Do You Really Want Effective Altruism?

There is something fundamentally different between the way nature solves complex problems and the way human beings currently do. When nature wants to solve problems beyond a certain level of complexity it uses networks. From whatever evidence we can find that remains of the billions of years of the history of this earth, we don’t see even one example of nature having solved complex problems like vision, or flight, or cognition, through single cells. Nature solves such complexity through multi-cellularity (through networks of cells). Nature also solves the problem of exponentially increasing outcomes through networks. As an example, from the two gamete cells that combine to form a fertilized ovum only able to travel centimeters within its mother’s womb, grows a blue whale consisting of quadrillions of cells that can swim thousands of kilometers through the world’s oceans. In achieving this, nature exponentially increased the outcomes achievable by the cells of the whale.

Human outcomes, whether sustainable economic growth or access to affordable education, are achieved through products and services delivered by individuals and/or organizations. Inserting a network based approach when solving human problems that might be faced anywhere along the entire product or service life cycle from research and design to manufacturing takes a different way of thinking about things. If humans were asked to manufacture something like a whale we might attempt to calculate how much raw materials it requires, how much plankton, how much water, how much oxygen, etc. But nature didn’t’ manufacture the whale in a single project. It manufactured a self-sustaining network of cells that self-assembled additional networks to become the whale.

Nature’s method for solving complex problems through the design and manufacture of organisms is somehow able to explore every region in the available design configuration space of those organisms. Nature conducts this exploration in order to find the optimal definition of any problem it is trying to solve, and in order to find the optimal solution for that problem. Humans on the other hand are more goal directed and look at problems on much shorter time scales. We aren’t able to see problems and we aren’t able to look for solutions except through the limitations of our cognitive biases. We also don’t have a universal way of looking at problems and solutions that allows us to solve problems today in a way that will help us solve problems a hundred million years from now. Nature on the other hand has continuously adapted life over at least one and a half billion years, continuously using we believe to be the same principles.

If we consider every cell to be an experiment that nature has conducted in problem-solving, or even if we consider every organism consisting of anywhere up to the quadrillions of cells in a blue whale to be such an experiment, the odds that the solutions to complex problems don’t involve replicating nature’s ability to find networks of cooperating interventions are effectively infinity to zero. In other words, the odds are the countably large number of multicellular organisms that have ever lived on earth with designs that have solved complex problems like vision or cognition, in comparison to the potentially zero number of single-celled organisms that have done so.

In a way this isn’t new. It is well known in the academic literature that networks of cooperating solutions can achieve far greater impact on any targeted outcome than stand alone solutions (which to all intents and purposes essentially compete with each other). Yet it is also well known that network based approaches are never used in tackling our greatest economic, environmental and other challenges. Of all the organizations tasked with solving our greatest and most existential issues, not one single one of those organizations has policies that even permit networks of solutions that self-organize to select the optimal definition of the problem, regardless of whether such self-organizing approaches might exponentially increase impact on targeted outcomes. Since no organization currently has a universal way of measuring the fitness of networks of interventions to achieve their targeted outcomes, they couldn’t find such solutions even if they wanted to.

Luckily, replicating nature’s problem-solving strategies in human systems, doesn’t require any major changes in organizational policies, social infrastructures, and evaluation metrics. Because self-sustaining networks of new organizations with new social infrastructures using new evaluation metrics can just as easily be created. The possibility of leveraging network effects ensures that by cooperating to do collective good those networks can reliably out-compete existing organizations that compete to achieve their own self-interest. When those new networks become large enough they might reliably subsume the old. All that might be required is to spread the concept far enough for such organizations and networks to spontaneously form and self-assemble.

Given the evidence from the uncountable number of experiments that nature has performed within the process of evolution, given that nature has not found any other solution, it’s critical to build clarity and mind share around the understanding that any approach other than creating the capacity to methodically introduce networks into the problem-solving process for complex challenges is tantamount to saying “I don’t want to solve this problem”. It’s also important to draw a distinction between the networks which research in many areas like climate change research, public health, complex systems, and AI development, already rely on, and the ability to systematically search for networks with greater fitness at achieving some collective outcome, as well as the ability to objectively assess that fitness. This is critical because the fact that we collectively continue to put 100% of our research funding into interventions that from this perspective are non-networked, and that (at least according to this model) cannot reliably solve the problems we say we are trying to solve, means that we are effectively trying to solve very different problems.

If the claims in this article are true, nothing is as important when it comes to societal problems as this network based intelligence or “collective intelligence”, and focusing on anything else is tantamount to prioritizing things other than solving those problems. For example, we might have an ideology which assigns the blame for societal problems on corporate greed. But from the perspective of this model, just like the greed of each single cell is not the reason that a billion single celled organisms can’t self organize into the network of a billion cells in a bird in order to solve the problem of flight, gaining the capacity to self-assemble social impact projects into networks that multiply their targeted outcomes also requires an infrastructure based on tools like generalized network theory and collective intelligence to facilitate that cooperation. While it’s easy to see this is true in a biological context, it is difficult to see how this could possibly be valid in a societal context where individual interests can indeed act as barriers to collective action. However, on considering that matter more closely, by definition, a system of collective intelligence with general problem-solving ability at the group level must have the capacity to potentially solve any group problem in general, and therefore must be capable of reliably optimizing collective outcomes even in the face of the power dynamics, differing ideologies, varied resource distributions, and other challenges that human societies are fraught with. From this perspective, it is limiting to view the network-based approach as only offering a valuable lens for viewing societal issues. It is much more.

There are of course limits and potential drawbacks to a network approach, as any approach will have its limitations and potential downsides. Even in nature, single-celled organisms still exist, which likely reflects the fact that some problems aren’t best solved through complex solutions. In the case of a network-based approach, one might consider issues such as the risk of creating overly complex systems, coordination challenges, and issues related to decision-making and accountability in a network. However, all of these are issues related to the fitness of a network to achieve collective outcomes. Having a model for the “fitness” of a network based system addresses all of these issues.

Because of the person-years of suffering that accumulates as a result of every day that our most pressing societal problems go unaddressed due to the lack of such a network based approach, or for every year that human potential is wasted due to the lack of solutions capable of scaling to the point that they might nurture that potential in every individual without excluding some due to lack of resources, any other way of seeing the problem might be tragic.

Networks have already been applied to complex societal problems based on an approach called Human-Centric Functional Modeling (HCFM) that represents the human systems through which we perceive and otherwise interact with the world as moving through graphs called functional state spaces, where these graphs can also hypothetically be used to describe the behavior of any living system, whether as simple as homeostasis or as complex as cognition. In these functional state spaces, all problems are defined as the lack of a path from one functional state to another, resulting in the insight that all adaptive processes through which nature solves problems are potentially generalizations of the same solution in functional state space. This approach allows us to model the collective problem-solving ability of networks of systems in terms of their ability to navigate their collective functional state space, and it allows us to see how networks create the potential to exponentially increase that ability. This in turn has led to a model for group decision-making called a General Collective Intelligence or GCI platform that copies this same solution. GCI is important because arguments can be constructed using this functional state space approach to support the prediction that GCI creates the capacity to exponentially increase the general problem-solving ability (intelligence) of groups.

HCFM might be a profoundly important innovation in that it is a function based approach to modeling that simplifies such systems to the point that drawing deep insights about those systems is reliably achievable. Whether comparing human problem-solving processes to natural ones is entirely fair or valid, and whether human problems and their contexts are fundamentally different from the issues that nature addresses, if problems in one functional state space occupied by one natural system has analogies in some functional state space occupied by human systems, then we can indeed learn from nature’s processes, without having to assume that all aspects of nature’s problem-solving strategies can or should be replicated in human systems.

An experiment has been designed to test the claim that by representing organizational processes using the graph of a functional state space, and by using that representation to discover networks of interventions, it is possible to exponentially increase impact per program dollar on any collective goals, whether AI safety, or access to affordable healthcare or education, or achieving sustainable economic growth. This experiment is the tip of the iceberg. Biological and social systems appear to be fundamentally different in many respects. While nature operates under the laws of physics, chemistry, and biology, human societies are driven by a plethora of subjective factors such as individual motivations, cultural norms, historical contexts, and economic considerations, making any direct comparison between the two systems appear to be a gross over simplification. However, from the perspective of HCFM, functional state spaces can be used to describe physical matter, chemistry, and biological systems. Human societies also have representations in the functional state space of the individual cognition, as well as in the functional state space of the collective cognition. To test the claim that any solution to a problem defined in one functional state space can potentially be generalized to apply to another, the next stage in this experiment is to validate the claim that functional state spaces can be defined to represent problems in disciplines such as mathematics or physics, and that doing so could exponentially increase our capacity to solve problems in those disciplines as well, as it is predicted to do within the cognitive domain through GCI.

Stay tuned for the announcement of an upcoming webinar that will issue a call for participation to anyone in a country that provides access to research funding, and who is eligible to be funded as a Primary Investigator, so they might help organize interested participants to conduct a multidisciplinary project to independently test this claim.

There are a number of concrete examples of network/collective intelligence based solutions projected to radically increase impact on societal problems per program dollar. For example, one proposed program introduces a self-supporting network of local businesses to fill local demand for cotton based goods in order to boost livelihoods in the agricultural sector. An analysis of this program suggested that it could feasibly increase agricultural livelihoods per program dollar spent by up to seven hundred and fifty times. But still somehow the message hasn’t been spread that if these claims are true, then nothing is as important when it comes to societal problems as collective intelligence and network based approaches, and that focusing on anything else is tantamount to not solving those problems, and tantamount to wanting to prove one’s ideology about altruism rather than wanting altruism to be optimally effective. Network-based problem-solving approaches don’t need to be universally applicable and don’t always need to be the optimal choice for this to be true. Because collective intelligence simply implies the capacity to reliably discover such solutions where it is optimal to do so.