Slashdot Mirror
Report: Google Partners With Ford To Make Self-Driving Cars (yahoo.com)
An anonymous reader writes: A new report from Yahoo Autos says Google and Ford plan to announce a partnership to build self-driving cars. "By pairing with Google, Ford gets a massive boost in self-driving software development; while the automaker has been experimenting with its own systems for years, it only revealed plans this month to begin testing on public streets in California. Google has 53 test vehicles on the road in California and Texas, with 1.3 million miles logged in autonomous driving. By pairing with Ford, the search-engine giant avoids spending billions of dollars and several years that building its own automotive manufacturing expertise would require. Earlier this year, Google co-founder Sergey Brin said the company was looking for manufacturing partners that would use the company's self-driving system, which it believes could someday eliminate the roughly 33,000 annual deaths on U.S. roads." Automotive News reported on the same plans independently, saying, "It isn't clear whether Ford would design a purpose-built vehicle for Google or supply a standard production car fitted with the sensors and computers that the car needs to guide itself down the road."
FORD = Fixed Or Repaired Daily or Found On Road Dead
Now if only Ford could make good cars. Or is this an attempt to suppress driverless cars by making them intentionally terrible?
Given that they'll be testing in California, is there any chance those cars will be electric?
You know you want one.
I still don't understand, why we don't have self-driving trains already — the task is so much simpler with one-dimensional roads, no size/weight restrictions on the necessary equipment, and full control of the signs and signals — without having to teach the computer to understand, what's meant for humans...
In Soviet Washington the swamp drains you.
It would be funny to shoot at it just to see how it reacts.
I think that should read "we're going to kill those 33,000 people in new and interesting ways."
Either way, the odds are good the Ford front end will yet again be cribbed from the Aston Martin style book.
- Political clout. Ford has plants in many states, and a network of dealerships in every state. And many loyal truck and car owners who vote.
Anything else? Well I did mention the network of dealerships, right.
1. There was no way Google was ramping up a manufacturing line. That would have been too risky and costly. It's always been about selling the software IP, just like Microsoft knew the money was in software licensing not the commodity hardware. Vehicles will be defined by software in the future. Expect Apple to jump into this game as well.
2. The pod car (sans steering wheel) Google has been testing would never have been marketed to consumers; it was a service vehicle for the taxi and delivery industries.
3. Autonomous vehicles will mostly be electric and aimed at urban areas where short travel distances don't require exceptionally high-capacity batteries and high-concentrations of pollution can best be targeted for reduction.
Just what we need, the Google Pinto!
This is cool, but it would be great if they could start by having Google replace SYNC.
https://en.wikipedia.org/wiki/...
If you post as Anonymous Coward, don't expect a reply.
I still don't understand, why we don't have self-driving trains already
We already do. You could have found that out in under 20 seconds on google. It's being implemented on normal railways too.
the task is so much simpler with one-dimensional roads, no size/weight restrictions on the necessary equipment, and full control of the signs and signals
I think you may be underestimating the complexity of train operations. Nevertheless the cost of a person to operate the train is much smaller in comparison to a car. Trains already do have a lot of automation and are getting more all the time but the financial potential of automation isn't nearly as large as with automobiles.
This is the only partnership that would let them get away with using the Model G name. It was inevitable.
We put that dress on a pig. It's going to run a muck and shit all over place, and everyone is going to say,"Well. Yeah. But it's silk."
And is having a Ford required or could people with other cars join in too?
"it's understood the venture would be legally separate from Ford, in part to shield the automaker from liability concerns"...because it may eliminate 33,000 deaths/year "someday".
This is not showing a whole lot of confidence for the short term, however. If there was a high confidence that the cars actually were safer, and that the financial risks were manageable, then Ford would just build the cars without a liability shield.
Ford and Google will take the profits, and the public can assume the risks.
If Nexus line is any indication, that will switch through every car manufacturer?
4wdloop
An average taxi driver will rack up 70,000 miles a year, so any individual taxi driver who has been in work for 20 years has more miles under his belt than Google's entire project.
Step outside the technological backwater known as the USA (sorry, but you are now) and you'll find self driving trains have been around for decades in europe and the far east.
What is the train-driver ("engineer") supposed to do, that a computer-program can not do?
Train engineers are effectively a pilot very much like the pilot for an airplane or a ship. Much of what they do can be and has been automated to some degree. In a few cases there are no human operators but not always. Engineers are useful both technically and economically. Full automation of any kind has to cover enough corner cases, do all the things that human operators can and be low enough priced to justify the high up front costs. Do a little studying on what it takes to operate and pilot a train safely. You'll find it's quite a bit more complex that you would ever guess. The economic value of automating the remaining bits is harder to justify because the returns aren't amazing.
A train driver gets paid salary, health-benefits, and retirement plan — not insignificant monies.
You have to divide that salary by the value of what is being transported. A train carries FAR more cargo than anything except a ship so the unit cost of their cargo is generally lower than for a car. Nobody is staying engineers are cheap but rail is economically very competitive for a reason. Warren Buffet bought BNSF railroad because it is very profitable and likely to remain so. Much more so than pretty much any trucking company out there.
I do not deny, that a self-driving car is useful. But the hardware and software required to achieve that is vastly more complex.
Yes it is more complex but that doesn't mean that the automation for trains is simple. That is a separate issue from what the economic opportunity is. You could fully automate every train in the US and it wouldn't make a huge difference from what we have now. Railroads would make a bit more bank and engineers would have to find another occupation. Fully automated cars on the other hand would change daily life as we know it. Yes it's a harder problem to solve but there is FAR more money to be had in solving it.
Automating trains would've been the logical first step — the way black-and-white TV was a precursor to color.
If the economics to automate railroads fully was available and made economic sense it would have been done already. Humans are still in the loop for a variety of reasons, mostly economic, some technical and a few political.
I would even go into the apple store for one of these bad boys
...oh wait...
But would not a computer be cheaper and better still?
Not necessarily and the reasons are usually based in economics. I run a manufacturing plant for my day job. Much of what we make in our plant can be substantially automated. The technology already exists and I could write a check to buy it tomorrow. But I don't most of the time. Why? No return on investment. For automation to make sense a few things have to happen. 1) The automation has to do a job with adequate or better competence than the people it is replacing. 2) The automation has to deliver unit costs lower than the wages that would be paid for the lifetime of the project. This has to include the opportunity cost of tying up that capital, the cost of financing, the loss of flexibility, and the risk of the project. It doesn't matter if you are producing widgets or passenger-miles or transporting cargo. You have to be able to recoup the cost of the automation. If you can't then there is no point in even developing it, much less buying it.
Generally speaking any kind of automation requires a relatively large amount of production volume to justify the higher up front costs and loss of flexibility. It's trivial to show cases where it does and doesn't make economic sense, both for manufacturing and for automated transportation. Computers are demonstrably not always cheaper. Full disclosure, I'm a certified cost accountant. I do these sort of calculations for a living. Computers are NOT necessarily cheaper or better.
Drive a gord, the greenest vehicle on the planet. Grow 'em in your back yard and put wheels on 'em. Sorry. I'm just in that kind of mood.
For all intensive purposes, "whom" is no longer a word. That begs the question, "who cares"?
but the per-unit costs should be substantially lower.
The salary of the train drivers is only a small fraction of the total operative costs for the train company.
Same also for a potential computer.
Whoever/whatever you put on the command of a train isn't going going to play a big part on the price of the ticket.
Companies have no financial incentive to switch from one to the other. Maybe they'll spare a small fraction of percent of the price.
What will probably happen in the near future is probably what happens in some airplanes and a few high-speed trains:
as they are only a small fraction of the global cost, they come (relatively) cheap and you can end-up using both.
Computer that handles the routine work, and human overseers to check for unforeseen situation when the computer needs help.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
Ford is actually a decently managed company--maybe the only major US Automaker not to declare bankruptcy during the financial crisis.
I've never had an especially good experience with their cars. The last Ford I rented was so short inside that sitting down, the top of my head hit the ceiling. But their trucks are okay.
Either way, though, a self-driving car is a significant competitive advantage, and google has both experience and a lot of mindshare in the space.
the car drives itself to the repair depot even fortnight to replace the parts that have fallen off, worn out or were substandard to start with.
makes it feel like you have a good, reliable car!
had a ford years ago. people would say the spare parts are cheap. i'm not surprised, seem to get through 10x the parts of my uk built japanese car!
A higher quality car since they will have to stop running windows on it.
I prefer the "u" in honour as it seems to be missing these days.
Makes sense. Ford has expertise and factories to mass produce cars. Google has the AI, computer vision, and other necessary expertise to make the self-driving part work. Google is not a manufacturing company at heart and it doesn't make sense for it to become one, at least not right now. (That could change in the future, as it has for Microsoft.)
Some hard questions remain. Should Google find an exclusive manufacturing partner or offer Google Cars that are made by more than one company? How should the cars be branded: Ford, Google, Google Car by Ford, Ford featuring Google Car, or something else?