The Pre-Singularity Economic Environment
Accelerating change is a subtle and complex topic. As Rolf Landauer and others have noted, universal processes of computation, independent of specific technical platformseach of which are outmoded increasingly quicklyappear to be on a path that involves no physical limit to their continued exponential growth for the foreseeable future. If this is true, they are doing something very different from social and politicolegal systems, which follow much more cyclic and human-dependent dynamics.
Economic systems, which show cyclic and chaotic patterns on top of a base of accelerating inflation-adjusted productivity, are the most important bridge between the cyclic biological and the (presently) exponential technological world.
Many individuals are beginning to recognize the dramatically exponential rate of change and, in the long term, market valuation increase that exists in the computer and computer-driven industries. Many analysts and managers are beginning to recognize that every business, regardless of its product or service, can be evaluated as a better or worse generator, user, and architect of information and information processing systems.
Yet in this environment of increasing recognition of the computational nature of business, any new computational technology (such as the internet, or wireless, or MEMS) will be periodically exploited by large groups of self-interested participants to generate colossal "bubbles" or "hype waves" of temporary market valuation, followed by inevitable crashes. These societally-generated boom-and-bust fluctuations are superimposed on top of the real longer-term picture of accelerating growth. Such promotional bubbles and crashes bring instability to a market that is already chaotic due to ever more rapid technological change. Does all this suggest that future markets may one day become too volatile for individual investors?
This does not appear to be the case for several reasons. Investors learn to reduce their risk by strategies such as diversification, long term (10+ year) holds, and professional fund management. Marketmakers and legislators also learn to decrease volatility, by improving the rules. It has been proposed, for example, that investors in all new IPO offerings be required to leave their money in the new company for a minimum of 12-36 months, to eliminate the current incentive for institutional players to "pump and flip" new stock offerings. Such reforms would go a long way to reduce the Ponzi schemes that continually emerge in imperfect, plutocratic capitalist societies, yet they are slow to emerge, and there are penalties (lower returns, capital migration) for the capital markets that implement them first. Effective market legislation, like campaign finance reform, is a haphazard process, but increasingly likely as global transparency, social quantification, and digital democracy advance in coming decades.
But even in absence of better regulation today, we may note that it is always the earliest hype waves in any new industry that are the most overblown, and the earliest technology and automation revolutions that are the most economically volatile and damaging to individual businesses. The parties who have lost investment assets in the bubble, always the large majority of the parties involved in market-rigging schemes, slowly learn discrimination in the presence of positive and negative hype, thus dampening the waves. A similar discrimination is quickly learned in the adoption of new technology -- again, after making the mistake of either ignoring it or overreacting to it in early periods.
This does not suggest that hype waves and countering "negative hype" recessions, driven both by self-interest and technology disruptions, will ever disappear. There are many who legitimately profit from both hype waves and recessionary busts. A hype boom in any sector changes the rules of business temporarily, by selecting for a subset of companies which have temporary mass support, and thus attempt to become good at continually purchasing other companies with their inflated stock (consider Cisco or JDSU in the 1990's). This may in turn keep their valuation higher for a longer (but still limited) period, speeding the emergence of larger and more complex companies in that sector. While such conglomeration can clearly stifle innovation at one level, the standardization and convergence which results may stimulate it at another, and occasionally such efforts increase the rate of true growth in an industry that the public considers particularly valuable. Likewise, periods of expansion and infrastructure growth, especially in deflationary and counterrecessionary products and services, can be most affordably accomplished in a recessionary phase, when costs for this expansion are uncharacteristically cheap.
Yet the record of the last century seems to show that despite the chaos of successive new technologies and the regular interference of new hype waves (both manic and depressive), we have seen only temporary positive or negative deviations of market value away from the long-term picture of a smooth exponential curve of accelerating growth in inflation-adjusted valuation. Furthermore, the deviations due to disrupting technology and hype have clearly become either more time-compressed or less dramatic (and occasionally, both) the more complex our market-based society becomes.
For example, the Great Depression of 1929 lasted about a decade in the U.S. and was quite severe. Yet a few 1920's paychecks invested broadly and simply in the technology giants of the day, like Ford or CTRC (the precursor of IBM) would be worth multi-millions today, even accounting for intervening depressions and bankruptcies. Furthermore, all subsequent U.S. depressions have been both much shorter and less extreme. The Japanese experienced a 10 year "Lost Decade" in the 1990's, but this depression was itself a response to a previous boom, and was quite gentle (Japan has simply, and perhaps temporarily, lost its recent claim to top economic status). The U.S. has recently (2000) exited a five year temporary and unrealistic boom in internet and technology valuations, and we should therefore expect a corresponding corrective (and again unrealistic, relative to the long term curve) depression in those valuations. But if history is our teacher, this disruption will be even gentler on our society (if not on particular companies or technology investors!) than past temporary deviations from the exponential curve, and should itself last no longer than five years (e.g., 2005), if not less. [2005 Note: The next mini tech boom occurred even more rapidly than this, in fact. 2009 Note: The U.S. may now be in our own five to ten year "Lost Decade" due to our major politicolegal mismanagement.]
Understanding this big picture of the pre-singularity environment allows one a large measure of immunity against the innumerable manias and panics of the daily business, investment, and technology communities.
It has been suggested that the introduction of human-independent computational systems (AI's) might represent a disruption on a scale never seen before, such that human-based methods of valuation will become irrelevant. I would agree that human-centric perspectives will at some point be replaced by, and must evolve into, a "transhuman" perspective, but I would argue strongly against this being perceived as fundamentally disruptive by the human beings involved, for many reasons, better introduced in my forthcoming book.
When we learn to model accelerating change as disruptive only on some scales, and continuous on others, when we realize that the technological singularity will represent only a partial cognitive singularity for human actors, then we can use the benefit of history to understand the constrained trajectory of the financial market as a complex system. In this light, perhaps the easiest way to forsee why the markets which presently exist to solve human problems will not be destroyed once technology has surpassed unmodified humanity is to examine the record of computational technology's effect on society to date, as we have briefly done above. The more computationally complex any system becomes -- whether it be a society or a self-organizing AI -- the more self-balancing, integrative, and convergent it becomes.
The bottom line seems to be that even with the continual disruptions and volatility we should expect in the economic world in coming years, these disruptions will cause decreasing real damage to the markets and societies they spring from. Therefore we will continue to witness accelerating levels of investment in and performance from the technology sectors, and particularly the computer-dependent technology sectors, within all developing (change-permissive) countries.
Bottom line: It's going to be both a satisfying and continually surprising century.
Now that the first great dot com boom has busted, and the land-grab fever has subsided, we have an opportunity to take a clearer look at just what is presently being accelerated in the business domain. Finding the true measures of economic productivity, in a society driven by ever more intangible assets, is a constant challenge, central to all our market capitalizations and investment activity.
First, we must mention computational capacity. Exponentiating measures of processing, memory and long term storage, inputs, outputs, and bandwidth are all highly predictable, apparently developmental aspects of the acceleration. Billions are invested in new business plans based on these assumed capacities, and in general, these short term (1-3 year) mostly proprietary extrapolations (e.g. PriceWaterhouseCoopers annual technology surveys) turn out to be quite accurate. This is the often overlooked realm of futurism (what I call "research-predictive future studies") that gets too little billing, even though it is consistently useful and surprisingly accurate.
While new computational capacity makes new applications and business models viable at particular future dates, it is always much less clear that the new solutions will be needed in the marketplace, particularly when aimed at change-saturated consumers, or even at time-pressed businesses. Much more predictable is that the technological infrastructure itself will need these new computational capacities to optimize its own development, as technology never saturates its own need for additional computation. Indeed, technology's demand for excess computation seems to steadily increase the more autocatalytic and evolutionary its structure becomes. Thus businesses like Cisco, Intel, and Xilinx are in very special positions with long term value propositions, as long as they can continue to reinvent the hardware platforms they provide for technology development.
There are also new modes of computation. Wireless is today a leading example of this, and wireless performance is on its own exponentiating curve. Customer adoption of the new platform, however, is far less predictable, as it is influenced by usability, reliability, and network effects (e.g., how many other people have fax machines, or cellphones) that add significantly more uncertainty to the prediction problem.
Additionally, there are new developmental paradigms, such as evolutionary computation (evolvable hardware, genetic algorithms and classifier systems, etc.). Such alternative approaches to complexity development have their own minimum set of gestational conditions needed before they can begin their own autocatalytic replication, variation, and selection for the solution of ever larger classes of human problems, and thus benefit from the economic investment cycle.
Perhaps the dominant technological story of the coming decades will be the rise of sophisticated network computation (e.g., Google, and its successors).When the IBM PC emerged in 1981, several pundits expected that networking them in the next few years (using those old, rubber phone cup modems) would be even more important than running applications. Unfortunately, they were about fifteen years too early in their predictions--the complexity of networking, and its then-severe bandwidth limitations, kept this application constrained on the PC platform, relegating it instead to specialized devices (fax machines) and social applications (BBS's). Yet the first company to finally solve this problem in a way that consumers perceived as simple, America Online, was able to unleash a strong (and delayed) exponentiation effect.
Today, as network computation has reached its critical developmental stage (one that evolutionary computation will apparently reach in a decade or so hence), the dominant story of business acceleration will involve businesses that add political, social, economic, and technological intelligence to the PC and server network.
Companies like Yahoo (for its time), and now Google (using a superior distributed and scalable computational search architecture) have done a lot to improve the utility of the internet. Nevertheless, we remain in the still-very-early, still-very-stupid, essentially flat stage of the network's developmental trajectory. Much more interesting things are in store. Read Paul Ford's fun, illuminating, and somewhat time-accelerated future scenario "August 2009: How Google Beat eBay and Amazon to the Semantic Web," for a rough idea of what's on the horizon for the basic infrastructure of network intelligence in the next few decades.
In particular, as I have written elsewhere, it appears that the next major developmental paradigm for planetary computation will be a Conversational Interface, to emerge some time circa 2020, in the present estimation of several who work in this field. It is worth understanding this, and asking ourselves how we and our institutions will be changed in a world where we talk to our machines, supervising them as they do a growing range of increasingly useful, algorithmically-defined tasks.
Many, many other stories will be central to the transition, of course. Envisioning them, creating them, and betting on them is a fascinating and deeply rewarding challenge. We wish you great success!
China, with more than 10 times the U.S. working population, a much higher GDP reinvestment rate (50% vs. 12%) and a much faster GDP growth rate, will clearly continue to ascend in economic and political importance in coming decades. Given this knowledge, a number of our readers have questioned whether China will be poised for world leadership in the post-CI environment, some time after 2020. We have certainly seen other empires (Rome, Spain, Portugal, France, UK) rise and fall from positions of leadership.
Will the U.S., with its tremendous debt, distractions of affluence, and substantially smaller technical populace also lose its 1) economic, 2) politicolegal, and 3) sociocultural edge? This is a fascinating question that we will briefly address in three parts.
First, it does now (2002) seem clear that with regard to leadership in the production of economic value, as measured by total GDP and GDP growth rate, we can expect continued acceleration and increasing leadership from the developing Asian (and to a much lesser extent, Indian) economies. As long as we remain in the pre-singularity era, we may expect that human ingenuity, even more than machine automation, will continue to be the primary driver of technological and economic change.
The teeming masses of creative people in all the world's pro-development countries are presently disconnected and underutilized. I would argue that the world's economy is in the process of discovering that in the late Information Age, far more global productivity can come from hooking up the most motivated of the Bottom Three Billion to the emerging world network than from pumping more education or development dollars into the economies of the First World. Most developed world brains are already going as fast as they can (have reached a "saturation" plateau, as determined by natural human limits and the present sophistication of our technology), and are actively looking for ways to slow down and enjoy the fruits of their labors. But most humans in the developing world have yet to taste those fruits, and in the most change-permissive cultures, they are powerfully driven to achieve a higher standard of living.
In hindsight, I expect that we will percieve the 21st century as a period where significant leveling of the 20th century's GDP divides occurred, a time when increasing numbers of Third World countries (but unfortunately, still only a minority of them) became key players in the global technological economy, beginning in a range of outsourceable domains (e.g., commodity manufacturing, technical support). This leveling seems due in large part simply to the maturity of the networks (transportation, information, communication, sociopolitical) connecting the world's countries, and the new practicality of creating transnational enterprises on an increasingly technological planet. Just as the transcontinental railroad allowed U.S. wealth and culture to become bi-coastal in the half-century following, so too can we expect the world's economic and technological wealth to be increasingly evenly distributed across all countries that are most desirous of playing the globalization game.
The economic hardships presently being faced in the United States and other first world countries, as we see production jobs move to the developing world, presage a new era where the average standard of living in the most rapidly globalizing blocs of the third world (Asia, India, certain countries in Latin America, Europe and even a few nations in Africa and the Middle East) must rise more rapidly than the already saturated economic conditions of the First World.
I don't have good data on this assertion yet, just isolated examples a strong intuition that national income disparities, on a planet of finite sphericity and increasing linkage density, must follow pendular dynamics of divergence and convergence. Let me know if you'd like to work on a paper in this area.
Secondly, with regard to politicolegal leadership, I don't think the propostion of Asia as world leader has credibility, at least for the next half-century. From my experience, world political leadership is driven more by social developmental histories, and the favorable cultural and geographical conditions that drive that development. For a brief primer on that topic, I'd suggest Jared Diamond's Guns, Germs, and Steel, 1999, for a valuable longitudinal view on the conditions that have historically driven development in societies.
A major assumption is that Fukuyama, in End of History, 1992, is essentially correct in seeing capitalist liberal democracy as a developmental destiny for all modern governments (though what he didn't realize then is how short this "End" period would be, prior to the emergence of the technological singularity and planetized global intelligence). Liberty-promoting, individualistic Western democratic capitalism, with increasing social moderation and safety nets, is the last form of politicolegal structure we are likely to see before human surpassing intelligence emerges.
Given this assumption, I suspect it will take several more decades for Asia, even as it gains economic leadership, to develop all the institutions and social contracts (see Hernando de Soto, The Mystery of Capital, 2002 ), and to learn all the slowly-determined social lessons that have brought the West to our current level of development. Where will the U.S. be by then? Further politically developed or in social decline? The choice seems ours to make, as far as I can tell at present.
In the interim, while China, Korea, Japan, and the less developed Asian tigers are developing their capacities for broadly inclusive social leadership, it seems likely that the U.S. will continue to have a "most favored culture" status for at least two or three more generations (40 to 60 more years), driving a substantial (but declining) fraction of the world's most liquid foreign investment, and the immigration of many of the world's best innovators, to our tolerant "mongrel shores." Europe, prioritizing nationalist and socialist values over economic progress, has had its own difficulties by comparison with us, though the EU is clearly a learning organization. Thus it remains possible that the U.S. could lose its politicolegal leadership not to Asia, but to the innovative policies of some of the European drivers, such as Germany, or Scandinavia several generations ahead.
Certainly 40 or more years from now the tide of world political leadership may start shifting noticably to Asia, or not, but I think that battle is ours to win or lose in coming years. Empires don't fall quickly, and history shows that if they maintain military might, as we most certainly will, they don't need to produce much in their capitals beyond protocols, political rulesets, and cultural standards that are considered fair and attractive to all concerned.
What about sociocultural leadership? There I think the future is clear. Western culture, with all of its benefits and flaws, increasingly wins in democratic societies the world over, as individuals choose freedom and comfort, and voluntary collectives, over control and responsibility to the state. After generations of toiling, we want to relax and let our machines serve us, and be free to pursue lifestyles that are less productive and far more casual than the past. But we have to bring development, education, and employment to all cultures for them to see the value of such cultural freedoms. To date we've been slow in rising to that responsibility.
If we wish to maintain our politcolegal leadership, we have a great challenge ahead, to keep crafting our democratic society into an increasingly multi-lateral, accountable, transparent, tolerant, and fine-grained powerhouse of individual freedom, creativity, and economic and social opportunity. It is also clear that most opportunity that the U.S. is involved with, from this point forward, will be global in nature, not simply local. This insight is ours to collectively realize, or to lose in bits and pieces to other nations in the struggle.
At the same time, technology will continue on its own amazing and increasingly self-determining developmental course. I presently expect that circa 2050 to 2070 we'll enter an Autonomy Age, a time when our self-improving technological and electronic systems have become an even more important issue than numbers of creative human brains or political leadership, with regard to realizing all the kinds of futures that humans deeply care about. Again, we will see.
Let us know your opinions on these fascinating issues.