Author Archives: Patri Friedman

While a first-order model might see the human brain as a general-purpose computer, coordinated, centralized, acting with a single voice, we now know that this isn’t so.  The brain consists both of hierarchical layers ("a man riding a dog riding a lizard", as John Brunner put it), and of many parallel modules.  To a large extent, our unified view is an after-the-fact interpretation.

One way to deal with this is to think of each of these modules as a different voice, providing input that can be analyzed by the conscious mind as part of making a conscious decision.  For example, Rational Recovery, a method for dealing with drug and alcohol addiction, involves identifying the addictive voice (the one that says "I need a drink"), and then consciously labeling it as "other" (the "beast" in their terminology), rather than "self".  In geek terminology, one might look at this as finding a dysfunctional module and labeling all of its output as questionable, to be subjected to extra critical review in the synthesis phase of decision.

In practice, identifying these voices is difficult, because often modules communicate, not clearly and distinctly, but in broad swaths of emotion that infuse every fiber of one’s being.  As an IPC system, it sucks.  But with practice, we can learn to identify these messages and their sources.

Continue reading "Modules, signals, and guilt" »

Scott Adams has a suggestion for overcoming bias:

When Dilbert launched in newspapers, the response was underwhelming. In the early years, it wasn’t a workplace strip. It was about Dilbert’s life in general. He just happened to have a job. I was surprised to learn, via my e-mail, that readers loved the relatively rare comics featuring Dilbert in the office. Personally, I didn’t think those were my best work. My ego told me to do it my way. My readers told me I was wrong.

What the hell do readers know? After all, they aren’t syndicated cartoonists, and I was, albeit in only a few dozen newspapers. But this time, fortunately, I ignored my ego, changed the focus of the strip to workplace humor, and it took off. …

I’ve come to call this ego-driven behavior the “loser decision.” I don’t mean it as an insult.  It’s an objective fact that life often presents us with choices where the comfortable decision leads nowhere and one that threatens your ego has all the potential in the world.

You need a healthy ego to endure the abuse that comes with any sort of success. The trick is to think of your ego as your goofy best friend who lends moral support but doesn’t know shit.

I don’t think it’s quite fair to be so down on comfortable decisions – discomfort is, after all, a cost.  But it is an upfront cost, and given our irrationally high discount rates, we are likely to understate the net present value of the decision.  Also, my experience is that people tend to consistently overestimate the amount of discomfort involved in doing something new, and once they start, it usually isn’t that bad.  (Which I have trouble seeing an evolutionary explanation for – any ideas?)

So I mostly agree with Adams, but my solution is a bit different:

Continue reading "Goofy Best Friend" »

Bryan Caplan posits that people treat irrationality as a consumption good, in that they choose to buy it when its cheap.  Irrationality is cheap when holding wrong opinions is unlikely to have negative impact on our lives, such as when we are opining about things we have no control over (eg politics).  Since opinions on these things can make us feel good, there is value to buying them, and so bias is rife.

A few years back, I noticed a striking example of this in my own thought processes.  I have long been a libertarian, and I realized an alternative way to look at why governments perform poorly, which gave a strange new idea about how they could be made to do a better one.  Coming up with this solution made me suddenly see all the problems with conventional libertarianism – the views which I had but recently held.  And not just small problems – I went from thinking that if only enough people could be converted, or a state taken over, a stable libertarian government could be created, to thinking that it would never happen, and if it did would quickly devolve back into big government.

Despite my purported values of objectivity, my subconscious had sneakily purchased some irrationality.  The irrationality was cheap, as erroneous political beliefs have little impact on my life, and it was valuable, as it gave me hope that the society I find most appealing might someday exist.  It was only when that value disappeared, because I saw an alternate way for the society to come about, that I was able to rationally examine the belief.

This suggests some techniques for reducing bias.  If it is important to you that something be true, imagine an alternative which gets you the same positive feeling, and see if the first thing is still believable.  Or imagine yourself in a position where your opinions on grand social issues actually matter.  One of the great things about betting markets is that make many types of opinions matter, thus forcing people to more carefully consider them.  If the blowhard at the bar really thinks Iraq is such a mess, ask him to place a bet.  If he won’t, everyone (including, most importantly, himself) knows that he doesn’t really believe it.  If he does bet, he’ll quickly learn whether he is overconfident about his judgment.

On the other hand, perhaps it is best for our mental health that not all of these attics be subjected to the light of critical examination.  There is some value to illusions – particularly those our subconscious sees as cheap at the price.