I worry that driving will become cost prohibitive if driverless cars have a certain amount of adoption. Insurance companies will say "use driverless, or you pay X times more".
Well, the risk assessment/rate adjustment system is already in place. At least insurance companies will not do it for *political* reasons, they will define the insurance rate on the basis of existing risk assessments. When I started driving I paid 250% of the standard insurance rate. Then it was reduced every year, until I now pay 30%, after X years of not having any accidents. From what I heard, a British insurance company is pondering to put automated vehicles at 90%, if they ever hit the market.
I have worked in "process automation" for 30 years now. Including some basic driverless factory transport vehicles.
I'm pretty confident the *technical* problems of automating vehicles on 99% of highways can be solved in the next few decades. +
What I *don't* believe is that drive less vehicles will be able to navigate some of Europe's inner cities OR some of the more remote dirt tracks that we still have. (Both which in need to drive pretty regularly) My guess is that in my daily commute there are at least three spots where any "rational" AI would come to the conclusion "it's not save to proceed" and just stop there.
At least that should be made simple by software. The meeting time should be UTC in the database, and both the client making, and the client receiving the meeting data should handle the conversions.
On the other hand, the synchronisation between Outlook and Sharepoint doesn's even handle DST correctly. When we tried that, our team birthdays wound up being from 23:00 to 23:00 in one half of a year.
> The situation has gotten so dire that IT admins in many corporate environments are waiting for as long as six months before they are certain that it is fine to get the staff to move to the "newer" major software update.
I have been in Corporate IT for close to 20 years now. That we lag 2-3 years (sometimes even more) behind software version is pretty much standard.
Quite often we even move to major Version X only when major Version X+1 becomes available for the "public beta testers"
Then you say "username incorrect" or "password incorrect" as appropriate. You generally do the username lookup first anyway so the logic could be short circuited.
The thing is, the system has no way of knowing which case is appropriate. It could be the password is wrong, or it could be the password is right, but the user name is mistyped to another existing user.
Heh. Before I read your post I thought those pictures in the article where adds for something completely different, and was going to complain that there wasn't a single screen-shots of that alleged new winamp-like player anywhere in the article.;-)
I find the whole "X or Y" discussion a little funny.
If you have a system X that can prevent some cases of problems, and a system Y that can prevent some other cases of problems, then the logical solution would be to combine those systems, so that when of of the systems doesn't detect an oncoming accident the other does.
That that is so often overlooked, I suspect, is because people are not really interested in preventing problems, but more on putting the blame on one of the systems. Which gets a pretty hard to do, when two systems co-pilot the car.
No wonder the earth branch of Watto's Junk Shop Emporium went broke, seeing how little he seems to have sold. It was even a bigger loss than all the guys trying to stick him with worthless Republican Credits.
Since the battery was invented in 1745, and the ICE in the 1790s, I don't see much of an age difference, and the possibility of efficiency improvements for both,
I remember..... No Internet at all, browsing through library cards to find information... .. Having to decide which city to call long distance to connect to CompuServe... .. logging into the Library of congress to look up things via telnet... .. Listening to the modem connect sounds... .. being disconnected from all that, because grandma picked up the phone in the other room... .. dialling on a rotary phone... .. trying (and failing) to transit C64 datasette tapes via a phone line... .. Damn, I'm old.
A huge chunk of what self-driving vehicles have to deal with is the unpredictability and persistent rulebreaking of human drivers.
This is where it gets interesting. There I see a lot of parallels to my "normal" software projects. Usually, you have: The business rules, the software, the users. And to get something to work you usually have to tweak all three at some point.
Here you have the traffic rules, the AI and the humans. And at some point all of them will probably need to be tweaked. (Even if at some point all vehicles would be AI controlled, it might take longer to replace all pedestrians with AI controlled ones)
Another funny thing I just remembered: ~20 Years ago I worked in a factory with autonomous floor-borne vehicles. People behaved almost completely different towards them than they did towards human operated forklifts. But it worked. I wonder if this time around we can get a "unified" behaviour system working, or if it's also going to settle on a "Ah, this is a human car, I have to react such-and-such" vs. "Ah, this is an AI car, I have to react such-and-such" system, too.
Exactly. When I look at the picture in the article, I see quite a few things I as a human driver would have made "differently" from the shuttle AI:
- Not closely approaching a big rig that is backing up at all. - When I see a big rig backing up into the spot I'm at, I back up too, if possible.
There could of course be thousands of hours of QA and discussions if those decisions are technically "right" or "wrong", but the fact remains that they prevent a big rig crashing into my car.
... to the same shop with the same computer with the same problem three times in a row, then something is seriously wrong.
Or, as they say: The difference between something that CAN break, and something that CAN NOT POSSIBLY break is, that when something that can not possibly break breaks, it's impossible to get at and fix.
I have seen roughly the same. While we have a somewhat bigger company, the "IT Part" is about 15-20 people.
about 8 of them, including me, have been in the company about 15-20 years. Then we have a few people "rotating through", to see if we find some additional "good" people. But, never, ever was someone "fired" one-sided.
When, after 2-3 years and 4-6 "performance reviews" (which, in our case are both a feedback of the manager on how the employee works out, AND a feedback by the employee on how the manager works out) it was somewhat obvious that it wasn't working out in one of the directions, all the people who left left on amiable terms.
Of course, the problem is, the people who DO the firing have no clue how to disable access most of the time.
Only the person they are firing has that knowledge in most cases where management is stupid enough to "fire" them. It's like firing a pilot while the plane is still in the air, or a fire-fighter while your house is still burning.
There was a huge hype then, too, that "expert systems" would soon replace doctors and military strategist, etc...
The only difference being, that back then very few people or organizations could afford the machines that weren't really able to pull it off, but today everybody can buy a machine that can't really pull it off.
Slashdot Mirror
Plus, even when Unicode-Support is completely broken, "normal" umlauts schould still work, I hope
It's the Baden-Württemberg class
Have you tried talking to a scientist?
Yes, actually. ;-)
https://www.youtube.com/watch?...
I worry that driving will become cost prohibitive if driverless cars have a certain amount of adoption. Insurance companies will say "use driverless, or you pay X times more".
Well, the risk assessment/rate adjustment system is already in place. At least insurance companies will not do it for *political* reasons, they will define the insurance rate on the basis of existing risk assessments. When I started driving I paid 250% of the standard insurance rate. Then it was reduced every year, until I now pay 30%, after X years of not having any accidents. From what I heard, a British insurance company is pondering to put automated vehicles at 90%, if they ever hit the market.
I have worked in "process automation" for 30 years now. Including some basic driverless factory transport vehicles.
I'm pretty confident the *technical* problems of automating vehicles on 99% of highways can be solved in the next few decades. +
What I *don't* believe is that drive less vehicles will be able to navigate some of Europe's inner cities OR some of the more remote dirt tracks that we still have. (Both which in need to drive pretty regularly) My guess is that in my daily commute there are at least three spots where any "rational" AI would come to the conclusion "it's not save to proceed" and just stop there.
At least that should be made simple by software. The meeting time should be UTC in the database, and both the client making, and the client receiving the meeting data should handle the conversions.
On the other hand, the synchronisation between Outlook and Sharepoint doesn's even handle DST correctly. When we tried that, our team birthdays wound up being from 23:00 to 23:00 in one half of a year.
> The situation has gotten so dire that IT admins in many corporate environments are waiting for as long as six months before they are certain that it is fine to get the staff to move to the "newer" major software update.
I have been in Corporate IT for close to 20 years now. That we lag 2-3 years (sometimes even more) behind software version is pretty much standard.
Quite often we even move to major Version X only when major Version X+1 becomes available for the "public beta testers"
Then you say "username incorrect" or "password incorrect" as appropriate. You generally do the username lookup first anyway so the logic could be short circuited.
The thing is, the system has no way of knowing which case is appropriate. It could be the password is wrong, or it could be the password is right, but the user name is mistyped to another existing user.
I'm talking about behavior patterns of intelligent and successful individuals, and adopting a dead platform is not one of those patterns
Yeah. That's why George R.R. Martin will never have any success with that funny books he has written on an that old typewriter-platform.
Heh. Before I read your post I thought those pictures in the article where adds for something completely different, and was going to complain that there wasn't a single screen-shots of that alleged new winamp-like player anywhere in the article. ;-)
O, and the number one facilitator of illegal transactions in prison: cash. Should definitely be outlawed as well.
I was under the impression they were already working on that. (Like phasing out the 500€ Bill next year...)
So on one hand they want to put a minimum-size limit on phones, on the other hand they work on a maximum-size limit for cash.....
He. I usually say the old "I take problems and caffeine, and turn them into solutions and sarcasm"
I find the whole "X or Y" discussion a little funny.
If you have a system X that can prevent some cases of problems, and a system Y that can prevent some other cases of problems, then the logical solution would be to combine those systems, so that when of of the systems doesn't detect an oncoming accident the other does.
That that is so often overlooked, I suspect, is because people are not really interested in preventing problems, but more on putting the blame on one of the systems. Which gets a pretty hard to do, when two systems co-pilot the car.
No wonder the earth branch of Watto's Junk Shop Emporium went broke, seeing how little he seems to have sold. It was even a bigger loss than all the guys trying to stick him with worthless Republican Credits.
Battery technology is currently in its infancy,
Since the battery was invented in 1745, and the ICE in the 1790s, I don't see much of an age difference, and the possibility of efficiency improvements for both,
I only believe it has left "our backyard" when we get a ticket for flying a vehicle without a license in public interstellar space.
I remember ... .. No Internet at all, browsing through library cards to find information...
.. Having to decide which city to call long distance to connect to CompuServe...
.. logging into the Library of congress to look up things via telnet...
.. Listening to the modem connect sounds...
.. being disconnected from all that, because grandma picked up the phone in the other room...
.. dialling on a rotary phone...
.. trying (and failing) to transit C64 datasette tapes via a phone line...
.. Damn, I'm old.
Especially if they can't hear you, because their phones are buzzing to loud....
A huge chunk of what self-driving vehicles have to deal with is the unpredictability and persistent rulebreaking of human drivers.
This is where it gets interesting. There I see a lot of parallels to my "normal" software projects. Usually, you have: The business rules, the software, the users. And to get something to work you usually have to tweak all three at some point.
Here you have the traffic rules, the AI and the humans. And at some point all of them will probably need to be tweaked. (Even if at some point all vehicles would be AI controlled, it might take longer to replace all pedestrians with AI controlled ones)
Another funny thing I just remembered: ~20 Years ago I worked in a factory with autonomous floor-borne vehicles. People behaved almost completely different towards them than they did towards human operated forklifts. But it worked. I wonder if this time around we can get a "unified" behaviour system working, or if it's also going to settle on a "Ah, this is a human car, I have to react such-and-such" vs. "Ah, this is an AI car, I have to react such-and-such" system, too.
Exactly. When I look at the picture in the article, I see quite a few things I as a human driver would have made "differently" from the shuttle AI:
- Not closely approaching a big rig that is backing up at all.
- When I see a big rig backing up into the spot I'm at, I back up too, if possible.
There could of course be thousands of hours of QA and discussions if those decisions are technically "right" or "wrong", but the fact remains that they prevent a big rig crashing into my car.
Although I wonder if the fix would be just to be brave enough to just remove the calculator. Maths is an outdated technology anyway...
... to the same shop with the same computer with the same problem three times in a row, then something is seriously wrong.
Or, as they say: The difference between something that CAN break, and something that CAN NOT POSSIBLY break is, that when something that can not possibly break breaks, it's impossible to get at and fix.
I have seen roughly the same. While we have a somewhat bigger company, the "IT Part" is about 15-20 people.
about 8 of them, including me, have been in the company about 15-20 years. Then we have a few people "rotating through", to see if we find some additional "good" people. But, never, ever was someone "fired" one-sided.
When, after 2-3 years and 4-6 "performance reviews" (which, in our case are both a feedback of the manager on how the employee works out, AND a feedback by the employee on how the manager works out) it was somewhat obvious that it wasn't working out in one of the directions, all the people who left left on amiable terms.
Of course, the problem is, the people who DO the firing have no clue how to disable access most of the time.
Only the person they are firing has that knowledge in most cases where management is stupid enough to "fire" them. It's like firing a pilot while the plane is still in the air, or a fire-fighter while your house is still burning.
.... that "The year of the Linux Desktop" is already here, just as "The Year of the Porsche Car" is.
There was a huge hype then, too, that "expert systems" would soon replace doctors and military strategist, etc...
The only difference being, that back then very few people or organizations could afford the machines that weren't really able to pull it off, but today everybody can buy a machine that can't really pull it off.