« Detached Lever Fallacy | Main | A Genius for Destruction »

August 01, 2008


There was insufficient discussion of Occam’s razor:

(1) At the beginning of time the universe was in the simplest possible state (minimal entropy density). Why?

(2) Popper showed that an infinite number of theories is compatible with any given set of finite observations. Mere algorithmic shuffling to
calculate Pr(B) probablities according to the Bayes formula won't help
much. Successful (ie. real-time) induction needs the principle of Ocaam’s razor to ensure the priors are set correctly. What does this say about the limits of Bayesianism?

(3) Occam’s razor is not an add-on, it’s an all-pervasive principle of science. Favoring the simple over the complex is equivalent to saying that some representations are more elegant than others, which actually means that there’s a general set of aesthetic principles underlying Occam’s razor. As computer hacker Paul Graham pointed out, the same set of aesthetic design principles keeps popping up again and again. Why?

Has there been any discussion of 'rationality' under the influence of alcohol?
I don't know if this is a topic 'frowned upon' by the readers and writers of this blog, but i thought it was worth mentioning. I often find myself either almost 'more rational' or more willing to discuss issues of importance with little inhibition ( key term here) under the influence of alcohol.

Any comments?

Phew. That took a while to write.

I couldn't easily find statistics on this) so I rely on my personal observations(of mostly Southern Finland) and I might be suffer from confirmation bias here, but it has for a long time seemed to me that people are more likely to exceed the speed limit when it is lower. Does anyone know of research on this? It seems most obvious when comparing behavior between speed limits of 80 km/h, where it seems relatively common to drive closer to 90 km/h, and 100 km/h or 120 km/h limits, where most seem to drive at close to that speed or maybe even a bit under the speed limit. (My assessments are based on my car's speed meter, which based on several measurements at roadside checkpoints

If this observation is correct, there are a few interesting points:

1) The speed limits are supposedly based on assessments of safe speeds for those roads(and noise when they are close to housing)
2) at least in Finland the penalties for speeding are based on fixed excesses for certain ranges, e.g. 16-20 km/h excess for speed limits of at most or over 60 km/h, and it is often cited as a rule of thumb that the police don't mostly care about excesses of less than 10 km/h except maybe at limits of 30-40 km/h. A fixed excess is more likely to be punished at lower speed limits, so why would people speed more often at lower speeds?
3) the relative effects on time for reacting and braking of a fixed increase in speed are greater at lower speeds. Do people's assessments of safe speeds differ from "official assessments" more at lower speeds and if so, why? This is especially interesting as in Finland roads with 100-120 km/h speed limits are wider and mostly have much better visibility ahead and to the sides, even when taking into account longer times to stop.

At the beginning of time the universe was in the simplest possible state (minimal entropy density). Why?

Because at that point there is no information remaining about any earlier time. If anything happened before it reached minimal entropy density, all the evidence has been erased.

Whatever was that "Mundane Dishonesty" post?

When is Eliezer's book coming out? What topics will be covered? I want to go back and read some old of his posts, but I get computer headaches; so if some of this stuff will be in book form soon, I would prefer to wait. Eliezer, can you provide any information? Thanks.

Favoring the simple over the complex is equivalent to saying that some representations are more elegant than others, which actually means that there’s a general set of aesthetic principles underlying Occam’s razor. As computer hacker Paul Graham pointed out, the same set of aesthetic design principles keeps popping up again and again. Why?

Because, all else being equal, something with many necessary parts is easier to break than something with few.

Medicine vs Science
I am an MD/PhD with very strong interests in both molecular biology and medicine (currently thinking of residencies in neurology, psychiatry, neurosurgery, cardiothoracic surgery, endocrinology, immunology, or infectious disease).

Now, I know many people on this blog have the opinion that medicine is not only *not* science, but out-right witch-doctory... IE, most docs are in violation of the hippocratic oath of 'Do No Harm,' by acting *at all* when they should not (harm both to the patient and the economy).

On the other hand, many people here have also expressed disgust with the current state of academic science, inculding its many biases and out-right fraud, which mars the field and makes it difficult to make anything out of the literature (first hand experience tells me this is infuriating indeed). There is also a generally unstated belief of many scientists that anything other than the best minds shouldn't be bothering with it *at all.*

To Quote Enrico Fermi, "There are two types of physicists: the very best, and those who shouldn't be in the field at all. Any theoretician who isn't the best is a fraud, a pretender."

Certainly there is more room for creativity in MoBio, yet the field is plagued by useless grad-students who don't know what to do with themselves, let alone an NIH grant...

So, the question for discussion: Who should become a doctor and who should become a scientist and how/when can they know this? Are there any basic criteria that should be laid out for young, intelligent college students when deciding a field? What *should* all those useless grad-students be doing with their lives other than suckling off the teat of the government and stamping on the shoulders of geniuses? What can *we* do to encourage our friends to a career that will make them happy and contribute something to society?

This question is mostly for Hopefully Anonymous, to some extent Caledonian, and anyone who views their own existence as their highest moral calling: Where does quality of life rank in your view? There are so many kinds of existence you could have - happy, fulfilled, miserable, fraught... And, at least in my experience, quality of life can be strongly affected by interpersonal interaction, how happy/fulfilled those around you are, the overall character of your society and culture, and so the state of other people in general. (Do you disagree?) Obviously, you don't even get to attend to happiness unless you exist in the first place, but is there any balance, in your view, between actions that lead to your continued existence, and actions that lead to a desirable existence?

Josh, my hope is to get started on editing the collected posts into e-books (which might also be put up at Lulu or some such) no later than the end of August. Don't hold me to that, though. I want to finish writing the long sequences before I switch to lighter posts and collection-editing mode. The popular book would be an additional project after that.

Sign up to get notified of this sort of thing at:


Many early "flying machines" had feathers and a beak. They didn't fly. Same principle. The only reason it sounds plausible is because flying seems so mysterious that you can't imagine actually, like, solving it. So you imagine doing an exact anatomical imitation of a bird; that way you don't have to imagine having solved the mystery.

Re-reading the AI new-timers post, and got thinking about the above analogy. Although the planes we build don't bear much resemblance to birds, they owe them a huge amount:

- demonstration that flight was possible
- a basis for studying Bernoulli's Principle, angle of attack for wings, and other theories invaluable to human flight
- nature's fliers have numerous strategies and tactics that have been borrowed or adapted for human-designed fliers

Please don't unravel this analogy too far, it doesn't stand up brilliantly, but can 'intelligence' not be substituted for 'flight'? We only know of one system of General Intelligence in the universe. I'd like to know how much weight cutting-edge AI research puts on brain scanning.

Brain modelling is a Really Hard Problem, granted. But hey, so is AI, right? Feel free to slap me down if I just plain don't get it, but surely a strategy of coding/theorizing and brain scannning 'meeting in the middle' is the best approach here? What's the evidence suggesting that neural modelling is too far behind the curve to be of much use (as suggested in comments to A Premature Word On AI)?

Some random thoughts (kindly intended) on LF's comments:

"Now, I know many people on this blog have the opinion that medicine is not only *not* science, but out-right witch-doctory..."

While we're quoting physicists, I believe Richard Feynman once said that if one is ill, he should go see a witch doctor. Perhaps he was being ironic.

"There is also a generally unstated belief of many scientists that anything other than the best minds shouldn't be bothering with it *at all.*"

This is ambiguous. What does "best mind" mean, what are the temporal and social constraints to its identification, and who is in a position to judge? Who could have guessed that a certain patent clerk working in Basel around the turn of the 20th century would go on to change our conception of space and time?

"Who should become a doctor and who should become a scientist and how/when can they know this? Are there any basic criteria that should be laid out for young, intelligent college students when deciding a field?"

I'm not sure there is any response that takes the form of some a priori rule, or even rule of thumb. Some say to follow your muse; others say follow the money. Human brains seem fairly well adapted to making up plausible explanations after the fact, but not terribly impressive at forecasting anything as varied and complex as the course of a individual's career. But there is a real need for clinicians who are skilled in the expression of empathy. A factor for consideration?

Could an AI which was unable to model a human mind deceive us? Persuade us? Understand that we perceive it as a threat and will try to turn it off?

If not, does that narrow the space of dangerous AIs to a more comprehensible subset?

Alan: "What does "best mind" mean, what are the temporal and social constraints to its identification, and who is in a position to judge? Who could have guessed that a certain patent clerk working in Basel around the turn of the 20th century would go on to change our conception of space and time?"

Unfortunately, a brief stint in academia will reveal to you that scientific administrators make this determination *all of the time* using their own criteria, which are probably better than a hard and fast set of rules, but still leaves so much to be desired... Being different is still unacceptable, even in a field as seemingly eccentric as science, unless your results so far overwhelm the cultural standards pushing for your ejection. And by results, I mean publications, NOT good science or sound lab-practices... Who ever thinks about *that*? My PI's blood has almost boiled off several times this summer at high-ranking associates' simply *horrid* to the point of dishonesty laboratory practices. I have only met two scientists who have had this reaction, and both were alpha-females struggling to keep their grants among the big dicks in the field who will say anything for a paper...

Interesting claim about alpha females. It's odd... a casual examination of scientific history and of people who I know suggests to me, though weakly, that intellectual honesty has the opposite pattern from verbal intelligence and so many other domains. Women seem somewhat less intellectually honest than men on average but they seem to have a higher variance leading to more women than men at the positive extreme.

As for "the best minds", a more reasonable approximation is probably that you shouldn't participate in a particular field of science unless its clear enough that you are better (combination of greater specific talents, dedication, intelligence, and intellectual honesty, in no particular order) than the average for your field that overconfidence isn't a plausible explanation, or alternatively, good enough to contribute to the conversations had by the very best people in the field save 2-5 outliers.

Of course, there are lots of fields of science (and of medicine), and some of them are in desperate need of dedication (to persevere against dishonest opposition) and intellectual honesty and can make due with merely 99th percentile intelligence and mildly exceptional talent.

GREAT quote about stomping down the shoulders of giants!

Thanks. I suppose I could always just fix my printer.

As for the grad students, good question. One answer is that some of them could do genuinely useful science in somewhat less prestigious fields than they actually work in. The theoretical physicists could do applied physics, the applied physicists chemistry or economics or evolutionary dynamics, etc. This only applies in so far as the necessary skills for each field are the same and the only difference is level of prestige/difficulty/intensity of competition, but this is at least somewhat true. Likewise, many people can contribute to science by going off the beaten path and exploring the less explored fields, such as the neurotransmitters other than dopamine, seratonin, adrenaline, glutamate and acetylcholine or the immune cells other than T and B cells and maybe macrophages. For other people, there is always engineering, or invention, or using one's skills as a rationalist to do an extrordinary job at running a small business, or architecture or teaching elementary or high school or working 40hrs/week for 20 years then making cool videos on You Tube, or marrying money and turning your connections into a philanthropic career or playing professional poker, among many many other options.

Intellectual honesty should produce better results in the long run, but it may produce fewer results in the short run, meaning that the long run never comes.

Eliezer, that subscribe page is insane. Those checkboxes overlap semantically something fierce, and I'm not sure if they combine with OR or XOR. I just went ahead and checked 'em all anyway, which I hope makes sense.

When to Write? When to Sing? When to Get Down to Business?

Continuing along the line of inquiry of who should do science, let's ask who should be involved in the arts. My knee-jerk reaction is, "Why EVERYONE of course! DUH!" But this is clearly not how the world works... While Fermi might say only the best physicists should be, I don't think the same applies to artists in really any field at all... Bob Dylan couldn't sing, and look at the masterpieces he produced! Some of the most interesting art has come from people with perceptual problems. There's a wonderful museum in Queens called 'The Living Museum' where psychiatric and brain-damaged patients come for 'art therapy' and produce paintings, sculpture, and poetry... Truly *amazing* stuff! Like, much of it was not only interesting conceptually, but actually incorporated very skilled technique and talented composition... Anyway, sorry for the tangent...

Back to my point, it seems that there is a high degree of pressure in our country, especially among the elite, to consider 'amateur' endevours at art as worthless or much worse, *embarrassing,* like if you can't do it *very* well, you shouldn't be doing it at all... This seems very destructive to me, as any talents of this sort require cultivation, and even the act of cultivating them seems to bring happiness to the creator...

On the other hand, the big con I see is time-eating... Is mediocre-talent cultivation worth the time of someone who could be further increasing a different talent that said person would do much better at with the same effort? I would think not, but... Any takers?

If amateur efforts are to be encouraged, how should we go about doing that? Do you tell your friend that her lop-sided sketch of a puppy is good? A good 'try'? Or do you say, 'You know, you should stick to the guitar...'

Julian, the fact that checking the "Eh, just tell me [...]" box checks all the other boxes indicates to me that it's an OR.

My assorted thoughts on the worthless grad students / amateurism issues follow below (I am supposing that a long, personal comment is acceptable in the context of this "Open Thread" discussion; if not, then do delete and we can take this to Nick's forum if desired)--

These sorts of questions are of great personal interest to me, as part of the reason I voluntarily left UCSC (undergrad) last year was that I didn't feel like I was doing anything useful. (Although the fact that I, um, had a nervous breakdown probably has something to do with it, too.) I reasoned thusly: if one is going to engage in any sort of endeavor, one should have some clearly defined purpose. I can see two general purposes for school: on the one hand, there is intellectual edification and knowledge for its own sake, and on the other hand there is vocational training and credentialing. In the modern University, these two purposes are bound up within the same institution--but I'm not sure they should be. The University is full of students who throw together crappy together papers at the last minute in exchange for a grade. The students don't want to write the papers; the TAs don't want to grade them; they're not going to be read by anyone else. It seems like such an enormous deadweight loss. Is this really helping us (run an economy, get to the Singularity, whatever)? I think of the countless thousands of students who are forced to write papers on Shakespeare. Now, of course, Shakespeare is a great writer, and many papers deserve to be written about his works--and they have. The world does not need any more papers about Shakespeare by uninterested seventeen-year-olds! It's been done!--and so you have to wonder what could be done with all that sunk cognitive firepower. Maybe some of those students could be writing literary analyses of (say) Greg Egan, instead. (Cf. Eliezer's "Lost Puropses," and James Somers's "Us and Them." These posts might have played a causal role in my decision to leave the University. Might, because intospection is imperfect, and I haven't read Judea Pearl yet.)

So I figured that when it comes to intellectual edification, I do that on my own with books and blogs--and as for getting a nonSafeway job, I'm pursuing an eighteen-month A.A.S. in IT (with an emphasis in network security). So my plan, then, is to ace my classes at this tech school, get a useful, decently-paying job, and in my free time I can study the things I'm truly interested in. So am I doing things right, then?--is it the rest of the world that's crazy, rather than me? Or is all of the above a clever post facto rationalization excusing my inability to get myself to do my homework? (I don't think it is, but if it hadn't been for that nervous breakdown, I'd probably still be at the University and would have declared my Philosophy major by now.)

I agree with Laura that the creation of art is not just for specialists. It's true that only so many people can do cutting-edge research, and similarly, only so many people can write a bestseller. Yet the space of all possible art so incomprehensibly vast!--are we to disregard this seemingly boundless depth of riches, just because not everyone can be famous (cf. Eliezer's "Joy in Discovery")? I myself have a distant dream of writing a novel about a specific cluster of ideas which I wish to tell the world. (Distant, because at the moment I'm still one of those worthless hypocrites who has ideas of writing and is continually jotting down assorted bits and scraps of thought without ever synthesizing them into something readable.) A lot of times I tell myself that I can't let myself think overmuch about publishing, because then I'll really never get anything done. I think that deep inside I have something very (subjectively objectively) important to tell the world--but as a rationalist, I won't delude myself into thinking that the world is dying to hear it. Everyone dreams of writing a bestseller--what makes me think that I'm special? My answer to this is that (from the world's perspective) I might not be special--but I can't let that matter to me. "Fame" is a chimera, an artifact of how we label our axes. The work must exist (eventually)!

Michael- I'm interested to know what you mean by saying women are more dishonest in your experience. Do you mean lying to their friends or in papers? Because in my experience, women are much more likely to lie to their friends about how they feel, but much less likely to lie about their (own) results. Many more of the men I've worked with have been willing to exaggerate their results and their significance at presentations and in papers/grant proposals, building themselves up and even comparing themselves to known geniuses, than the women, who tend to be more modest and shy about these things... If anything, it seems to me that women are too worried about a) being wrong and having it pointed out to them, and b) displeasing these very men. This seems to make them much more cautious about what they will say about a result, and though they probably are right to be cautious, it won't gain them any respect... unless of course they're also screaming bloody murder about why the man down the hall should be fired for fraud and giving damned good evidence... Women who won't exaggerate and won't bust balls end up relegated to pretty, mindless lab-techs whose ideas are either routinely ignored or subsumed by a PI... I think what women tend to lack, and I'm not arguing nature-nurture here-just a pattern, is confidence in their own ideas and abilities... Then again, when even the rationalists are telling you to go save puppies and raise babies, it's damn hard to live in a reality apart from the one in everyone else's mindspace...

ZM Davis- I don't know your situation, but I hope you are comfortable and happy and write that book soon! =)

There's been a lot of discussion about dividing people's skills and abilities along gender lines on this blog... I'm actually much more interested in knowing what people's personalities tell us about their expected performance. Does anyone here have a strong preference for a personality classification system that seems useful? Meyers-Briggs gets quoted a lot, but I personally find it to have no predictive value and to often be at odds with my own intuitions about people, which have thus far served me well. It also seems too easy for people to shift among the different Meyers-Briggs classifications depending on their mood that day, which indicates it's not a stable predictor of anything. Perhaps there is a better personality test used by corporations? What are some obvious indicators for classifying people that we meet but cannot test?

Big Five is pretty much the standard I think.

Competitive types, found in abundance among the elite intellectuals and professionals, may tend to commoditize achievements and relationships. This is not a value judgment, but rather a statement regarding a kind of perspective. Are competitive types drawn to elite academic pursuits and the professions, or are professionals and elites required to operate competitively in order to survive and prosper?

Owing to the efforts of such people in the world, thankfully, rather lower functioning individuals (among whom I should count myself one) get by. Who, for instance, would feel comfortable knowing that their aircraft mechanic takes a Zen or laissez-faire approach to maintenance and repair? I should prefer to patronize mechanics, phyisicians, accountants, engineers and others who are borderline obsessessive compulsive in their professional activities.

But what may be adaptive traits in an academic or professional setting may also be maladaptive in a personal setting. There seems to be something profoundly unphilosophical about speaking of acquiring or quantitatively or qualitatively superior car, library, house, etc. If happiness, whatever that means, resides in the pursuit or acquisition of exterior achievements or objects--or the fickle evaluative judgments of others--then happiness will always remain contingent, prone to continual degradation and mostly beyond reach. Socrates wrote that the unexamined life is not worth living. But what of the over-examined life?

Thanks, Steven, for the reference! I'm surprised I've never seen the big five before, since they seem much more useful than Meyers-Brigs.

Alan- for a data point of one, I wasn't always as competitive as I am now. I actually used to (and still do)to a large degree find competition disgusting- I just realized what was necessary. When I was a kid, I used to deliberately lose every-other chess game so the other children would like me... I win one, you win one. As you might guess, I was very low on the social ladder back then.

With the coming Olympic Games, I think an excellent post idea would examine the overrated areas of sports and patriotism

For men, sports is effectively a drug, allowing users hits of exhilarating testosterone given victory. Furthermore, fan behavior is dependent on the expected outcome of the match, making upset losses even more violent than usual.

In one related OB post, Patri Friedman expresses his sports disinterest in passing, noting sports' zero-sum nature.

As for patriotism, it shares the in-group/out-group biases with sports fandom, where the in-group selection is by and large luck of where you were born. The Olympics in total represents a series of zero-sum games performed for the glorification of national status.

I am an INTP on the Myers-Briggs test. I feel strongly that this is a meaningful and true statement. I feel that classifying me as any other type would be incorrect. I have read enough about both INTPs and other types to feel fairly confident that this is not confirmation bias or the Barnum effect, but a legitimate predictive claim. My classification as an INTP comes from both personal introspection and multiple formal tests (albeit internet tests rather than professionally-given tests.)

I also find that the Myers-Briggs is _useful_. I am pretty good at guessing what type people will be, and I find that when I keep their personality type in mind I am able to better understand and appreciate their differences from me. I feel a strong sense of _explanation_ when I recognize previously unexplained behavior as manifestations of someone's type.

I can certainly behave in ways that contradict my type, but I find that I do so in ways that could be predicted from my test results. I am not a particularly strong I, so I will sometimes act more like an ENTP. But not as often as I act like an INTP. I am a really strong N, so I almost never act like an ISTP. And I certainly never act like an ESFJ.

However, I also recognize that I have absolutely no idea what the different scales on the Myers-Briggs actually refer to. The Jungian psychology behind the test seems very odd, and almost mystical. I also recognize that empirical data are very mixed when it comes to the test.

So am I simply being fooled by the unreliability of anecdotal evidence? Or does the Myers-Briggs really track important correlations, even if the mechanisms by which it works are unexplained?

This seems to get at some important questions with social science in general. What does it mean for a social scientific theory to "work"? How can you measure predictive power in anything but highly controlled conditions, and are those conditions anything like the real world? Does the fact that the theoretical justifications for Jungian Typology are almost certainly wrong effect how we should treat the test itself? What would it mean for a psychological theory to be "right"?

Peter, I prefer to think of Myers-Briggs as an elegant formalization of folk psychology. You could get similar predictive results by looking up words like "nerd" in the dictionary; but with Myers-Briggs, you also get a pretty- looking formalism with "equations" such as Ti + Ne = INTP.

>Because at that point there is no information remaining about any earlier time. If anything happened before it reached minimal entropy density, all the evidence has been erased.

JThomas, that only pushes the question back to another level: why is information/time correlated with entropy density? Listen carefully folks, because right here in this post I'm going to spell out exactly what is wrong with Bayesianism.

*To recap: Occam's razor is an all-pervasive principle of science required to narrow the theories under consideration down to a manageable number of possibilities, based on the principle that 'the simple is favored over the complex'. What needs to be explained:

(a) Give me an exact technical definition of Occam's razor?
(b) Exactly why does Occam's razor work?

I suspect that even the Bostrom's, Hanson's and Yudkowsky's will find it very very difficult to answer (a) and (b). And as long as (a) and (b) aren't answered, there's a huge gaping hole in epistemology.

The problem with Bayesianism is simple in retrospect: It only deals with finding externally observable patterns. It does not consider the *semantics* (meaning, representations) of these patterns -that is to say: it does not explain how these patterns represent *knowledge* - and knowledge is *not* the same thing as Shannon information!

Occam's razor is not just concerned with finding patterns, it's *also* concerned with the actual *meaning* or *semantics* of the theories under consideration. And that's what puts it beyond the scope of Bayes.

Does anyone have more advice along the lines of What You Can't Say for dealing with beliefs impopular enough to carry severe social penalties? If you had, say, ten, then would you just pick the most important one to advocate and carry the rest into the grave lest they all besmirch each other by their presence?

Has there been any discussion of 'rationality' under the influence of alcohol?
Same question for caffeine.

Also, second the request for weekly open threads.

Prediction markets seem to be an extraordinary good tool for political parties to avoid selecting candidates with hidden skeletons likely to fall out of their closet (the John Edwards problem).

Robin, you might find this interesting:


I wonder if religion can be considered a cheaper way to signal that you care than health care?

JThomas, that only pushes the question back to another level: why is information/time correlated with entropy density?

My answer wasn't about that. I'm saying that if you have a real honest-to-goodness big bang then you can't get information about anything before that because any such information has been destroyed.

If it turns out that I'm wrong about that and you can get information passing through a big bang about times before, then that is evidence that the big bang was not the beginning.

But without any information about times before, there's no evidence that there was a time before. So you can say that the last big bang was the very beginning and nobody can prove you wrong.

JThomas, that only pushes the question back to another level: why is information/time correlated with entropy density?

But here's an answer to your new question. Suppose that you accidentally latch onto something that looks mostly organized. Like a whole lot of molecules traveling in mostly the same direction at mostly the same speed.

It's rare that things go from a state that looks more disorganised to a state that looks less disorganised because there are so many ways that look disorganised to us and so few ways that look organised. But it must happen sometimes that things start to look organised, even by random chance, and we pick one of those times.

Then if you keep looking at it, the appearance of organisation will start to fail. The molecules will have "random" molecular motion that you can't predict, and it will begin to have results you can see. The slow ones and the fast ones will start to separate. The things moving in slightly different directions will spread out. Eventually the appearance of order will go away, and it might be a very long time before you see something else that looks orderly.

That's all there is to it. The illusion of order falls apart over time, and it doesn't come back except as rare events, so it looks like time is correlated with information density. Because the information we can pay attention to consists of very unlikely states that change to things we find harder to track.

Wow, Obama is killing on the prediction markets:



I feel so retarded. I didn't know that stuff like this existed. Spent too much time restoring male sexual function. Overall, would you guess that technically savvy people are more likely to vote for Obama, and if so, would the fact that they knew about prediction markets explain the huge margin? If techies are not more likely to vote for Obama, then the actual margin is much higher, and makes a bet on him seem like a sure thing.

FYI, I've gone through and deleted all or part of many Caledonian comments in which he explicitly, or by Gricean implication, had myself or others holding opinions we do not hold. I'm going to consider this deliberate trolling from now on, and since Richard Hollerith fell for it due to having not read one post, I'm assuming that this is in fact damaging new readers who may come in not knowing the actual positions that Caledonian misrepresents.

Things that sound sane enough to be someone else's comment will make it through. Classic Caledonianisms will get deleted. We'll see how much that leaves. Caledonian, feel free to start your own blog and link there, or post the original versions of your comments to ni.codem.us.

If you post a reply to Caledonian's comments and the original Caledonian comment is deleted, the moral is that you shouldn't respond to trolls.

Wow, look how low Biden was until the last minute. Some people must have made killings when his value went up and order of magnitude or two.

I see the Obama/McCain gap is holding steady at 60/40, despite a few news cycles of narrative that Obama is in "serious trouble" and that this race has "tightened up".

Also, I predict there's easy money to be made by betting Romney will be the VP nominee for McCain (his price seems strangely low for such a strong Veep frontrunner, and will probably climb to 90+ once his selection is announced). I may put my money where my mouth is and put a large bet on Romney.


For the record: Caledonian is not a troll. He's one of the more thoughtful posters on this site.

I notice that when a post topic has a comment within the last 10, and thus is visible under "recent comments", knowledgeable readers more frequently notice that the topic is unresolved or has open questions requesting answers, and they frequently add a very informative comment, starting the process all over. However, if new posts, or even one hot one, are published, all evidence of the previous post discussion may be gone in as little as 15-30 minutes. For instance, today, David J. Balan obviously spent a good deal of time constructing a lengthy, and erudite, comment on "Good Medicine in Merry Olde England", and before it could be discussed, 2 new posts came out, comments were rapid, and there was nothing to indicate that his topic still required educated responses (at least, I, and I suspect he, would appreciate it if there were). Could someone who has followed this blog longer than I please comment on an acceptable remedy? I've seen posters put something like "hello" in a new comment just to get it back on the list, but that seems dorky (and probably not within the rules). Is that what is done?

Urologist, the problem has been discussed here, and I share your frustration.

Eliezer, would you rather I didn't provide Caledonian a forum, or do you think the audience is small enough that it's not a problem?

If Eliezer said yes, you'd start censoring ni.codem.us?

Obama an overcomingbias.com reader?


Nick, I think if you create your own forum, Caledonian is your own problem and it's up to you to decide what to do about him. It's a free country, etc. The direct answer to your question is that I have no problem with it.

I think smart people are spending way too little time thinking about how to create forums that allow for better discussion, relative to time spent actually discussing.

I am an MD/PhD with very strong interests in both molecular biology and medicine

There already exists plenty of knowledge relevant to protecting human health, but the process by which the knowledge flows from the knowledgeable to the patient is extremely inefficient. Consequently, a competent rationalist is better placed to make a contribution to human health as a clinician than as a researcher. It seems though that the clinician's job is more frustrating and stressful than the researcher's job now that insurers no longer defers to the judgement of clinicians the way that they did 20 years ago.

If you believe that the problem is getting the information to flow from the researchers to the patients, then a clinician could improve the lot of her patients, but that's not aiming high. If you want to make a difference, you have to affect other people's patients. Researchers have the advantage that they're supposed to affect a lot of people. But if problem is clinicians, a clinician has the advantage of understanding them, and maybe their trust.

But pure research probably isn't useful. I think I have a lower opinion of the quality of medical research than Richard, so I think there's room for a lot of improvement, but if the current system generates junk, why should a good researcher expect good research to win? It's a knowledge transmission problem either way.

The comments to this entry are closed.

Less Wrong (sister site)

May 2009

Sun Mon Tue Wed Thu Fri Sat
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30