This (from March 09, 2006) was a precursor to this...
--------------------------------------------
Every now and then I see articles about the "future of the car", like this one. I am trying to figure out what is so smart about the mindset of the people who think these things up. Leaving comfort, speed, entertainment, information, fuel efficiency and environmental impact aside and just focusing on safety, what are these "futurists" thinking?
They have two general ways of thinking about this. The first is: make the car bigger, heavier and built of stronger metarials so that it can withstand crashes and keep you alive. The obvious result is an evolutionary arms-race between Hummer, next years "Uber-Hummer", the 2008 "Super-Hummer", 2009 "Ultra-Hummer" and 2010 "Hyper-Hummer". What's next - a Sherman tank? More and more violent vehicles, lulling their owners into a false sense of safety, promoting their beligerent aggression, and killing more and more people unlucky enough to be able to afford (or do not want in principle to own) the latest behemoth.
The second strategy is to employ electronics to replace driving skills. You are reading this, thus you have some experience with computers - do you really want them to be in charge of life-and-death decisions? Is it possible for a computer to be able to correctly respond to every possible situation - out of an infinite number of possibilities - that can arise in traffic? I didn't think so....
When I think of an ideally safe car, I think of driving something embedded inside bubble-wrap packaging, with styrofoam peanuts. In case of a crash I want an air-bag to deploy not in my face, but between the two vehicles (or the car and the tree), amortizing the impact and gently bouncing back.
Imagine a highway on which all the cars have, on the outside, the consistency of marshmallows. The worst thing that can happen if you bump into somebody is that you get a finger raised at you. Doesn't that sound saner?
- Log in to post comments
More like this
Actually I trust computers more than humans. After all, one stunning engineering success happens to be the Mars rovers.
They were only supposed to be good for 90 sols, now they've been running for over 1,000 sols.
But I understand, humans program the computers so something is bound to happen. Even the Shuttle has a colloquium of machines that vote on everything. And they make mistakes too.
But the one thing computers can do better than humans is manage and analyze multiple streams of information. That's the key difference.
We have circuits operating vehicles right now. Modern accelerators seem to be simply rheostats, so that if a stray voltage shows up -- say an unexpected static electrical discharge -- the throttle may go full on, even though the gas pedal hasn't been touched and the driver is standing on the brake with both feet.
This gets blamed on driver error -- "He musta got confused an pushed the wrong pedal."
My neighbor had put her SUV in reverse to back out of her parking spot and was moving her foot from the floor toward the brake pedal, her normal practice, when the engine suddenly revved up and smashed the vehicle into a building. She was okay, but scared badly. The SUV was totalled. And the news didn't carry the story of a runaway vehicle hijacked by an errant circuit.
It is perfectly possible to create hardware and software that is far more reliable than humans. It just isn't practice at Microsoft or any of its competitors, so you don't see this on your "computer"--that is, the one computer out of the thousands you interact with continually that is clunky and unreliable enough for you to remember that it's a computer.