Having just returned from Amsterdam yesterday, I’ve got a pile of mail and chores that I should be working on. But I wanted to record my impressions of Kevin Kelly’s keynote presentation at the “Next Web” conference while my memories are still somewhat fresh. Bottom line: if you’ve been feeling jaded, skeptical, or even cynical about the hype surrounding Web 2.0, you should take advantage of any opportunity to hear Kelly’s perspective. It’s optimistic, refreshing, and downright uplifting; and he’s a great speaker, with a good sense of humor, too.

Kelly points out that today’s Internet/Web is only about 4,000 days old — if you start counting from when the Web first became widely available to the general public in 1995, when Netscape released its browser. The prospect, ten years ago, that we would have what we do have would have been considered preposterous; if someone had predicted the free availability of Google Maps, for example, it would have been dismissed as science fiction. And he argues, fairly persuasively, that the next 4,000 days will bring about changes and improvements in Internet/Web technology ten times as great as what we’ve already seen.

This “10X” theme is something Kelly repeated several times during his presentation. Information, he says, has been growing at the rate of 66% per year for the past decade, while physical production has only been increasing at an annual rate of 7%. Thus, information about any artifact, X, is growing ten times faster than our production of X itself.

This growth rate is roughly the same as the hardware growth rate generally referred to as Moore’s Law — i.e., it doubles every 18 months. And Kelly argues that that’s all we need to know in order to ride the economic wave of improvements and benefits that will follow; similarly, he says, if we had really believed in Moore’s Law when it was published in 1965 (April 19th of that year, as it turns out; see the Wikipedia entry for more details), that would have been enough to get rich, even if we didn’t know whether to invest specifically in IBM, Apple, or Microsoft. (I understand his philosophical point, though I’m not sure I agree with the specifics: vast fortunes have been both made and lost based on well-timed or ill-timed investments in specific technology companies.)

The distinction between the physical manifestation — i.e., the “atoms” — of X, versus the information — i.e., the “bits” — about X is an important point. As Kelly observes, the pundits of Web 1.0 portrayed atoms-versus-bits as a binary schism; and they frequently asserted that the entire world of atoms was undergoing a complete, revolutionary transformation to a world of bits. Web 2.0, says Kelly, is a convergence of these two worlds: we still make and sell atom-based widgets (as Amazon and eBay clearly demonstrate), but those widgets have more and more bits embedded within them, and those bits “speak the language of the Web.” So when the Internet enthusiasts began telling us a decade ago that, sooner or later, even our toasters and refrigerators would have IP addresses, they were basically right: everything we build is, or soon will be, Web-enabled. But Kelly acknowledges that it will probably take another 5-10 years for this to play out.

As a result, he says, we’re just beginning to see the “information planet.” Every object that is produced, and each of us, and everything we do, is part of the Web. We are wiring the entire planet together, a phenomenon that will take form during our lifetime, which thus represents a unique epoch in the history of mankind. If this sounds like over-hyped hyperbole, Kelly argues the opposite: the Web, he feels, is actually under-hyped, at least in the long run. He suggests that we’ve gone from the “crazy” moment of the dot-com bubble, to today’s “golden age” of excitement about the things we can really build and do.

But this doesn’t mean that everything old — and, for example, everything associated with Web 1.0 — will disappear. Kelly argues that technological revolutions don’t actually die: all of the old stuff remains available for a long, long time. I’m not completely convinced of this, because it seems like I’m constantly finding “old” things that I really cherished, but are no longer available. But Kelly showed us some examples that he had found on the Internet, such as new parts for the original century-old “Stanley Steamer” automobile, and new productions of the original, unchanged versions of various household and farm implements that had first been marketed in a Sears & Roebuck catalog from a century ago.

Thus, Kelly argues, the original Web and Web 1.0 will continue for a long time; and Web 2.0 will be built on top of that, to eventually be followed by Web 3.0, etc. Web 1.0 won’t go away, but will simply become the “legacy” foundation upon which successive generations of technology will be built. This is an argument I can support: after all, much of the mainstream business world today still runs on billions of lines of COBOL code, some of which is emulating an earlier generation of assembly-language code, some of which is emulating programs written in the 1950s for the IBM 1401. Kelly didn’t bother mentioning all of that historical detail to his audience in Amsterdam, but perhaps he was trying to reassure them that their Web 1.0 skills would not necessary become completely obsolete and useless in the near-term future. Yes, today’s Web 2.0 techies need to know the details of Ajax; but we still need a lot of old-fashioned HTML, and not all of it can be generated automatically.

One part of Kelly’s presentation that I found particularly interesting was the subject of economics — and specifically, how can businesses (and entrepreneurs, and venture capitalists, and investors) make money with Web 2.0, in a world where “information wants to be free,” and where many of the startup Web 2.0 companies are offering their products and services to the public at no cost at all? As I mentioned in yesterday’s blog posting, this was an issue that clearly bothered several people in the conference audience (including me), because sooner or later, most of us have to figure out how to make enough money to pay the rent and put food on the table.

Well, Kelly argues that the origin of the Web (going back to a decade ago) is analogous to the “gift economy” of the annual Burning Man event in the United States. I think he may have been stretching a point here, since Burning Man is a localized event that takes place for only a week each year, but in any case, Kelly argues that it’s entirely natural that people build/donate/contribute things on the Web for free. Wikipedia is free, and hopefully always will be; and open-source software is, and always will be, free — by definition.

To a lot of people and a lot of traditional business organizations, that doesn’t make sense; it’s not surprising that the world’s richest man (Bill Gates, for those of you who don’t keep up with such mundane things) reacted to the early descriptions of the “information-is-free” ethos of the Internet by exclaiming, “We (Microsofties) aren’t Communists!” But in today’s Web 2.0 world, we’re beginning to accept the notion that not everyone does everything for money (except Bill, perhaps, but even he is giving his billions away). Yes, I still need to earn enough money — the old-fashioned way — to pay the rent and put food on the table; but if there are a few extra hours in the day, I may be motivated to do something for no money at all — like this blog posting! — and contribute it to the Internet community because I have a passion for the activity, or because it’s fun, or because I want to share it with others, or because I think it will enhance my reputation.

But not everything on the Web is necessarily free forever — and if the Amsterdam audience really focused on this part of Kelly’s presentation, I think a lot of their (and my) concerns would have been eased considerably. Kelly argues that there’s a natural economic “cycle” involved, from free stuff to commercial stuff, and back to free stuff once again. In the beginning, when the urge to create and innovate is driven by passion and fun and the desire to share “cool” stuff with everyone else, the stuff we create tends to be given away free (assuming, of course, that we’ve got some other mechanism to generate enough money to pay the rent). That was even true in the early days of the PC, when hobbyists and hackers developed early word-processing programs like Electric Pencil, as well as the myriad software creations developed in university environment.

I mention Electric Pencil, because it was first released in 1976, about the time that many of us were trying to figure out what we could do with our newly-acquired Apple II’s, Altair 8800’s, TRS-80’s, and NorthStar Horizons, besides programming games of tic-tac-toe in Basic or assembly language. But Electric Pencil was eventually overshadowed by a commercial product that some people (not me!) still remember fondly: WordStar. And WordStar was, of course, eventually overshadowed by Microsoft Word. Kelly’s point is that when early, prototype, rough innovations — often created by lone developers who care little about the nuances of user-friendly interfaces, etc. — gradually get enough support, they transform into commercial products whose features and functions can be optimized and maximized. You had to pay real money to get a copy of WordStar, but it had more features than Electric Pencil. And you have to pay a lot of money for Microsoft Office … and it presumably has several thousand more bells and whistles than WordStar, even if I don’t use most of them.

But this often creates a problem for at least a portion of the marketplace, especially given the typical commercial disease of “featuritis”: we have bloated products, with zillions of features, whose cost exceeds what many members of the global marketplace can possibly afford to pay. It’s very easy for those of us in North America, Western Europe, and other parts of “advanced” society to forget this — but it really is true.

I first encountered this reality in January 1992, when I had the good fortune to be invited to a United Nations technology conference in Cairo. During a lunch break, I was quizzed by a delegate from one of the North African countries, who apparently assumed that because I was an American, I could serve as an official spokesman for the American software industry. “Why don’t your American software companies provide more service and support for customers in developing nations here in Africa?” he asked, in a slightly confrontational tone.

Annoyed, I replied somewhat bluntly, “Perhaps it’s because there’s so much software piracy here in Africa. If you didn’t steal our software so often, perhaps we would be more interested in providing support.”

The delegate looked stunned — not because I had been so rude, but because I had demonstrated a complete ignorance of the economic realities involved. “Perhaps you don’t realize,” he said patiently, as if explaining basic arithmetic to a young child, “that your country’s ‘office’ suites cost about $400. That’s roughly equal to the salary for one year for a college-educated person in my country, someone like a teacher who might be able to use a word-processor if his school was lucky enough to have a computer.” His argument was simple: the intellectual creation of a word processor was vital to society, and it was simply wrong for anyone (including whoever had written the software for the word processor) to charge that much money for it. And since that was wrong, it was right to make copies of the software, especially since the copy did not “diminish” the asset of the person holding the original version of the software.

I began to object, but a colleague and friend, Tom DeMarco, who was attending the conference with me as the only other American delegate, diplomatically reminded me that there had been a time when Americans behaved in an equally self-righteous fashion. Though I had learned about it in school, it was something I had conveniently forgotten: back in the mid-18th century, the British textile industry did its best to protect the inventions of people like Richard Arkwright, Edmund Cartwright, and others through patents and other legal mechanisms; among other things, experts in textile technology were forbidden to emigrate from the British Isles, lest they help launch a competitive industry in one of the Colonies (or, god forbid, a hostile nation like France or Spain, or one of their colonies). Nevertheless, there were numerous incidents of mills and factories being burned, looted, and spied on by competitors. Hence, paranoia was rampant, and patent protection was taken quite seriously.

Enter Samuel Slater. Born in Derbyshire, England in June, 1768, Slater become involved in the textile industry at the age of 14 when he was apprenticed to Jedediah Strutt, a partner of Richard Arkwright and the owner of one of the first cotton mills in the area. Slater worked for Strutt for eight years and rose to become superintendent of Strutt’s mill. It was in this capacity that he gained a thorough understanding of Arkwright’s machines.

Believing that textile industry in England had reached its peak, Slater emigrated secretly (i.e., illegally) to America in 1789 in hopes of making his fortune. Others with textile manufacturing experience had emigrated before him, but Slater was the first who knew how to build as well as operate textile machines. By the way, lest you think this was the illegal act of an isolated individual, it should be noted that America was paying rewards for textile-manufacturing information at the time.

If all of this had happened in the late 20th century, the intellectual theft probably would have involved smuggling a floppy disk with drawings and engineering details — or simply zapping it from England to America via the Internet. But in the late 18th century, Slater used a much simple, brute-force approach: he simply memorized the details of the machinery, the procedures, and all of the other details required to build a textile mill. With funding from local investors and assistance from skilled artisans, Slater built the first successful water powered textile mill in Pawtucket, Rhode Island in 1793. And thus began the American textile industry.

But I digress… Kelly’s argument, as I had been explaining, is that creative new stuff on the Internet starts off as a freebie give-away, then gets commercialized into an “optimized” product that people pay for, and that vendors profit from. But eventually — and I think this is a crucial point — the product becomes a commodity, at least in an intellectual sense, and so it gets transformed back into a free product once again, typically in the form of open-source software. There’s less and less reason for someone in the Third World to bemoan the economic distress of paying $400 for a word-processing office suite when you can get one free from Open Office, or from its smaller cousin, Writely. And, as noted in the discussion above about the atoms-versus-bits dichotomy, the free-versus-commercialized distinction between software products is not a binary, all-or-nothing distinction: there’s also the world of shareware, which allows you to get a semi-professional version of a product, with either simplified features or specialized features, for a very modest sum. A quick check of Version-Tracker, for example, shows a list of some 85 shareware word-processing products for the Macintosh platform; my favorite is Tex-Edit Plus, which sells for the princely sum of $15.

None of this has to be explained to techies in startup companies, for they’re already busy throwing free stuff up on the Internet to see if everyone else likes their ideas as much as they do. But for the older generation, the ones trying to figure how this Web 2.0 stuff relates to their traditional view of the capitalistic marketplace, Kelly’s advice is “follow the free.” By that, he means they should operate a rational business enterprise as if their products were free. Open-source software companies have understood this all along: they’re obliged by the various forms of open-source licenses to give away the software for free, but they can charge for training, support, textbooks, t-shirts, and coffee mugs. And the vast majority of Web 2.0 companies understand this in another form: they provide services on the Web for free, and hope they’ll recoup their expenses through advertising click-throughs.

But that’s still a hard sell when you’re talking to a CIO, CEO, or profit-center manager at an airline, bank, insurance company, or any other company selling traditional widget-based “atoms.” Kelly told the story of KLM, to whom he offered the “follow the free” advice several years ago. They dismissed his idea as crazy, but their revenues in recent years have paled in comparison to RyanAir, which announced that they had given away 3 million free seats last year. So here’s another way for CIOs and CEOs to think about all of this: imagine how you’ll respond when your competitor starts giving away its products and services for free. It’s one thing to cut your prices by 10%, but if your only competitive advantage is price, it’s pretty hard to compete against someone offering the same product as yours, for a price of zero. Thus, Kelly’s advice to today’s publishing industry is: copies are, or soon will be, free. So let go of things — not just digital “things,” but atom-based things, too, since China (and various other countries) can counterfeit or pirate almost anything.

Of course, it’s not all that radical to propose giving away a product for free (or almost free) in order to charge money for ancillary products and services. Gillette understood this 50 years ago, with the notion of giving away razors, in order to charge for razor blades. In the pre-digital photographic era, Kodak figured out that it could essentially give away its cameras in order to make profits from the film; even today, Kodak offers half a dozen different “one-time use” cameras, and one of their current products is actually a digital one-time use camera.

But can this idea be carried any further? Kelly predicted, though he didn’t provide much in the way of details, that we might someday see open-source products like cell-phones, cameras, televisions, and automobiles — and that perhaps these products will come from China. The point here is that the raw materials that make up such products are relatively inexpensive, and the greater cost component is the design, the embedded software, and the information-based infrastructure that “surrounds” the atom-based product. If all of that software-based stuff was open-source and free, then perhaps we could sell a car for a thousand dollars instead of $10,000. Or, perhaps if we were really clever entrepreneurs from, say, Saudi Arabia, we could give away the cars for free, and just charge for the gasoline.

There’s another reason that companies can afford to give away more and more of the software “stuff” for free: they themselves can get it for free from the customers, and from millions of other people out on the Internet. As Kelly pointed out in his presentation, there are more and more situations today where external users are doing the work that was previously reserved only for internal employees (one good example is Eli Lilly’s Innocentive program, which involves 30 companies, 90,000 scientists, and rewards of up to $100,000 for R&D inventions submitted by the contributors). Especially in the area of software, users contribute enhancements and plug-ins; users sometimes provide testing, debugging, and service (e.g., voluntary technical-service “help desks”). Users do some of the marketing (last time I checked, several years ago, Amazon had 50,000 “associates” who plugged their books with links to the Amazon website). And, says Kelly, users can even finance some of the new products and services; I’ll take his word for that, though I don’t have any examples that I can provide.

Branching off in what first seemed like a different direction of his talk, Kelly suggested that the link (i.e., URLs) and the tag will be remembered as the greatest invention of the past decade; I’m not sure whether he meant “greatest” within the narrow boundaries of the Internet/Web, or the entire computer field, or perhaps all of science. In any case, he says that this is part of the overall evolution of the Web: it began with digitizing information into bits; that was followed by linking things on the Web with URL hyperlinks. The current stage of evolution is collaboration, using tags; and the next stage will be the addition of semantics, to provide some structure to the vast amount of information out there on the Web.

The current evolutionary phase of collaboration, according to Kelly, consists of three different types; the first is “inside” collaboration, of which Google’s link-based-”recommendations” form the basis of their powerful search engine. The second type, he says, is collaboration between organizations, of which API’s and mashups are the obvious examples. And the third type is “aggregations,” of which he suggests digg.com as a prime example. (A little later in his presentation, Kelly suggested another example of aggregation: if all the washing machines in the world were connected to the Internet and could aggregate their experiences, they would do a much better job of washing our shirts; and they could offer us some great advice on what kinds of shirts to buy for various uses and environments.)

It’s the third form of collaboration that Kelly seems most interested in; he referred to the concept of “deep collaboration,” in which hordes of disconnected Internet users might come together to create open-source collaborations for things like the design of automobiles, cameras, cell phones, and other such things. Kelly suggests that there’s a natural tension between the non-collaborative, top-down “controlled” design of products (of which Steve Jobs’ design of the iPod is a prime example), and the grass-roots, bottom-level ad hoc collaboration (of which Wikipedia is a prime example). While there may be various situations where one extreme or the other is “better,” Kelly argues that the greater power is at the “bottom” level of ad hoc collaboration. The bottom level, he argues, will always take things farther than you thought; consider, as an example, Wikipedia (with 1,241,000 articles) versus Encyclopedia Britannica(with 80,000 articles). So, says Kelly, every industry needs to ask itself, “How far can we go towards the bottom?”

From my scribbled notes, I’m not quite sure how Kelly got to the next point in his presentation, but it had to do with asymmetry: he observed that it’s much easier, in today’s Internet environment, to download stuff than it is to upload stuff. But because production of “stuff” is becoming more democratized, we’ll have to shift the mechanism for uploading stuff onto the Internet to make it a lot easier; to some extent, that’s what Web 2.0 is all about. (By the way, there’s a great new book on this topic called Democratizing Innovation, by MIT’s Eric Von Hippel; it’s downloadable free here.)

One place where “symmetry” doesn’t hold true, says Kelly, is attention. That is, the only real scarcity on the Web is the attention of the user who surfs from one Web site to another — individuals only have a limited amount of attention. In a similar vein, connected back to his advice on copying is: the only values that cannot be copied are things like immediacy, insight, attention, and trust.

Moving toward a conclusion to his presentation, Kelly suggests that in the “goodness of time” (perhaps he meant what lawyers refer to as the “fullness of time” (?)), the Web will own every bit of information that humans (and machines) produce. That’s not yet true today, because there’s all kinds of digital information (spreadsheets, word processing documents, databases) that exists on PCs and mainframes; but sooner or later, he says, it will all be on the Web. Thus, ultimately we will have one big planetary computer; and it will be much more vast than Google’s current “farm” of 500,000 unreliable servers (actually, 450,000 servers, according to a June 14, 2006 article in the New York Times entitled “Hiding in Plain Sight, Google Seeks an Expansion of Power“). Kelly rattled off a series of statistics that I’m not sure I wrote down correctly: we have a billion PC chips connected to the Internet today; a million email messages per second; 8 terabytes per second of Internet traffic; 65 billion phones calls, etc., etc., etc. And within another decade, we’ll be dealing with numbers at least an order of magnitude larger.

What will all this mean? As Kelly puts it, we think we’re just clicking on Google today, but we’re actually programming it and telling it stuff. “As we search it,” he says, “it searches us.” I couldn’t help thinking of the parallels between Kelly’s vision of the future, and that of Ray Kurzweil’s The Singularity is Near: when humans transcend biology. It’s a breath-taking, and somewhat scary vision of the future, but it’s one that our children will experience — for better or worse — in their lifetimes.

Despite the length of this commentary, I should emphasize that Kelly actually covered a lot more ground in his presentation, and illustrated most of his themes, observations, recommendations, and arguments with terrific illustrations and graphics. But, as was true for all the presentations at the Amsterdam conference, none of this information was provided in advance to the attendees; and though the conference organizers have promised that all of the presentation materials will be available on their web site, it’s not there yet. Thus, I can’t promise that my review of Kelly’s presentation is complete, accurate, or even faithful to the overall themes and philosophies that he espouses. But I did the best I could, and it’s probably the best you’re gonna get, unless you buy one of his books or listen to one of his presentations.

Speaking of which: Kevin Kelly has written roughly half a dozen books, of which the one most pertinent to this presentation was a prescient 1998 book titled New Rules for the New Economy: 10 Radical Strategies for a Connected World. You can check out his Web site to see where/when he’ll be speaking somewhere near you; you can also read one of his provocative articles about the future of books in a May 14, 2006 The New York Times article entitled “Scan This Book!“; and you can follow his current thoughts on his semi-blog. If his presentation in Amsterdam was any indication, then any or all of these investigations will be well worth your time.