Chris X Edwards


Review: Misbehaving

2016-07-15 16:57

Review of the book "Misbehaving - The Making of Behavioral Economics" by Richard H. Thaler.

I have complex feelings about the field of economics. Note that I didn’t say the "science of" economics. I pretty clearly don’t believe that word is appropriate, "dismal" or not. Formal study of economics was a requirement for my university degree and since those innocent beginnings I’ve had serious philosophical doubts about the field. I remember thinking at the time that I should be able to convert the knowledge I was receiving in an economics class directly into cash. Either that or what exactly is the point? I have taken engineering classes which I felt I could indeed convert directly into a conceptual design for a reasonably safe bridge. I don’t feel this way about economics at all. At its best I feel there may be elements of truth in the field’s orthodoxy but to me it mostly seems like astrology which also contains elements of truth and also has not made me a penny richer.

I remember sitting in those economics classes looking at supply/demand curves etc. and getting an unsettling sense that they were tacitly doing something the culture of physics is quite honest about; they were talking about spherical cows in a vacuum. I think that physicists don’t mind this because such simplifications often lead to a deeper understanding of the situation. Physicists gloat over the fact that real cows falling through the earth’s dense atmosphere can actually be approximated by an amalgamation of spherical cows in a vacuum. This sort of thing is even true for computer science too.

Unfortunately for economists who styled their art after such thinking, this is not how economics works. Useful economics must primarily ask questions like, "Why the hell would someone throw a cow from an airplane?" Only in the last few decades has economics started to turn to such questions. It’s been a radical transformation. So radical that economics has bifurcated into "economics" and "behavioral economics". I believe it’s only a matter of time before "behavioral economics" changes its name to "economics" and the old economics is forgotten like the old geology that couldn’t acknowledge plate tectonics for half a century despite overwhelming evidence.

This book is basically the inside story of this transformation from the perspective of one of its leading revolutionaries. Thaler does seem about as well qualified as anyone to tell the tale. First, not only can he appreciate that real human beings are generally the participants in economic activity, he also knows that real humans do not like reading the kind of prose normally written by economics professors.

Richard Thaler does also seem to be one of the most important economists to the field of behavioral economics. If any one person created the field it would be him. (Of course I just read a history of the subject written by him so maybe I’m biased.) The only people perhaps more important to the field are Tversky and Khaneman who, despite winning a Nobel prize in economics, are ostensibly and revealingly not economists.

The book is a superb introduction to behavioral economics. Beyond the interesting history of this new field, it contains many fascinating insights into how markets behave. Or don’t. It contrasts the new thinking with old ideas like the "efficient market hypothesis" of traditional economics. Here is Thaler and EMH’s creator, Eugene Fama, having an excellent discussion about it. The EMH basically says if you see $10 lying on the ground it must be an optical illusion because if it were real, someone would have already picked it up. I am sympathetic to the idea that beating the market (i.e. reliably finding unclaimed money) is extremely hard and perhaps impossible. But I believe that if it’s possible or impossible, the reasons mostly involve the problems of irrational thinking. It’s difficult for me to conceive of how a steaming load of irrational thinking can create efficient markets (as Fama allows may be possible).

I think what I like about this book is that it really nicely chronicles the field of economics going through a serious philosophical rethink. I am more interested in the philosophy of the topic and I see many parallels with my main philosophical interest, probability. Not only are humans easily tricked into believing in obvious incorrect future outcomes, but how we even think about the concept of predicting future outcomes is extremely tenuous. This business about economics (behavioral or Newtonian) lacking predictive power is obviously important. It may be that economics can be defined as: the philosophy of looking at the world in a way that if a reliable rule to predict the future is discovered, the rules instantly change. I have to say I lean that way, but, obviously not entirely. If I did, it would be utterly pointless to even think about economics.

Microsoft Linux 2

2016-07-07 09:05

A while back, I pointed out that Microsoft had quietly created their own Linux distribution (as an offered VM image in Azure). Although I’m a little late noticing this, it seems Microsoft has doubled down and truly created their own distribution which El Reg rightly calls Microsoft Sonic Debian.

This Linux system is basically a special setup to control switches and network technicalities in their Azure cloud service. It is apparently not available to the public but it is used by Microsoft internally.

I’ve always pointed out the irony of my employer not using the popular and capable operating system it developed. I think Microsoft choosing to not use a Microsoft operating system for its needs goes beyond mere irony. It’s more like the pope using a rhetorical point of Richard Dawkins.

If Microsoft doesn’t automatically think that Windows is the right OS, you probably shouldn’t either.

To Whom It May Concern

2016-07-07 10:26

I’ve previously mentioned that I’m somewhat of a grammar enthusiast. I’m more of a language and communication enthusiast really but this does come with a bit of grammar pedantry, usually just enough to embarrass myself when I reread old things I’ve written.

For future reference, I wanted to clarify my thoughts about a particular relic of English grammar, the crusty old dative pronoun whom. I quite like this word and occasionally enjoy having some fun with it. I personally can not help but mentally copy edit text I hear (Bo Didley - "Whom Do You Love?", ACDC - "Who Made Whom?", Ghostbusters - "Whom you gonna call?", etc…)

But here’s the thing, as those examples illustrate, real English speakers are usually perfectly ok with not using whom "correctly". My personal rules are simple.

  1. If you definitely know what you’re doing and you want to use whom and it is 100% correct, use it.

  2. If you’re writing for a prestigious widely read publication that has a well established history of punctilious grammatical perfection, then learn the rules well enough to apply #1.

  3. If you’re using a set piece that most everyone knows, don’t go mucking that up by changing it. In other words, don’t use something like "to who it may concern" or "for who the bell tolls".

  4. If none of the previous rules apply, don’t worry about whom! Use who always. No one (you care about) will mind.

If you follow these rules you’ll not run afoul of what I do consider a serious English faux pas: using whom when it is not correct. It is always better to err on the side of not using it. To add a superfluous whom says, loudly, that you tried to be a fancy English communicating sophisticate and failed.

Here’s an an example of the usually immaculate Economist violating this dictum.

"However, Brexiteers have jumped on anyone whom they think is exaggerating the economic impact of the decision to Leave."

This type of construction is very tricky and catches otherwise careful writers off guard. The mistake is easy to see if you remove the (almost parenthetical) "they think" clause.

"Brexiteers have jumped on anyone whom is exaggerating the impact."

The sentence could have clumsily been stuffed with a correct whom if it was rearranged into something like this: "However, Brexiteers have jumped on anyone whom they think of as exaggerating the economic impact of the decision to Leave." In that arrangement, it’s the object of a preposition, one of the surest signs that whom is correct (as in the examples of rule 3).

As I said, just be cool about it and if you take the trouble to learn and internalize the whom rules, consider personal grammar editing a fun little hobby that you can (usually!) keep to yourself. Otherwise treat it like quantum physics, i.e. assume that some people must know how it all works but you don’t have to.


Procrastinating By Diverting Trolleys

2016-06-29 10:28

This MIT article tells me something I already knew as a cyclist.

"Most people want to live in in a world where cars will minimize casualties… But everybody wants their own car to protect them at all costs."

I have noticed that!

But let’s not dwell on that issue. Rather let us use it as an example of what’s going on with the current state of the art of autonomous cars. All this tepid trolley problem philosophizing is getting tedious. You know what else has philosophical implications involving the trolley problem? Trolleys. And if a mine needs some mine carts I’m guessing the miners spend very little time worrying about the exotic situations where this tool could get into ethical quandaries.

I’m starting to think that all this attention on the moral dimensions of extremely unlikely situations is a type of mental fatigue that prevents people from doing the truly difficult thing, actually making this robotic technology feasible. The really interesting lever you should wonder about pulling is the one that makes lidar drop in price. Etc.

I used to read Hacker News. A lot. I think the best way to describe why I suddenly stopped would be to say that the reasons were similar to why I’m going to stop reading "news" about autonomous car trolley problems.

3D "Printing"

2016-06-08 09:13

I was surprised to see this article declaring the reasons "why home 3D printing never lived up to the hype". I was surprised because this is a very unpopular opinon. I know this because it’s been mine for the last 10 years. About 10 years ago I watched with interest as the RepRap Project filled space in the nerd news. At first glance I knew that this project was doomed. (Your RepRap hasn’t had any offspring, right?) I knew this for several reasons.

A long time ago, in a galaxy far, far away I was the entire manuacturing engineering department for a large (150 people, 150k sqft) machine shop. I did that for seven years. Although I have a relevant university engineering degree, that time in the shop learning from the machinists really opened my mind to how real manufacturing was done. And it’s not with 3d printers. If 3d printers were going to supplant traditional manufacturing it would have happened no later than the mid 1950s when the chemistry of thermoplastic resins (a.k.a. hot glue guns) was well enough understood.

The reason for this is that 3d printing enthusiasts today completely miss the point. There was a revolution in manufacturing. It was very, very important and profound. But it was not making flimsy plastic parts, slowly, using hot glue guns. The technology to focus on if you want to follow the trajectory of real progress in manufacturing is computerized motion control. The avalanche of hype for 3d printing is quite properly deserved by the concept of computerized motion control. What would make sense is if all instances of the words "3d printing" were replaced with "affordable computerized motion control". Then the hype would make perfect sense.

Over twenty years ago, I made quite a hobby of designing and building prototypes for a computer controlled positioning system. What would such a system position? Anything, including hot glue guns if that’s what you wanted. But just as interesting would be a laser or water jet cutter or a high speed spindle (from Dremel to wood router) or a welder. In fact if the focus were properly on computerized motion control you could be "3d printing" epoxy or concrete or chocolate mousse.

The problem seems to be that we’re now in a generation that is rapidly losing its hands-on practical people. Enthusiasts of 3d printing seem more influenced by the replicators on Star Trek than by actually tinkering in a garage their entire lives (and a manufacturing culture and tradition unbroken since the dawn of the industrial revolution). For example, the label "3d printing" to describe additive manufacturing is a glaring misnomer. The word "print" comes from the Latin "premere", to press, and if anything should be called 3d printing it would be sheet metal presswork. Of course there already is a well-established and adequate vocabulary for that art (though it’s being lost, e.g. my spell checker didn’t know that "presswork" was a word).

Another reason I was unmoved by the excitement for 3d printing was that I could easily refute one of its most important claims. The latter day enthusiasts of additive manufacturing like to remind us that with this kind of manufacturing, products are possible that can not be produced any other way. I understand that excitement because I felt it too - in 1992. It was then I first saw a demonstration of SLA. That is a strange acronym, which I’ve never understood, for stereo lithography. What the young people today would call, "3d printing". At the supercomputer center where I work, there are several display cases filled with ancient SLA relics from this time. They "printed" molecules and topo maps and human brain models and mathematical esoterica and it looks like loads of fun. But eventually they ran out of cool fun pet projects and we are left with some shelves of dusty unusual knick-knacks. I personally was absolutely thrilled at the prospect of designing some incredibly cool things that could only be created with an additive process (e.g. a sphere inside a sphere). But decades later I have not thought of a single thing that would be even mildly interesting that I could not also manufacture with an assembly or subtractive techniques. Easier and better.

It’s strange to me that the internet’s hype has turned its gaze to these hot glue dispensing machines. Just looking around, if there’s one thing the world didn’t need an improved production technique for it would seem to be cheap Happy Meal plastic trinkets. Why not focus attention on benchtop milling machines? Seriously, these are very cool. With a small desktop milling machine you do actually have a fighting chance of being able to build a copy of the machine itself. You can even clamp a hot glue gun to the spindle if you’re into that sort of thing.

Just remember the important new thing is really affordable computerized motion control. If we start concentrating on that, I’m certain good things will happen.


For older posts and RSS feed see the blog archives.
Chris X Edwards © 1999-2016