Byrne's Eye View

In a recent essay, Bruce Schneier comments on how we deal with security threats by preventing the last attack -- that after the Oklahoma City bombings, we worried about far-right white loners with access to fertilizer until about the day before twenty Middle Eastern men hijacked some aircraft. He pointed out how ridiculous this was given that we form our first impression of attackers during the attack itself, which is when our information is as bad as it's ever going to get.

For example, the Columbine killers weren't creepy, friendless goths; they had friends, and they behaved like typical adolescents. Of course, they spent an unhealthy amount of time playing video games and shooting guns -- but they did that with a group of friends who blatantly don't fit the moody psycho lone male killer stereotype. As far as I can tell, their gun fascination was typical right up to the moment that it included shooting fellow students.

So it's ironic that so far, Cho Seung-Hui fits the incorrect assessment of the Columbine shooters. He was lonely, violent, and weird, and the only reason he didn't stand out more was his shyness. At least Cho showed some blatantly worrisome obsessions with his awful writing -- but if this and poor interpersonal skills is enough evidence to consider someone a potential killer, we need to keep a close eye on Chuck Palahniuk and Brett Easton Ellis.

The best conclusion I can think of is that the people who are prone to going on shooting rampages or committing terrorist attacks are all outliers. They're bad data. We don't have enough information to usefully profile the people behind these attacks -- all we can hope to do is minimize the damage and accept that some forms of technology are deadly, and that they'll end up in the hands of murderers.

I just read Francis Fukuyama's Our Posthuman Future. I picked it up because I'd heard the author's name and because I'm a fan of posthumanism and singularitarianism and other associated philosophies (one of my atheist friends said he used to like the idea of the singularity until he realized how religious it sounded -- which, as far as I'm concerned, is one of its best features: Ray Kurzweil is an eschatological prophet for secularists). I had no idea Fukuyama was vigorously opposed to the whole notion and worried that genetic engineering would destroy humanity.

He certainly sets forth some powerful arguments -- that it would redefine humanity out of existence, that it would create a genetically-enhanced aristocracy, that we don't know the effects of messing with our genome until decades later -- but his analysis has just enough depth to miss the obvious refutations.

For example, Fukuyama considers the possibility of fiddling with a child's genes to raise his IQ a standard deviation or two, for a fee. Some rich parents would unhesitatingly do this; many poorer parents wouldn't be able to afford it. Suddenly, intelligence would be (even more) heritable, and each rich generation would be even smarter and richer, creating a permanent aristocracy. All of this is true -- Fukuyama gracefully ignores the fact that aristocracy is a bad idea because it's arbitrary. If I can be pretty sure that Joneses are, on average, 50 IQ points smarter than Smiths, there's no problem with it -- it just makes it slightly easier to identify smart people.

Fukuyama's solutions are, in fact, even worse than his analyses. His theory of law is that we create some behaviors that are Permitted and some that are Forbidden, based on what we, as a society, believe ought to be universally allowed or unilaterally banned. That's not how laws work -- laws impose extra constraints, and they change behavior, but they're created, enforced, and obeyed (or broken) by humans, so absolute success is absolutely impossible. We as a society have deemed tobacco and marijuana harmful. We express this by charging a financial tax on tobacco use, and occasionally assessing a temporal tax (i.e. a prison sentence) on marijuana use. This doesn't get rid of either practice -- it just changes the incentives (confining tobacco smoking mostly to the already addicted, and putting marijuana distribution in the hands of people who are already criminals and thus have the least to lose). Western governments could probably enforce restrictions on human cloning and some edgier forms of genetic research. If we do, progress will still happen and discoveries will still be made -- and to take advantage of them, you'll have to be friends with Vladimir Putin or Hu Jintao.

Fukuyama's strongest criticism of biotechnology is that we may end up changing humanity to the point that it's no longer recognizably human. Once again, he makes an obvious mistake -- the humans of 2107 will be unrecognizable to us. But consider that in 2007, I'm typing this essay into a five-pound object that contains ten days of music, several hundred thousand words of text, and has the capacity to instantly transmit this essay to an audience anywhere in the world. Would this be comprehensible to a farmer in 1907? Would it even make much sense to an executive in 1957? You have to move up another twenty-five years or so before you can even imagine people who would be informed enough to realize that this is cool, rather than just bizarre.

It's easy to scare yourself by imagining your self of today confronting the world a few decades from now. But Fukuyama is 55 years old -- by 2057, he'll have spent nearly half of his life in what we today can consider the future. If he's having trouble accepting the next few decades all at once, he might do what the rest of us do and take the future a day at a time.

I just finished Nassim Nicholas Taleb's Fooled by Randomness, which is one of those books that lets you trade a little happiness for a lot of knowledge. His thesis is that we overestimate the role of skill and seriously underestimate the presence of luck, because we're wired to see causality and consistency when there's only randomness.

That's a good thing to keep in mind, especially in the parts of the book that focus on short-term news and short-term price swings rather than long-term trends. The other day, I bought a stock that, for no particular reason, jumped 6% two days after I purchased it. I'd like to imagine that this move confirmed my investment thesis and demonstrated my skills, but a 6% move on essentially no news is probably just a random fluctuation that could have just as easily gone the other way.

The book has a few weaknesses, though. Taleb insists on treating investment as a competition and a psychological exercise, and while it can be both of these things, I treat it as simply the act of exchanging a sum of money for something of slightly greater value, with the hope of realizing that value. From that perspective, the question isn't "Who is the best?" or "How do I measure my skill compared to my luck?" It's "What can I do to get the highest possible return -- and which actions are likely to increase my returns?"

Taleb and I would agree that the person with the highest five-year return over the last five years is not necessarily the best investor -- and is probably the most leveraged and thus the most likely to eventually blow up. But if he and I both considered an investor with slightly above-average returns (15% per year, say) who, before buying stocks, thoroughly read their financial statements and only purchased stocks he felt would grow faster than the market but which traded at a lower price/earnings ratio, Taleb might attribute the extra return to luck -- I'd attribute it to effort. I'd hope that if we make it more explicit ("There is value in knowing what you're buying" versus "Learning about the stocks you're buying is a waste of effort,") he'd agree -- but you wouldn't know it from the book.

One last flaw: he attributes Microsoft's success to path dependence -- once you've picked a software package, the cost of switching to another is high enough that you may stick with an inferior product despite knowing about better opportunities. This is ironic, inasmuch as Microsoft's biggest innovation was that you could mass-produce software and sell it for multiple systems, instead of crafting it bespoke. This brought prices down by a few orders of magnitude, and played a major role in making computers ubiquitous. Meanwhile, they offered their office applications for several operating systems (DOS, Windows, Apple, OS/2), and sold operating systems with several paradigms at once -- graphical user interfaces (Windows), a proprietary command-line interface (DOS), and even Unix (Xenix).

This played a major role in their success. Before Microsoft, buying a computer meant buying into a single paradigm and selling your right to switch cheaply. After Microsoft, buying a computer probably meant being able to use native formats and programs -- or switching into the nearly ubiquitous Microsoft options.

And then they stopped. Instead of breaking down path dependence, Microsoft started fencing off their path and locking the gates behind them -- they sold their Unix product, they dropped Internet Explorer support for Macs, and they kept tight control over Outlook so the cost of communicating with a Microsoft user was, approximately, the cost of becoming a Microsoft user yourself.

And then Sun Microsystems started supporting the OpenOffice.org project, which creates a Microsoft Office clone that runs just about anywhere. And Novell created Evolution, which does about the same thing for Outlook. Both companies saw an opportunity to break down Microsoft path dependence -- they followed the classic entrepreneurial equation (10% of a market that's 90% smaller than it used to be is a lot better than 0% of the market that existed before). Path dependence explains some of Microsoft's success, but a lot more of it was thanks to getting mediocre software done before the best software was ready -- and making some slick deals with IBM, too.

(Don't even start with Taleb's other example.)

I'd be surprised if a book that tried so hard to be iconoclastic didn't contain at least a few errors. Taleb wrote it because people see signal where there's only noise -- which makes it a bit ironic that his faith in the ubiquity of noise made him blind to some pretty blatant signals.

Speculative trading is based on the 'greater fool' theory -- no matter how ridiculous it is to overpay for a stock, the next buyer will offer an even more ridiculous price. It's the height of arrogance to assume that you won't be greatest fool, especially when the person you hope to sell to is almost certainly employing the same strategy.

But I wonder if traditional investing is any less arrogant. After all, value investors are simply hoping to be the lesser fool: since the scope of our ignorance in the market is almost unlimited, the best edge we can hope for is to be a little less wrong. But even if we're certain about the stock, how can we be sure about our certainty? A speculator assumes that someone out there has a little more money and a little less sense; a value investor assumes that nearly everyone had less sense or didn't have the capital to make prices reflect their logic.

There's an obvious difference, though: value investing works as long as people have been wrong in the past, since the only way for a price to be irrational now is for it to have been mis-traded before. Meanwhile, a speculator doesn't necessarily care about how smart investors were in the past, as long as they're less so in the future. So an investment strategy is more than just a money-making algorithm -- it's a statement about whether or not we're getting more reasonable.

In a world of easily-accessible knowledge through Google and reasonably accesible expertise through an increasingly open higher-education system, I feel safe shorting ignorance.

All hail economic populism: Yukos, a company built through rigged auctions of state assets, is being sold. To the state. In a rigged auction.

Incidentally, the Russian media are hilariously subservient to the Kremlin:

For the information of Condoleezza Rice, who despite being Secretary of State of her country, continues to demonstrate an ignorance of world affairs at a shockingly consistent level, the notion that the Kremlin exerts a grip on the media is a fairy tale invented in the gardens of Washington. As correspondent of the English version of Pravda.Ru, director and chief editor of the Portuguese version and collaborator for three other Russian publications, two of these being official media organs, I have frequently asked for guidelines from the Kremlin on what line to follow.

The answer: "We are afraid we cannot give you any guidelines as you request. You will have to write the truth, after checking your sources, obviously", or words to this effect every time the question is posed.

As journalists we have a collective responsibility to present the facts for our readers to decide where the truth lies. Presenting the facts can sometimes hit obstacles, such as in the reporting of terrorist attacks and it is here that the international press has decided to take up the story, twist it round, exaggerate it and regurgitate it for gullible and unthinking readers to swallow hook, line and sinker.

Certainly, the Kremlin imposed some limitations on reporting in certain circumstances, these being when the information pertains to secrets of state, not to impose limitations on freedom of the press, but to protect state interests, to name one example, not giving terrorists secret information.

I can't imagine an American media outlet bragging that every time they ask The Party for The Party Line, they get rebuffed.

Pleasant: a lawyer who made huge sums working for penny-stock fraudsters now works full-time spotting penny-stock frauds.

he searches the Internet for clusters of seemingly unrelated companies that use the same obscure accountants, lawyers and underwriters, and share the same mysterious offshore investors.

Sounds worth automating. If only the SEC's search function wasn't about ten years behind the most mediocre search engines.

Cognitive Daily suggests peer review for blogs, which is a funny thing to suggest for a medium that is 1) radically decentralized, to the point that even if you have a good policy you won't be able to make anyone adhere to it, and 2) radically decentralized, so critics have as much authority as whoever they're criticizing.

See? That's all the peer review blogs need.