« Zefram Cochran's To-Do List | Main | More on Alternative Energy »

Economic Model for the Singularity

Over on Warrior Class Blog, Will Brown poses a strategy for an economic model of the Singularity. He begins by putting some definition in place around the idea of money.

Money provides a transportable mechanism for assigning value to things under variable circumstance, both in the present and in predicted future circumstance. It offers a mechanism for determining the worth of something under varying circumstance relative to other things. It further creates the means for arriving at a mutually acceptable exchange of things real and ephemeral between disparate people. It also, and here we arrive at an often little recognised consideration, creates the motivation for recording these valuations for future (or distant) consideration. In other words, money gave rise to writing.

Personally, I think money is a means of increasing MEST density. As John Smart explains it:

Life's history has been doing more and more (universal computation) with less and less (physical resources, MEST per standard computation), and we will soon be doing almost everything with virtually nothing. I call this driver MEST (Matter, Energy, Space, and Time) -compression, -efficiency, or -density, and it appears to be an unrealized developmental attractor for all complex systems.

Where Will asserts that money was a driver behind the development of writing, I think John would say that both money and writing are products of this ongoing process of doing more and more with less and less. Writing encodes knowledge so that it need not be memorized and can be shared over time. Writing can encode information related to transactions, while money, as it evolves, comes to encode the transaction itself.

The first transactions that humans engaged in were straightforward bartering. So many bushels of grain equals so many big varieties of fish or some larger number of smaller fish. Barter is awkward and time-consuming. As Will points out, the introduction of coins represented a kind of economic Singularity in its own right -- value was abstracted (encoded), providing a huge leap in the number of transactions that people could engage in.

But that was just the beginning. Let's jump ahead to the Renaissance and the introduction of paper money. The original bank notes allowed a merchant to complete a transaction without actually having to lug chests of gold around. Just as chests of gold were a huge improvement over barter -- a chest of gold is much easier to transport than its equivalent value in cattle or slaves -- bank notes were an enormous improvement over chests of gold coins.

Moving into more recent times, checks represent a MEST compression relative to cash, credit cards a MEST compression relative checks, and PayPal a MEST compression relative to credit cards.

The evolution of commerce parallels the evolution of technology and the ethical evolution of human society, which itself may simply be playing out the script handed to it by this tendency to increase MEST density.

Take slavery. Ultimately, it was abandoned (by the West, anyway, and most of the world) because economic enticements to work and adopting a model of wage slavery were more productive. This struggle was vividly dramatized by the American Civil War. Why did slavery come to an end in the South? Because the slave-powered Confederacy was no match for the industrialized Union. The Yankees had MEST density on their side.

It wasn't just that slavery was wrong and needed to end. That was true, but it wasn't enough. Slavery was just as wrong when Spartacus led his rebellion against the Roman Empire. But in that pre-industrial world, the Romans had the greater MEST density working for them.

We can find parallels in ethical developments such as environmental consciousness and animal rights. Societies tend to adopt these ideas as soon as they can "afford" it; that is, as soon as their implementation would not hinder the ongoing compression of matter, energy, space, and time in overall economic activity.

Because of the increase in MEST density that it represented, industrialization was more productive than pre-industrial individual contribution. Today, post-industrial (even de-industrialized) individual contribution may prove more productive than industrialization ever was. From slave to serf to factory worker to middle manager to freelancer, our freedom has grown not because the world has become nicer, but because the road to niceness and the road to MEST compression appear to be on a parallel course. Let's hope that continues to be the case.

Of course, not every step in the sequence has been a net positive move. A hunter-gatherer may have enjoyed more freedom than his great-great-great grandchild living as a slave in a proto-city-state. And a farm laborer may have lost freedom and quality of life by taking a factory job (although he gained something in the deal, or he probably wouldn't have made the move.) But overall, we seem to be enjoying more liberty as a result of the continuous MEST spiral.

Where does that eventually take us? Will writes:

[I]t seems likely to me that the principal “item” of exchange beyond our physical selves (our interactions with each other) will be the products of our imaginations.

I like the sound of that. But if we buy and sell the products of our imaginations, what currency will we use? Cory Doctorow has provided one possible answer. If reputation and imagination are what drive the future economy, the best advice you can give a young person would be to be creative, be a good person, and be effective at whatever you choose to do.

Interestingly, that's pretty much the advice I would have given anyway.

UPDATE: Will Brown comments:

My basic point, Phil, was that we will continue to need some mechanism for exchange whatever our individual abilities may allow us, at least for as long as we remain recognisably human.

I think I have to dis-agree with John Smart's concept (or at least your presentation of it) as being wrong ways 'round though. MEST is a process that occurs when conditions permit, not that creates conditions themselves (we can argue the metaphysics of the effect of potential, but my position is that percieved benefit is the actual driver and that MEST is simply an expression of refinment of that perception to added, previously unpercieved, benefit). Thus, money gave rise to accounting (record keeping anyway), which gave rise to writing, which together put in place the means for the MEST process to occur.

Your Spartacus example supports this, I believe, since the "slave army" handily defeated several Roman armies in the course of the rebellion. As long as the slaves continued to advance, Rome's MEST density potential couldn't be realised on the field of battle. It was only after Sparticus and company attempted to conquer a part of the republic (to persue the same MEST density potential) that Roman forces defeated them.

I think that something quite like the present e-bay model will become the dominant trade mechanism in the coming years, which will require some form of universally accepted transactional medium like money. Why re-create the wheel? That said, I also think much of what will be traded will be personal services (and those transactions will likely make present-day stock derivitive trades look simplistic by comparison) and the ephemeral "products" of our imaginations, like entertainment, education, consultation and simple advice. I am less concerned by the question of what currency will be used than I am that some form of currency will be used. Given that stipulation, a strategy to transition from the present model to that future model seems useful.

I think my presentation of John Smart's ideas might have been overly brisk. In the piece I linked, he presents a highly controversial idea in his Second Law of Technology:

2. Humans are selective catalysts, not controllers, of technological evolutionary development.

Read that in the contect of his First Law of Development:

1. The universe is an evolutionary developmental system, with both a rigged long-term developmental outcome and an unknowable short-term evolutionary path.

So the perceived benefit may be the apparent driver for any particular instance of progress -- generally represented by an increase in MEST density. We develop things (technology, social structures) that increase MEST density because of the benefits these things bring us, not because we have any particular interest in increasing MEST density. But if we take a broader view, we see that we (human beings) might well be a developmental optimum in a universe that was already hard at work increasing MEST density for billions of years before we existed.


Phil, this seems a more suitable place (if you recall our prior short discussion) to discuss my interpretation of what "the Singularity" would be. Namely, at a perhaps too simple level, it's a phase change from a less inefficient use of resources (not just physical resources but also algorithmic ones) to a more efficient (not necessarily optimal) one. Here, I use "phase change" in the sense of a physical system, namely there's an abrupt change in one or more properties of the system. I'm not sure what would be suitable properties, but the "MEST density" idea above is a good start and perhaps would suffice on its own as an indicator of significant phase changes.

I think we're in general agreement that current technology development is adding a lot of capability that isn't well used. Moore's law still holds but computers aren't becoming exponentially more valuable or more useful as a result. Human longevity is steadily improving, but at a snail's pace given the trillions of dollars thrown at the problem. Human education is still viable, but there's been a long trend towards specialization for centuries and no sign of it stopping. In other words, we haven't reached the limits of our current system, but we've built up this huge amount of capability and have a very narrow understanding of how to use it. I think this is somewhat analogous to physical systems like supercooled fluids where the system is stable but past the point of a phase change (ie, below the freezing point of that liquid at the given conditions). Even a small nudge (like the introduction a particle of the substance as a solid) can trigger the phase change.

Anyway, with the ability to build something more intelligent than us and to iterate that process comes the advent of "the Singularity" and the nudge that begins crystalization of a new phase, maybe. Here's where I start to diverge from the standard viewpoint.

First, there need not be the Singularity. Ie, no abrupt changes, no accelerating change (well anything beyond exponential rate), no point beyond which things become unknownable. It just might be like the development of computers, a smooth exponential increase in MEST density (or your favorite metric) over decades or centuries, with the destination known well ahead of time.

I see two possible reasons. The limitations of the physical universe may be hard even for a higher level of intelligence. Unlikely, but even nanotech and virtual entities need energy and need ways to dump heat. There's even limits on how much information you can cram into a physical region (it's proportional to surface area). Physical systems will probably remain slow. More likely, the need for control will slow development. Even if things get way out of hand for humans, it doesn't mean that everyone will be out of the loop. After all, you really don't plant some nanotech in the backyard and just hope, while it's implementing the restructuring of the Virgo Supercluster, that it got around to building that cool palace you wanted. Everyone wants control over what happens. Some will have this capability in at least a limited fashion. This desire and more importantly capability for control will prevent a lot of Singularity scenarios, IMHO.

Second, there's the matter of the Singularity as a single point. It need not be a single point in space-time at which the Big Thing happens. Phase changes can be messy with different systems shifting at different times much as the components of a chocolate chip cookie can melt (and/or burn) at different temperatures. Some things might not change at all. My bet is that a power engineer from today will still be able to understand the energy collection mechanisms of tomorrow. No one is going to develope a better classical algorithm for sorting ordered lists (and the quantum version is probably explored well enough that that isn't going to improve much either). Meatspace transportation isn't probably getting much better than it currently is. You'd still probably walk a lot.

Third, there's the matter of prediction. A common concept is that the Singularity is this barrier past which future events can't be predicted. This seems ok until you ask "By whom?" Maybe the 14 year old junior high school student in Iowa isn't prepared, but that doesn't mean that the World.Net.Mind isn't on top of the situation for the next few years. It might even already have things scoped out till the end of the Universe, though I imagine even plans of superintellects are superfluous at those timescales.

Fourth, there are various concepts that accompany Singularity predictions that I just don't like due to my upbringing as a mathematician. For example, "accelerating change" as propagated need not be "non-linear". A particular pet peeve of mine is the interpretation of Moore's law. I find Moore's Law to be quite linear. Yes, exponential growth is classic accelerating change. After all, the rate of change *and* the acceleration are both proportional to the amount of change which as noted is exponentially increasing. But the proper scales for Moore's Law are time and the logarithm of transistor density (or the log of the number of computations per second, if you use that variant). After all, it's a big deal to add a transistor to a five transister IC. You've increased transistor density by 20%! But it's just not as exciting when add one more to a few billion transistors.

Moore's Law instead should be seen as the manifestation of a phase state where humans fairly rapidly built up computers and where one could extend the line and see that amazing things were on the horizon. Before that, "computer" was a job title. After all, who knows? But there seems to be plenty of life left in parallel processing and other distributed computing concepts.

And to grind this ax a little more, in the reasonable scale of time versus log-transistor density, Moore's Law isn't "accelerating change". True accelerating change would in my opinion be relatively short term since it is faster than exponential growth (I believe it's called "superexponential growth" though that term is used in unrelated ways). In fact, this sort of growth so unusual, that it probably indicates some sort of phase change. Eg, some physicists use it to classify economic conditions as "bubbles" and "anti-bubbles" (superexponential growth and decay respectively). Anything that resembles the Singularity would probably have substantial superexponential growth of key indicators (like MEST density).

Finally, here's my take on what the Singularity would look like if it happens. I think computational artificial intelligence is the real driver. Even if the labrats take over, their computers will be where the real action is. I see this for several reasons. Manipulating objects in virtual space is virtually cost-free. You can simulate expensive, complex systems and substantially cut the cost of developing new technology. No other technology has this degree of synergy with everything. Huge amounts of data can be stored and manipulated. Computer systems scale. That means once you implement an action in a computer, that action can be repeated on a ridiculous scale. Maybe it needs a lot of hardware or some coders to maintain it, but little else has the same orders of magnitude utility of the computer. The raw capabilities are astounding.

Often one considers a measure of intelligence to be the ability to solve complex problems. However, there are stupid algorithms like brute force (routinely steered by much more intelligent heuristic tools) and genetic algorithms that can solve amazing problems. An intelligence which can interact with programs on the timescale of the program would be powerful indeed.

My take is that there would be a substantial improvement in the utility of existing computers. The demand for computational power might be so fierce that someone developes nanotech manufacturing just to meet it. The effectiveness of capital would jump tremendously. For example, these days, if we threw a few trillion dollars at longevity, we'd probably extend lifespace by a number of years maybe decades. During the Singularity, that same money would solve the longevity problem and fulfill a bunch of transhumanist dreams on the side. Well, if things turn out that way.

I don't know if it's really relevant, but the acronym "MEST" was coined by L. Ron Hubbard and I gather it has a specific connotation with respect to the Scientology religion. So we might not be interpreting the term in the way that the author meant. Having said that, I don't think there's much room here for confusion.

Karl --

Great stuff. Just a couple of quick responses:

- Moore's Law is not accelerating change; Moore's law enables accelerating change, as does all technological development. In a very short period of time (relative to the rest of human history), we've gone from horse-delivered mail to telegraph to telephone to IM. Moore's Law is one of the enablers of this kind of progress.

- The phase change idea is right on the money. John Smart uses the example of the driving of the last stake on the transcontinental railroad. In that one moment, the time required to travel from the East Coast to the West Coast went from weeks to days.

- My understanding is that MEST, both the the concept and the acronym, pre-date L. Ron Hubbard's use of the term. These kinds of associations are often made; many believe that Eric Drexler coined the term "nanotechnology," when in fact he did not. I note that Wikipedia has this wrong. If anyone has a reference to the actual origins of MEST, I'd be interested in hearing about it.


I really dig your blog and particularly the thinking and analysis related to this post. :)

A few semi-random thoughts:

1. Money seems like one of the early Interactive Communication Technologies (ICTs - technologies like cave paintings, chants, whistling, songs, language, books, the telegraph, radio, television, the world wide web, metaverse) and fits nicely on a developmental continuum enabled by ongoing MEST(I) Compression.

2. Money allows for MESTI "teleportation." (just a more technical way of stating what has already been stated)

3. The ongoing evolution / development of ICTs, enabled by local MESTI Compression, is fueling exponential growth in Technology, Communciation, Information and potentially Human Intelligence.

4. Uber-Speculation: John Smart's Second Law: "Humans are selective catalysts, not controllers, of technological evolutionary development." makes sense from a biomimicry / cosmomimicry standpoint. We are compulsively uncovering new concepts on a distributed basis. And we are not conscious of the broader implications and the "big WHY?" But over time, as we evolve our perspective, we are controlling more and more of evolution and will continue to do so. The ramifications of this are interesting. Will we evolve to understand and empathize with the whole universe?

5. The ongoing MESTI Compression of money is allowing for better quantification of neurons and is facilitating broader networking and increasing the value of neural networks. I can't wait for some neuroeconomist to link brain behavior to MESTI compression and broader evolution.

6. Money and writing are both ICTs that exist interdependently on a developmental continuum.

Keep up the AWESOME work!!


Post a comment