TimS comments on How can we get more and better LW contrarians? - Less Wrong

58 Post author: Wei_Dai 18 April 2012 10:01PM

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

Comments (328)

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

Comment author: Onelier 19 April 2012 05:06:19AM 8 points [-]

Stream of consciousness. Judge me that ye may be judged. If you judge it by first-level Less Wrong standards, it should be downvoted (vague unjustifiied assertions, thoughtlessly rude), but maybe the information is useful. I look first for the heavily downvoted posts and enjoy the responses to them best.

I found the discussion on dietary supplementation interesting, in your link and elsewhere. As I recall, the tendency was for the responses (not entrants, but peoples comments around town) to be both crazy and stupid (with many exceptions, e.g., Yvain, Xacharaiah). I recall another thread on the topic where the correct comment ("careful!") was downvoted and its obvious explanation ("evolution works!") offered afterward was upvoted. Since I detected no secondary reasons for this, it was interesting in implying Less Wrongians did not see the obvious. Low certainties attached since I know I know nothing about this place. I'm deliberately being vague.

In general, Less Wrongians strike me as a group of people of impaired instrumental rationality who are working to overcome it. Give or take, most of you seem to be smarter than average but also less trustworthy, less able to exhibit strong commitments, etc. Probably this has been written somewhere hereabouts, but a lot of irrationalities are hard to overcome local optima; have you really gone far enough onto the other side? Incidentally, that could be a definition for x-rationality (if never actually done): Actually epistemically rational enough that it's instrumentally useful. Probably a brutally hard threshold to achieve and seems untrue of here, as I believe I've seen threads comment.

I was curious about the background of the people offering lessons at the rationality bootcamp, and saw some blog entry by one of them against, oh, being conservative in outlook (re: risk aversion). It was incredibly stupid; I mean, almost exclusively circular reasoning. You obviously deviate from the norm in your risk aversion. You're not obviously more successful than the norm (or are you? perhaps I'm mistaken). Maybe it's just a tough row to hoe, but that's the real task.

Personal comment: I realize Dmitry has been criticized a bit elsewhere and the voting trend doesn't support generalization to the community at large, but my conversation with him illustrates what I generally believe about this place. I knew more than he did. I said enough that he should realize this. He didn't realize it and shoehorned his response into a boring framework. I had specific advice to give, which I didn't get to, and realized I was reluctant to give (most Less Wrong stuff seems weak to me).

A whole lot of Less Wrong seems to be going for less detail, less knowledge, more use of frameworks of universal applicability and little precision. The sequences seem similar to me: Boring where I can judge meaning, meaningless where I can't. And always too long. I've read about four paragraphs of them in total. The quality of conversation here is high for a blog, of course, but low for a good academic setting. Some of the mild sneering at academics around here sounds ridiculous (an AI researcher believes in God). AI's a weak field. All round, papers don't quite capture any field and are often way way behind what people roughly feel.

Real question: Do you want me here?

I like you guys. I agree with you philosophically. I have nothing much to offer unless I put some effort into it (e.g., actually read what people write, etc). No confusion: You should be downvoting posts like this in general. You might want to make an exception 'cause it's worth hearing a particular rambling mindset once. My effort is better spent elsewhere (I can't imagine you'd disagree). I can't see anything that can be offered to me. I feel like I was more rational at age 7 than you are now (I wrote a pro and con list for castrating myself for the longevity and potential continuity of personality gains; e.g., maintaining the me of 7). A million other things. I'm working on real problems in other areas now.

Comment author: TimS 19 April 2012 02:20:48PM *  8 points [-]

A whole lot of Less Wrong seems to be going for less detail, less knowledge, more use of frameworks of universal applicability and little precision. The sequences seem similar to me: Boring where I can judge meaning, meaningless where I can't. And always too long. I've read about four paragraphs of them in total. The quality of conversation here is high for a blog, of course, but low for a good academic setting. Some of the mild sneering at academics around here sounds ridiculous (an AI researcher believes in God). AI's a weak field. All round, papers don't quite capture any field and are often way way behind what people roughly feel.

This. A thousand times this. As a lawyer, LessWrong pattern matches with people outside a complicated field who are convinced that those in the fields are idiots because observers think that "the field is not that complicated."

That said, "Boring where I can judge meaning, meaningless where I can't." is an unfair criticism. Lots of really excellent ideas seem boring if you had already internalized the core ideas.

Comment author: Will_Newsome 20 April 2012 02:46:42AM *  8 points [-]

Reminds me of part of a comment on Moldbug's blot, by Nick Szabo:

[legal reasoning]

It's a disciplined and competitive (dialectic, in the true original sense of that term) use of analogies, precedents, and emergent rules, far more sophisticated than normal use of analogy and metaphor. I learned it my first year of law school and it's a radically different kind of thinking I had never encountered before in school. The Bayesian bloggers seem to be completely oblivious to it, and to the tremendous value of tradition generally. That makes them, from my POV, culturally illiterate and incompetent to opine on law or politics. Yes, legal training also made me stuck up. :-)

If you can't afford law school, you can learn most of what you need to know from Legal Method and Writing by Charles R. Calleros and a first year law school common law casebook (Torts, Property, or Contracts).

The extremely short description of legal or scholastic reasoning is to think of a proposition or dispute as Schrodinger's Cat, both true and false at the same time, or each party at fault or not at the same time, or the appropriate dichotomy. Then gather all the moral or legal disputes that are similar to this one. Argue by analogy for each side both from the facts of those prior disputes and from the informal rules ("holdings") implied by the decisions resolving those disputes. This kind of reasoning allows a lawyer to anticipate an opponent's as well as their own argument in a case, and allows a judge to appreciate both sides of an argument, the latter also crucial, but often absent, in reasoning in about politics, morals, and the more complex areas of science, which in absence of this kind of discipline is dominated by confirmation bias and lack of understanding of other points of view.

Law also has a sophisticated set of qualitative probabilities I've blogged on, which imply not just degrees of truth but various aspects of gathering evidence, burdens of proof, and so on. The scientific method derived in large part from the Continental law of evidence, with which Galileo, Leibniz, etc. were intimately familiar having studied law. But legal reasoning, or scholastic reasoning as it used to be known, is still capable of covering a far wider swath of the human experience than scientific reasoning which is really just a special case and applies well only to hard evidence or the hard sciences.

I've been studying the history of common law lately due to Nick's influence, after which I'm gonna read the book he recommended. I notice that his description of legal reasoning is very similar to how I use my chess subskills for rationality.

Comment author: TimS 20 April 2012 07:51:59PM 2 points [-]

The extremely short description of legal or scholastic reasoning is to think of a proposition or dispute as Schrodinger's Cat, both true and false at the same time, or each party at fault or not at the same time, or the appropriate dichotomy. Then gather all the moral or legal disputes that are similar to this one. Argue by analogy for each side both from the facts of those prior disputes and from the informal rules ("holdings") implied by the decisions resolving those disputes. This kind of reasoning allows a lawyer to anticipate an opponent's as well as their own argument in a case, and allows a judge to appreciate both sides of an argument, the latter also crucial, but often absent, in reasoning in about politics, morals, and the more complex areas of science, which in absence of this kind of discipline is dominated by confirmation bias and lack of understanding of other points of view.

This is a moderately reasonable model of litigation, but it isn't complete. For example, Thurgood Marshall litigated separate-but-equal in the law school context specifically because every judge has a gut feeling of how to compare law schools, which just isn't true about other educational institutions. In law school, I heard the apocryphal story that the law for the State of Texas argued that the new segregated law school was just as good as UT Law School, and Justice Clark - a graduate of UT - passed a note to a colleague that read "Bullshit" That's clever lawyering and has nothing to do with arguing from precedent.

Further, not all law is litigation. The legislature empowered to make new laws that have no relationship to old laws. In short, there's a fair amount more to the practice of law than reasoning by analogy, even if reasoning by analogy is an important skill for a lawyer.