thomblake comments on Beware of Other-Optimizing - Less Wrong

79 Post author: Eliezer_Yudkowsky 10 April 2009 01:58AM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (117)

You are viewing a single comment's thread. Show more comments above.

Comment author: Eliezer_Yudkowsky 10 April 2009 03:08:44AM *  19 points [-]

I tried reading your blog posts and couldn't (allergic to your style), but I'm sorry to inform you that you haven't reached the level of universal generalizations as yet. The stories you make up to explain why your tricks work are not the deep answers which constrain both the rule and the exception; from other sciences I have learned what true general models of the human mind look like, and your explanations, I'm afraid, are not in that class. The fully general art of combating human akrasia has not been invented by you. Your clients are only the ones for whom your techniques happen to work.

I hope that having discovered some tricks that work for some people is enough honor for you; and that you do not need to claim that your tricks work universally in order to value them. And that it does not wound you too deeply, if there are some people for whom your advice does not yet work, and who you do not yet understand. This is not the counsel of despair: study the exception and the rule, and you may find the deeper law.

Of course you could decide that I'm just being lazy. (Laughs.)

Comment author: pjeby 10 April 2009 02:37:46PM *  12 points [-]

I'm sorry to inform you that you haven't reached the level of universal generalizations as yet.

My generalizations aren't, for the most part, in my blog posts, nor in most of my for-pay material, actually. Abstractions don't help most people take action. The only really important "theory" on my blog is The Multiple Self, which was where I first realized that I was being stupid to assume that my conscious mind had ANY direct control over my actions, given how late consciousness appeared from an evolutionary perspective.

Most of the other generalizations my work sits on top of can be found in General Semantics and NLP, anyway... they just don't help much in their raw form.

But here is a useful generalization: if you test autonomous responses, you can create techniques that work. If you're not testing, or not making use of your autonomous, involuntary responses (both mental and physical), you're utterly wasting your time.

More than half of my early blog posts are wastes of time, in precisely that sense. They were written long before I learned how to shut up and test, as it were.

The fully general art of combating human akrasia has not been invented by you.

Heck no. I've really only specialized in chronic procrastination and personality sculpting. Fighting akrasia was a label that people here applied to my work. I don't really believe in akrasia, anyway -- a better description would be anosognosia of the will. (That is, we explain our behavior as akrasia or failure of will, because we don't understand that our will isn't singular. And we do it for the same reason we see gods in the forest -- our built-in projections of mind and intention. When applied to self, they produce prediction errors.)

I hope that having discovered some tricks that work for some people is enough honor for you; and that you do not need to claim that your tricks work universally in order to value them.

I've discovered very little, actually. Most of what I've done has also been invented by other people (as I've sometimes discovered when somebody says, "hey, your stuff is kind of like author X"). All I've really done is systematize some things so that they're more teachable and repeatable, and try to replace mystical explanations with mechanical ones.

In fact, I would go so far as to say that all I've really done is take a very narrow subset of NLP and CBT that can be self-applied (by most people) with self-testing and don't require physical presence to be taught, and throw in a few heuristics about what to look for and what to apply them to.

Of course you could decide that I'm just being lazy.

For you to test something basic, like a submodalities of motivation exercise from an NLP book, would take you maybe 15 minutes... only slightly longer than it took you to squirt ice water in your ear. ;-)

Now, personally, I'm not sure if such an exercise would work for you. I've never been really good at doing submodality work on myself, though I'm okay at guiding others through it. But you need to understand that having trouble accessing something like your submodalities on your own, doesn't mean they aren't there.

What I'm getting at is that individual idiosyncracies only affect what techniques you'll be able to usefully self-apply; not what techniques will actually work. There are NLP techniques I still can't currently self-apply, and have to have someone else walk me through in order to do them, because I can't think about the technique and do the steps at the same time. That doesn't mean the technique "only works for some people" -- clearly I have the hardware the technique operates on, I just have limited fluency in accessing that particular hardware.

Similarly, there are techniques in my repertoire that some of my clients can't self-apply; I have to walk them through, or they have to use a recording or some sort of external aid. Some of these issues go away with practice, some don't.

The major insights I've had regarding self-help material is not that "some things work for some people and others don't" -- it's that:

  1. Some people can learn a particular technique from a particular book, and others can't,

  2. Some people can do a particular technique on their own, and others can't (although they may be able to learn to), and

  3. Self-help books usually barely whisper some of the critical mental and physical distinctions needed to make a particular technique workable, while most people have too many existing preconceptions shouting in their head, for any of those whispers to be heard.

The #1 most important thing in doing virtually any self-help technique worthy of the name is being able to pay attention to your unconscious, automatic responses, without adding voluntary thought or anosognosiac explanations on top of them. And in my experience, it's the hardest thing to learn to do on your own; and as far as I can tell, nobody (not even me) has made a systematic attempt to teach it. (So far, I just point it out to people when they're doing the wrong kind of thinking.)

But if you can pay attention to your responses, and you are disciplined about testing those responses, you can invent your own techniques. That's what I did, for a while, and then I started going back and re-reading self-help books, using the idea of testing my autonomous responses to validate which ones worked, and the skill of paying attention to autonomous responses in order to apply them in the first place.

And what I've found, for the most part, is that virtually all self-help techniques work for something, if used correctly. It's the "used correctly" that varies immensely from person to person.

Even some techniques that I thought were utterly stupid (e.g. EFT and Sedona) can be made to work, and I learned some interesting things from them. Mostly what I've noticed, though, is that the people teaching them have a tendency to leave out (or say only in a whisper), certain things that you need to make them work, or they fail to explain the common failure modes.

(The common failure modes are very similar, btw, across a wide variety of techniques; mostly they amount to trying to do things by willpower or conscious analysis that can only be accomplished by waiting for an autonomous un-willed response.)

Anywho... if you want to find universal models, I recommend you skip my blog and go straight to the source: your own brain. Start observing the responses you don't control -- the almost-subliminal flashes of memory and sensation that occur in response to pondered questions or the thought of taking a particular action. Experiment for yourself, and find out whether these responses are repeatable in response to the same stimuli, and what techniques actually produce changes in those responses, and your resulting behavior.

Comment author: thomblake 14 April 2009 02:30:07PM 1 point [-]

When applied to self, they produce prediction errors

Correction:

When applied to self, they produce predictions. Sometimes these predictions are wrong.