Google points finger at human after robo car accident One of Google's self-driving cars has driven into another car. But Google says it wasn't driving itself at the time. As first revealed on Jalopnik, a Google robo-Prius hit someone else's non-robo Prius earlier this week. But according to a Google spokesman, the accident occurred while a human was driving the Googly automobile …
The title, the tone, and the closing statement clearly show that the author is biased either against google, against new technologies, or both.
Of course the fall back system is flawed -- the fallback system is a human. Humans are responsible for hundreds of thousands of accidents a year. Clearly google is not okay with that, they are trying to validate technology to fix the problem. Way to be a cowardly critic sitting on the side lines offering no suggestions or alternatives, only a negative, ignorant attitude.
Learn to be open-minded and consider the costs of technical innovations over the centuries. Go watch Fight Club. Go get a new job -- it's journalists like you that make the news such a joke.
Re-read your post three times, and I still can't get past the feeling that it was written by the robo-driver. The indignation, the plea for understanding of "new technologies", the disdain for human drivers. Oh, and the "go watch Fight Club" (you stupid messy humans you)
Own up - driving all those miles gave you lots of time to "think about things", right?
"....Of course the fall back system is flawed -- the fallback system is a human....." Ah, but in that statement you have identified exactly the reason why we should treat such systems with suspicion. The robo system was designed by humans, and - as humans are fallible - ergo the systems is only as reliable as the person(s) that designed it.
What's worse is that the designer isn't in the car when it crashes so they're slightly less motivated to pay attention!
A worrying trend is that insurance companies are beginning to see automated car systems as a way of reducing accident claims. The old "computers don't make mistakes" attitude of the unthinking policy makers will make it very difficult for an individual driver to prove that the automatics were at fault. I'm not seeing any commitment to equip systems and accident investigators with the tools (eg independent black boxes that the police and owner can read, not just the manufacturer) they need to be able to diagnose a systems fault. Without such things the 'driver' is likely to get the blame every time. Not for me thank you!
Not ignorant journalism, acerbic. Sarcastic. Very jaded, and occasionally acrimonious. This is The Register. If you don't understand this, I'd suggest that you either have a very low reading comprehension or are exceptionally new to the site. Is this the first article you've read on El Reg? It's the first post on the NixKnacks account - I think most of us can draw some obvious conclusions from that alone.
-d
Humans have vision, senses and reactions far in excess of any machine.
Humans drive perfectly well when they're in a fit state to drive, when they're not being distracted by phones and gadgets.
Computers need to be programmed to handle all the unexpected things that can happen on the road. Do you really think a self drive system will cope with a brick in the road? or glass?
Will it avoid massive potholes? will it know about level crossings? will it know about speed bumps?
This post has been deleted by its author
"Humans have vision, senses and reactions far in excess of any machine."
Are you kidding?
Machine reactions are many times better than any human could ever hope to achieve. The best reliable reaction time a human is typically capable of is about a fifth of a second. A current Antilock brake system will react 40 times a second. Human senses can't see black ice, they cant see in infra-red or millimetre wave radar (which is useful for spotting things like bricks and potholes even in pitch black). A machine will be just as alert at 3 in the morning after 5 hours of motorway driving as it is doing a 5 minute run to the shops.
"Humans drive perfectly well when they're in a fit state to drive."
over a milion deaths and about 40 million injuries a year worldwide suggest otherwise.
The Google car has managed 160,000 accident free miles, I am pretty sure that it encountered a reasonable number of bricks, pieces of glass, potholes and other assorted unexpected hazards in that time.
The fact of the matter is that while there is a lot of work need to program and test a computer controlled car to make it very safe in all likely situations. Getting one to be safer than the average human driver with current technology is a pretty easy challenge.
I'll do a quick fun test with you if you like, I can dig out a 486, and we'll rig it up to a light sensor, then you can show everyone just how much faster you are at reacting to something than a machine.
You also seem to be making the assumption (which I think is pretty wrong) that humans will react correctly when they encounter something that they aren't used to. Humans need to be coded too, it's just that it's a bit easier (currently) to teach a human something, assuming of course (which I also think is quite wrong) that it goes in first time and that they apply it correctly when it does.
This post has been deleted by its author
"Honestly, I was a little bit sad when I saw this thread and the number of downvotes for the OP. What I saw is a community of IT professionals, instead of being at least a bit enthusiastic about a technology that's *actually* "magical and revolutionary", immediately attack the one person that argues with the tone of the article, ostensibly because he *dared* to defend a megacorp as a side effect."
If you really want an honest set of answers to that...
1) The OP showed EVERY sign of being a shill. The account was created THAT day, it made ONE post, and it was to defend a corporation in the middle of a bad PR problem, with no apparent understanding of the culture of the local site. Oh, and it caught this article fast enough to do all of the above faster than the regular readers could make a post. Obvious shills (even if by some chance they aren't, they just manage to act like it) get downvoted. It's an internet user reflex; to get upset about it is like spitting into the wind.
2) The Google auto car is NOT "magical and revolutionary". It is technological and evolutionary. Honestly, they are still pretty damn cool, IMO. However, the google auto-pilot tech had NOTHING to do with the downvoting of the OP. The OP was downvoted because he was an apparent shill.
3) http://www.thefreedictionary.com/shill
-d
This post has been deleted by its author
aha brilliant. How is pointing out that humans are flawed any sort of revelation? Surely this piece is pro-Google or certainly pro-technology. I thought so, but then again I'm only human.
Can I sign up to the google list for the gbody yet? I want my brain uploaded into the android gbody when my current one becomes useless. This is obviously what google are really working toward. :)
Well, as we say in the aviation industry, the most dangerous component on an airplane is the nut in the cockpit.
Incidentally, this is not new technology, MIT were doing this 20 years ago using neural nets programmed on SPARC workstations to drive around the campus in MIT. It will be interesting to see what patents the googlebots try and take out on this one!
Paris, she knows all about nuts in the cockpit.
Woah - my Yank-dar just went crazy.
Here's a tip: .co.uk means this is a UK site, and that means Satire, Irony and other forms of humour will be used.
That article - for el'Reg - was actually EXTREMELY factual, with hardly any opinions at all.
The comment at the end was IRONY - go look it up, numbskull.
And humans still don't know why they need to be destroyed. Honestly, my cyborgs wouldn't have left a scratch on the other vehicle - because they would have rammed it off the road and blown it into pieces! Ha-ha-ha! What do you mean I sound like Metal Mickey when I laugh?
Start aiming for the fsckers when you see them riding around - entertaining? Perhaps. Just wrong? Probably. Stick a real and pointy finger (or fender) in Google's eye? Fscking hilarious!
Yes sir, officer. Just a test of electronic reaction time and response of backup system - it failed. Of course Google will pay damages, that's what I'm here for. <slowly edge away from borrowed Prius and haul ass>
If only this wouldn't be so true. When (not if) automated cars turn up, there will be some out there who, just for the sheer hell of it, will to cause them to have accidents, brake suddenly and do anything else unwanted. Doubtless some will have hidden agenda (the tech nay-sayers) but others just for the thrills, malicious or not.
Just don't do it in your own car. These things are armed with video cameras. You would be charged for failure to signal at minimum or reckless driving/endangerment on the upper end of the scale...which would be enough to pin the damage on you.
I could imagine three lanes of commuting robocars being chased out of the way by an old rogue 1960s to early 1970s American land shark. :) If the lasers and cameras do their job, it would be a quick way to get through rush hour traffic since they would all swerve to avoid the collision.
Where can I find one of these wonderful self-driving cars ? I WANT ONE !
Imagine a driver backup system that can maintain speed and distance in traffic, has a faster reaction time than me, would prevent me from merging when there is a car in my blind spot, knew every speed limit on every road, and maybe, just maybe, a nightvision system for avoiding pedestrians, bicycle riders, and deer at night. (Yes, sometimes people walk in the road, or ride their unlit bicycles at night - go figure).
Google may be evil (think spawn of Satan and Damien rolled up into one), but their car is cool.
Here's a scenario ...
Googly vehicle develops a software glitch (these things happen, believe it or not) and starts to drive straight down the tailpipe of the car ahead.
Human sees the problem, disconnects auto-driver and slams on the brakes - too late to stop the impact, but enough to turn a full-scale collision into a minor fender-bender.
Google PR does much the same sort of damage reduction by pointing out that a human was driving 'at the time of the collision'.
So a highly relevant question should have been 'For how long before the collision was the human in control?'
We had that in Germany, but unfortunately our companies got overtaken by beancounters just before it got into production. (Beancounters dislike innovation because products which haven't been available before, haven't sold before, and therefore won't sell in the future)
http://www.youtube.com/watch?v=ID0JlT_3HY8
I get your concern, and I am just as worried about our ability to focus on one thing, and perform "deep" research - the kind that takes focusing on a problem 16 hours a day for weeks on end until the answer "reveals" itself. But would us loosing driving skills be any real loss? As people have pointed out we really were never "built" for this kind of work to begin with, and don't we have better things to do with our time?
How many weeks of our lives are spent learning this skill presently? I'd much prefer that millions of people learn medicine, philosophy, or even just how to program entertaining games or write stories than such a waste of a humans time.
Not many of us remember how to make fire anymore, and I'd say for the most part thats a good thing
I wouldn't expect we would loose these skills until the cars are good enough for us to not ever need them. At that point human intervention would consist simply of pushing the red "power off" button, where the car would mechanically be disabled and brakes gently applied to a stop.
[This mechanical "power off" button should be on all AI devices just to keep the digital overlords from being able to run us all off cliffs and into poles at the singularity point*]
*Yes I know that's not the common definition of the Singularity - but really you think Digital sentience has any need for you meatbags?
Was the Prius body completely designed by a robot? because it looks totally bland and uninspiring, if I ever manage to get enough money for a first car you can bet your bottom dollar it won't be a Prius.
With the way some humans 'drive' their vehicles on the road I'd feel safer to have computers do the driving instead, because for starters the code controlling the car would have many many many more hours learned experience driving on roads than humans do after passing their driving test, and at least they'd be programmed to flat out refuse to drive on public highways without MOT, tax & insurance
(gets back on electric bicycle)
Human nature being what it is, we'd p ourselves at the thought of being a passenger in a motor vehicle which was 100% under control of a computer program, but don't realise (or give a damn) if we haul ass between two airports then the airplane we are sat in is 100% controlled by an auto-pilot, which is entirely possible.
We'd instead want a respectable driver like Jeremy Clarkson up-front, presumably on the basis that he's sharing the same thoughts about not wanting to be extinguished as you are.
You'd also wet your knickers if we were told that the surgical operation you are going to have was going to be using a scalpel which was under computer control. And yet that surgery couldn't be done otherwise because it requires a level of control which wouldn't be possible with Jeremy Clarkson in control.
Hmm, under the scalpel with a choice of either Jeremy Clarkson or a computer. That'd be a toughie.
Consider a future where these cars are proved to have a much lower accident rate than people, what then? Does society allow them? Does society REQUIRE them??
And what about liability? Accidents will still occur. With human drivers the blame is clearly with the human (usually), but with a robo-crash the blame is less clear. Will survivors of the deceased get to sue the car maker every time it happens, despite the proven superior safety factor of such cars? " Your honor, my clients placed their safety and very lives in the hands of this companys' product, and that trust was unforgivably violated."
I don't see robo cars EVER being accepted, simply because it would expose the makers to enevitable ruin, unless they became universal and required, with shield laws for the car makers in place. That would require everyone to give up their personally controlled cars. Ain't gonna happen.
Airplanes are already flying themselves these days: landing, take-off and in between.
Someone else in this thread already mentioned, the pilot is already the worst error point (but still necessary IN CASE).
For devices in an early stage of release their success rate is pretty phenomenal (and this is coming from a pretty staunch google hater.)
I would wager the opposite, give it 5-10 years and most cars will be auto-driven.
And yes at least in America you will still have the option to drive yourself. Of course assuming you have the Money and ellite status to risk the more realistic and inevitable lawsuits for getting in an accident while operating your own vehicle "HE was actually driving himself officer!" no more driving for the poor, but you bet the richies wont give up their penis extensions
Oh, and you did catch in the article where these are ALREADY legal in California and Nevada right?
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
