![]() |
| Smart's 'Laws' of Complex Systems |
|
Outline |
||||
| Laws of Development [1][2][3] Laws of Information Theory [1][2][3][4]
|
Laws of Technology [1][2][3][4] Laws of Prediction [1][2][3] |
|
||
|
These edicts aren't scientific laws, of course, but they may be important developmental constraints on Earth's complex systems (e.g., highly statistically probable, on average, in all Earth-like environments), potentially universal moral principles, or at least useful rules of thumb. There are many more than these that we could propose, but I consider the following particularly important to keep in mind as we construct a foresighted 21st-century systems theory. Other authors have also championed several of these. I've made some attributions where known, and will make additional ones as memory serves and as readers point me to prior citations. I hope you find them useful. Let me know of any others you'd highly recommend.
|
| Laws of Development | 1. The universe is
an evolutionary developmental system, with both a rigged and statistically
predictable long-term developmental outcome and myriad unknowable, unpredictable
short-term evolutionary paths. 2. Inner space, not outer space, is the apparent
constrained developmental destiny of increasingly complex systems in the
universe (also known as the "Law of MEST Compression, MEST Efficiency,
or MEST Density"). 3. We are engaged in an asymptotic approach to
universal computational limits, and an apparent, effective computational
closure (also known as the "Law of Locally Asymptotic Computation"),
a form of path dependent developmental optimization at the universal scale. |
| Laws of Information Theory | 1. The Cosmic Watermark
Hypothesis: Informational Intelligence (average distributed complexity
(ADC)), a product of two-way communication in a collective of evolutionary
systems, grows superexponentially at the leading edge of local development. 2. The Empirical Ethics Hypothesis: Informational
Interdependence (breadth and depth of symbiosis, or non-zero sum
interactions) a product of two-way communication in a collective of evolutionary
systems, grows superexponentially at the leading edge of local development. 3. The Child-Proof Universe Hypothesis: Informational
Immunity (ADC resilience to catastrophe), a product of two-way communication
in a collective of evolutionary systems, grows superexponentially at the
leading edge of local development. Immune systems work very well, in general, and even in those instances where they fail, they are generally quite benign in their damage to the network, though their failure can be devastating to the individual. In example after example, the immune learning which occurs with any catastrophe always seems to statistically increase the average distributed complexity (ADC) of the local network, if not the individual. This hypothesis has valuable implications for ways we can use our growing understanding of the lever of immunity to aid the stable development of our increasingly human-surpassing technological intelligence. 4. The Incomplete Universe Hypothesis: Informational
Incompleteness (a zone of intractability) is a permanent feature of local
computation. |
| Laws of Technology | 1. Technology learns
ten million times faster than you do. 2. Humans are selective catalysts, not controllers,
of technological evolutionary development. It would be futile, for example, to try to stop the global adoption of the wheel, or electricity, or computing, or human-competitive autonomous intelligence (A.I.), or even the ability of a handful of motivated individuals to be able to engineer superpathogens in their basement in 2100, as all of these appear to be statistically inevitable technological developments within the network of human civilizations. Nevertheless, we have the power to locally delay (with regulatory or social adoption "speedbumps") and even temporarily regress any particular developmental outcome (as Japan and China did with handguns for several centuries, for example), and to create our own local evolutionary pathways to these eventually inevitable capacities, and to reward the emergence of environmental conditions that make these capacities nonthreatening when they finally do emerge. Thus we might ensure the emergence of A.I.'s that have been incrementally proven to be safe via stepwise development, we might spur the global ability to manufacture and deliver effective antidotes to any biological pathogen by 2050, and we might catalyze the emergence of enough global development and transparency to prevent most individual terrorism attempts from emerging, while simultaneously providing fine grained assurances of individual liberties and meaningful employment for those who seek it. In the same manner, as we come to realize that even humanity as a whole does not control the technological world system, we can nevertheless strongly influence the evolutionary path of a range of harmful technological applications (e.g., nuclear weapons proliferation, CBW research, pesticides and pollutants, first generation nuclear power technology), while accelerating the development of balancing and beneficial technologies (e.g., communications, computation, automation, transparency, immune systems R&D), and phasing them in in ways that improve, rather than disrupt, human political, economic, and cultural systems. 3. The first generation of any technology is often
dehumanizing. The second generation is generally ambivalent to humanity.
The third generation, with luck, becomes net humanizing. It is a constant challenge to the designers and users of any technology to seek ways to minimize the duration and extent of the negative externalities we so often see with any new technological deployment. Yet even with our best intentions, we seem to take three steps forward, two steps back, six steps forward, two steps back the eternal dance of accelerating change. Those who would criticize a technology as dehumanizing and unacceptable would do well to realize that developmental advances have always always been associated with disruption and some degree of dehumanization, as we learn to adapt to the new order of things. Fortunately, the faster and more intelligent our technology becomes, the greater the social standard we can hold it to, and the sooner we can move it from dehumanization and disruption to enhancement in its net effect. A recent example is takeback legislation (cradle-to-cradle recycling of manufactured goods) a third generation of manufacturing that has increased the sustainability of European manufacturers without significantly impacting their competitiveness. There are good arguments that sustainable takeback programs would have been impossible in a world without supply chain automation, recycling automation, and other technological advances, but there is a time when such advances become affordable, and it is incumbent upon us to recognize when that time has arrived, and to advocate for the next generation to to emerge. 4. Technology should serve, self-actualize,
and improve people and their cultures, not degrade, addict, or enslave
them in 'structural violence.' Perhaps the most obvious example would be the automobile, a tool most of us must use to compete in the modern world (we have little choice in the matter), and yet one that claims 40,000+ lives in the U.S. and 1.2 million+ lives in the world every single year. Leaving fossil fuels aside for the moment, which also have great health, environmental, and political costs, and focusing solely on safety, just a little thought applied to the issue makes it clear that we could make many low-cost improvements to our automobiles and the political-legal structure around them that might cut these terrible costs in human lives to half of their present daily toll, or less. Clearly intelligent machines will be driving us, with vastly lower fatalities, just a few decades (or generations) hence. But what can we do in the meantime? The list of presently unutilized technological aids to this problem, as for so many other social problems, is quite long. Consider modifications to the car, such as four point harnesses, internal occupant sleds, crash webbing, internal airbags, bumper airbags, telemetry-assisted braking (where sharp braking in one car induces braking in all cars in the vicinity), and even helmets (which some of us would wear if they were retractable, and if there were an insurance break for wearing them). There are many potential modifications to the environment (rumble strips, lower speed limits, etc.) and to legal requirements (drivers ed, driver training, license renewal). Some of these should be required, some should receive R&D and prize money to stimulate innovation, some should be subsidized with insurance incentives for their use, some should be promoted in drivers ed, and some left to the free market. The public apathy that exists today with respect to the safety of automobile technology is itself a clear form of structural violence, as the true social costs of the technology are hidden from the citizen, the putative ultimate decider in our democracy. Such apathy will only change when voting citizens are allowed, and incentivized, to realize the real ongoing cost of such technologies to our culture. Imagine a country with the foresight to pay for the production of two minute accident reports, delivered to all car navigation computers, summarizing all the fatal and critical accidents that have occurred in the driver's county since the car was last driven, the currrent trend in such accidents per capita and per miles driven, and some brief safety lessons that may be drawn from them. Now imagine a culture that required the in-car watching of such reports on at least a weekly basis in order for the car's ignition to turn on. Would that be a new form of enslavement or a tool for social improvement? Clearly these are issues for a democracy to decide, but we are far today from even considering them. While we can point to partial exceptions, from Singapore to Scandinavia, the modern social system promotes apathy on such issues in virtually all countries, and the transportation machine continues its killing unnoticed. We can make a similar case for structural violence in many social problems, as I will do in the second of my forthcoming books. If we seek to understand the universe and our place in it, we must recognize that technology is not only by far the most rapidly learning and the most powerful system in the human environment, but also a system that we can craft to serve, self-actualize, and improve us every day. We appear to be inevitably and progressively handing off the mantle of highest intelligence to our technological successors, but we remain responsible for our own continued improvement, as individuals and as a species. When we ignore that responsibility, when we succumb to technology's many distractions, addictions, and outright enslavements, we deny our future and remain impoverished. |
| Laws of Prediction | 1. The more things
change, the more some things stay the same. 2. Most prediction is a predictable failure. For example, at the molecular scale, human development is intrinsically unpredictable. But step back to see the big picture, and after you've seen one human life cycle you've got a good idea how developmental (not evolutionary) events will proceed in the next. And after you've seen a multiplicity of developmental cycles, at a range of matter, energy, space, and time scales, you've got a good idea what kinds of developmental events are occurring in your local environment. I can't predict which software company will be dominant in 2030, but it is a good bet that they all will be running the most sophisticated CUI network in existence. I can't tell you what computer architecture will come after MOS, but I can predict it will be vastly more MEST compressed and efficient. And in a controversial astrobiological example, while you would go broke quickly trying to predict the exact shape of humanoid life forms on other Earth-like planets, or the styles of cars that will sell best in those worlds, you can make an excellent developmentalist bet that those planets must all produce computationally dominant humanoids, that the humanoids will all be highly likely to have two eyes, bilateral symmetry, jointed limbs (possibly with an average of five fingers on each limb), and large number of other predictably convergent developmental features. Furthermore, there are great developmentalist arguments that all such planets will be very likely to invent internal combustion-based, automobile-like machines as swarm computing time-compression devices, that the dominant car body plans will involve four wheels, and that the environment must include a vast number of other universal technological archetypes, or developmental optima, such as electronic computers. And if you find any of that hard to believe, you're in good company. I'll do my best to address these issues in my book. 3. Long-term predictions of computationally-dependent
processes tend to be socially unreasonable. That's our own loss, and it impoverishes us in the present wherever we lack sufficient social foresight into the inevitable mechanisms of accelerating change. We at ASF will do our part in coming years to attempt to rectify our species' cultural proclivity for ignoring the historically unreasonable growth of computation.
|
| Additions? Missed Attributions? Disagreements? I look forward to your
comments. Sincerely, |