Reading "The Singularity is Near"

I have been working my way through Ray Kurzweil's "The Singularity is Near" (TSiN) over the past few days, having been the fortunate recipient of a review copy. The book might alternatively be titled "The Modern Futurist Consensus: a Review" or "Damien Broderick's The Spike: the Extended Remix." Those of us who have haunted transhumanist enclaves in the past few years (or more) are unlikely find new ideas here, but the book serves a most useful purpose in bringing the best and brightest of transhumanist, futurist themes and thinking all together under one roof, in a popularist manner, with a unifying, easily-marketed theme. It's been done before - by the aforementioned Damien Broderick, amongst others - but not quite as comprehensively. This sort of book is something of a necessary precursor to wider advocacy and education in today's culture; a pleasant irony, given the subject matter, and one could debate where in the present S-curve in the evolution of futurist thought TSiN fits.

My own two cents thrown into the ring say that the class of future portrayed in TSiN is something of a foregone conclusion. It's quite likely that we'll all be wildly, humorously wrong about the details of implementation, culture and usage, but - barring existential catastrophe or disaster - the technological capabilities discussed in TSiN will come to pass. The human brain will be reverse engineered, simulated and improved upon. The same goes for the human body; radical life extension is one desirable outcome of this engineering process. We will merge with our machines as nanotechnology and molecular manufacturing become mature technologies. Recursively self-improving general artificial intelligence will develop, and then life will really get interesting very quickly. And so forth ... the question is not whether these things will happen, but rather when they will happen - and more importantly, are we going to be alive and in good health to see this wondrous future?

As you might guess, my criticisms of TSiN center around the timeline predictions for development of new technologies, the acceleration of the rate of discovery, and the management of complexity. I made a stab at discussing this last item recently in connection with Arnold Kling's comments on TSiN (which are well worth reading, by the way):

Progress towards general (and/or strong) artificial intelligence (AI) - a grail for many transhumanists and other futurists - has been slower than we'd like. The level of difficulty has been consistently underestimated in the past, and I see this as one part of a larger underestimation of any form of complexity management. You may recall seeing this idea put forward in a variety of 1990s writing on the topic of nanotechnology; the production of millions of nanorobots wasn't thought to be as hard as the process of controlling and managing those nanorobots in a useful fashion - strategies for information processing are as much the key to future medical technologies as nanoscale and molecular manufacturing. Complexity is hard, both to manage and estimate in advance.

Now replace "nanorobot" with "human cell" and that's where we are today with biotechnology. Biological systems - such as your body, or even just a small piece of it - are immensely complex. The reason researchers can make meaningful progress today with medical technology such as gene therapies and stem cell research is that they are, effectively, tweaking settings on existing machinery that largely handles the complexity management itself. Our grasp of how things work - based on our ability to process information and build the tools required to gather information and effect change - is now adequate for this task, just as it is almost adequate to guide existing biological machinery to build replacement tissue and organs in a useful, controlled manner. But it seems to me to be a very large leap - in terms of managing complexity - to go from where we are today to reach the point of, for example, replacing biochemically complex systems within the body with artificial substitutes. Or reverse-engineering the brain, that sort of thing.

Kurzweil's commentary on types of complexity in TSiN is a good read - and one of the better explanations for the layman I've seen - but it seems a little disconnected from the actual business of dealing with complexity in ways that matter. My take on it all is that science is largely the process of discovering keys to complexity; by this I mean finding algorithms, recipies or methodologies that enable us humans to understand and manage complexity that would otherwise be beyond us. To take an applied example, manipulating stem cells through comparatively simple procedures enables scientists to perform tasks - the regeneration of age-related tissue damage - that they cannot even monitor in detail, let alone control. A simpler and more abstract example would be the mathematics and physics of atoms, comparatively simple equations that we can use to describe very complex collections of objects and behaviors.

We humans are in the process of building tools that enable us to create or meaningfully interact with ever-greater complexity, and computers are at the heart of it, but this process is one in which our individual, unaided capacities for complexity management are not increasing. Humans are still humans as of this decade, and the keys we utilize have to be useable at our level. I view the speeding of progress as part and parcel of building a larger capacity for discovering and utilizing the keys to complexity. This, as Kurzweil makes the case in TSiN, is a process that is growing exponentially, and we are moving out of the timespan in which exponential growth appears more linear.

There is one important area of complexity management in which we seem to be making little headway, however: the organization of humans in business and research. For all that we can now accomplish with faster computers and enormous leaps in telecommunications, we don't seem to have made significant inroads in getting large numbers of humans to cooperate efficiently. As Arnold Kling points out, that the excessive use of Who Moved My Cheese? is even in the running as something to try is not a good sign. I've been involved in more technological attempts to improve efficiency in large organizations, and the state of the art is not pretty - nor especially effective in the grand scheme of things.

I am prepared to go out on a limb here, as I have done before, and say that business and research cycles that involve standard-issue humans are incompressible beneath a certain duration - they cannot be made to happen much faster than is possible today:

I'm dubious about large reductions in the length of business or research cycles through technology while humans are still in the loop. You can certainly make the process cheaper and better, meaning that more attempts at a given business or research model will operate in parallel, but there is a point past which the length of the business cycle cannot be easily compressed. That point is very much a function of the human element: meetings, fundraising, decisions, organizational friction, and so forth - all very time-consuming and proven very resistant to improvements in the time taken.

This is not to say that they cannot be made cheaper. But cheaper doesn't equate to faster business and research cycles; rather, it means that any given problem will be tackled by many more parallel attempts. The professionals are joined by skilled amateurs, the priesthood dissolves, and everyone with a will to work gets in on the action. In this sort of a market, any given problem (what business model works, how does this disease process kill people, what does this biochemical signal do) is more likely to be solved in a single cycle of innovation. Biotechnology is not too many years away from this state of affairs, a repetition of what is currently taking place in the software development industry. If matters become cheap enough, people will be willing to risk ventures and research on incomplete solutions, on untested business models, and thus shortcut the existing cycle - but all to many forms of development are not vulnerable to this sort of shortcut. The answers cannot always be guessed or jumped to on the basis of incomplete work.

Back in the deep end, expensive projects mean conservative funding organizations, which means organizational matters proceed at a slow pace. This is a defining characteristic of our time: we have blindingly fast rates of research and technological advances once the money is on the table, but the cycles of business, fundraising and research are still chained to the old human timetable. I regard this incompressibility of the business or research cycle - the fact that a given iota of progress cannot be accomplished as fast as technology allows because of human organizational factors, and there is a certain minimum length of time taken to accomplish this iota of progress - as a form of limit on exponential growth, one we are now hitting up against.

Kurzweil's Singularity is a Vingean slow burn across a decade, driven by recursively self-improving AI, enhanced human intelligence and the merger of the two. Interestingly, Kurzweil employs much the same arguments against a hard takeoff scenario - in which these processes of self-improvement in AI occur in a matter of hours or days - as I am employing against his proposed timescale: complexity must be managed and there are limits as to how fast this can happen. But artificial intelligence, or improved human intelligence, most likely through machine enhancement, is at the heart of the process. Intelligence can be thought of as the capacity for dealing with complexity; if we improve this capacity, then all the old limits we worked within can be pushed outwards. We don't need to search for keys to complexity if we can manage the complexity directly. Once the process of intelligence enhancement begins in earnest, then we can start to talk about compressing business cycles that existed due to the limits of present day human workers, individually and collectively.

Until we start pushing these limits, we're still stuck with the slow human organizational friction, limits on complexity management, and a limit on exponential growth. Couple this with slow progress towards both organizational efficiency and the development of general artificial intelligence, and this is why I believe that Kurzweil is optimistic by at least a decade or two.

So how does this all fold into healthy life extension? Well, physical immortality is one obvious product of singularity-level nanotechnology, biotechnology and complexity management. There are no known barriers in physics to the construction of nanomedical systems capable of simultaneously managing, repairing - or replacing - every cell in our bodies. Even something as complex as the sum of all your cells can in principle be kept in the best possible shape for as long as you like - "all" it takes is knowledge, the future tools of nanoscale engineering and powerful enough computers. But when do we get there? This is the question, and it is one that shapes the actions of futurists and transhumanists. There are many who believe that the best sort of activism and advocacy for the future - even for healthy life extension - is in the area of artificial intelligence, because making self-improving intelligence arrive earlier will lead to all other currently pressing problems, such as age-related degeneration and death, being rendered trivial in the mid to long term. Obviously, I'm not in that camp: I'm sufficiently dubious about Kurzweil-like timescales - based on my views as set forth above - to think that we need to be tackling the problem of aging first.

Technorati tags: , , , ,