Friday, September 26, 2008

how much do humans think, really?

Some realizations manage to be both obvious and mind-blowing, such as "much of the time, the majority of people don't apply thought to problems and actions". It explains a lot about human behavior, particularly for someone like me who at times has trouble correctly interpreting what others do. To clarify, "apply thought" refers to collecting data with a minimum of bias, analyzing and synthesizing the data, reaching a conclusion, creating a feasible plan that corresponds to the conclusion, and executing the plan. Since this is a description of the way in which people ascertain and manipulate reality, many examples exist of using it or not, across the wide spectrum of situations from choosing lunch to the most far-reaching decisions of rulers in the highest-ranking echelons.

One or more of those steps is usually skipped or botched. For instance, people substitute the conclusion step for the data collection step: "I don't need to hear anything more, I've already made up my mind". Or they disregard the analysis and synthesis step by reacting, shooting from the hip, doing a gut-check, winging it. They're ruled by their previously-trained mental associations to the stimuli. Or they mess up the plan creation step with faulty assumptions about how reality will respond to them. Or they know the plan but they do something easier instead.

On the other hand, I'm not naive enough to seriously suggest that the "apply thought" approach is all that anyone needs, all the time. It's slower, it can't operate when the data is insufficient, it requires effort, it's fallible simply because the thinker is fallible, it can prompt people to excessively trust their past conclusions and plans. For all these reasons, especially "slower" and "requires effort", people could quite reasonably argue that the cost outweighs the benefit to apply thought to tiny and small-term questions; I often agree. I even grant that large contingents of the population are, mostly through no fault of their own, ill-suited and disinclined to it. I'm also fine with the concept of other mental approaches as strong supplements.

The bit that depresses me is when humans just don't apply thought (as described here) to the questions that actually are complex and long-term, sometimes despite appearances. Those questions unquestionably merit it. And the mind-blowing fact is that someone who doesn't apply thought is someone who cannot be reasoned with. Arguments, details, etc. are all useless. They're speaking a different cognitive language, more or less. All someone can do is ensure they are "triggered" properly, not engaged in meaningful conversation.

The failure to apply thought may be most subtle, and yet the most apparent, in the case of habits. A habit is a succession of tiny and small-term conclusions and actions, and the succession as a whole is pivotal and long-term. Hence, the refusal to properly apply thought to one out of the bunch is minor and inconsequential, but the repeated refusal to properly apply thought to the aggregate is potentially catastrophic. The key is to perceive the invisible stakes stretching through time, not the minuscule visible stakes: this is the analyze and synthesize step. (I'm sure all casino gamblers realize that every "game" is by definition designed so most "players" lose, and in fact if someone devises an excellent strategy to win then the casino accuses him or her of "cheating"?)

It's revealing that a related expression, "just one more", is such a common catchphrase. A startling number of heart attack victims either don't change their behavior, change it too little, or change it for too short of a duration. When I'm most discouraged about humanity's future, I wonder if we're trapped in a deadly medium. Humans, or at least a few of them, apply thought sufficiently to empower themselves. At the same time they, or at least a few of them, don't apply thought sufficiently to handle the power well. (I'm not referring obliquely to the ethical use of power. I'm referring merely to using power non-stupidly, i.e. not being short-sighted.)

Saturday, September 13, 2008

you might be a Myers-Briggs INTP if...

...your immediate response to "INTP" isn't a variant of "What the...?" but "That one's easy: int pointer!"

If you then go on to start musing about how the INTP enforces the type of its referent, and how it handles overflow, well, there's even less doubt...

Thursday, September 04, 2008

the Algorithm will be incorrect

...and we the users (or maintenance programmers) write the pertinent definition of "incorrect". Software and hardware might execute flawlessly the exact steps proscribed by a set of exhaustive written policies. If that is the case, good jorb and a round of gold star stickers to everyone responsible. Still, the algorithm is incorrect each time it produces a different result than we think it should.

The way to cope is to plan ahead for inevitable incorrectness. As much as is reasonable, assumptions should be flexible. Options should be open-ended. Overrides should be available; users should be advised and protected from mistakes, but some users have their own good reasons to request something "incorrect".

A perfect algorithm will be incorrect. More poetically, a perfect algorithm is too good for this imperfect world.