|
|
|
|
|
|
|
|
|
Posted Wednesday, April 30, 2003 IBM Through Intel's BinocularsAndrew Grove is interviewed on what is important:
I love the picture I get in my head of IBM strategists staring at the world through the wrong end of the binoculars. Thanks to Lawrence Lee for the pointer.
. . . . . .
Sucky Social Software, part 2Now there's this one.
Looks more like emergent idiotocracy to me. Winning a stock-picking contest requires aberrant behavior. The winner in this period is unlikely to beat the market subsequently. Instead, take the advice described here. I'm seeing a disturbing trend. Remember when "push technology" was the Next Big Thing (it made the cover of Wired and Business Week the same month), but all the applications sucked? That concept wound up being the Edsel of the early dotcom era. I think that if this thread makes it to "Sucky Social Software, part 6," I'll be ready to retract my favorable article and instead declare social software an empty fad. . . . . . .
The Real EntrepreneursChris Charuhas talks about the real life of an entrepeneur. For example,
In my book, I call this phenomenon "the early divorce." Also, real entrepreneurs don't come from McKinsey. The McKinsey consultant-turned-entrepreneur was one of the failed fads of the dotcom bubble. . . . . . .
Moore's Law and SurveillanceLawrence Lee points to this article, which says,
David Brin's The Transparent Society was written two years prior to 9-11. He saw technology trends as inevitably enabling more surveillance. His book influenced my thinking a great deal. "You can't encrypt your face" would have been a good subtitle for Brin's book. . . . . . .
The Case for Text AdsJacob Nielsen argues that text ads are only suited to search engine sites, because
I strongly disagree with this analysis. I think that the behavior on both search and non-search sites tends to be "leave and come back." That is, you click on a link, but either because you right-clicked in the first place or you use your back button or history button, you return to the site that sent you, in case the original site has some more links that are interesting. Furthermore, as a non-search site, I have the option of placing text ads at points where it is appropriate to leave the site. We did this all the time at homefair.com. We would have a mortgage calculator where the user entered inputs, clicked a submit button, and what came back was a page with results. On the results page would be a text ad from a mortgage lender. When the goal is click-through to advertisers, text ads rule. In that regard, the distinction between search and non-search sites is much less significant than Nielsen suggests. . . . . . .
Posted Tuesday, April 29, 2003 Sucky Social SoftwareI don't know why Brad DeLong links to this or this. In my view, both of these efforts give social software a bad name. Creating a "tool" that lets you form a random group is not solving a real problem. My challenge to would-be developers of social software: articulate the problem that your software will solve. "We need a tool to..." is a lousy way to start a problem statement. The problem should be a problem even if your tool does not exist. If you can only describe the problem by referring to your tool, then you don't get it. . . . . . .
Posted Monday, April 28, 2003 Not Ready for Prime TimeThis was already slashdotted, so you may have read it already. (I got it from Howard Rheingold.)
The article goes on to describe several initiatives driven by surveillance and security concerns. It makes the progress in technology sound much too rapid for my taste. It's not that I'm against surveillance technology per se. It's our social and political institutions that are not ready for prime time. I believe that we need a new Constitution that strengthens the checks and balances against excessive government power. Even if our existing Constitution will suffice, I would like to see an awful lot more discussion, debate, and awareness. Instead, the stuff that I thought was decades away seems to be sneaking up on us as already. . . . . . .
Dan Bricklin on Digital CopyingIn this post, he says,
And in this one, he says,
The sad thing is that the music industry is not just arguing for a legal position that fails to appreciate the economics of digital copying. I actually think that they believe their own legal spin when they make their business decisions. They are blaming consumers for problems caused by the music industry's own inability to adjust its business thinking. As Bricklin points out, the software industry lives with the same economics, and it manages to thrive. . . . . . .
Not that I'm in favor of spamBut does this sound like it's really a good thing?
I wonder what sorts of babies are going to be thrown out with this bathwater. I'm a little bit worried about privacy protection, but I can get past that issue. I'm a lot worried about consumers losing convenience or choice. See this concern, for example. Or this one. UPDATE: Kevin Werbach has some good thoughts, including "Spam solutions that involve fundamentally changing email, such as charging to receive messages, will never catch on."
. . . . . .
Posted Friday, April 25, 2003 From Music Distributor to Software CompanyEMI appears to be making the transition.
To me, the interesting thing is that "partner sites" is plural. That suggests to me that EMI has opened up an interface to its catalog, offering it at a wholesale price, and letting others figure out the retail market. If only all music publishers would to this. Pointer from Lawrence Lee.
. . . . . .
Posted Thursday, April 24, 2003 False Positives and Terrorist ScreeningZimran Ahmed argues that when an event is rare (such as having someone turn out to be a terrorist), screening is difficult. He says that it leads to a lot of false positives, but that does not necessarily indicate a bad screening program. I want to amend his point slightly. When you use statistical filtering to classify something into one of two buckets, you can set a target either for the rate of false positives (calling someone a terrorist suspect who in fact is ok) or for the rate of false negatives (calling someone ok who in fact is a terrorist). With terrorist screening, a major reason that you will have a lot of false positives is that you have a low tolerance for false negatives. Go to the airport, for example, and you will find that you are a false positive. When you get in line, they treat you as a terror suspect as they search and scan you. That is because they believe that they must do that to avoid the false negative--letting someone on the plane who has the motive and the means to hijack. They act as if a false negative has infinite cost relative to a false positive. That relationship between the cost of false positives and false negatives does not always hold. For example, with spam filtering, classifying a legitimate email as spam is more costly than classifying a spam email as legitimate. Accordingly, you would set a very low tolerance for false positives, which means that you end up getting more false negatives. Other things equal, the false positive rate depends on the proportion of positives in the population, as Zimran says. But the screener can adjust the classification algorithm to achieve any desired rate of false positives--subject to the trade-off that with fewer false positives you get more false negatives. The screener should try to balance the relative cost of the two types of error. Finally, the better the screening algorithm, the lower the number of false positives and false negatives. That is what the knee-jerk opponents to data-driven screening fail to appreciate. You have to think in terms of the fact that screening is taking place all the time. At this moment, each of us is either being treated as a terror suspect by the authorities or not. Prohibiting the use of databases in this screening process will not reduce the false positive rate. It will instead increase both the false positive rate and the false negative rate. . . . . . .
An Anti-war PleaI argue against initiating trade hostilities with France, Germany and Russia. Don't Smoot the Weasels. . . . . . .
Posted Wednesday, April 23, 2003 A Surveillance ApplicationFrom a story on news.com:
But wait until the apples and corn flakes start to raise the privacy issue. . . . . . .
Telecom RegulationMy friends at TechCentralStation, where free markets meet technology, continue to take a peculiar view of a "free market" in telecom. Today, Bruce Fein makes the case for regulatory Federalism, which to me means turning loose all the Lilliputian local officials to keep the Bells tied down. It means full employment for lawyers, consultants, lobbyists, and other parasites. The result? Thomas Hazlett uses Michael Heller's phrase "tragedy of the anticommons."
You probably have to read Hazlett's whole article to get the idea. To put it concisely, clear property rights are crucial in order for capitalism to work. Facilities-sharing and "regulatory Federalism" are antithetical to clear property rights. My prediction is that we will get none of the purported benefits from regulation. On the other hand, when the Bells start to collapse due to natural market forces, they will come running to the taxpayers for protection, making an all-too-plausible case that the government owes them a bailout in exchange for forcing them to live with facilities-sharing and other regulatory nonsense. . . . . . .
Stock Options, Once AgainIt seems that stock options are going to be expensed. Holman Jenkins writes,
In the context, I cannot tell whether he is being sarcastic. My view is that stock options are not well structured as incentives. Managers should receive bonuses that are contingent on specific corporate performance objectives. Stock price should be part of the performance measure, but only relative to other stocks. The reason that we see stock options is because all other forms of compensation, including stock options indexed to the market, are expensed for accounting purposes. In fact, this is a case where the accounting treatment is anything but irrelevant--it is the whole reason for the popularity of stock options, which is why the companies that rely on the them are fighting so hard on the issue. Jenkins concludes,
But the politicization was brought about by the high-tech community. Accountants have wanted to expense stock options all along. Only Congressional intervention prevented this a decade ago. If Jenkins wants to endorse an apolitical accounting standard, he should endorse expensing stock options. . . . . . .
Social Software, IIRay Ozzie also sees potential.
That sounds like the issue-resolution application that I described in my essay on the topic. . . . . . .
Posted Monday, April 21, 2003 Private-sector Surveillanceis booming. I don't think that we will find it possible to stop it. The best we can hope for is audits of the surveillance systems and exposure of abuses. In the case of government surveillance, I think that the attempt to stop technology is not the right way to go. Formal checks and balances, with powerful auditing functions, are our best hope for keeping government surveillance from being desctructive. . . . . . .
Social SoftwareIt sounds like an important buzzword, but none of us is really sure what it means. Here is my attempt to describe it. . . . . . .
Posted Sunday, April 20, 2003 The Privacy IssueMichael Kanellos captures our contradictory impulses.
Eric Norlin argues that to protect privacy you would have to start building restrictions into the Internet. Your impulse is to want to be able to Google anybody you want, but not have people be able to Google you. As David Brin pointed out in The Transparent Society, that is not a sustainable position. Thanks to Doc Searls for the pointers. . . . . . .
Posted Friday, April 18, 2003 Sounds Like Moore's Law
. . . . . .
Posted Wednesday, April 16, 2003 How Artists Can Get PaidDan Bricklin has an essay on the topic of alternative ways for artists and musicians to get paid. He argues that with mass-media based business models,
I think that a lot of what is going on under the heading of copyright enforcement and Digital Rights Management is an attempt (doomed to fail, in my opinion) to hard-code the superstar business model in an environment that is ripe for alternatives. Bricklin's essay spells this out quite well. . . . . . .
The Elastic EconomyIt's nice to have the military phase in Iraq over with, so I can go back to writing economics essays, like this one.
. . . . . .
Posted Monday, April 14, 2003 Data Mining or ESP?David Reed thinks that data mining to find terrorists will be as fruitless as ESP.
I expect that data mining will prove useful. However, it would be good to have the government agency audited by skeptical folks with an understanding of statistics--folks like David Reed. . . . . . .
Little MediaWhen Larry Lessig and others lament that Big Media is winning, they need to look at reality. For example,
The article also talks about artists creating their own record labels. I think this is going to be a big trend in the next decade, and not just in music. I'm thinking of self-publishing if I ever do another book. Especially now that I see how little an established publisher has to offer (long lag time between writing and publication, minimal publicity efforts). . . . . . .
Posted Sunday, April 13, 2003 Surveillance Nation, Con'tTechnology Review has part 2 of its story on surveillance systems. One theme is that one way to deal with this technology is to set up rules for use. For example, you could have a set of cameras that together can take a clear image, but you are not allowed to assemble the image unless there is a crime under investigation. . . . . . .
Posted Wednesday, April 9, 2003 The PC is not the PlatformHere is one article on how the phone has replaced the PC as the platform of choice for innovation.
That sounds like a level playing field to me. In the same vein, Dan Gillmor writes,
I think that this concept of a personal universal controller is what will turn out to be important. In what Howard Rheingold calls "the era of sentient things," we are going to encounter all sorts of interactive devices. Just as the Internet needs the browser, the era of sentient things needs a personal universal controller. Some of this trend was visible at least 3-1/2 years ago, which is when I wrote,
Thanks to Lawrence Lee for the first two pointers. . . . . . .
Reconstructing IraqI suggest that there be no rush to privatize Iraq's economy. Instead,
. . . . . .
Posted Tuesday, April 8, 2003 Micropatronage
Pointer from Blogcritics from Instapundit. A year and half ago, Dan Kohn predicted that micropatronage would emerge as a business model for digital content. . . . . . .
Posted Monday, April 7, 2003 Lessig SightingHe's interviewed on econlib, the folks that host my other blog. Talking about copyright and stuff.
I've said that my views are Lessig's with a minus sign. But in the case of the sentence quoted, his views are mine with a plus sign. I don't really have too much to say on the copyright issue these days, because sister Donna has it covered so well. . . . . . .
Develop Recreational Drugs?Brother Zack has an interesting proposal.
I'll run his idea past the high school students in my "technology and society" class. I'm sure they will have some interesting things to say. Actually, I'd be surprised if drug companies are not already doing this. With the success of Viagra, it seems to me that the line between recreational and therapeutic drugs has already gotten pretty blurry. And speaking of drug development, welcome to new brother Derek. . . . . . .
The Government as CustomerMichael Schrage points out how the government can be a frustrating customer for companies selling information technology products and services.
He points out that this process puts small, innovative firms at a disadvantage. My experience is that his description of government procurement of IT services applies to a lesser extent in large corporations. For an individual corporate executive, it is safer to fail conventionally than to succeed unconventionally. Basically, every market segment presents its challenges for a small company with an innovative product or service. Large organizations chew you up in meetings, making the cost of sale very expensive. In my book, I recommend that a small company should not get trapped in the large organization's research cycle, unless you can charge the large company a consulting fee for the hours you spend going to meetings. If you can't charge the large company for your sales presentations, then you'll end up with a lot of time and travel spent, and nothing to show for it. Smaller, aggressive companies are faster to make decisions, but they will drive your pricing down. They are not going to overpay, because they are too smart, too poor, or both. The market I think is best to sell to is the established small business market--restaurants, real estate companies, and so on. They can afford to pay for a good product, and they don't have any more time than you do to waste in meetings. The challenge with this market is gaining the trust of the owners. They are in tough, competitive businesses, where they hear a lot of bull--many of them are bs artists themselves. You can't sell them on technical superiority--you have to deliver a bottom line result with your product or service. Thanks to Lawrence Lee for the pointer to the Schrage piece. . . . . . .
Posted Friday, April 4, 2003 Filtering and ErrorsHere is an example of poor use of mathematics.
It is MacDonald who is right and Ball who is wrong. Think of it this way: what do we do without a database? Then we will have a larger error rate and even more investigations. Or, implicitly, we will not have any terrorist investigations. In that case we save the cost of the investigations (both to the government and to the people who would have been investigated). That's wonderful. Except that now we pay the price of more terrorist incidents, which might be $100 billion each (look at 9-11). In order to know whether an error rate is costly or not, we need to know what the alternative would be. If the alternative is a higher error rate and/or more terrorism, that puts the error rate of a database in perspective.
. . . . . .
Posted Thursday, April 3, 2003 Tragedy of the Commons, ContinuedOne month after this spectrum conference, the audio-video archive that was originally promised is still not available. I would venture a guess that somebody promised to put it up for free, and then backed out. It does not necessarily work to rely on people to do something for the good of the commons, without compensation. . . . . . .
Copyright 2002-2003 Arnold Kling. All rights reserved. Terms of use |
|
|
