Thursday, October 31, 2002

FCC and open spectrum

I like Michael Powell. I liked the sort-of-libertarian-with-a-sense-of-humor portrait of him in a recent New Yorker article (thanks for the link, DF). I like the way he kinda grins and does nothing, infuriating crusty incumbents and opportunist entrants alike. It seems he gave quite a remarkable speech on reworking radio spectrum pointing out (quite correctly) that current spectrum usage is grossly inefficient.

While I'm no fan of unregulated spectrum, there might be more 802.11x like bands we can free up, and current spectrum owners should certainly be encourages to sell of what they're not using efficiently now. Will we see perpetual leases, with no use restriction, and a free spectrum aftermarket? Sadly, I doubt it, but I think more radio bandwidth is going to become available under Powell.

Wednesday, October 30, 2002

Is technology maturing?

HP is on the record stating that the technology industry is maturing. While this may provide cover as Carly reduces the once great innovator into a Microsoft VAR, in these recessionary days it's worth pondering if the tech sector has any growth left. And what does "maturing" mean anyway?

Historically, demand for computing cycles has proven incredibly elastic. The cost of computing has gone down dozens of orders of magnitude from the 60s, but total real expenditure on computing has continued to rise. Note that this is quite remarkable -- if the price of gasoline went down we wouldn't drive around more because gas was cheaper, we'd take our savings from gas and spend them on other things. So I guess one good measure of maturity is "have we entered the inelastic part of the demand curve"? In other words, have we run out of new uses to put ever cheaper computing to or are we going to take our savings from falling computer prices and use them to buy other stuff?

A good way to think about this is to ask "is technology becoming more durable?" It may be strange to think of something both as immaterial and as hard wearing as "software" as a "durable" good, but we do decide whether the latest version of Windows or whatever is worth buying, and if the program just isn't improving much any more it makes sense to postpone the purchase.

Economists model durable goods as "stock and flow", and while I don't want to get into the maths I will say that "stock" is the amount of that good out there (installed base) and "flow" is the amount of new goods needed each year to cover growth and "depreciation" in the stock.

Growth is, well, new demand from increasing populations or whatever, but "depreciation" is what you need to make to replace all that existing stock that gets worn out. This include both physical wear and tear, and "Pt - Pt-1", which is how much the value of last year's product (Pt-1) has fallen compared to this year's (Pt). Another way of interpreting "Pt - Pt-1" is "how much better is this year's product compared to last year's?" If there has been alot of improvement, Pt - Pt-1 will be big -- compare the price of a new computer to a five year old machine on eBay.

If good depreciate very rapidly, like strawberries, they're non-durable. If goods depreciate less rapidly, they're more durable. If computers are no longer improving at the rate they were, Pt - Pt-1 is becoming smaller and computers are becoming more durable. This means the industry is maturing.

J Bradford Delong (U Berkeley) actually has some data on this stuff. It turns out that real investment in computers is at an all time high -- 21 time as much as in 1989. But nominal investment in computers is only 2 times what it was in 1989. So our total expenditure on computing continues to rise even as the price continues to fall. It looks like we are still in the elastic part of the demand curve and technology is not yet maturing.

I don't have information on particular sectors within "technology", but you might be able to mine eBay to see if the difference between a 2001 and 2002 computer is smaller or between a 2000 and 2001 computer. Remember: as Pt - Pt-1 become smaller, the good becomes more durable and the industry "matures".

(Here's a quick application: Before the return of Jobs, Macs held their price pretty well, meaning that they had "matured" and weren't getting any better (Pt - Pt-1 was small). People claimed this meant Macs were better value, but all it really meant was that they were stagnant, so it's unsurprising that people chose not to buy them. Once Apple began to improve its products again, the value of old Macs began to plummet. PCs, on the other hand, seem to be becoming more durable as people are opting not to buy the fastest and most expensive Intel chip any more and are buying cheaper systems, spending their savings elsewhere).

Tuesday, October 29, 2002

Krugman v Kling on income

Arnold Kling has an excellent commentary on Krugman's recent NYTimes piece complaining about increasing income inequality in the US.

I remember, back in the day, when Krugman's MIT site and Phillip Greenspun's homepage were high points of the Internet. Sadly, Krugman has degenerated into a complete hack, and his NYTimes article doesn't even read as if it was written by an economist. (Topel, labor economist at U Chicago, felt pretty much the same way).

Topel also pointed out that the very tip of the income distribution is made up of atheletes/entertainers and CEOs (top .0001%), and technological change and stock ownership explains why their incomes rose so dramatically, and that these aren't bad things. He also points out that the rich have become richer, and the poor poorer, but the two are not linked and that economics has nothing to say about this being societally harmful.

Monday, October 28, 2002

Microsoft v. Linux

A friend of mine was interviewed by some Microsoft strategy folk a few days ago, and they were grilling him on Linux. The Microsoft employee felt that since MSFT has more man hours to throw at the operating system, it will always be better than Linux.

This is nuts. The interviewer was an ex-consultant, so he probably knows little about software development and has never read Brook's excellent "Myth of the Man Month"--in short, man hours don't scale linearly. Secondly, even if Windows OS is better, Linux reduces Microsoft's ability to price it high, and if users cannot absorb Windows improvements fast enough, they'll switch anyway even if Microsoft's product is better AND improving more quickly. Finally, the interviewer made some bizarre comment about Linux being a niche operating system (servers, embedded, etc.) Linux is just a (commodity) instantiation of a client/server OS, applications themselves are being written to the Internet operating system. In that context, Windows is a niche operating system just as much as anything else.

Good words for the boom

Michael Lewis has a really nice article defending the Internet boom from the villification it's going through now. And I feel disdain towards the smug bankers/consultants/MBA students who sneer at Internet companies now when they themselves were 1) selling services to the companies before while 2) hating themselves for not joining in the boom themselves. Lewis points out that democratic capital is good, and there was lots of that in the 90s and technological innovation is good for the economy, even though producers (and their shareholders) tend not to capture much of the benefits for themselves.

Sunday, October 27, 2002

Happy Halloween

Kat and I went as Magritte's ""Son of Adam" to the annual Belgian Student's Halloween party at U Chicago. A good time, as always, and playing to the home crowd netted us second place in the costume competition.

Wednesday, October 23, 2002

UK Indie labels compete

I guess recording labels in the UK are not bound by the CARP flat rates, so the Indie labels there are competing on price for web radio distribution in the US. If CARP rates had simply been fixed as a ceiling, maybe Indie labels in the US would be doing the same thing.

Tuesday, October 22, 2002

Moore's law and accounting

I think David Reed went to Harvard Business School, so perhaps he can be forgiven for not understanding accounting. In this article he claims that since telco equipment follows Moore's law and becomes obsolete quickly, depreciating it (in accounting terms) is wrong and leads to under investment.

First of all, accounting measures of depreciation don't drive business decisions (actually, accounting measures of anything don't drive business decisions much). I will talk more about the economics of durable goods in a future column, but essentially the way it works is that if a good depreciates rapidly (ie. the price for a one year old version of the good on eBay is *much* lower than the price of a new one) you better anticipate earning lots of money from owning that good to justify going out and buying it now.

In telco, the Telecom Act of 96 means entrants can freeride on capital investments made by incumbents, which has lead to neither making capital investments. Unfortunately for Reed, this will be fixed by giving back local monopolies to incumbent phone carriers, not a change in accounting rules. And I don't see him arguing for that.

Microsoft has stopped making it easy

My friend in Europe followed up on yesterday's letter, talking about how VB.NET shares none of the features that made VB so popular (easy, quick, gets the job done) and how Microsoft may be making an error in abandoning this segment. Sean Martin (IBM, Advanced Tech Group) also wonders what's going to enter that low end niche.
Yes, the transition to .NET may be traumatic for many VB shops. This is one of the issues we've been discussing here.

If non-OO to OO is a bigger transition than non-microsoft to microsoft, then going from VB to Java is not necessarilly harder than going from VB to VB.NET or to C#. This is the idea at least. However, it may be that familiar Microsoft tools and brand names will significantly sugarcoat the the transition, even if the fundamental change is real and substantial.

To be honest, it seems like an error for microsoft to push everyone into .NET in a way that seems simply to abandon all this territory in the low-end niche. As I said, I don't know what the concentration curves look like, but I've heard the same stat as everyone else that there are more VB developers in the world than anything else. Presumably that's not because V & B are magic letters. What I'm saying is, maybe VB developers are the most common type of developer for the same reason bad novelists are the most common kind of novelist -- that is, because it's easier that way.

If Msoft imagines they can bully the world into using technologies which are harder, and more expensive in terms of talent and training, in order to produce exactly the same lightweight standalone enterprise apps, then they must be making a mistake. If they imagine that because of web services, internet, basic shifts in technology, etc., there is no longer going to be a place for very low-end development, that's probably also wrong. Cheap-and-easy-development is surely as important a feature as robustness, security, etc., to many organizations.

I feel pretty sure about this. I've observed the growing shock on the part of the CEO as he slowly realizes that careful OO development is simply slower than VB, which *already* felt slow to him because he doesn't know enough about development. If Microsoft is planning to force a wholesale upgrade of their low-end developers then they're throwing away one of their comparative advantages, which is their dev tools. (The dev tools are a kind of a bridge between their strength in apps and their weakness in the enterprise apps the tools create.) And anyway, this upgrade might be simply impossible.

Monday, October 21, 2002

Microsoft makes it easy

A friend of mine working in Europe wrote in describing his experiences with IT out there. He points out what many tech-heads, through arrogance, ignorance, or whatever, refuse to get: it's important to make things really really easy. I saw this often at IBM. Although they now pay lip service to making better developer tools, their IDEs are still not as good as Microsoft's and Lotus Notes makes it pretty plain how well they understand Human-Computer Interaction. Maybe J2EE will create better IDEs, Microsoft will certainly continue to improve their's.
I am advising the CEO of a small software company, which is now basically a VB shop (20-30 people), on the transition to J2EE, to .NET, or maybe a little of each. They've decided they want to do some kind of shift to java, and now I'm setting up trainings on technology, best practices, etc..

Here is what I notice:

People know much less out here on the margins. It would shock you. This goes for management as well as techies. Mgmt knows less about proprietary vs. open, microsoft vs. non-microsoft. Techs know less about development methodologies, basic engineering, alternative technologies. At first I was astonished, but now I see that their clients are on the same knowledge level so deals still happen.

But overall this plays to Microsoft's advantage.

Management likes MSoft partly because it projects respectability and talks like a business. The Neal Stephenson 'command line' article that originally talks about M's OS addiction being like Apple's hardware addiction, also talks about M's bourgeois appeal. That observation seems pretty believable from where I stand. But -- ignorance is a fickle patron. Their clients may not know exactly what java is but they think it's trendy so they want to buy some of it and ask for it specifically. Also, they don't like paying licenses and maybe they feel angry at Msft because they're frustrated with computers (like everyone else who uses them).

This may be the smaller factor. Microsoft's advantage with unsophisticated tech may be bigger.

Out here, tech likes microsoft because it's easy. Microsoft has great IDE's and VB development is quick and easy. You're right Msoft makes great apps. That carries over into their development tools, and the integration between their DB's, IDE's, etc., so that you can whip together a lot of stuff on Msoft tech very quickly. This gives them an advantage in small intranet and enterprise apps, even if their servers are dodgy or expensive.

Also, when you add it all up, I'm not convinced it's an open-shut case if microsoft costs you more than, say, linux and java.

In javaworld you don't need to pass microfost licensing costs along to your customers. BUT, maybe development takes longer and costs more per hour, because the tech is harder, and because the tools and overall development environment is not as well integrated. I've also followed close discussions of this on places like, and it's not a no-brainer.

(Personally I'd recommend developing in a well-integrated environment like Websphere + WSAD, then deploying ontro free platforms like Tomcat/Apache. Sort of a half-half solution.)

Lots of these hacky VB apps stumble and certainly don't scale, but most software fails so it's hard to pin it on the fact that it's sloppy rushed VB work.

I don't know what the distribution curve of global IT sophistication looks like. But I wouldn't be surprised if the least sophisticated make up most of the people, and my context which was at first very surprising to me is in fact close to the norm.

Perhaps the transition to .NET may disrupt some of this, since it's fully OO and hard and slow like java, at least relative to vb6. Too soon to tell. I suspect Msoft excellent IDE's and platform integration will still make it easier to develop for.

Friday, October 18, 2002

Crazy rumour

Word on the street is that Microsoft is going to buy (a stake) in Siebel. Why? 1) Microsoft has no channel to sell into enterprise, and Siebel is quite good at that and 2) Microsoft needs an application that will drag .NET into enterprise, and Siebel's CRM apps will do that. Look for what billg announces at the upcoming Siebel conference (Oct 20-23).

Windows isn't good at operating systems

Paul Wolpe wrote in disagreeing with my assertion that Microsoft makes poor operating systems. He argues that NT is actually quite good, and given time, it will be even better. I spoke with Sean Martin at IBM about this (Sean is one of the advanced engineers there and a really smart guy) and he had good things to say about the way Microsoft pumps out code. Given how hard code production is, and how bad most companies are at it, I concur--Microsoft software is actually quite good compared to most other commercial software.

So why they insist on lumbering it with their bad OSes I don't know.

Microsoft is bad at operating systems the same way Apple is bad at hardware. (Note that I write this on my 800 Mhz iBook which I deeply, deeply love. But I would love it more if it had a fast x86 processor.) Just as Apple hindered its wonderful operating system in the 1980s by clinging to proprietary hardware, Microsoft is hindering its wonderful applications now by clinging to proprietary operating systems.

Having an inhouse team of developers, no matter how talented, is just not the best model for producing and maintaining a kernal. Wolpe mentions GNU/Linux in his post, but then dismisses it. Balmer doesn't, he's on the record as saying Open Source software is the #1 threat to Microsoft. And he's right, in terms of performance, scalability, robustness, and cost GNU/Linux beats the pants of anything produced in Redmond.

So why doesn't Microsoft exit the OS business (the way Apple has and simply ship their quite good enterprise applications on Lintel blades? It's because they're addicted to the operating system business, and that will prove their downfall.

Thursday, October 17, 2002

Microsoft on GNU/Linux

Microsoft is quite good at applications, but quite bad at operating systems. It wasn't until working at IBM this summer that I appreciated how easy to use, simple, and user friendly Microsoft software was (shocking, but true). Microsoft's biggest obstacle to penetrating enterprise computing is their insistance on running SQL Server, BizTalk, MSFT CRM etc. on the lousy .NET Server instead of Lintel. I asked a VP of .NET if this would ever change, and once he had finished laughing he told me "no."

But the failed Passport identity system seems to be crossing over to Unix, which is pretty wacky. They really must beleive it has no chance.

Tuesday, October 15, 2002

Don't knock selling bananas

A few weeks ago, the Register ran an article complaining that the capital markets destroy research because they won't fund True Engineering. MJ wrote in asking what I thought of this.

From my time hanging out around tech companies, I've noticed that engineers beleive salvation lies in the Right Technology and have little patience for whatever those MBA marketing guys waste their time doing. I would argue that history demonstrates how having the Right Technology rarely determines who wins, but distribution (what marketers call "channel") does.

No piece of engineering is an island--it sits in a "solutions stack", all of which must be in place, to deliver something useful that a customer is willing to pay for. Engineers might recognize a hardware, OS, application stack, but from a customer perspective it actually extends all the way up through education, training, service, technical support, etc. Now, this stack can be provided by one company or many. Apple integrates alot of it, while the PC market has Microsoft monopolizing the OS layer (and Office software) while having competitive markets for hardware, education, training, service, technical support etc. The guys who own the scarce parts of the solution stack get all the profit in the industry.

By the way, did I mention that distribution was part of the solution stack too?

My point is that financial markets only care about money, not technology or distribution, but don't knock how important the distribution part is to making money. This single minded focus on lucre is a good thing because it means they're quite happy funding anything consumers will buy (like Napster of TiVo) without getting bogged down in nostaligia. It so turns out that technology and its distribution channel need to coordinate quite strongly (ie. distribution and technology are often co-specialized assets). IBM research has invented lots of neat stuff over the years, but its distribution was often optimized for other things so they were never able to capitalize on their inventions. Xerox invented the future of computing at PARC, but again had no co-specialized distribution assets that could take it to market.

So don't underestimate distribution. Commercializing technology is hard. Really really hard. Hard hard hard. It often makes sense to build a whole new company, essentially one big specialized distribution asset, to get that technology to market. Which brings us back to the capital markets. While people accuse Wall Street of being short sighted it really isn't (although it gets over-emotional--just look at how far it was looking into the future when valuing Internet stocks!) The market is very hard nosed and will evaluate R&D spending by 1) will it come up with anything useful and 2) when it does come up with anything useful, will the company make money off of it? Sadly the answer to 1) is: probably not (Microsoft) and to 2) is: almost certainly no (Xerox, IBM, Apple, IBM, Motorola, IBM).

But do not lose heart. The above just means that public financial markets are not a good source of financing for speculative technology investment, although they support developmental applied research (Intel) handsomely. But this is the way it should be -- you really don't want lots of start-ups working on crazy ideas to try and IPO so they could test them out (and despite what anyone thinks, speculative technology does not suddenly become free if it is inside a big company). Fortunately, there are lots of financing sources standing by to develop speculative technology, including government grants, VCs, Private Equity, Angel Investors, etc. If financial markets seem to favor distribution these days it's only because distribution makes more money. If HP wants to withdraw from systems research and focus on becoming a Microsoft VAR that's fine with me -- their R&D didn't bring great products to market.

What's really killing the technology sector right now isn't lack of financing (remember, VC firms are struggling to make good investments and are having to give money back because they can't find opportunities) it's lack of demand. Consumers are not that excited about PDA/Phones, PVR/DVDs, or digital media/PC combos (not that *some* consumers aren't excited, but it's not an Internet-esque stampede). Business are not excited about CRM, ERP, EAI, or ERM, and are struggling to make their numbers at a time when the bottom line really matters to shareholders. There is lots of patient capital out there funding new technology, but right now, dissatisfied, burnt out customers aren't buying.

Kill the CD

Since CDs are easy to rip, the music industry wants to shift to a crippled media format that will forbid that. Unless the CD (and MP3, and PC) are banned, these formats will fail.

Monday, October 14, 2002

RIAA vs the Turing Machine

Ed Felton writes how the US Government thinks "the general purpose computers is a threat, not only to copyright but to our entire future." Now, while the PC is often too complicated to use, and some users may benefit from a simpler, restricted machine, the desktop Turing Machine is still critical to innovation. But last year when I mentioned this to Randall Picker (U Chicago Law prof and smart guy) he struggled to see why.

Given the market success of most information applicances, the general purpose computer is not going anywhere unless outlawed by the Government. Microsoft, with Palladium, is already locking up the Windows PC, but there are many other platforms more focused on empowering end users.

Sunday, October 13, 2002

Lessig's afterthoughts on Eldred v. Ashcroft

It is very worth reading Lessig's notes after arguing Eldred v. Ashcroft.

Saturday, October 12, 2002

Hobbyists and copyright

There's a nice op-ed in the NYTimes, answering Valenti's argument for extending copyright "Who is going to digitize these public domain movies?" Answer: Us.

Friday, October 11, 2002

Laboratory economics experiments

I wanted to update my last post on the 2002 Nobel Prize for Economics. A Slate article on it pointed to an experiment where people seemed willing to force others to give money to another party, and pay for the priveledge, even if they would not give money to others directly.

The article mocked this result as being weird, but I think it illuminates how limited people's understanding of the consequences of their actions are. For example, people often support rules and regulations that take money from one group (i.e. car makers) and give it to another (i.e. steel manufacturers) even though it costs them money (through higher car prices, which in turn reduce the amount of money available for *all* *other* *workers*). This type of wealth destruction is known as dead weight loss, which is pretty obscure outside academic circles. 50 years ago, one economist got so angry at this widespread ignorance that he wrote an entire book talking about nothing else. "Economics in one easy lesson" by Henry Hazlitt is a fun read, which I recommend to all.

Certainly around taxes and regulation, people do seem very willing to make laws which impoverish themselves and confiscate wealth from one party to give to another. I think that people do not realize that government spending comes from taxes, or how law can act as a tax and take money from one group to enrich another, and how this harms society as a whole.

Thursday, October 10, 2002

Nobel prize for Economics, 2002

When I wrote about last year's Nobel prize for economics I ended up getting a cease and desist letter from MIT Press and had to take it down. *Sigh* For those who remember, last year's prize went to information asymmetry, which showed how people would not trade if they could not be convinced that they weren't being cheated. A nice and interesting wrinkle on Chicago School micro.

This year's prize goes to Behavioral Economics. There are two schools of "behavioral economics", one which focuses on empirical research and the other which takes the classical micro model into new areas.

First, some context. The standard microeconomic approach looks at how a rational individual, who tries as best as he can to look into the future, goes about maximizing his "utility." While it's easy to make jokes about how unrational people are in real life, this approach is unparalleled in its power and generality, and by that I mean it explains a great deal of many different sorts of behavior. Quantum Mechanics aside, I struggle to come up with any other theoretical model that has had quite the material success of the microeconomic rational utility maximizer. This school of economics was invented at Chicago in the 50s.

One school of "behavioral economics" takes the standard micoeconomic rational actor and extends the utility function to include all kinds of goods that go beyond wealth. For example, racial prejudice can be thought of as a consumption good where an individual exchanges money for prejudice (by not hiring black labor even though it's cheaper and better). This school of behavioral economics was also invented at Chicago by Gary Becker (who won a Nobel Price in 1992), who has done amazing work in this area with Kevin Murphy. Through this, they've come up with powerful and insightful models of personal taste, addiction, the family, crime, etc.

The other school of "behavioral economics" does away with the rational actor and instead presents real people with real choices in a lab and sees what they do. One of the leaders in this field, Richard Thaler, is also at Chicago (along with Barberis) and focuses on "unrationality" in the financial markets. It is for this school that Kahneman and Smith won their prize this year. (Note: Kahneman was the thesis advisor of a good friend of mine at Chicago).

Of course it's hard to argue that people are perfectly rational, but you can make the utility function in the the standard microeconomic model pretty broad, including alturism, (unstable) preferences, and lots of other things and come up with good, robust predictions about the world. Whenever you deviate from this sort of thing you end up with someone losing all their money unless they become "rational" again, so the model is pretty self correcting. I don't think the experimental behavioral economics gained much traction until it found irrationalities in the stock market (which, when publicized, were arbitraged away and no longer exist).

The main criticism of the experimental school of BE is that, well, people perform strangely in labs. When Thaler points to some small, market irregularity and calls it the tip of the iceberg, Gene Fama (Chicago, invented modern Finance) says "No, that's the whole iceberg!"

So I'm pretty torn about this year's prize. I guess I'm skeptical of how useful lab experiments are in generating useful data, but people are wacky and market anomalies remain (for example, there is too much volatility in the stock market, and there are financial booms and busts). Useful stuff may come out of experimental BE in the future, but I'm still trying to wrap my head around the standard microeconomic model.

Wednesday, October 09, 2002


This converged MSFT device is just the sort of nonsense I feel will fail in the marketplace. At least Microsoft dropped some of the Digital Restriction Management flaws.

Eldred v. Ashcroft, Day 1

I found some good notes on the first day of Eldred v. Ashcroft, but I have no idea how to interpret any of it. Ultimately, the length of copyright is something that needs to be set by society via Congress, not through the courts. As O'Conner pointed out, bad policy is not unconstitutional. But retroactive extension absolutely *is* unconsitutional as it clearly does nothing to "promote arts and science" and so should be beyond Congress' power to mandate. I was glad to see Breyer get sensibly "economic" on this.

Tuesday, October 08, 2002

Goodeasy for Windows

Someone's trying to put together a Goodeasy for Windows. Nice! (Via Mark)

CARP modified

Since I've written about CARP before it seems appropriate that I comment on it now. (CARP, incidently, is the royalty system set by the US Copyright Office that sets a fixed rate for web radio.) It seems there's been a small carve out for small webcasters that sets rate according to complicated accounting defitions of costs and percentages etc. (details here). These ammendments may help some small webcasters, but they don't do much to inject competition into the online distribution of streaming music, because the (lower) rates are still fixed, and not ceilings. Sadly, I heard neither the RIAA nor webcasters argue for a rate structure that just set the ceiling, and let content owners compete for distribution.

Monday, October 07, 2002

Spectrum is scarce

USS Clueless wrote up a good article talking about how CDMA cellular technology squeezes more data into less radio spectrum through "spectrum sharing". Spectrum sharing (forcing processors to clean up noisy signals) is one of the key technologies open spectrum advocates invoke when arguing that all spectrum should be unregulated. I've long been skeptical of these assertions, so I asked Steve Den Beste if spectrum sharing technology can, indeed, make spectrum non scarce. He answered:

Spectrum usage is subject to Shannon's work. Spectrum is a finite resource and if too many people try to use the same spectrum then no one gets anything out of it.

Technology can't fix this. It's inevitable because of the laws of physics.

Shannon's work is in part based on the Second law of Thermodynamics.
To do: bone up on Shannon.


There's a nice article on how DVDs are not only outselling VHS tapes, but they're grossing more than movies. This is changing the how movies are made -- stuff is now being filmed for the express purpose of it ending up on the DVD.

The economics of all of this are pretty neat. Tapes cost about $25 to make, degrade with use, and cannot store much. DVDs cost about $1 to make, don't degrade, and can store lots. Given that demand for movies seems to be pretty elastic, lowering their cost greatly increases the amount people consume. This, combined with the fact that DVDs don't degrade, means that not only is consumption higher, but it also shifts from renting to owning. The article ends by suggesting DVD prices may fall to $10 a pop, making them impulse buys. The most efficient distribution network for video may end up being supermarket checkout lines.

Friday, October 04, 2002

Regulating fair use

I think Lessig, at some recent conference, pointed out that most things people did with content was unregulated--that is neither protected under fair use or controlled by copyright. Such uses include lending a book to a friend, listening to a CD multiple times, or watching a DVD at a friends house. The RIAA and MPAA's fight against unauthorized copying often involves regulating previously unregulated activities--and assigning new rights to copyright owners. Now, economics does not care who owns something, just that people are free to trade stuff, so giving content owners perfect control is economically efficient, but makes consumers unhappy by raising prices on them. Assigning ownership to new (previously unregulated goods) is best done through a political process, which is essentially what the Boucher bill is about.

Wednesday, October 02, 2002


As mentioned in an earlier post you can argue that Microsoft's XBox is more than some overgrown kid/exec's dream of a fun job because it, like Palladium, is a DRM system for secure, restricted content deployment, and as such competes against Sony's PS2. Moreover, since PS2s can run Linux, and you can plug a keyboard and ethernet cable into them, the idea is that console and PC are "converging", just like everything else ultimately will.

Essentially, the thrust was that since content owners didn't want their stuff played on open systems where it could be copied, PCs were going to have to have a special "closed" mode if they wanted to be platforms for new media (post-CD, post-DVD). For Microsoft, this special closed mode is called "Palladium".

How important is it to play DVDs or CDs on personal computers (or a game console)? How important will it be in the future? Given how consumers want PCs to be open and content owners want media players to be closed, how much will these domains "converge"?

For the home market, there is little benefit in converged devices. The consumers does not need to make many tradeoffs between buying a game console, and DVD player, and a PC. I bet that most PS2 owners also own a DVD player and that their willingness-to-pay for the PS2 would be about the same if it did not play DVDs. Same thing with a PC that plays DVDs.

This does not hold in the mobile market, because there are very real tradeoffs between different features and functionalities. It make sense to pick a slower processor for longer battery life on a laptop, but not for a desktop PC. Similarly, carrying a PDA, MP3 player, cell phone, and digital camera is cumbersome for people so convergence makes sense there also. Unfortunately, combined devices are expensive, difficult to use, and don't do anything particularly well, so while in the long run it makes sense for these to converge, in the short run I see ever more *divergence* as manufacturers churn out many models that suit the widely varying needs of consumers.

So back to Sony/PS2 vs Microsoft/XBox. The game console is no threat to the PC (although a $200 GNU/Linux box might be).

Tuesday, October 01, 2002

Microsoft on campus some more

Another Microsoft exec (super nice guy) came to speak to a smaller group of students at Chicago today. We were discussing what stuff gets decided by engineers, and what gets decided by business folk, and I asked about Palladium, .NET, and "Trustworthy Computing". I was told that all of the important platform strategies are worked out by Billg and his core engineers.

Microsoft on campus

For some reason, Microsoft wants to hire many MBAs this year, so is out in force on Chicago GSB's campus. The managers present are all very nice, but where I want to talk about Palladium, "Trustworthy" Computing, and .NET integration, they're talking about MSN Shopping. In general, they don't seem to know about the other stuff, so I'm guessing that the real corporate strategy is not handled by these guys. One thing that is interesting is that last year I used to get blank stares whenever I mentioned Linux, but this time everyone knows what it is.

iSync, Palm, and Now-up-to-date

Last night I got my Visor to sync with Now-up-to-date (the best calender in the business, better than Outlook, better than iCal) and the new Apple Address Book using a combination of iSync and Handspring's Palm desktop Hotsync manager. I was very impressed by how gracefully iSync handled conflicts between different devices -- very nice!