Archive for July, 2014

Swarm Intelligence and Capitalism

July 31, 2014 1 comment

Recently I revisited an article that I posted once upon a time on Swarm Theory (Decentralized Problem-Solving). This is the notion that a self-governed network of entities operating on simple heuristics (e.g. bees in a hive, ants in a colony) can solve certain problems more effectively than central authority. I received a comment from the CEO of a software company in Finland that harnesses Swarm Theory within specially designed social networks to solve problems.

As I was searching around for more information on Swarm Theory, I came across this quote on a blog: “SWARM OR HIVE INTELLIGENCE: Communism Without The Corruption! […] Individual Capitalism Vs. The Collective! One Thing Is Certain: Capitalism Cannot Be The End Result For Humanity Or We And Very Likely Earth Itself Are Doomed!”

Interesting perspective.

But at the same time, there’s an interesting point to be made: this quote has its logic exactly backwards. Sure, a leaderless structure like an ant colony brings to mind the visual idea of a commune. But Communism as we’ve experienced it (and Socialism for that matter) are associated with central planning, which is the opposite of Swarm Theory. If you want to look at the economic system that bears the closest resemblance to Swarm Theory, you’ll be looking at good old Capitalism.

The Invisible Hand = Early Crowdsourcing


Since Adam Smith described the “Invisible Hand” in 1776, arch-capitalists have ascribed a magic, almost religious quality to the economic distributions within a free market. As supply lowers, prices rise. How does the system know? How does wealth accumulate with those who are meeting a strong demand? How is investment capital finding its way to the most promising opportunities? Is it God’s will that it be so?

The “invisible hand” is an instance of Swarm Theory, nothing more. The motivation to provide value in exchange for monetary reward is a simple heuristic and economic participants (workers, owners, investors, etc) are the automatons who follow it. Therefore the strengths and weaknesses of the various Capitalist systems of the world can be understood by how effectively they replicate a basic swarm system.

In his book Business Stripped Bare, Richard Branson coins the term Gaia Capitalism to describe the more environmentally and globally aware form of Capitalism to which he subscribes. This promotes the idea that there are many manifestations of Capitalism that have varying degrees of concern for long-term thinking or preservation of common goods. Should a Capitalist system have rules? Should it be transparent? Should it reward individual actors? Do regulations always make it less efficient? We can understand these questions by understanding how they apply to swarm systems.

So let’s revisit the characteristics of an effective swarm system:

  1. It is self-organized
  2. Its actors follow simple heuristics, though different classes of actors may follow different instruction sets (foragers vs patrollers, etc.)
  3. Actors must act in a diverse fashion (i.e. bees don’t search for a new home be all flying in the same direction)
  4. There is a communication mechanism by which information is shared with all actors
  5. Actions must be self determined, without imitation, cohesion or fad-following

It also must be understood that the collective result will not be perfect, but it will be ever-improving. Just because bees agree on a location for a new hive doesn’t mean that the resulting location is perfect. The quality of the new location depends on the terrain that was explored (i.e. luck), and the system that has evolved to determine consensus. Notice that it does depend on the outsized reconnaissance skills of any one super-star bee.

Notice also that Swarm Theory works to solve the problems of the collective, e.g. how to propagate and defend an ant colony, or how to allocate wealth to the providers of value and quality investment. They do not exist for the purpose of enriching individual actors at the expense of others. To be sure, there are often status hierarchies within swarm systems, and individual actors are motivated to act by their own survival, but the nature and essence of the system is to propagate the entire community.

Answering Hard Questions on Capitalism


Now, assuming the analogy between a swarm system and a Capitalist one, and also that our goal is to make the most productive system that we can, lets draw some conclusions.

Should Capitalist systems have anti-trust laws, and how vehemently should those laws be enforced? Well, are swarms more effective when actors are acting diversely or in concert? Diversely. Acting in concert through collusion or anti-competitive measures weakens the system by limiting the diversity of opportunity. The dynamic becomes indistinguishable from central planning.

In healthy Capitalist systems, should all trades take place on transparent exchanges (i.e. removing the dark pools of investment banks). Well, are swarms more effective or less effective with transparent communication? They’re more effective. In fact, they depend on the systemic aggregation of collective information so all actors can make informed decisions. Keeping information secret for private advantage weakens the system by disallowing all the actors from making effective decisions.

Do central regulations always make a Capitalist system less efficient? This one is harder to demonstrate with analogy, and it also depends on the time frame that one is talking about. I interpret “regulations” as rules and safeguards. As part of its heuristic logic, a forager ant will not leave the colony to forage for food unless it come into contact with at least four patroller ants within the space of ten seconds. This lets the ant know that it’s safe to forage. It might be more efficient for the colony in the short term if the forager ant would search for food immediately, without this safety procedure. In the short term, the chances of running into food might be higher then the chances of running into an anteater. But the colony will pay a heavier cost (population count) every time this gamble doesn’t work out.

Yet, this procedure evolved organically. It was not enforced from a central authority. It’s hard to find analogies to central rule enforcement within natural swarm systems. These systems, by definition, have no central authorities. Instead, this might be a question for digital automatons in labs.

Still, we can see from this demonstration that the strongest, most effective Capitalist systems are not necessarily the ones that provide the greatest enrichment to individual members. They will more likely be the ones to harvest most effectively the wisdom of the swarm.