Cosma Shalizi

See also, this classic review: "A Rare Blend of Monster Raving Egomania and Utter Batshit Insanity" http://bactra.org/reviews/wolfram/ (originally at http://www.cscs.umich.edu/~crshalizi/reviews/wolfram/)

A great post about the same problem but from a Math/CS perspective, and as fun to read: http://cscs.umich.edu/~crshalizi/weblog/918.html [Cosma Shalizi's "In Soviet Union, Optimization Problem Solves You]

Always cool to see articles trying to tease apart correlation from causation. That said, I am automatically suspicious of any publications mentioning power laws, since they make for attractive stories and can easily be fit to data which is not actually generated from a power law. See http://cscs.umich.edu/~crshalizi/weblog/491.html for an excellent discussion.

Whenever anyone mentions NKS I'm always reminded of Cosma Shalizi's brilliant and hilarious review, 'A Rare Blend of Monster Raving Egomania and Utter Batshit Insanity'. http://www.cscs.umich.edu/~crshalizi/reviews/wolfram/

The best review of New Kind of Science: http://www.cscs.umich.edu/~crshalizi/reviews/wolfram/

For an equally scathing but more contentful review, it's hard to beat Cosma Shalizi's at http://www.cscs.umich.edu/~crshalizi/reviews/wolfram/ -- I love the title: "A rare blend of monster raving egomania and utter batshit insanity".

That's an anecdote, not a study. The criticism I leveled is at twin studies in general, not twin studies . Heritability is very specific measure, and it does not mean "genetic". Zip codes are highly heritable; number of arms is not heritable at all. I refer you to a great article on the subject of what "heritable" really means in this context: http://www.cscs.umich.edu/~crshalizi/weblog/520.html

> You'd think that such a smart guy would have realized... This is the best theory I've seen yet to explain his behavior: http://groups.google.com/group/sci.math/msg/37f51351f5340e58?dmode=sou… It makes a lot of sense, in light of the man's biography. See this: http://www.cscs.umich.edu/~crshalizi/reviews/wolfram/ for an informed critic's perspective on the details.

The scientific debate is much more complex than you give it credit for. Yes, there really are people who believe that IQ, or (to be more sophisticated) g, really has validity. I personally view it as convenient shorthand, but not real. However this is not a universal position. For a take on why it is probably not valid, see http://cscs.umich.edu/~crshalizi/weblog/523.html. But whether or not you accept that IQ is meaningful, the following fact remains true. On a wide variety of abilities that we can measure, including most specific mental attributes (working memory, processing speed, etc), males have much higher variability than females. The average gender differences are small, but men are over-represented in both tails. That fact is true both when abilities are measured in isolation, or on a test that aggregates them, like a standard IQ test. See http://www.edge.org/q2008/q08_10.html#cronin for an example of where this is discussed.

This is a wonderful comment. However, what psychologists know as g is the result of over-interpretation of a descriptive method known as factor analysis (which is very similar to principal components analysis). http://cscs.umich.edu/~crshalizi/reviews/flynn-beyond/ Major thanks to Cosmo Shalizi who opened my eyes to these issues. Also, the SAT grew from IQ tests, so it really isn't surprising that the two measures are correlated, given that the questions for SAT were probably kept on the test because they correlated with IQ tests. That being said, the work of Kahneman (And the new paper which I should probably read instead of commenting on HN) is pretty rock solid as far as it goes. It is worth noting that there is a converse position in this field, that of Gerd Gigerenzer who argues that these heuristics exist because they are useful, and only go wrong in artificial situations. www.cogsci.msu.edu/DSS/2007-2008/Todd/environments_that_make_us_smart.pdf Personally, I incline far more towards the views of Gigerenzer than those of Kahneman, especially given that Gierenzer and colleagues attempt to model the mind given their theories computationally, which is something that psychology could do with more of. Full disclosure: I'm a psychologist who's very frustrated with the lack of statistical sophistication and interpretation in my field.

The quick answer is "no, it doesn't, nothing does": http://cscs.umich.edu/~crshalizi/weblog/491.html The better answer emerges if you think about the example a little more closely. The big power law brands driven by customer appreciation are in media -- people love Lady Gaga (say), so she sells gazillions. Nobody loves yellowtail or sutter home. Many people love Screaming Eagle or Chateau Lafite. However, wine is an agricultural product; you cannot produce more bottles of a great wine than you can grow in some very particular fields in a year. The way scary hit-driven industries work is that you sell things way above marginal cost to everyone who wants them; in wine, since there's actually scarcity, very few people who love a brand can get as much as they want of it, so the prices at the top are insane (recent vintage screaming eagle is indicated at $2-3k per bottle), but the brands are smaller. I'd guess the way to think of it is like the music industry before recording devices or electricity -- sure there are stars, but because supplies are limited nobody can by a megastar.

Many, many things aren't a power law, but are widely reported to be due to sloppy statistics and a deep misunderstanding of regression on log-transformed data. The problem is this: people frequently log-transform empirical counts of things (like library book circulation) and then fit a line to the distribution. Since the line generally fits very well, with very high (but ultimately meaningless) R-squared values, they rush off to write a paper. I could cite at least a dozen very "important" papers that gush over power law relationships in this or that. Instead, I'll let Cosma Shalizi's excellent blog post on the topic do the talking and citing: http://cscs.umich.edu/~crshalizi/weblog/491.html It's an excellent read for anyone reaching for a log-transform. EDIT: I didn't realize that Cosma was the author of the linked piece. The other blog post by him I link here is also highly recommended.

Reverse-engineering from the article, the original statistical result was probably something like, "Variations among managers in technical skill account for very little of the variations among managers in performance reviews." One way to get that result would be for technical skill to be unimportant. Another way would be for the variations in technical skill within the sample to be much smaller than other relevant variations. Given that (según the article) the managers were selected for technical skill, but not for the other qualities, it seems almost certain that this was the case. This might sound like a bug that could be avoided by better statistical analysis techniques, but it's really not. You really want to know what causes the actual variations in performance in your sample, not what could conceivably cause variations in performance. Surely there are hundreds of things that would cause even bigger variations than what's in your actual sample — you could have managers who kill all their employees, or managers who are interned incommunicado in Guantanamo, or managers who obsessively turn every project into an investigation of fluoride contamination of their precious bodily fluids, or managers who are actually undercover headhunters from a competitor trying to steal away all of your best talent. Cosma Shalizi has an excellent post about the pitfalls of multivariate statistical analysis at http://www.cscs.umich.edu/~crshalizi/weblog/520.html; although he's talking…

Here's Chomsky on the same issue: http://www.cscs.umich.edu/~crshalizi/chomsky-on-postmodernism.html A quote: "It's entirely possible that I'm simply missing something, or that I just lack the intellectual capacity to understand the profundities that have been unearthed in the past 20 years or so by Paris intellectuals and their followers. I'm perfectly open-minded about it, and have been for years, when similar charges have been made -- but without any answer to my questions. Again, they are simple and should be easy to answer, if there is an answer: if I'm missing something, then show me what it is, in terms I can understand. Of course, if it's all beyond my comprehension, which is possible, then I'm just a lost cause, and will be compelled to keep to things I do seem to be able to understand, and keep to association with the kinds of people who also seem to be interested in them and seem to understand them (which I'm perfectly happy to do, having no interest, now or ever, in the sectors of the intellectual culture that engage in these things, but apparently little else). Since no one has succeeded in showing me what I'm missing, we're left with the second option: I'm just incapable of understanding. I'm certainly willing to grant that it may be true, though I'm afraid I'll have to remain suspicious, for what seem good reasons. There are lots of things I don't understand -- say, the latest debates over whether neutrinos have mass or the way that Fermat's last theorem…

Wolfram is damn smart. I'll give him that. He also made a pile of money on Mathematica. I'll give him that as well. It was classic "solve your own problem" entrepreneurship. He needed a better computer algebra program, so he wrote one, then started selling it.[1] But he's a crank. He wrote a couple good papers in the 80s, then hasn't done anything decent since. He published A New Kind Of Science in 2002, which was supposed to change the world; except that half of it was nonsense and the other half was stolen from other people. (http://www.cscs.umich.edu/~crshalizi/reviews/wolfram/) Feynnman (kinda) calls him out on this: Complexity research doesn't need its own institute, because complexity research is a pseudoscience that means whatever you want it to mean, like how the homepathy quacks have latched on to the word "quantum". 1: Wolfram Alpha is the same thing. Wolfram wanted a natural language parser, so he wrote one. Except it doesn't work worth a damn, because natural language parsing is a hideously difficult problem, which probably requires strong AI.

This is probably not what you're looking for, but while trying to figure out if you could still download a free copy of Drake's out-of-print Fundamentals of Applied Probability Theory... [HINT: The answer is yes: http://ocw.mit.edu/OcwWeb/Electrical-Engineering-and-Computer-Science/… ...I found this potentially interesting page full of references: http://www.cscs.umich.edu/~crshalizi/notebooks/probability.html I post it here so that I can find it again later, but I fear that this is not the answer to your question. These appear to be more probability theory than nuts-and-bolts statistical analysis [1], and they may require more math background than I have (and I managed to handwave my way to a solid state physics Ph.D.) Try browsing around the MIT OpenCourseWare site. One of the great things about online courses, even if you don't take the courses, is that they often publish syllabuses that contain a lot of good textbook references. [EDIT: Ooh, look, if you follow links from the page I mentioned earlier you get here! http://www.cscs.umich.edu/~crshalizi/notebooks/teaching-statistics.htm… ] --- [1] Not that I can reliably tell the difference. I know just enough about probability and statistics to know that I should probably study them some more. ;)

I don't know Wolfram either, but I have read reviews of A New Kind of Science by experts who are competent to judge, and they all say that the book is full of old ideas that Wolfram claims sole credit for. The only really impressive bit in it was a proof that was done by one of his research assistants. Wolfram sued the guy to keep him from publishing until the book came out. He's also sued other academics who have so much as dared to sniff around what he considers his territory. So, yeah, he's clumsily arrogant in ways that are especially odious to his academic peers, and he doesn't have the sort of ground breaking accomplishments that might incline people to overlook such things. Is that a reason to hate the guy? No, I agree with you, don't be hatin'. It's unbecoming, and a waste of time. (Hecker's website is down, so I can't say if that's what he was doing.) If you really are defining yourself by who you hate, you are a pathetic individual who should be mocked yourself (but not hated). Wolfram is a brilliant, brilliant guy, and his company has written some pretty cool and impressive software, after all. What hast thou accomplished, O mocker? But I don't believe in giving arrogant people the benefit of the doubt just because we don't have a couple hundred years of perspective on their ideas either. The number of people who warrant that kind of treatment can be counted on the fingers of one hand, and they were almost always overlooked because they were out of the…

Two wonderful companions to this article are "Thinking Intelligence is Innate Makes You Stupid" http://lemire.me/blog/archives/2007/12/03/thinking-intelligence-is-inn… and (especially) Cosma Shalizi's article on the malleability and heritability of IQ http://cscs.umich.edu/~crshalizi/weblog/520.html

For a contrary opinion about the book see http://www.cscs.umich.edu/~crshalizi/reviews/wolfram/ (yes I've read the book and mostly agree with that review)

I don't actually know that much about Wolfram or A New Kind of Science, this was my first exposure to either: http://www.cscs.umich.edu/~crshalizi/reviews/wolfram/ Would people who know the book or the man say this is accurate? It's all I associate with the name now.

Anyone interested in NKS might also be interested in Cosma Shalizi's review which details several substantial problems with the book. http://www.cscs.umich.edu/~crshalizi/reviews/wolfram/

> most stuff measured about humans is distributed like a power law twitch http://cscs.umich.edu/~crshalizi/weblog/491.html http://cscs.umich.edu/~crshalizi/weblog/390.html

Here's a 2007 article on the same topic for those like me without the ability to get past the Science paywall: "So You Think You Have a Power Law — Well Isn't That Special?" http://cscs.umich.edu/~crshalizi/weblog/491.html

Dawkins is perfectly aware of this. In fact, he gives arguments of your flavor in The God Delusion. See also http://cscs.umich.edu/~crshalizi/Dawkins/viruses-of-the-mind.html He is aware that his approach is not optimal in terms of changing people's minds, but I'm not so sure he cares. http://www.youtube.com/watch?v=-_2xGIwQfik

I see there is a recent press release about that ridiculous and poorly evidenced claim. (The haphazard survival of fossils means the confidence interval around such an estimate must be quite large.) A more nuanced view of the relationship between brain size and brain capability can be found in a Scientific American article that quotes some of the leading researchers on the issue. http://www.scientificamerican.com/article.cfm?id=does-brain-size-matte… As another reply has already pointed out, the Flynn effect http://cscs.umich.edu/~crshalizi/reviews/flynn-beyond/ is an empirical observation that IQ scores in large, pan-national data sets that have been gathered by standardized methodologies have been rising all over the world for most of the last century.

Or maybe not: http://cscs.umich.edu/~crshalizi/weblog/491.html

But intellectual labor can become an industrial product. Not all products of intellectual effort are theorems and mathematical proofs. And please read this before citing Goedel's Incompleteness Theorem: http://cscs.umich.edu/~crshalizi/notabene/godels-theorem.html

And the companion review http://www.cscs.umich.edu/~crshalizi/reviews/wolfram/ kind of completes the cycle.

Nogwater nails it below. Specifically, I'm think of things like http://cscs.umich.edu/~crshalizi/weblog/491.html and http://cscs.umich.edu/~crshalizi/weblog/656.html

On that note, here's one well-written but scathing review by a UMichigan professor who goes into that among other things: http://www.cscs.umich.edu/~crshalizi/reviews/wolfram/