by Tyler Cowen
on March 14, 2012 at 2:04 pm
1. Driverless cars will minimize the need for traffic lights.
2. Can this method help you avoid getting stuck on problems?
3. Very good piece by Josh Barro.
4. Very good comment thread on neuroimaging.
5. There is no great stagnation (in bookmarks), short video.
Re: #2. Nice to see that psychology has finally reinvented the mathematical and computer science concept of decomposition. This method has existed for what, three hundred years now?
A word to the psychologists: decomposition, which is a technique every programmer around the world uses to aid in abstracting complex systems (computer programs), only works when the sub-parts are largely independent. Which, for any complex problem worth talking about, is “never”.
I’m all for mocking psychologists and lauding math, but really there is a tad more to this technique as described. The key idea is not the decomposition, but of challenging your descriptor for components. The claim is that switching descriptions and mental models to something use-agnostic unlocks additional potential uses for the components.
I wonder where the traffic engineers think all the pedestrians will go in the traffic light free, stop sign free future.
From the original article: “we won’t need traffic lights at all (or stop signs, for that matter). Traffic will constantly flow, and at a rate that would probably unnerve the average human driver. The researchers have modeled just how this would work, as you can see in the animation below.”
People should be able to cross at will, and the cars will stop for them. (In fact, very old science fiction stories predicted that people would just randomly wander out in traffic wherever they want, since they know the cars will stop.) Note that this is far off in the future, where the assumption is that all cars are driven by computers instead of ugly bags of mostly water.
I wonder if we will move people away from crossing at the intersection to crosswalks set up in the middle of a block.
Cars cannot violate the laws of physics. If you try to cross while a car moving at 100 km/h is 10 feet from your position, good luck with avoiding being hit. No amount of AI can prevent that. Crossing at any location you like and lack of regulations regarding it are a pure fantasy.
No one can protect you against jumping in front of a car going 100 mph, but the car certainly can determine your speed and trajectory and modify its route to account for possible changes in your direction/velocity.
Right, but that actually requires enormous amounts of computing power (probably 10^3 or 10^4 times more than what is required for mere AI-driving systems) and very, very good algorithms. We won’t have that any time soon. Besides, even the best system will make mistakes occasionally. So, again, the whole “cars will drive themselves at 100 km\h and that will be safe” is simple bullshit.
In a busy city centre there are so many pedestrians that allowed to cross at will, they will simply block all car traffic. Actually I would like that, a city for people, not for vehicles.
1. I ran across that video awhile back, pure awesome. And it’s nice to see Andrew branching out from forensic gynecology into traffic engineering.
Did anyone notice the simulation uses an intersection of two twelve lane roads.
How many of these are around? Anyone know?
Re #1: Driverless cars aside, a lot of traffic lights are counter-productive and could be removed already.
1. Humans can do that if the crossing is designed properly. It’s called a roundabout.
Driverless cars could be one way out of the Great Stagnation. Alexander Field in “A Great Leap Foward” talks about how the expansion of roads and commercial trucking helped create a massive boost in total factor productivity in the 20’s and 30’s.
Driverless autos won’t be as revolutionary as that development, but it could allow for much more efficient and flexible use of trucking by American retailers, distributors, and manufacturers. For example, cars driven by computers will decrease transport times of goods to the extent that they eliminate or reduce traffic jams.
#3. Let’s assume most of the current tax rates remain unchanged. How many decades or how many presidents will it take for economists and pundits to refer to current tax rates as something other than the “Bush Tax Cuts”? Will our great-great-grandchildren still be complaining about the “Bush Tax Cuts”? And will they continue to assert that when tax rates are increased this will have no effect on the economy but will simply increase tax revenue in direct proportion to the rate increase?
#1 is actually a video I made while flying over the intersection using my personal jet-pack.
Why cannot we start with cars that drive themselves only on the interstates? It seems an easier problem.
Comments on this entry are closed.
Previous post: Model this (the Goldman guy)
Next post: The bottom line
Email Tyler Cowen
Follow Tyler on Twitter
Email Alex Tabarrok
Follow Alex on Twitter
Subscribe in a reader
Follow Us on Twitter
Marginal Revolution on Twitter Counter.com
Get smart with the Thesis WordPress Theme from DIYthemes.