Carlson out at Fox

[hammerofglass]

I don't think I could ever feel safe in a robot car. I've had two different GPS units tell me to turn off a bridge and take the river. This was back in the early models of course, but left an impression.

[McAvoy]

That one-off flatlander joke got waaay out of hand.

What's a flatlander joke?

Local slang, wasn't thinking about where I was.

A flatlander in north-central Pa is anyone who doesn't live in the mountains, from southern Pa ot Philly, or most often New Jersey. Bad driving is part of the stereotype which is why I made a crack about NJ drivers. Although what I said is completely true of my own experience.

Ridgerunner used to be the word for locals but that fell out of use before my time.

From my experience PA drivers are the worse. You have yet to live when you have to dodge incoming cars driving on the wrong side of the road... Happening too often on 95. Or see drivers make a hard right across multiple lanes to make their exit at the last second. Or doing 100 MPH on 295 while weaving in and out of traffic.

Or my favorite. Doing 55 in a 65 in the far left lane (fast lane) and literally holding up dozens of cars behind him.

Robot cars cannot come fast enough.

As an aside, I often wonder: what will humans in a hundred years time think of today? And imo, the idea of letting humans lose on vehicles capable of 150 MPH powered by explosive and polluting liquid without any computer oversight will be one of those things that children will look back on and laugh at us for. It'll be so alien to them. So foreign. Its like us trying to imagine a hundred thousand coal fires burning over streets full of horse and cart traffic.

So over the weekend, tanker truck full of fuel caught fire under a over pass around Philadelphia on I-95 which caused one side to collapse and the other side structurally unsafe.

So the main highway to Philadelphia is literally cut in half now. There is literally talk on the radio even by the conservative talking heads for workers that can work from home, should work from home now. Used to be different.

[Madner Kami]

I don't think I could ever feel safe in a robot car. I've had two different GPS units tell me to turn off a bridge and take the river. This was back in the early models of course, but left an impression.

One of our local bridges had to built anew, due to flooding damage. The old bridge's ramp was reconstructed into a bicycle lane, leading onto the new bridge in two very tight S-turns, barely 2m wide. For years, we had retards trying to get up that way and logically getting stuck. Some on the ramp, some managed to somehow get their cars through the double-S turns and only got stuck on the much wider bicycle-lane on the bridge. All of them failed to update their maps.

As for trusting robot drivers, you can't. They've become quite good, but are bad at dealing with unpredictability.

I myself once fell asleep in my Civic on the Autobahn. I was passed out for approximately 15 minutes. My beloved Aoba kept me in the lane, adjusted the speed based on the signs (I know, because there's a limited section with a stationary flasher on that route and never recieved and mail) and finally slotted me in behind a truck, driving a cozy 80km/h. Despite that, I'd never do that intentionally, because I know what it can and can not do and there are some incredibly mundane and frequent things it always gets wrong.

[Riedquat]

I don't think I could ever feel safe in a robot car. I've had two different GPS units tell me to turn off a bridge and take the river. This was back in the early models of course, but left an impression.

My take on them is that they're nowhere near as ready as some people seem to think they are. I am rather sceptical about the ability of one to work out how to get to my house (down a rough track, with a good chance of it falling in to the canal), and there are a few roads around with gates on them.

The sorts of errors they'll make will be different from the ones humans do. It is possible though that that'll overall be fewer errors. However (and this is just my suspicion, not actually backed up by data so take with as large a pinch of salt as you want) is that they may well be safer than the small minority of drivers who cause a large majority of serious accidents, the drunks, the lunatics showing off and so on. This may indeed lead to an overall decrease in such accidents, by decreasing the risk from that group a lot but increasing it slightly for everyone else.

[clearspira]

I don't think I could ever feel safe in a robot car. I've had two different GPS units tell me to turn off a bridge and take the river. This was back in the early models of course, but left an impression.

My take on them is that they're nowhere near as ready as some people seem to think they are. I am rather sceptical about the ability of one to work out how to get to my house (down a rough track, with a good chance of it falling in to the canal), and there are a few roads around with gates on them.

The sorts of errors they'll make will be different from the ones humans do. It is possible though that that'll overall be fewer errors. However (and this is just my suspicion, not actually backed up by data so take with as large a pinch of salt as you want) is that they may well be safer than the small minority of drivers who cause a large majority of serious accidents, the drunks, the lunatics showing off and so on. This may indeed lead to an overall decrease in such accidents, by decreasing the risk from that group a lot but increasing it slightly for everyone else.

There is nothing a human can do that an AI cannot. Its just a matter of time. And when that time comes you will have a driver that doesn't drink, doesn't take drugs, doesn't speed, doesn't sleep, doesn't undertake, has perfect reactions, can drive in any conditions, and -probably- won't get lost.

[Riedquat]

There is nothing a human can do that an AI cannot. Its just a matter of time. And when that time comes you will have a driver that doesn't drink, doesn't take drugs, doesn't speed, doesn't sleep, doesn't undertake, has perfect reactions, can drive in any conditions, and -probably- won't get lost.

Very, very wrong. We're nowhere near the point that AIs can do everything humans can do. Oh, theoretically it's probably possible, but actually doing so remains firmly in the realms of science fiction. It probably won't happen in my lifetime. I certainly hope it doesn't, a world where we mistrust our fellow human beings, are so scared of them and ourselves that we don't do anything except sit around as useless lumps of infantile flesh, Wall-E style, is a dystopian nightmare.

[hammerofglass]

There is nothing a human can do that an AI cannot. Its just a matter of time. And when that time comes you will have a driver that doesn't drink, doesn't take drugs, doesn't speed, doesn't sleep, doesn't undertake, has perfect reactions, can drive in any conditions, and -probably- won't get lost.

Very, very wrong. We're nowhere near the point that AIs can do everything humans can do. Oh, theoretically it's probably possible, but actually doing so remains firmly in the realms of science fiction. It probably won't happen in my lifetime. I certainly hope it doesn't, a world where we mistrust our fellow human beings, are so scared of them and ourselves that we don't do anything except sit around as useless lumps of infantile flesh, Wall-E style, is a dystopian nightmare.

You probably could automate at least the highways entirely with current tech if the cars in each segment were all directed by one mainframe. Although it would cost so much to implement in both money and political capital that it would never happen.

[Madner Kami]

Very, very wrong. We're nowhere near the point that AIs can do everything humans can do.

Except for a few things, expert systems flat out are better than the common human in a given field. Our advantage lies in our ability to do everything good enough at the same time and adapt on the fly.

[Riedquat]

You probably could automate at least the highways entirely with current tech if the cars in each segment were all directed by one mainframe. Although it would cost so much to implement in both money and political capital that it would never happen.

Sure, there are some things that could be (not that I want them to be, automation of things human beings are quite capable of is high up my list of things I loathe), but the point is that, overall, AI is still nowhere near humanity in terms of what it can do.

I think the problem really lies with the term. We're used to the idea of artificial intelligence as seen in science fiction, so a lot of people equate what we can do now with that when they hear the term AI, but they're two very different things. It would probably be better to have a different term for current capabilities and technologies (Mass Effect's VI maybe?)

[Thebestoftherest]

Hey should we have a thread for AI alone?