post by Eliezer Yudkowsky (Eliezer_Yudkowsky)
"All my life I have been intensely repelled by the idea of 'making an effort'. I hate this idea today as much as I did as a child. I don't know why I hate it so much; I just do."
-- Raymond Smullyan, The Tao Is Silent
In the Hollywood version of rationality - or even the Traditional rationality that was passed down from supervisor to grad student in ancient days before Bayesianism - rationality is a great strain, a great effort, a continuous battle to coerce your mind into a desired shape. Spock, the archetype of Hollywood's concept of rationality, represses all his emotions.
And this great effort, they conceive, is virtue unto a rationalist. The more effort you expend on forcing yourself into the mold, the better the rationalist you must be. It's like working extra hard at your job, as demanded by the Protestant work-ethic. If the one works long hours - sweating, getting ulcers - surely the one must be worthy of praise?
This, I think, is an instance of a Lost Purpose. People see that successful folk must sometimes make an effort, and so
they conclude that effort of itself is virtuous whether or not it
I am reminded of an AI koan from AGI '06, where the discussion turned (as it often does) to defining "intelligence". A medium-prominent AI researcher suggested that an agent's "intelligence" could be measured in the agent's processing cycles per second, bits of memory, and bits of sensory bandwidth. To which I replied with a quote from Dijkstra:
"If we wish to count lines of code, we should not regard them as 'lines produced' but as 'lines spent': the current conventional wisdom is so foolish as to book that count on the wrong side of the ledger."
Surely (I said), an agent is less intelligent if it uses more memory, processing power, and sensory bandwidth to accomplish the same task?
This reply was due, in no small part, to my having read Raymond Smullyan's The Tao Is Silent at the age of sixteen. Raymond Smullyan
is a mathematical logician, a great composer of logic puzzles, and
sometime Westernized Taoist. Though I disagree with much of The Tao Is Silent, I would count "just the parts of the book I liked" as one of my most important formative influences as a rationalist.
In particular, it was in The Tao Is Silent that I first encountered the
Taoistic principles of spontaneity, working with rather than against the nature of things, human goodness as a distinct phenomenon from the heavy weight of dutiful moral obligation, and above all, wei wu wei, "acting through not acting".
Smullyan's Taoism was more inspiration than instruction, but it was important inspiration. I matured as a rationalist while keeping firmly in mind that my "rationality" was not measured by how much effort I expended on proper thinking, but rather how little.
You can see this same view manifested in these lines from The Simple Truth:
"You have to throw in a pebble every time a sheep leaves through the gate?" says Mark. "Take out a pebble every time a sheep returns?"
Autrey nods. "Yeah."
"That must be really hard," Mark says sympathetically.
Autrey brightens, soaking up Mark's sympathy like rain. "Exactly!" says Autrey. "It's extremely hard on your emotions. When the bucket has held its level for a while, you... tend to get attached to that level."
A sheep passes then, leaving through the gate. Autrey sees; he stoops,
picks up a pebble, holds it aloft in the air. "Behold!" Autrey
proclaims. "A sheep has passed! I must now toss a pebble into this
bucket, my dear bucket, and destroy that fond level which has held for
so long -" Another sheep passes. Autrey, caught up in his drama, misses
it; so I plunk a pebble into the bucket. Autrey is still speaking: "-
for that is the supreme test of the shepherd, to throw in the pebble,
be it ever so agonizing, be the old level ever so precious. Indeed,
only the best of shepherds can meet a requirement so stern -"
"Autrey," I say, "if you want to be a great shepherd someday,
learn to shut up and throw in the pebble. No fuss. No drama. Just do
Long ago - I think I must have been pretty young - I decided to move my limbs with maximum "efficiency", to save effort. I might even have been thinking of Vulcans.
So I tried what my youthful mind wordlessly conceived of as "efficiency": I tried to move my limbs in perfectly straight lines as quickly as possible, with corresponding sudden stops and sudden starts.
"Efficiency" didn't feel very efficient. The sudden starts and sudden stops took effort. Moving my hand in a straight line forced my elbow and shoulder to move in strange curves.
You can buy books that teach this same life lesson, but they use a lot more pages.
Now this is scarcely Taoism, at least so far as philosophical premises are concerned. According to authentic Taoism, you can exert no effort at all while accomplishing all worthwhile things. This seems to me around as plausible as an agent that achieves its utility function using zero computing power and is therefore maximally intelligent. The only way you could do it is if the agent assigns constant utility to all outcomes, or if the utility function's maximum is set by sleight of hand to wherever the universe goes anyway. This may be why I am not a Taoist: "A maximally intelligent agent with zero computing power and no utility function" sounds like a good metaphor for the Tao. I object to a metric of intelligence that makes me dumber than a rock.
According to Taoism, everyone ought to act in accordance with their natures. One can scarcely see how it could be otherwise. I think this religion only appears nontrivial because of selective failure to consider all its consequences.
In any case, my own nature is to make certain efforts even if they seem unpleasant. Therefore I have no objection to making an effort now and then.
But one should not think that force, effort, and control are virtuous unto a rationalist - that would book them on the wrong side of the ledger.
Addendum: Before anyone else points it out: Yes, I know that my critique of Taoism is appallingly simplistic and that the Taoists are well aware of it. That doesn't make it wrong.
Comments sorted by oldest first, as this post is from before comment nesting was available (around 2009-02-27).
comment by Yoshi ·
2007-12-23T05:06:33.000Z · LW(p) · GW(p)
Perhaps you should write someday about choosing examples. In certain circumstances the computing power wasted is a good measure of intelligence, for example if one considers chess programms.
comment by James_Bach ·
2007-12-23T05:15:39.000Z · LW(p) · GW(p)
Thank you for this post. I'm in the process of writing about my system of self-education, which has two interesting elements I haven't heard anywhere else: A) it requires no self-discipline whatsoever, B) it is centered on the feelings of learning, rather than artifacts and techniques of learning (those latter two things are interesting, but they orbit the first).
One of the things I try to explain is how to embrace certain mental behaviors that seem bad, such as procrastination. I procrastinate extensively. I am procrastinating right now (I'm "supposed to be" writing a chapter on the history of buccaneering and how that relates to intellectual buccaneering).
Procrastination helps me learn with less effort. It's too long to explain fully here, but one way I use it is called "springboard procrastination" which is the phenomenon of trying to work on one thing, and feeling your mind aggressively push you into another thing. I once thought that was a shameful thing, to let my mind push against my will, but I eventually discovered that by rolling with that impulse, I could get lots of things done. I read more, I write more, I exercise, I am quite productive while avoiding the work I'm "supposed to do".
I also use a technique I call "procrastinate and push" which means I keep coming back and knocking on the door of my mind, trying to work on the problem of the day, and when my mind tells me to go play a video game instead, I just do that. But I come back a little later, and a little later. I go through these cycles without any sort of bad feeling (unlike when I was younger and disgusted with my inability to bring my mind to heel). Eventually, usually shortly before a deadline, my mind relents. Often progress is very rapid after that.
These experiences have caused me to explore lots of way in which I can make good progress without the feeling of making an effort. One of my mentors in this, Jerry Weinberg, recently wrote a book about a relaxed way of writing called the Fieldstone Approach, which he first put into words while coaching me on this stuff.
Anyone interested in reviewing my book prior to its publication should contact me directly. I will soon need a few bright people to critique it.
comment by Mike_Kenny ·
2007-12-23T16:32:03.000Z · LW(p) · GW(p)
I loved the Tao is Silent too. You seem like you're on the same page as Smullyan to me! Recall the chapter in which three people talk about making an effort--one everything is easy for, one person everythign is hard for, and another is in the middle, IIRC. I took it as different people have different natures, and trying to change those natures is a bit like being the Horse Trainer who harms the horses in his effort to make them 'better'. Edison was an inefficient workhorse perhaps, whereas Tesla was a more efficient type. But they both achieved great things.
comment by Michiel_Trimpe ·
2007-12-23T17:23:44.000Z · LW(p) · GW(p)
Here's my perspective on that utility function you hold so dear:
Imagine what would happen if a neural network in the wild developed a utility function through evolution. Does eating from the tree of the knowledge of good and evil sound familiar?
Now imagine some neural networks gaining read/write access to their own utility function, once again evolving beyond good and evil, abandoning desire and living without effort, which is simply frustrated desire.
comment by briarandbramble ·
2007-12-23T18:30:19.000Z · LW(p) · GW(p)
"Surely (I said), an agent is less intelligent if it uses more memory, processing power, and sensory bandwidth to accomplish the same task?"
Be careful not to confuse the intelligence of the agent with the intelligence of the person who created the agent.
Two agents that accomplish the same task are equally intelligent. The fact that one agent uses more memory and the other uses more styrofoam peanuts is irrelevant.
comment by Recovering_irrationalist ·
2007-12-23T19:00:39.000Z · LW(p) · GW(p)
I hope this post is the start of a series. My main problem is not managing to actually do what I know perfectly well in my head would be the rational thing.
I'm suspicious, but much of this rings true. Please tell or link to more, I found nothing on your blog.
comment by poke ·
2007-12-23T20:52:45.000Z · LW(p) · GW(p)
I think the Taoists (in your simplified version) won the day here. Obviously you can't not act according to your nature but, as you proved with your efficient movement experiment, you can act according to horribly incoherent theories about your nature or not act according to them and do better. What the world is trying to tell you (screaming, shouting) in all the examples you gave is that "intelligence" is a horribly incoherent theory about your nature. It means "doing less" exactly insofar as it is just identical with the human behavioral repertoire; it means having greater resources (memory, processing power) exactly insofar as it is a misplaced analogy with the properties of an mechanical artifact. The human behavioral repertoire cannot be improved upon because there is nothing but itself with which to compare it; it is simply a product of evolution. (Overcoming bias is self-annihilation.) What the Taoists grasp is that you do better to recognize this than not (and that's the only way to do better).
comment by Barkley__Rosser ·
2007-12-23T22:10:32.000Z · LW(p) · GW(p)
Raymond Smullyan used to pull dimes out of my ears when I was a kid, no kidding.
My old man, a friend of Smullyan's (hence his access to my youthful ears), used to
argue that a major motivation for mathematicians was "laziness," a desire to figure
out ways to solve problems with less effort.
comment by pdf23ds ·
2007-12-23T22:45:18.000Z · LW(p) · GW(p)
James Bach, RI, PC: I think the technique James described is basically a way of dealing with mild ADD without guilt. People with more severe ADD won't have as much luck with the technique or other forms of behavioral modification, and should probably look for medical treatment. (At that point behavioral modification could help you reduce the required dosages or whatnot.) That's where I am at the moment, anyway. Really, consider the possibility that the desired level of productivity is just out of your reach without medication. I know it's a complicated subject.
comment by JulianMorrison ·
2007-12-24T02:39:35.000Z · LW(p) · GW(p)
The action through inaction thing is not about having no preference between alternatives. It's about choosing an alternative by settling yourself into a pattern where your preference is the unforced consequence.
Ron Paul (aside from his chances or the advisability of his politics) is a perfect demonstration of Wu Wei. "The highest type of ruler is one of whose existence the people are barely aware [...] The Sage is self-effacing and scanty of words. When his task is accomplished and things have been completed, All the people say, 'We ourselves have achieved it!'"
Paul is able to lead and direct a spontaneous organization of thousands because of a history and a nature that he has expended much time cultivating. He has worked, but I suspect it wasn't "effort" - because he was acting in accordance with his nature. He has made himself an exemplar of his philosophical theory, and so others are attracted and organize themselves around him. He doesn't have to command from the center, because people already know what he would want - and they agree. So they just get to work.
(For those who think I'm over-praising RP, consider that I'd describe Osama Bin Laden in the same terms.)
comment by Barkley__Rosser ·
2007-12-24T21:28:28.000Z · LW(p) · GW(p)
Yes, there is even a stronger motive for computer programmers than for just plain old mathematicians. Real money is involved in shortening those programs.
Regarding Taoism and economics, there has long been a dialectic in Chinese culture and philosophy between Taoism and Confucianism. The former is "the scholar out of power," opposing state power over the economy and society, practiciy we wei and so on; while the latter is "the scholar in power," supporting hierarchy and imperial state power over the economy and society in a supposedly harmonius balance.