Slashdot Mirror
Ford Has An Idea For An Autonomous Police Car That Could Find A Hiding Spot (jalopnik.com)
Ford has submitted a patent application to the U.S. Patent and Trademark Office for an autonomous police car that could function "in lieu of or in addition to human police officers." From a report: Now, companies always file patents for technology that may never get made, but an autonomous police cruiser seems like the logical conclusion to the development self-driving cars. But damn is it weird to read about. The patent, describes how the hypothetical car would rely on artificial intelligence and use "on-board speed detection equipment, cameras, and [it would] communicate with other devices in the area such as stationary speed cameras."
In a world of self-driving cars, there will not be any speeders. OPPS!
Ford could partner with Roomba to 'clean up' blighted neighborhoods and provide security theater in more affluent neighborhoods.
Ford could also partner with Google to provide Live up-to-the-minute Street View updates for subsequent off-line analysis for crimes and offenses.
Ken
... Ford tries to raise its share price by making fatuous announcements using flavour-of-the-month automated car meme.
It's not an invention if it doesn't exist! The patent office should reject it!
I'm picturing police cars that leave the station parking lot and drive to where they can hide. ;)
If you are going to cut out the police officer from the equation, why bother making it so complicated?
And in jurisdictions where photo radar can't be implemented because of legal restrictions, the same factors that make photo radar illegal would also outlaw automated police cars.
File under 'M' for 'Manic ranting'
Police say it all the time. If you've done nothing wrong, then you have nothing to hide.
So why do they need to hide? What are they hiding? What are they afraid of?
Isn't the whole point of policing to increase public safety? Isn't the point of enforcing the speed limits to increase public safety? If a visible police presence is seen on a roadway, that alone will deter most drivers from speeding. The stupidest drivers who speed anyway will then get ticketed for speeding.
Police might object that the number of tickets they write would diminish. But isn't that the whole point? If you think that ticketing is a source of revenue then you've already gone down the wrong side of a slippery slope that leads to all kinds of crooked behavior by police. Next police start to think that all sorts of crime should lead to revenue. Lesser and lesser infractions lead to assets seized until at least no crime is needed at all to justify just robbing people for no reason. And this already happens in some places. Police will stop and rob people who have done nothing wrong except for merely being out of state. Seize their money and send them on their way.
Writing tickets is not a goal in and of itself. The goal is to get people to stop speeding. Not to raise money. If a visible police presence stops most speeders, then the job is being done on a better and larger scale than not being able to ticket every single speeder. Hiding is a sign that police ARE doing something shameful and wrong.
I'll see your senator, and I'll raise you two judges.
What you should be patenting is a flying police car.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
The point of being human is that humans have true consciousness, freewill and can make choices using novel creativity.
Either service the machines or become fuel for their operation. Your choice. Be as creative as you want in choosing. You have ten seconds to decide.
I'll see your senator, and I'll raise you two judges.
The problem is that they are determined to make it work with the technology we have. They think they can write a set of rules for driving in the real world. I don't think they really believe a ruleset will handle everything, but they are willing to put the onus on the owner of the car for the rest.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
How will they deal with source code and calibration log requests? as if they can't give them our or the calibration has not been done in a log time that ticket get's dismissed
More patents hinder innovation. Especially stupid patents. Or patents not supported by innovation or actual invention.
And people wonder why the US has dropped off the list of most innovative countries.
I'll see your senator, and I'll raise you two judges.
I think they believe that they can make good VC money working on autonomous cars because they think it is cool and want to avoid real work, and the tech industry is swamped with excess money. They should start spending that money on projects that actually improve technology, rather than chasing projects that will never work and we don't need. Are we suddenly going to run out of people that drive cars and trucks? How about spending money on fixing the huge security holes in our software? Too boring and too much like real work. We would rather discuss going to Mars and AI.
50 years ago, we hadn't landed on the moon yet.
30 years ago, the first smartphone hadn't been invented yet.
Hell, 10 years ago, home use 3D printing wasn't really a thing
Are there a lot of technological hurdles to overcome before we have good self-driving cars? Ones that even your grandma feels safe using? Sure.
But to say, flat out, that we're never going to reach it is asinine.
Mr. Hu is not a ninja.
You have 30 seconds to comply!
Ford is building the precursor to ED-209. Will the modern car soon feature a telemetry system that sends data back to the manufacturer? which then have a back door for LEA's to monitor and automatically send you speeding tickets - or at least send the autonomous cruiser to investigate and eventually gather the evidence to ticket you?
It is better to be the hammer than the anvil.
Actually, I don't have an iPhone. I refuse to buy one... but I digress.
No, progress is not inevitable. But we're not at the end of it yet either. There are self-driving cars now. Are they great? Not really. And we're probably never going to have autonomous Formula-1 racing. But progress is being made in the field.
Sure, if there was only one company looking at self-driving cars, then it would be much less likely to happen in the short or long term. But there's more than one company looking at it. And competition in the field drives (if you'll pardon the pun) progress.
Mr. Hu is not a ninja.
IT IS A FUCKING WISHLIST
Inventions are patentable.
Wishlists and vague descriptions of features are not. At best, they can be prior art preventing the patenting of an actual invention some time in the future.
"What you should be patenting is a flying police car."
Prior art.
Flying police cars are called police helicopters.
Autonomous flying police cars are called police drones.
As long as the car talks with a sarcastic tone and is a Pontiac Firebird.
I'm a good cook. I'm a fantastic eater. - Steven Brust
I cannot wait for the New Detroit.
https://www.youtube.com/watch?...
Remember kids, if you're not paying for the service, YOU ARE THE PRODUCT THAT IS BEING SOLD.
You are really, and truly, an idiot.
Why do you keep banging on about Moore's Law? Do you even know what it is? It says nothing about computing performance or technological progress. All it says is how many transistors can be crammed on a die. There is a lower lower limit to how big a transistor can be since you can't make it smaller than the electrons that flow through it.
This just means that future advances will be in different avenues. 3D chips, multi-core systems, optical interconnects, efficient thermal usage. All of these things are being developed today. Some of them won't pan out. Some of them will.
You are not even a Luddite. At least they admit that technology will progress as much as they dislike it. You are like the living embodiment of the argument from ignorance. "Progress won't happen because I can't figure out how it could!" Maybe you can't figure it out because you are an idiot. Did that ever occur to you?
I'm a good cook. I'm a fantastic eater. - Steven Brust
He lies in wait near the villages;
from ambush he murders the innocent.
His eyes watch in secret for his victims;
like a lion in cover he lies in wait.
He lies in wait to catch the helpless;
he catches the helpless and drags them off in his net.
Prove anything by multiplying Huge Number times Tiny Number
Nature is always building better idiots. :)
I remember when my company got a big SAP system sold to us (An SAP salesman infiltrated our leadership) and we were told, "well, it's not supposed to meet your business needs, you are supposed to change your business practices to conform to SAP". And the executives like this, because it meant our system would be just like everyone else's at the cocktail party -- no more questions or embarrassment.
Self-driving cars could easily go the same way, with politicians forcing us to conform to the limited capabilities of the "AI", mainly so they can all feel like part of the in-crowd and enjoy a good cigar smoking.
Strange things are afoot at the Circle-K.
First will come autonomous cars for the public.
Then there will be no need for autonomous police cars as outlined in the patent request.
Some simple code that detects if a driver takes over manual control of an autonomous cars and breaks any driving laws, the autonomous car simple locks the doors and drives to the closest police station.
It all becomes self policing.
Autonomous fire engines and police cars will be needed, but will be used to quickly navigate the streets at safe high speeds.
Remember, once all cars are autonomous we don't need things like stop signs or traffic lights, the cars will communicate with each other and setup appointments to pass through intersections at speed making slight adjustments to miss the cross traffic in the intersection.
Eventually, no person will need to drive any vehicle.
This will also eliminate most personal owner ship of cars. People will use uber like services that use autonomous cars to get around when needed. Houses will no longer need garages, car dealers will go out of business, auto repair shops will convert to fleet maintenance facilities for the fleets of cars used by uber like services and government organizations. There will be no more truck driving jobs, ot taxi drivers. It will be more convenient for food delivery services, when you place your order on line an autonomous car will arrive at the restaurant and the food will be placed in the car for delivery to your door. No tip needed for a driver.
This will be one of the most disruptive technologies in a very long time once it becomes ubiquitous.
Are we suddenly going to run out of people that drive cars and trucks?
We're not going to run out, but I'm sure we can find something else for the US's 3.5 million truck drivers to do once we automate them away. Maybe we'll have them mine coal or manufacture buggy whips.
He's getting rather old, but he's a good mouse.
So it can autonomously park behind Lil' John's cocktail lounge. Big deal. Our cops have been doing that for years.
Have gnu, will travel.
Every time a science fiction writer publishes their book, they should also just fill out a couple dozen patent forms in the process. Of course the 'Epstein Drive' won't actually be invented before the sci-fi patent expires, but hey at least there will be prior art on the books... Or if you have the money I guess you could just patent anything you want right now, as long as you are sure to include the notion that some sort of included "AI" will just make it work.
...we're probably never going to have autonomous Formula-1 racing.
It seems to me that track racing would be a great application for autonomous driving. Response time, optimized acceleration/braking, few distractions, etc. What makes you think it would be so much tougher than navigating traffic?
He's getting rather old, but he's a good mouse.
While it's true that Moore's law is, essentially, dead this doesn't mean that computation can't get more powerful. There are lots of well known ways that would work. Most of them, however, require redesigning the algorithms. There's already been a lot of push into parallelizing things, but there could be a **LOT** more. There hasn't been because it was cheaper to rely on Moore's law.
FWIW, I think that the push into complex processor designs was a mistake. It lead to a local optimum that is quite far from the global optimum. What should be done is LOTS of simple processors, each with a SMALL cache of fast memory, and a much larger cache of persistent memory (so that it doesn't consume power and dissipate heat). The different CPUs should communicate via message passing and be programmed in a language that is adapted to this kind of computing. I'm thinking of something like the Erlang virtual machine implemented in hardware. If my guesses are correct, this design should be low enough in heat dissipation that 3D circuits are feasible without excessive work on head dissipation. It probably wouldn't even need water cooling.
This design is "sort of" like the ideas being floated for neural computers that keep showing up on the front pages, but I can't tell whether it's the same or not, because the descriptions are always so vague. They usually talk about "memristor" or some such, but that's just a particular technology that can be used to give non-volatile memory. IIUC any other non-volatile approach would work as well...though core memory would take up too much space, and that would slow things down.
I think we've pushed this "anyone can grow up to be president" thing too far.
No no, its impossible to have better idiots because of Moore's Law!
I'm a good cook. I'm a fantastic eater. - Steven Brust
While it's true that Moore's law is, essentially, dead this doesn't mean that computation can't get more powerful. There are lots of well known ways that would work. Most of them, however, require redesigning the algorithms. There's already been a lot of push into parallelizing things, but there could be a **LOT** more. There hasn't been because it was cheaper to rely on Moore's law.
FWIW, I think that the push into complex processor designs was a mistake. It lead to a local optimum that is quite far from the global optimum. What should be done is LOTS of simple processors, each with a SMALL cache of fast memory, and a much larger cache of persistent memory (so that it doesn't consume power and dissipate heat). The different CPUs should communicate via message passing and be programmed in a language that is adapted to this kind of computing. I'm thinking of something like the Erlang virtual machine implemented in hardware. If my guesses are correct, this design should be low enough in heat dissipation that 3D circuits are feasible without excessive work on head dissipation. It probably wouldn't even need water cooling.
This design is "sort of" like the ideas being floated for neural computers that keep showing up on the front pages, but I can't tell whether it's the same or not, because the descriptions are always so vague. They usually talk about "memristor" or some such, but that's just a particular technology that can be used to give non-volatile memory. IIUC any other non-volatile approach would work as well...though core memory would take up too much space, and that would slow things down.
The PS3 was built using this type of "cell processor" technology and they ran into performance problems. Granted, most were due to the underlying design, but some were due to the fact that scaling and the increase in delay as you get further away from the center is an issue. I'm not pretending to be an expert on CPU design, but the point is that there probably are fundamental issues with this type of processor that needs to be overcome.
http://www.redgamingtech.com/s...
My bet is that quantum processors or some combination of a traditional CPU with quantum components would be the next step.
Do we want Knight Automated Roving Robot. Because this is how we get Knight Automated Roving Robot.
The automation of highway robber-like behavior continues apace.
It's bad enough much human interaction amounts to throwing money at people until they go away, the bulk of "Hell is other people", but now we have to throw money at robots until they go away, laying the takings at the feet of their masters?
(-1: Post disagrees with my already-settled worldview) is not a valid mod option.
Perhaps there is an advantage over speed camera, but it is not obvious.
There is a huge drawback: a police car without policeman will be a nice target for vandalism
All it says is how many transistors can be crammed on a die. There is a lower lower limit to how big a transistor can be since you can't make it smaller than the electrons that flow through it.
No.
Moore's Law is about the *cost* per transistor and it does not matter how the cost per transistor decreases. The semiconductor dies may be larger, multiple semiconductor dies may be packaged together, or the transistors may be made smaller. The decrease in cost per transistor may even come at the expense of transistor performance which has happened several times.
https://www.youtube.com/watch?...
It's true that the communications will eat up speed, but the parallel execution will increase it. So you need to redesign the algorithms to optimize things differently. You can't use the same algorithms for this design and get good performance. But for most problems you can use different algorithms and do so.
Note that I talked about using something like the Erlang virtual machine as the machine language. You don't design things the same way for that kind of a machine. Hell, you couldn't even design things the same way for the CDC Star as for the CDC 7600. The architectures are too different. The proposed (cellular) design is more reminiscent of a GPU than of the current multiprocessor CPUs. It just isn't as specialized. Think of it as a hardware implementation of the Actor model, and you won't be too far off.
All that said, we appear to be almost *AT* a local optimum. Any change here is going to be locally down-hill. But I believe this is *far* from the global optimum.
I think we've pushed this "anyone can grow up to be president" thing too far.
Ron Goulart, "Into the Shop" I read it in 1970 (or 1964?) in the Magazine of Fantasy and Science Fiction, but since this topic keeps coming up, I bought myself a (used) copy of his book "What's become of Screwloose? and other Inquiries", where the story was reprinted. (The book has a stamp from the Seton Catholic High School. Gotta love those Jesuits, or whatever brand of Catholic bought this.) The story should be required reading for anybody who proposes this kind of nonsense.
Hardly,
Machine learning is just now starting to re-catch up to our current processing power. Metal neural networks still have a lot of catching up to do and we're making new progress with sensors every day.