

Re:great example of forward-thinking

[ks9208661]

Whether or not they want to build their own cars, I don't know, but what the hiring indicates is that they want to have and develop skills and knowledge in dealing with self-driving cars (i.e. the future). It would make short-term financial sense for them to buy self-driving cars from other manufacturers and then Uberfy the interface, but the caveat is that they will be giving this key supplier a lot of negotiating power, which could be detrimental to their long-term business. If instead they partner with an established car manufacturer to create an Uber line of self-driving cars, the balance of power is more equal. In both cases, they need to be a master of this technology, and they need to develop this expertise now, and not when the future becomes the present.

Goog luck to Uber

[al0ha]

Google was initially interested in helping Uber develop a global robotic car service, but announced a short while back that they were dumping that partnership in favor of developing their own service. And so I say Goog Luck to Uber; they're gonna need it/

Big data mining alternative to Google

[DrYak]

Question: How do you jump start research into car robotics when you're not Google and thus don't have a huge mass of knowledge about all the roads gathered by your google cars and google maps program ?

Answer: You get a ton of hipster to drive for you, record their trajectories/behaviours (remember "god mode" ?) and use their knowledge as a starting point to populate your initial database.

Re:Big data mining alternative to Google

[angel'o'sphere]

Or you simply buy a few hundred and later a few thousand self driving cars. No god mode needed.

and some one get's hurt they point to the fine pri

[Joe_Dragon]

and some one get's hurt they point to the fine print and say your own your own.

Just think of Sofia Liu.

Finally, mostly correct terminology

[YrWrstNtmr]

"ride-hailing", not 'ride-sharing'
Paid (taxi) drivers, vs some guy who can give you a ride to the airport.

But yes...no reason Uber would not be looking into this. Other than maybe a dubious cashflow situation...

Re:Finally, mostly correct terminology

[Fire_Wraith]

Not to mention the bad publicity they've been getting from some high profile incidents involving their drivers (allegedly) robbing or raping passengers.

Re:Finally, mostly correct terminology

[Anonymous Coward]

You try have tens of thousands/millions of "employees" and not have one do something wrong...

Re:and some one get's hurt they point to the fine

[ShanghaiBill]

Just think of Sofia Liu.

Look, I am sorry the kid was killed, but although the driver worked for Uber, he was not doing so at the time. People claiming that Uber is responsible are basically saying "Hey, they are a corporation, and they can afford it, so therefore they must be guilty."

and what the higher skill people who will pulling

[Joe_Dragon]

and what the higher skill people who will pulling 60-80 hour weeks with no OT pay.

Re:and some one get's hurt they point to the fine

[Joe_Dragon]

but the drivers own insurance said he was working for Uber at the time and we are not covering this and Uber tried to use fine print to get out it as well.

Just wait a for auto drive that has parts pushed out to many different contractors and sub-contractors that when something bad happens they all point to each other when you are sitting back with you bills racking up as the courts are fighting over who will pay up.

Re:and some one get's hurt they point to the fine

[Anonymous Coward]

but the drivers own insurance said he was working for Uber at the time and we are not covering this and Uber tried to use fine print to get out it as well.

You realize, I hope, that you're now staking the credibility of your claim on the honesty of an insurance company.

Re:and what the higher skill people who will pulli

[Anonymous Coward]

And struggling to make ends meet on salaries in excess of a million dollars a year.

welfare paid in Soylent green?

[NotQuiteReal]

a society where increasing amounts of people are simply unable to earn a reasonable living no matter how hard they're willing to work.

used to send them off to war, but that is no fun either.

It was the driver's fault!

[penguinoid]

I thought Uber's business strategy was to blame the driver for anything that went wrong, eg legal or insurance issues. But how will they blame the driver when the driver is their own AI?

Re:It was the driver's fault!

[plopez]

Software Engineers.

The point is to build them

[DrYak]

Or you simply buy a few hundred and later a few thousand self driving cars.

The problem is that currently you can go to the nearest dealer and buy them.
They don't exist yet. There are just prototypes being developed here and there.
They need to be developed (which requires having a huge database to learn from).

Also, the problem of buying car from another company (say Google if their robo car is the first to be mass produced), it that Uber would become dependant on Google's whim. If their future business model rely on a service powered by robo cars, it would a bit risky to entirely depend on an external company for said cars.

The point of Uber, apparently, is to beat others in the development of autonomous cars. Not to depend on anyone else. Make their own robo car business.

And they have a similar mass of useful data out of which to build the car's intelligence. Which was graciously provided by the hipsters using the service, who never consented to be part in an AI research in the first place.
The kind of data log that used to enable the controversial god-mode, can also be used to build a very precise model of "how are the driver managing to navigate inside city centers ?"

Re:The point is to build them

[angel'o'sphere]

We have self driving cars since a decade.

Not on the market but as working cars.

There is no AI or databases to learn from involved.

Well, limited AI in the picture analysis algorithms ofc.

The point of Uber, apparently, is to beat others in the development of autonomous cars. Not to depend on anyone else. Make their own robo car business.

Unless they can steal some know how somewhere that is impossible. I mean the beating. Even by hiring all the experts that have worked the previous decade for other car manufactors or companies investing into autonomous cars, they would need to start from scratch.

If their future business model rely on a service powered by robo cars, it would a bit risky to entirely depend on an external company for said cars.

It would not. Your assumption how self driving cars work, is wrong.