Now we are talking about every square foot of space monitored at all times by cameras?
At least in terms of monitoring vehicles, yes.
I was referring to Bateaux's claim that there will be no need for police or judges because AI will be able to identify every person with 100% accuracy from the way they walk. That assumes every square foot of the world is monitored at all times by camera. That isn't even a tech issue, its a question of whether humans will ever feel that making sure every criminal is caught is worth zero privacy, ever.
And its a perfect example how something which may be technologically feasible isn't necessarily likely and certainly not inevitable.
The advantages of self driving cars are not that their reaction times are faster, it's that they can be networked together. They're not trying to be safer in the way that individual human drivers are, they're changing the way the roads are used.
As far as safety, all they need to do is obey speed limits and other existing laws, maintain proper following distances, and pay attention at all times, and the accident rate would already drop by around 99%. We are probably only a few years out from that, no networking or rebuilding of roadways required.
Long distance travelers can form peletons for reduced wind resistance at higher speed. GPS tracking of other AI cars facilitates adaptive routefinding to avoid traffic. In busy urban areas, network connections to traffic lights automate signal cycles. Like with everything else technology does, the real benefits are not in playing the game better, but in changing the game.
And all of things will probably happen, and probably marginally help, temporarily.
Ultimately they are likely to have about as much impact on revolutionizing transportation as power windows did.
Again, its not a tech issue. There is a theoretical maximum of vehicles a given roadway can support at a given speed. Its smaller than the number of people who can be housed in a given area. As the advances you note start to happen, commuting gets easier - so more people start driving personal autos instead of taking transit, people stop bothering to carpool, people buy houses further from work, and very soon there are so many cars on the road that even with traffic patterns optimized in every way, cars are still averaging 15mph.
I think many of the things that you think are too hard, or too crazy, or too illegal, will eventually seem as commonplaces as freeway interchanges do to us now.
I do think we will eventually have robots that can do general handyman work, run an ANW course, and maybe even do abstract intellectual thinking. I just don't think those are coming in the next 5-10 years.
Seriously, I don't think we disagree much, if at all, people just keep reading absolutes into my comments that aren't there.
Great comments on car automation. I like the suspension memory and the peleton ideas. The efficiency of cars will increase dramatically. Imagine intersections where cars never queue at a red light because they moderated their acceleration in advance to reach the intersection when the light is green. I try to do this manually for some lights that I can see from a large distance, but it's pretty tricky.
I've thought about how theoretically you could put downward facing cameras on the front bumper that would allow suspension to adjust in real time for every bump. Which would be cool, maybe add a tiny bit of control and efficiency, mostly just improve comfort, but, again, would have as much impact on transportation as power windows.
The peleton idea already exists, in the form of "train". Its also often used by long haul truckers on long straight highways. Not saying that capacity given to cars wouldn't also marginally improve efficiency, but if that was a serious goal, cars would be shaped much more aerodynamically, which we have had the knowledge and tech to do for about 100 years.
The last part, again, we could already do, but usually choose not to. Its called "timed lights". Cities don't often invest in it, cause having regular lights is cheaper and easier, but when they do, you just drive at exactly the speed limit, and you are guaranteed to hit every green light down the entire street.
Trying to do it yourself on a street without timed lights is called "hypermiling", its part science part art, and has a big following of people (including MMM who wrote a post about it once, and myself, who was briefly a blogger for ecomodder.com). Unfortunately, no matter how good you are, many times it simply isn't possible with non-timed lights, and that would be no less true for a AI driver. Is a city which isn't willing to spend the money on a timed light system any more likely to spring for an entire network interface with cars?