|
|
|
|
|
|
|
|
|
Posted Wednesday, October 30, 2002 On-demand Computing?Here is IBM throwing away (euphemism) $10 billion to create something called on-demand computing.
Somebody explain to IBM the difference between the behavior of the cost of electricity and the behavior of the cost of computing. (Hint: it's a law that starts with the letter M) Time-sharing was a bust thirty years ago, and the economics have only gotten worse. If this concept makes a profit, then I'll retire as a pundit. I think it should be called IBM's no-demand computing initiative. . . . . . .
The Software Choke PointBrad DeLong mentions The Mythical Man-Month, which is a modern illustration of the law of diminishing returns. He writes,
I think that if computers are ever to get intelligent at the rate that people like Ray Kurzweil project, software is going to have to become self-writing and self-adapting. An important milestone will be getting to the point where I can buy a new computer and have it able to adapt all of my old data and applications automatically.
. . . . . .
Posted Tuesday, October 29, 2002 Krugman, Lewis, Kling, DeLong Paul Krugman's New York Times Magazine article last week argued that the top one percent's high incomes are a social problem, and Brad DeLong agrees with him.
In an essay that came out today, I disagree with Krugman and DeLong on this. If you're interested, go read it now. Then come back. OK? Well, the next week Michael Lewis' New York Times Magazine article argued that during the Internet Bubble it was the investing public who egged on the stock analysts rather than the other way around, and Brad DeLong disagrees with him. Once again, I take exception to DeLong's position. I followed the Internet Bubble very closely, from the very beginning. In fact, the Internet Archive still has a copy of the article I wrote just before Netscape's initial public offering, which originally was to be priced at $15 a share.
In the event, the IPO went off at $55 a share. At the time, there were no Internet analysts at all. I firmly believe that if you were to re-run history and have the stock analysts pooh-pooh Internet stocks throught the 1990s, investors would simply have laughed off the stock analysts. After all, the investors did not listen to Warren Buffett, or to me, or to many others who ridiculed Internet stocks. If anything, had the stock analysts criticized Internet stocks, this would have enhanced the feeling of invincibility that the Internet investors felt during the boom. The dominant phrase of the era was "They just don't get it." Investors would have happily told any Wall Street analyst who dared to question Internet valuations that "You just don't get it." In a ruthless, Darwinian fashion, the Bubble weeded out the analysts who were skeptical of the Internet and selected those who were bullish. As someone who saw Internet stocks as a bubble from day one, I believe that Lewis is absolutely right. . . . . . .
Posted Sunday, October 27, 2002 A Proposal for a Telco BailoutInteresting news from Europe.
Maybe someone would like to send the Europeans a copy of the "Fail Fast" letter. If the U.S. lets its telecom dinosaurs fail and the Europeans bail out theirs, my prediction is that this will be a huge advantage for us. . . . . . .
Read the Whole ThingEvery time I try to extract a quote from Michael Lewis' defense of the Internet Bubble, an even better one comes up in the subsequent paragraph. To pick a couple lines at random,
Think of computer and communications technology as a runner that has stumbled and is on the ground. Some of us think that it has the potential to pick itself up, dust itself off, and get moving again. It may happen sooner than you think. . . . . . .
Posted Saturday, October 26, 2002 Homeland Security Under Moore's LawRobert Steele's On Intelligence is a book about problems that need to be fixed in our approach to intelligence and security. One of the issues that Steele raises is the lack of integration of databases. From that perspective, the Washington Post buried the lede in this story. At the very end of the story, we find that
Steele would see this example as a microcosm of one of our Homeland Security weaknesses. Sooner or later, the government is going to figure out how to use Moore's Law as effectively as the credit card companies and other private sector data collectors. When that happens, I predict we will find ourselves confronted by a surplus of other laws that we do not want to be fully enforced. . . . . . .
Posted Friday, October 25, 2002 Treat the Bells as a Utility?Zimran Ahmed indicates what was wrong with the vision of the 1996 Telecommunications Act in this succinct statement.
The idea is that an entrant is supposed to be able to use the Bell's wires to provide services. This makes the entrant a free rider on the Bell's capital investment. Unless the central planners set a price for renting the wires that is exactly high enough to provide a rate of return to the Bell and exactly low enough to encourage usage by the other carrier. This piece by Duane D. Freese does not make me any fonder of the vision of regulated competition, even though that is what he is advocating. There are just too many minutia that the regulators have to worry about to get things right. . . . . . .
If I Were the Police, Con'tMy first reaction to the news that the sniper was caught was to retract everything I had written critical of the police. On reflection, however, I would not be so quick to pat them on the back. When I wrote
In my mind, I considered, but then dismissed, the possibility that he slept in a car. Surely, if he slept in a car, the police would notice. And they did! Muhammad was caught sleeping in his car in Baltimore on October 8, and the police let him go. Apparently, when the police are fixated on a white van, you can sleep in a different car in the middle of a sniper spree without arousing suspicion. . . . . . .
Posted Wednesday, October 23, 2002 For Moore's Law FansIntel is going to keep Moore's Law going, at least according to this story.
Actually, five times the speed in seven years is a little slow for Moore's Law, which would mean doubling every 18 months, but still... As an economist, I think about the implications of this as follows. Think of the economy as having two sectors--noncomputers and computers. The noncomputer sector grows at 1 percent per year, and the computer sector grows at 20 percent per year (because of quality improvement). Suppose that in 1987, fifteen years ago, the noncomputer sector was 99.5 percent of the economy, and the computer sector was only 0.5 percent of the economy. Then the average growth rate would be (.995 times 1) plus (.005 times 20), or about 1.1 percent. By this year, however, computers would be 6.25 percent of the economy, so that the overall growth rate would be (.9375 times 1) plus (.0625 times 20), or about 2.2 percent. But in another ten years, computers will be 27 percent of the economy. If computers are still improving at a rate of 20 percent per year at that point, overall growth will be (.73 times 1) plus (.27 times 20), or 6.4 percent! Ray Kurzweil is one of the few people willing to take Moore's Law seriously as a long-term proposition. If he's right, then according to my arithmetic the economic boom that lies ahead will be stunning.
. . . . . .
If I were the police, con'tI wouldn't beg the killer to call back so that we can negotiate. I'd say, "Call us when you're ready to turn yourself in. But if you mouth off again, or if you make demands again, we'll just hang up on you again." No, I would not expect him to call. And no, I'm not worried about what he might do if he gets angry. I'm worried about our community failing to mobilize its anger. (If you're not following the sniper story and want to know what I'm referring to, see Jim Henley's blog)
. . . . . .
Posted Tuesday, October 22, 2002 Yardeni is Back!Edward Yardeni is back on the web! (Maybe he never left, but I had a hard time finding him when he switched back to Prudential.) Yardeni is a big believer in the importance of information technology in the economy, and he also believes that the Internet and globalization are good for productivity and consumers but not for profits. When it comes to making predictions about the economy or the market, Yardeni likes to straddle--putting one foot on either side of the issue. This is a typical example of Yardeni-speak.
Yardeni seems to be staking out a position that it's a good time to buy stocks. However, he always gives himself an out, and in this case the "out" is a deflation scenario. . . . . . .
If I were the Police...I suppose I should be like everyone else I know, and when the question comes up say, "The police are doing the best they can" or "It's like trying to find a needle in a haystack." Or something along those lines. But instead I'm going to give my honest opinion about the sniper investigation. The police have put a lot of resources into searching for a weapon, a car, and "evidence" (hundreds of police doing "grid searches" for the latter). I would put those resources into searching for the killer. I assume that he goes to sleep at night. When he does so, it's somewhere in the area, probably within a 3-mile radius of this morning's shooting. It's probably an apartment. That doesn't narrow it down much--there are thousands of apartments around. But many of them house families, seniors, women, and others who can be ruled out as likely suspects. I would try to do a "triage" on the apartment residents--divide them into non-suspects, strong suspects, and in-betweens. Interview the strong suspects, and try to whittle down the in-betweens. Start with lists of tenants, which you could get from management companies. My guess is that you could use credit bureau information to do an initial triage. You also could look up information about credit charges. My guess is that not many tenants in Aspen Hill have charges in the Spottsylvania area--that might generate an interesting list. You could run the names and addresses against the database of social security and income tax collections. Someone who has worked in Aspen Hill and in the Spottsylvania area would be interesting. We need to turn the killer's passion for anonymity into a liability. We need to turn the concerns of the citizens into an asset. If I were the police, I would spend some time getting to know the residents of the apartment buildings. I would give a free barbecue with hamburgers and hotdogs at every building. I would invite several tenants to the barbecue and strongly encourage them to tell their friends. The publicity would spread by word of mouth. The residents who don't get the word might be an interesting list. At the barbecues, I would talk to the residents about how well they know all of their neighbors. Neighbors who are well known and reasonably outgoing can be scratched off the list of suspects. If there is a neighbor that nobody knows very well, that is interesting. This process should be focused *only* on catching the sniper. During one of the futile police dragnets, I heard a reporter on the radio say that "The good news is that police are catching all sorts of other violators." That is *not* good news. The police are supposed to be trying to catch the freaking sniper, not haul people in for having missed the deadline for their emissions inspection or whatever. Any police chief who encourages that kind of behavior on the part of his or her officers ought to be fired. . . . . . .
"Fail Fast"This letter to the FCC speaks for itself.
Read the whole thing, including the list of signatories. Although not many economists are included, I think it's fair to say that a lot of us are on board. Larry White, the economist praised in the letter, took a similar "let them fail" view of the Savings and Loan crisis fifteen years ago, and in retrospect it looks like we came out of that pretty well. . . . . . .
Posted Monday, October 21, 2002 Too Much Law Enforcement, Con'tSometimes, I think that everything I have to say in this blog was already written by David D. Friedman (son of Milton Friedman.), particularly in this draft. For example, on privacy, he writes,
This would seem like a strange view of the world. However, I know that in the past I have argued on the one hand that regulation can be unenforceable in cyberspace and on the other hand that laws might become too enforceable in realspace. Which echoes what Friedman is saying. But do I really believe it? Thinking about this now, I don't feel comfortable arguing for such a strong, technology-driven dichotomy between cyberspace and real space. It seems too much like what economists call a "corner" solution (what civilians call "all or nothing"). What feels better is to say that the ratio of cost to benefit will be higher for law enforcement in cyberspace than in real space. (Think of the benefits and costs of enforcing the law against song-swapping on the Internet compared with the benefits and costs of enforcing the law against stealing CD's from a record store.) Therefore, there will be relatively less law enforcement in cyberspace. . . . . . .
Too Much Law Enforcement, Once AgainWhat happens when Moore's Law makes it easier to enforce everyday law? I think it poses a serious problem.
It's one of my more provocative essays, and I think that the issue is timely. There is a place on the essay to post feedback, and I'd like to hear what you think. . . . . . .
Posted Saturday, October 19, 2002 The True BelieversSteven Den Beste questions the future of the Macintosh. The quick version:
However, according to Eric Hoffer, Mac users do not care about speed, price, or even availability of third-party software. They have sacrificed their lives to a cause. In his book about Apple, Hoffer writes,
Concerning the loathing that Macintosh users have for Microsoft and PC's, Hoffer says,
About Steve Jobs, Hoffer writes
(Eric Hoffer, The True Believer. All page numbers refer to the Perennial Classics edition.) . . . . . .
Wireless Last MileI was not aware of the concept of "underlay" until I read it in this paper by Kevin Werbach.
As we all know, the current regulatory environment of licensing spectrum to specific users for specific purposes is outmoded. Werbach tends to prefer the "commons" approach, while I am sympathetic to an alternative approach that allows spectrum owners to re-dedicate their spectrum to other uses, including shared spectrum. . . . . . .
Posted Friday, October 18, 2002 I'm slightly idiotarianEric Raymond is developing a manifesto, which is a work in progress (I find the latest version more palatable than earlier versions), on the subject of fighting terrorism. But he says,
The tone of the manifesto is exemplified by this sentence.
Raymond is like a guy in a barroom brawl who is being held back and snarls "Let me at 'em!" I do not want to hold him back. However, I do believe that we need to think through the issue of surveillance, not as a cure for terrorism or as an alternative to aggressive action, but as an inevitable consequence of technology. The costs of digital cameras , digital radios, and database technology are falling too quickly for this to be prevented. Right now, credit card companies know more about us than does the government. Similarly, surveillance is going to be accessible to some organizations. If the only organization that is unable to engage in surveillance is the government, our freedom will not long survive. We are going to need to work out a system of checks and balances. . . . . . .
Posted Wednesday, October 16, 2002 Too Much Law Enforcement, ContinuedDana Blankenhorn speculates that more widespread use of video cameras might have enabled us to catch the sniper sooner.
This is a good question, and there are many dimensions to it. One aspect that I'm trying to wrestle with is the fact that so much of our legal system seems to assume selective enforcement. Part of the reason that people don't want to apply technology is that they don't really want the laws to be enforced as written. That is why Congress always keeps the IRS computer systems and audit division under-funded, for example. Even leaving aside the privacy issue (as important is it is), better enforcement technology is not an unmixed blessing. We've gotten used to an equilibrium level of law enforcement, and better technology could disturb that equilibrium. . . . . . .
Felten's ChallengeEd Felten issued this provocative challenge:
Let me try. An analogy is with spoken language. The tendency is for a spoken language to be able to express any thought. Once a thought can be expressed in a language, trying to re-design the language so that the thought can no longer be expressed would be extremely difficult to do.
I'm guessing that the Washington types to whom Felten refers are people whose instinct is that movies, for example, can somehow be segregated and kept off of computers. However, once a movie can be represented in the computer's language of 0's and 1's, trying to redesign computers so that they cannot copy movies is like trying to redesign a language so that it can no longer express a particular thought. . . . . . .
Posted Tuesday, October 15, 2002 WWCD?Faced with the issue of song-swapping and copyright, what would Coase do? He might propose this idea, from Andrew Odlyzko.
Thanks to Kevin Werbach for the pointer. . . . . . .
Too much law enforcement?With this sniper thingy going on nearby--the biggest spree was in my neighborhood, near where I like to ride my bike--I've been thinking a lot about surveillance, law enforcement, etc. I've got enough musings for a really long essay. One issue is that as (or should I say, if) technology for detecting violations of laws gets better, we can be confronted with the problem that some laws or norms are better left unenforced or selectively enforced. For example, Doc Searls recommends this piece by Jon Udell on the topic of digital identity. At one point, Udell quotes David Weinberger as saying,
For example, police look the other way when people go a little over the speed limit. If you had computers doing the policing, they could not look the other way. What we really want is a "soft" speed limit. What Weinberger is saying is that there are many areas where we want the regulations to be soft. However, in a technology-laden environment, such as the Internet, this is difficult to implement. . . . . . .
Margin of SafetyBefore economists came up with the obscure jargon of "risk premium," Benjamin Graham talked about a "margin of safety." What is the margin of safety in the stock market today? I try to answer this question.
. . . . . .
Posted Monday, October 14, 2002 Broadband, Thin Case for SubsidyAdam Thierer sees government regulatory policy and subsidies for broadband deployment working at cross-purposes.
He has a cynical view of why legislators prefer subsidies to de-regulation as a way to promote broadband.
Finally, he says
He says that we will just have to be patient while the market sorts things out. I think that is wise counsel. I tend to get scared whenever anyone says that as a matter of public policy "We need broadband access now." We need lots of things now, but as a society we have limited resources. Anyone who thinks that broadband ought to get more of those resources might want to explain why telecom stock prices are telling a different story. . . . . . .
Posted Sunday, October 13, 2002 Secrecy and Intelligence FailuresHunter Thompson famously said that "When the going gets weird, the weird turn pro." In a world of the Internet and terrorism, someone like John Perry Barlow can be a professional in the field of intelligence.
A fascinating article, which I found by following a link from Robert David Steele. Perhaps the most important debate that this country can have right now is the debate over secrecy. In my opinion, we allow individuals too much anonymity (I am writing this in the midst of the DC sniper crisis). However, we also allow government agencies to shroud themselves in too much secrecy. Barlow emphasizes how dysfunctional the official secrecy dogma has become. I strongly suspect that law enforcement and security agencies use secrecy not to enhance their performance but to try to evade accountability--which surely weakens their performance.
. . . . . .
CDMA vs. GSMStewart Alsop argues that one of America's cellular technologies, CDMA, is superior to Europe's standard, GSM.
Of course, Steven Den Beste said the same thing three weeks ago. Advantage, Den Beste. . . . . . .
Posted Friday, October 11, 2002 Digital "Interference" OKedI am ignorant about the technical aspects of this FCC decision to allow digital broadcasting within existing AM and FM radio frequencies. But if one kind of digital "interference" is tolerable, why not other kinds? Maybe the FCC could allow what Kevin Werbach in a not-yet-published paper calls "underlays" that would allow something like WiFi to make use of AM and FM spectrum in a way that allows AM and FM radio to continue to function. Again, this is outside my area of technical expertise. For a cynical view of the politics of the decision, see Jesse Walker, who writes,
. . . . . .
Posted Wednesday, October 9, 2002 If Paper were a New InventionI've been reading Dana Blankenhorn's discussion of PDA's and e-book readers. He argues that potential customers are reluctant to buy because platforms are unstable. I think it's the other way around--the platforms keep changing because the manufacturers have not given the average consumer a compelling reason to buy. If all we had were e-books, paper books would be viewed as a marvelous invention. If all we had were PDA's, paper address books would be viewed as a marvelous invention (at least by those of us who have seen what happens when the backup battery fails on a handheld electronic device). I'm afraid that I am rather bearish on the e-book concept. I think that the application for e-reading is for much shorter pieces than books--maybe a 10-page essay at the most. And I don't see a market developing for short essays, unless you believe in micropayments, about which I'm also bearish. Jakob Nielsen says that clearer type on computer screens is a great productivity enhancer. I tend to think that clearer type is what consumers want, and e-books are not. . . . . . .
Posted Tuesday, October 8, 2002 Economics of the Wireless Last MileMy thoughts on this topic are assembled in this essay. Near the end, I say
Which sounds somewhat similar to the position of David Farber, as reported by Business Week.
Of course, this whole notion of letting the private sector allocate spectrum goes back to Ronald Coase, the Nobel Laureate whose work is familiar to FCC Chairman Michael Powell. Farber argues that the FCC could make better decisions if it had more independence from Congress. . . . . . .
Posted Monday, October 7, 2002 Is Spectrum Scarce?According to Zimran Ahmed, Steven Den Beste says so.
My guess is that David Reed would disagree. From a technical perspective, the argument is over my head. However, I'll bet that for practical purposes, the limitations that might be derived from Shannon will not matter. Even if water were free, people would not want to take a shower with infinite water pressure. I am guessing that even if spectrum were free, the demand for signals will not outstrip capacity. . . . . . .
Video on DemandRight now, the most economical way to deliver video on demand is via DVD's. As Zimran Ahmed points out,
So, even though the Net wins for music and text, right now DVD's rule for movies. My instinct is that somehow the Net will become the best distribution medium for video at some point, but we may be farther away from that point than I would have thought. Not good news for AOL TimeWarner's new strategy to turn the Internet into Cable TV. . . . . . .
If you think America's Telecom is Messed UpHere is Steven Den Beste's take on Europe's woes.
In explaining how it got to this point, Den Beste says that
In other words, an "industrial policy" approach to telecommunications, as advocated by John B. Judis, is not a good idea. Once again, Michael Powell comes across as wise for trusting decentralized markets. . . . . . .
BlogonomicsClay Shirky has caused quite a stir by invoking the the law of supply and demand.
Nick Denton begs to differ.
I agree with Shirky. Fundamentally, there are a lot of people willing to supply commentary at little or no cost. Thus, the equilibrium wage for commentators is going to be low. I made the same case about music creators.
Web logs are to traditional publishing what file-sharing is to the traditional music industry. The death of newspapers is as predictable as the death of record stores. As to Denton's view that "the stars will take a disproportionate share," that is a view that is tied to the economics of pre-Internet distribution. In the world of mass media, Britney Spears or Paul Krugman can achieve market shares and compensation relative to amateurs that far exceed the differences, if any, in talent and ability. As the Internet takes over, the huge concentration of rewards relative to abilities probably will disappear. . . . . . .
Posted Friday, October 4, 2002 Censorship and Congress: Man Bites Dog?Check out the first sentence of this story from Wired News.
The legislators think of authoritarian regimes as China or Saudi Arabia. Netizens think of the music industry...or Congress.
. . . . . .
Kiss AOL GoodbyeHere is a disturbing story about AOL's direction.
The vision of combining the Internet with television has been a total failure whenever it has been tried in the past, including many early attempts by Time-Warner. It is a vision driven by the corporate executive suite, that has absolutely no consumer momentum behind it. The irony is that AOL, prior to the merger, was a company driven by consumer research. AOL's executives did not attempt to impose a "vision" on people. Instead, the company gave people what its research showed that they wanted. The new regime at AOL apparently is determined to discard what succeeded in favor of a television-convergence model that is the most proven failure in the history of the Internet. You can kiss AOL goodbye. . . . . . .
Posted Thursday, October 3, 2002 VC-Funded ZombiesWhat happens when a company has enough funding to stay alive but no profits? In theory, the venture capitalist should shut the company down and try to put its cash elsewhere. But this apparently is not being done.
A fair guess is that most of these are zombie businesses, that will die once they run out of cash. I would think that the VC's are being remiss in not shutting more of these companies down and cutting their losses. . . . . . .
Posted Tuesday, October 1, 2002 Michael Powell Cites CoaseIn this New Yorker interview, Powell says,
Later on, he says,
My guess is that none of Powell's critics--certainly not John B. Judis--can cite the Coase Theorem correctly. And I would like to see what his critics believe he should have done on behalf of Worldcom or Qwest. (thanks to PremiumBlend for the pointer) . . . . . .
Copyright 2002-2003 Arnold Kling. All rights reserved. Terms of use |
|
|
