Major refinements to sharing information have come via language, writing, the printing press and the internet. In each case, more agency has led to more productive societies. The first 30 years of the web has been about divergence; about empowering individuals to explore their niche by connecting them with others who share it. In doing so, it has exponentially increased the amount of information in the world. Language models have recently arrived as a tool for synthesising that information, and promise to enable us to wield the full capacity of the internet from a universal interface.

“…the ‘message’ of any medium or technology is the change of scale or pace or pattern that it introduces into human affairs.” — Understanding media (McLuhan, 1964)

In the short term, language models are a new medium for creative expression. In the longer term, I’d like to argue that they are a tool for building a more cooperative internet, and with some luck, a tool for building more cooperative societies.

Language models synthesise collective knowledge

“Every day, we land on the internet without a map. Instead, search is the dominant wayfinding paradigm. It is the information equivalent of exploring the local area at ground level. A search is a hypothesis, an instance of trial and error. With enough searches, we can usually get where we're going. However, we lack the context of how everything fits together...” — Me (2021)

Language models are a synthesis tool. Their key feature is to compress information — to start with an internet of information, and reduce it to a consumable artifact. The first wave of capabilities that this makes possible — search, creative generation etc — are increasingly obvious by the day. What’s less obvious is what is happening behind the scenes, and why this is important.

Consider DALL-E or Stable Diffusion: start with a prompt, and finish with an image. To make this possible, models are trained on enormous amounts of images, each originally produced by a human in some form. Each image has a label that describes aspects of its style, composition and subject. Models extract features that are common in many images with similar labels. When you write a prompt, the model measures the relationship between the words you use and features it has extracted, and combines high-scoring features into a new image. In other words, models compress information by combining concepts from many images into something new.

Now consider what happens when an artist produces a piece of work. The techniques used — brushstrokes, colour schemes, shapes, materials (“features”) — are acquired as tacit knowledge through lots of practice over time. The art produced is some emergent combination of these techniques as a conduit for externalising an internal mental state. The artist’s mechanical process is to take accumulated information (techniques and experience), and combine it into art.

Language models are able to approximate art because the procedural mechanism is compression — merging features together. The difference is that language models are carrying out the combinatorial process at a higher level of abstraction. Rather than computing over a single perspective, as is the case with an artist, language models integrate over the experiences of many artists. Users can then manifest new images from the latent combinatorial space.

We all have internal mental states that we’d like to represent with art. Arguably, the difference between artists and non-artists is that artists have the tacit knowledge to make their ideas real. Language models close this gap, enabling ‘unskilled’ individuals to realise their ideas without having to spend years acquiring the techniques and experiences. To a limited extent, models subsume and compress artistic tacit knowledge and make it wieldable via prompts.

“…the content of any medium is always another medium.” — Understanding media (McLuhan, 1964)

Each prompt, then, is a recursive computation over increasingly abstracted human inputs, collapsing thousands of dimensions into a human-readable artifact. For now, there is nothing really ‘artificial’ about it. What is really happening is that users are empowered by collective intelligence — a medium within which to wield the combined skills and experiences of millions of artists around the world. Models are a mechanism to rescue knowledge from individual brains, and make it collectively useful.

Language models are a humanising technology for informational empowerment.

Upwards information and downwards causation

“Reading used to be reserved for the clergy, to hand down unquestionable Revealed Truths to the masses. Today, it's just what everyone does. Think about a society in which science is not reserved for the clergy, to hand down unquestionable Revealed Truths to the masses, but is just what everyone does." — Bret Victor

‘Art’ is just the current thing. The same compression mechanism is true of video, text, and a host of coming applications. The pattern in each case is to break open skills that were previously reserved for a select few, and empower the masses. This observation is certainly interesting in the context of prior transitions in tools for thinking together — language, writing, the printing press and the internet.

In the short term, language models are thus a new medium for creative expression. There is no doubt that jobs based on skills which are no longer scarce will be replaced. More importantly, the newfound agency is sure to create an entirely new and higher level pattern of work.

“…the new patterns of human association tend to eliminate jobs, it is true. That is the negative result. Positively, automation creates roles for people, which is to say depth of involvement in their work and human association that our preceding mechanical technology had destroyed.” — Understanding media (McLuhan, 1964)

A new medium for creative expression and the new opportunities it opens up is a compelling proposition. My conjecture, however, is that it is the second-order effects that are worthy of more focus.

We currently live in a world in which information asymmetry is profitable. To a certain extent, asymmetry is unavoidable — the world is complex, and we can only know so much. However, much of it is also engineered by competitive dynamics in markets and other institutions. The problem is that without openly sharing information, cooperation is impossible and we are left with the externalities of competition.

Language models suggest a mechanism for progressively incentivising agents to indirectly collaborate at multiple scales. The reason is thus. Training data is a key limiting factor for deep learning implementations, such that more and higher quality data is almost always better. Meanwhile, it will quickly become clear to individuals and organisations that what can be achieved via model augmentation far outstrips what can be achieved without.

My claim, then, is that models incentivise parties to share information — models tend towards better results with more training data, and users are incentivised to provide it in order to benefit from better compression. There is a group-level benefit to many parties contributing data to a model, which in turn can act as a group selection pressure for mechanisms that prevent free-riders and encourage contribution.

Over time, models will come to exhibit a kind of downwards causation. Information arising from many parties gets compressed to become reusable by all other parties, and all parties will come to rely on the compression because it produces better results than can be achieved otherwise. Models therefore act as a focal point — a reason to openly share information without needing to directly communicate with other parties. This mechanism changes zero-sum information asymmetry into positive-sum information sharing, where each party is made stronger by wielding the collective information pool.

“I propose we move from simple feedback to downward causation when components tune behaviour in response to estimates of coarse-grained, aggregate properties.” — Coarse-graining as a downwards causation mechanism (Flack, 2017)

When many competing entities are ‘tuning’ to the same information pool — in this case, the collective information pool — it can enable a higher level of organisation to consolidate from the bottom up. In turn, the mutual information advantage sets the conditions for the evolution of more cooperative relationships between lower level parties.

"The transition to the next stage occurs… by joining [parties] into a single whole with the formation… of a control system headed by a new subsystem, which now becomes the highest controlling device in the new stage of evolution." — The phenomenon of science: a cybernetic approach to human evolution (Turchin, 1977)

Consider a simple example in art. Let’s imagine that a large community of artists agree to continuously contribute their art as training data to an open-source diffusion model or some future descendant. The model compresses all contributions and makes them available for recombination. Each contributing artist now has the combined creative force of every artist in the community. My conjecture is that even for professional artists, the range of what can be achieved through this augmentation outstrips what can be achieved alone. This is probably true today, never mind in a few years’ time. If so, then there is a strong positive-sum mechanic at work — everyone contributing makes everyone’s work better. Community formation arises naturally, because there are strong incentives for cooperative relationships between community members.

At some point, a cooperative network empowered by shared models outcompetes its competitors because of the mutual information advantage. Individual network members have an information advantage relative to their external counterparts. There exists, then, a growing incentive to participate cooperatively rather than compete against the network. Within the network, the emergence of social norms and relevant monitoring mechanisms can ensure a stable base of cooperation, such that everyone contributes data and everyone can access the model.

“Viewed as a cultural evolutionary process, new entrants to the population would be more likely to adopt the preference ordering of those who obtained the higher material payoffs in the immediate past (Boyd and Richerson, 1985). Those who were less successful would tend to learn the values of those who had achieved higher material rewards (Borgers and Sarin, 1997).” — Collective action and the evolution of social norms (Ostrom, 2000)

Flack’s paper suggests that downwards causation via compression is not unique to language models, but instead a generalised mechanism in adaptive systems throughout the natural world (macaque social hierarchies, for example). Which is to say that if correct, it is extremely robust, scalable and has been continually selected for at many levels of the evolutionary hierarchy.

One might worry that everyone using the same models leads to a convergence on a local minimum, which would be anti-evolutionary. However, I think the opposite is true. The coming diversity of models and use cases is such that individuals and groups will use a plethora of different models in their own ways across a massive range of applications. The result is surely a dense, interwoven and constantly evolving combinatorial network integrating groups in complex ways, sufficient to overcome local minima and enable continuous evolution.

“…[parties] need not be in agreement about how to best to tune to these variables. The degree of agreement will depend on whether decision-making—in learning or evolutionary time—is influenced by other types of heterogeneity” — Coarse-graining as a downwards causation mechanism (Flack, 2017)

Language models are a substrate for more collaborative societies

“The various ways in which the knowledge on which people base their plans is communicated to them is the crucial problem for any theory explaining the economic process, and the problem of what is the best way of utilizing knowledge initially dispersed among all the people is at least one of the main problems of economic policy—or of designing an efficient economic system.” — The use of knowledge in society (Hayek, 1945)

Today, models are toys for creative expression. Tomorrow, they will be tools for increasingly autonomous technological production. If the central problem of economic society is the problem of productively organising agents with imperfect information, then a new mechanism of aggregating that information is likely to enable new economic structures.

It’s worth stating my philosophical assumption here that deep learning and related models are not, and are unlikely to become conscious beings. I think we are dealing with tools, not beings, that will quickly get better and eventually outcompete humans in all relevant productive capacities. But, productive replicators is all they will ever be. This is not to say that entrusting widespread economic activity to unconscious robots is without risk, but the distinction between tools and beings is a critically important one. Arguably, the reason that unconscious AGIs are risky is precisely because they are unconscious, and thus incapable of the social intelligence required to understand what it means to be ‘human-aligned’.

"…the cognitive capacity to represent the formal properties of mind differs from the cognitive capacity to represent the subjective properties of mind (Seager 2006). Thus a notional zombie Hyper-Autist robot running a symbolic AI program on an ultrapowerful digital computer with a classical von Neumann architecture may be beneficent or maleficent in its behaviour toward sentient beings." — The biointelligence explosion (Pearce, 2012)

My suggestion, then, is that models and their descendants can be viewed as tools with the potential to create more collaborative and more collectively intelligent economic societies.

To recap, it seems inevitable that models will become more effective in productive capacities currently entrusted to humans. Augmentation with models will therefore empower individuals and groups to do things they otherwise cannot do. Models, at least in the short term, are dependent on information arising from people, and more/better data makes more effective models. Effective models thus create a focal point for community formation, and groups at increasing scale will derive a selective advantage from model-mediated positive-sum information sharing. If so, then it further seems inevitable that the incentives of economic society shift towards an open information ecology — in which everyone openly and honestly shares information for collective benefit — and away from closed market information ecologies fixed on information asymmetry.

The most optimistic viewpoint is that this mechanism could enable coordination between groups at scale, for many of the same reasons. For example, many companies could coordinate with each other via a higher level model, or a complex of individuals, companies and even nations. Geoffrey West and co’s work indicates that there is no theoretical upper limit to information-dominated systems. If so, then open information ecologies coordinated by models could operate at community, city, national and planetary scale, with a selective advantage accruing to larger systems. As such, models may present focal points to overcome dangerous multipolar incentive patterns derived from information asymmetry.

A combination of cryptography and artificial intelligence may well be able to implement systems working on this principle. Of course, technological implementations are only part of the problem. Working on a collective principle of mass information sharing is a huge social step in a world dominated by competitive incentives. The idea that artists, companies and nations might openly and proactively contribute all internal information for collective consumption is frankly hard to imagine. Critically, it requires a shift in values away from individuals and towards collective networks at all scales, superseding individual incentives with collective incentives.

Multipolar traps are nothing new, and have been around for as long as life on Earth. Fortunately, stable solutions to multipolar traps have been continuously selected for at many levels of the evolutionary hierarchy. These are most visible in cases of major evolutionary transitions, in which groups of competing individuals cooperate to form new kinds of individuals at higher levels of organisation.

In search of solutions for the future, we can learn a lot from understanding what is common to the transitions of the past. At the very least, they show us that it is possible to establish stable cooperation at scale. Further, perhaps, they can provide clues for how we might establish an evolutionary trajectory towards intelligent coordination at planetary scale.

In this article, I aim to provide an accessible entry point to what the major evolutionary transitions have been, what we know about how they have happened and what has been common between them.

What is a major evolutionary transition?

Major evolutionary transitions represent rare points in evolutionary history in which groups of competing individuals came together to form a new cooperating whole at a higher level of organisation.

The most familiar example is the development of human societies. Way back in our evolutionary past, proto-humans lived primarily as self-sufficient individuals. As we fast forward through human history, however, we see the development of cooperative societies at increasing scale.

Of course, this has been a slow evolutionary trajectory. It likely began during confrontations between scavengers, establishing a precedent for cooperation to acquire food. The selective advantage of cooperating groups likely explains the emergence of hunter-gatherer bands, which we can observe as far back as 1.8 million years ago. From there, human ancestors begin to display evidence of cooperative breeding, which likely co-evolved with prosocial behaviours to enable advanced cognitive abilities like understanding of shared intentionality.

This progression laid many of the foundations for the evolution of language, and it’s here that we begin to see a major transition. Language provided a medium for the development of complex societies, leading humans to become increasingly dependent on the wider collective through the division of labour — something that is particularly visible in modern developed economies. What makes humans powerful is many people working together in real-time, accumulating information about the world over many generations. The ‘major evolutionary transition’ for humans, therefore, represents the protracted change from self-sufficient individuals to a mutually-dependent collective over many generations.

Human societies are far from the only example of a major transition in evolutionary history. Current consensus is that there have been six other major transitions, including protocells, prokaryotes, eukaryotes, plastids, multicellularity and eusociality.

While each transition is unique in its own way, there has been a concerted effort to analyse them together in the hope of extracting common principles.

Transitions occur when selection rises to the group level

Consider a population consisting of several groups competing against each other. Within each group, natural selection favours traits that maximise individual fitness regardless of their effect on group as a whole. Cooperative behaviours tend to incur an additional cost, either requiring more energy or a reduced proportion of a shared benefit. We should expect, then, that cooperative behaviours cannot evolve, because acting out of self-interest tends to have a higher payoff.

Of course, this is not what we see in practice. Broadly, major evolutionary transitions are characterised by the establishment of mass cooperation with limited realised conflict, producing a new individual at a higher level of organisation. Cooperation does evolve as a stable strategy.

The evolution of cooperation in the context of major transitions has been a controversial question over the past few decades, and I will avoid wading too far into the weeds. Instead, I will lean on the emerging consensus that there exists an additional level of selection which selects for cooperative traits that are advantageous at the level of the group instead of the level of the individual. The aggregate of such traits gives the group an advantage compared to other groups in the population. Group selection does not completely overcome individual selection, however, and selection instead jostles between both levels simultaneously — what evolves in the population results from their relative strength.

“Sometimes one level prevails, resulting in a homogenous population of public good providers (when between-group selection prevails) or free-riders (when within-group selection prevails). In other cases, frequency-dependent effects cause a mix of types to coexist in the total population.” — Generalizing the core design principles for the efficacy of groups (Wilson, 2013)

The majority of us have experienced life in a company, and so competing levels of selection is intuitive. Each of us has our own interests, like getting credit, getting promoted and making more money. But, we also know that we are ultimately better off when our team and the company as a whole succeeds, because it makes our individual position more valuable. As you pay attention, you’ll notice that these dynamics are at play in any group situation, and so we contend with the competition between individual selection (what is best for us) and group selection (what is best for the group) everyday.

Looking back through evolutionary history, these two forces have been in constant tension, shaping evolution at multiple levels of individuality simultaneously.

Group selection rests on the idea that there are some outcomes that are only achievable when parties cooperate, and that those outcomes provide an evolutionary advantage to group members. It is these outcomes that get selected for — synergistic interactions between individuals leading to the emergence of an advantageous group aggregate.

“If cooperation is to evolve, non-additive, or synergistic, fitness interactions are needed. If two or more cooperating individuals achieve something that a similar number of isolated individuals cannot, the preconditions exist"” — Major evolutionary transitions (Smith & Szathmary, 1995)

When the right conditions present themselves, group selection becomes the primary evolutionary driver, and major transitions to a new level of individuality are able to occur. Groups become collectives when traits that benefit the group are selected for above selfish traits.

This pattern is visible in all prior major transitions, and once complete, the resulting organism tends to become ecologically dominant. Eusocial insects, for example, began somewhere in the evolutionary past as an individualistic species, and at some point evolved to become a collective superorganism. Eusocial superorganisms now make up more than 50% of all insect biomass on Earth. Multicellular organisms follow a similar pattern. When it works, cooperation really works.

“Selfishness beats altruism within groups. Altruistic groups beat selfish groups. Everything else is commentary.” — Rethinking the theoretical foundation of sociobiology (Wilson & Wilson, 2006)

The concept of ‘major transitions’ implies that a new level of individuality not only enables superorganisms to outcompete less cooperative species, but opens up an entirely new evolutionary state space. The difference between cooperative human societies and the next smartest primates is not just that we outcompete them for resources — we exist within a fundamentally different space of possibilities. Similarly, multicellularity opened the door to all higher-order species in the animal, plant and fungal kingdoms; a possibility space simply not available to unicellular eukaryotes.

Integrating individuals through group selection thus represents a powerful mechanism for establishing new levels of the evolutionary hierarchy. However, precisely explaining how group selection comes to dominate has proved extremely challenging. The challenge emerges because in every prior case, competing individuals have not been capable of perceiving the ultimate benefit of cooperation. Thus, it’s not enough to say that cooperation was established because superorganisms ultimately outcompete selfish individuals, because many intermediate generations had to exist between a predominantly selfish species and the evolution of a cooperative superorganism.

While we do not yet have a complete account, there is a lot to learn from studying what we do know.

Phases of evolutionary transitions

Bourke has observed that transitions tend to progress through three stages, and this provides a useful frame. First, origin — there is an initial contact point in which a group advantage is established. Second, maintenance — mechanisms develop which cement cooperation by suppressing individual incentives. Finally, transformation — a new level of individuality is consolidated with sophisticated mechanisms to produce a new unit of evolution.

1. Origin via an initial cooperative benefit

To make major transitions possible, there must be some kind of initial advantage to cooperation — a ‘first contact'. In this phase, cooperation can be established without the need for sophisticated mechanisms, producing a positive selection pressure which enables those mechanisms to develop over many generations.

It is perhaps easiest to grasp this idea in the context of human ancestors. We might imagine that at some point in our evolutionary past, there were confrontations of multiple individuals hunting the same prey, or scavenging the same food. No doubt many of these early confrontations ended in violence. At some point, however, our ancestors would have worked out that if the group cooperates, then it can be to everyone’s benefit. For example, it’s easier to hunt for large prey, or to get food from hard-to-reach places. Variation in genetics likely predisposed some individuals towards primitive forms of communication, giving them the ability to gesture or form words. Individuals with these mutations were more likely to survive in groups of others who had a similar mutation, producing a kind of group-selected advantage that could lead to cooperation becoming more common.

The evolution of multicellularity can be thought of in a similar way. Multicellularity arose many times independently, giving us some indication that there was a significant selection pressure in the ancient unicellular world. We probably cannot know for sure, but it’s likely that the initial advantage was linked to size. Imagine all organisms lined up according to size. It’s always possible for there to be a new biggest organism, and if there is an ecological niche to be filled by a new biggest organism, then it’s likely that it will be selected for. Since being bigger makes it possible to isolate from the outside world and protect genes, it seems likely that some niches could be exploited better by larger organisms.

“In other words, the direction of the selection for size change depends solely on the ecological opportunities, and there is always room at the top. When the earth was populated with nothing but single cells, selection opportunities for multicellularity must have been inexhaustible." — The origins of multicellularity (Bonner, 1998)

Multicellularity and human societies are alike in that they represent fraternal transitions, in which each individual in the higher cooperative unit is initially the same. In these cases, the initial advantage of cooperation is most likely drawn from economies of scale — combining the same abilities across multiple individuals to achieve higher efficiency. This lays the foundation for division of labour to evolve later.

Fraternal transitions are in this way distinct from egalitarian transitions, which bring together individuals with different capabilities. In such cases, the initial advantage of cooperation is found in the emergent properties of combining different capabilities, where a division of labour is automatically established.

The transition to eukaryotic cells is a characteristic example. The general consensus seems to be that eukaryotic cells began when a prokaryotic cell ingested a mitochondria. Mitochondria produce lots of energy, and this advantage could be experienced relatively ‘immediately’ via a natural division in labour. The excess energy probably fuelled more experimentation with a higher number of genes, enabling a big jump in complexity relative to prokaryotic cells through the evolution of many different kinds of cellular compartments.

“Most organisms opt for the libertarian route, going it alone. Those that cooperate make two different kinds of alliances: egalitarian and fraternal.” — Cooperators since life began (Queller, 1997)

2. Maintenance via mechanisms that suppress free-riding

So, transitions happen when selection rises to the level of the group, via an initial cooperative advantage. However, group selection never completely overcomes individual selection, and the evolutionary pressure to act as a free-rider continues to exist. In this scenario, we should expect that free-riders do very well in populations of cooperators, since they get the benefit of cooperation without paying the cost. As such, we might expect that free-riders infiltrate budding cooperative populations and eventually become dominant.

Stable populations of cooperators are able to evolve because selection acting at the group level favours the evolution of mechanisms that suppress free-riding. In simpler terms, groups evolve rewards for cooperative behaviour and punishments for selfish behaviour.

For eusocial insects, achieving superorganism status required mechanisms to overcome many potential arenas of conflict, including sex allocation, queen rearing and male rearing. It’s no surprise that conflict is found primarily in reproductive traits — each individual is within its evolutionary incentives to ensure its replication, and so overcoming the individual drive to reproduce is key to forming a new level of individuality.

Mechanisms for controlling reproductive conflict in eusocial insects primarily include kinship, coercion and constraint. For example, workers in eusocial societies are generally capable of laying unfertilised eggs. However, workers that do so are otherwise unproductive and reduce the efficiency of the colony as a whole — a natural ‘tragedy of the commons’. To overcome this, reproduction is policed both by the queen (who retains reproductive dominance) and by other workers. In honey bees, it has been predicted that without policing, roughly 54% of workers would reproduce. With policing, however, only about 0.1% of workers become reproductive. Interestingly, ‘bottom-up’ worker policing (as opposed to ‘top-down’ queen policing) is about 98% effective.

In the transition to multicellularity, one can imagine a similar problem. There are economies of scale advantages to multicellular configurations, granted, but each cell begins as an independent replicator and will be selected to reproduce. How, then, do multicellular organisms overcome this paradox? Well, it’s likely that there was variation in the temporal dynamics of collective cell division among newly multicellular individuals. Some groups likely divided closer together, and others further apart. We observe that multicellular organisms tend to synchronise their replication, and it is likely that synchronisation was selected for because it reduced the chance of individual cells proliferating at the expense of others. As this was selected for, more efficient mechanisms for facilitating synchronisation evolved as a result. Synchronised cell replication likely cemented cooperation sufficiently to enable further evolution at the group level.

In human systems, mechanisms of suppressing conflict and selfish behaviour are everywhere. We continuously create systems of rules and regulations at every scale of human interaction — within families and friend groups, all the way up to international legal systems. Such systems of rules help to overcome game theoretic incentives enough to ensure that ‘individuals’ can predictably experience the benefits of cooperation, safe in the knowledge that free-riders will be penalised. Many anthropological studies show this process in early human societies, while Elinor Ostrom’s work shows how it remains at the foundation of effective cooperation today.

3. Transformation to a new evolutionary unit

With the suppression of free-riding, the groundwork is laid for mechanisms which complete the transformation to a new level of individuality. The newly emerging superorganism moves past a point of no return, where sophisticated mechanisms ensure a stable base of mass cooperation. It’s at this point that we no longer see the parts and begin to see the new whole — in all practicality, selection among the original parts is no longer a significant evolutionary force.

In the transition from single-celled to multicellular organisms, the integrity of the budding new whole is dependent on a strong division of labour, as epitomised in the development of the soma (“body” cells) and the germ-line (“sex” cells). The vast majority of cells are confined to somatic responsibilities in this arrangement, foregoing their ability to contribute to the next generation. However, each somatic cell technically maintains a copy of the genetic material required to reproduce, and would be within its evolutionary incentives to compete to participate in the germ-line.

In order to overcome this multipolar trap, cells have evolved epigenetic mechanisms which affect how differentiated cells are able to interpret and express information stored in DNA. The presence of different epigenetic factors explains why daughter cells arising in many different tissues can re-establish the differentiated functioning of their parent cells while having the same core DNA. Without epigenetic factors, daughter cells would not develop the specialised characteristics of their parents and would instead switch to inappropriate states that follow their individual reproductive incentives.

“Without efficient transmission of epigenetic information, the component cells of new multicellular organisms would have switched to inappropriate states that would have compromised the success of the individual as a whole. Epigenetic inheritance is thus one of the reasons why multicellular plants and animals retained their coherence in spite of the turnover of their component cells.” — The evolution of information in the major transitions (Jablonka & Lamb, 2006)

Over time, epigenetic information inherited by differentiated cells consolidates the division of labour. Differentiated cells can’t participate in the germ-line, because it would require erasing all the accumulated epigenetic information that enables their differentiation. Epigenetic factors thus act as mechanisms for consolidating a new level of individuality, ensuring that differentiated cells are largely only capable of synergistic actions that maintain group welfare — prior individuals move past a point of no return.

In eusocial insects, we saw that conflict resolution mechanisms like policing probably acted as intermediate mechanisms to facilitate cooperation. What we see in fully evolved eusocial society is a kind of caste system, composed of some reproductive and some non-reproductive individuals, continuing the pattern of extreme divisions of labour. As with multicellular organisms, the caste system in partnership with conflict suppression mechanisms largely ensures that individuals are, in all practicality, no longer capable of acting out of the group interest.

For human societies, we are still arguably undergoing the transformation phase. We’ve never been more dependent on the wider collective to live our lives, especially in the world’s developed economies. More complex economies introduce more division of labour — information becomes increasingly distributed and each individual becomes more specialised. In turn, we each become progressively more reliant on the wider economy to produce the vast majority of the things we use everyday.

One way to look at this is to ask yourself how much of your current daily lifestyle could be maintained if it was just you alone with no contact with society. For most of us, this should reveal just how dependent we are on the accumulation of knowledge, technology and social structure over generations. Here lies the transformation with respect to primate and early human societies — it is no longer practical for us to go backwards.

New mechanisms for transmitting information

Looking at the transitions through a sequence of stages, we continually hit on the concept of information and how it is transmitted between individuals. Perhaps the most interesting aspect of transitions is that each time, a new level of individuality is correlated with the emergence of a new medium for storing and transmitting information.

Prokaryotes are the first time in evolutionary history that we see what we commonly refer to as ‘cells’ evolve. Prokaryotic cells established symbiosis among networks of molecules to produce a new individual at a higher level of organisation. In order to replicate themselves as new wholes, prokaryotes had to discover a mechanism to send information about how to reconstruct themselves and their internal networks into the future.

The mechanism they discovered was genetic code (RNA), together with ribosomes to translate that code into proteins. Where before inheritance was iconic, genetic code made inheritance symbolic. Symbolic code was selected for because code can enable arbitrarily complex molecules to be encoded by a small alphabet of components. In doing so, the transmission of information moved from limited heredity to unlimited heredity with variation, unlocking a new state space of possible environmental adaptations.

Prokaryotes laid the foundation for eukaryotes, which brought multiple cells together to form a new kind of meta-cell with functionally distinct compartments. With the added complexity of replicating a meta-cell, a new mechanism of information transmission had to evolve. This came in the form of DNA. In prokaryotic cells, RNA does the work of both storing hereditary information and enabling/regulating the production of proteins. DNA evolved from RNA to establish a division of labour, in which DNA succeeded in storing information more effectively as a more stable molecule, while RNA primarily maintained its role of enabling protein production. In doing so, the process of transcription (“reading” DNA) became separated from translation (converting DNA into proteins) and DNA/RNA each became better in their respective domain.

To understand the role of transmitting information, consider evolution as a learning process in an environmental context. As time passes, variation and selection determine adaptations that are considered ‘fit’ in the environment. Information is a way to store the results of this process. To ensure that they maintain the evolutionary advantage they have acquired through selection over generations, organisms must find a way to store and transmit information about what works between generations. Information moves from a sender (the “parent”) to a receiver (the “child”) who uses it to reconstruct an adapted state, with a little variation.

“The transmission of information between generations, whether through reproduction or through communication, requires that a receiver interprets (or processes) an informational input from a sender who was previously a receiver. When the processing by the receiver leads to the reconstruction of the same or a slightly modified organization-state as that in the sender, and when variations in the sender’s state lead to similar variations in the receiver, we can talk about the hereditary transmission of information.” — The evolution of information in the major transitions (Jablonka & Lamb, 2006)

Notice that the transition to prokaryotic cells and eukaryotic cells aggregates information previously belonging to lower-level individuals. Each individual alone acquires that information over a long evolutionary history. If either individual alone wanted to develop characteristics of the other, they would have to ‘invent’ those characteristics by themselves. As a cooperative organism, however, each benefits from merging their evolutionary histories together.

This becomes important because new evolutionary adaptations have to be constructed from combinations of components that already exist. To get to complex adaptations, like complex vision, you first need many layers of underlying capabilities (light sensors, depth perception etc). Only when you have the underlying capabilities can you ‘unlock’ complex vision by combining them together. As prokaryotic and eukaryotic cells formed by merging individuals together, there were more underlying capabilities available to try out more complex combinations.

Consider, then, that there is an abstract combinatorial search space of all possible adaptations, where each one is reachable through certain combinations. The mechanism of evolution is to search through that space. For organisms to develop adaptations of greater complexity, they have to be able to build upon the information gained from previous generations. Of course, the number of possible combinations increases exponentially, and the search space quickly becomes very large.

"…even the simplest biological objects – the polymeric sequences of proteins and RNA - postulated to play a prominent role in the origin of life - are still combinatorially complex, existing within an astronomically large space of possible molecules that are just as physically plausible, but have not been selected to exist" — Assembly theory explains and quantifies the emergence of selection and evolution (Cronin & Walker, 2022)

Part of the ultimate role of RNA and DNA, then, was to merge separate information into mutual information. As symbolic languages, they achieved this because they compress a lot of information into a small alphabet of components. As we look out into the world now, much can be explained with the combinations of just four letters.

The compression becomes important for two reasons. First of all, compression enables the *regularities *to be found between information coming from multiple sources. By identifying regularities and discarding what’s not important, compression increases mutual information to enable further cooperation. In the case of RNA and DNA, this likely meant finding the behaviours/traits that were mutually useful at the group level, and reinforcing them. Through this increase in mutual information, “cooperation“, each individual becomes able to externalise information processing to the higher collective entity (“division of labour”).

Secondly, compression makes the language more generalised. Genetic base pairings are a generalised language for describing any protein and any combination of proteins, and are thus capable of describing a huge range of possible organisms. This generality makes the ongoing process of evolutionary search more efficient, and increases the likelihood of discovering useful combinations compared to other methods. Essentially, more of the total search space becomes available — the search space available to RNA was much larger than direct molecular replication, and the search space available to DNA was significantly larger again.

“Further evolution generalizes the system so that a hyperastronomically vast combinatorial space can be sampled by evolutionary search.” — Toward major evolutionary transitions theory 2.0 (Szathmary, 2015)

As we move forwards in evolutionary time, transitions are characterised by the emergence of mechanisms that transmit information with increasing speed and level of abstraction.

Epigenetics builds on the DNA paradigm by enabling cells to inherit state (“context”). Simply inheriting code is limited in its potential because all inheritors inherit the same thing. Epigenetics enables many different types of cells to inherit the same genetic information, while varying how they use that information. There is a slight distinction here from DNA, as it enables characteristics acquired during a lifetime to be passed onto the next generation.

Bees, meanwhile, are able to communicate with other bees by using a waggle-dance to indicate the location of a food source. Essentially, this involves walking forwards while waggling at an angle relative to the position of the sun, using the hive as a medium. The angle represents the location of a food source, while the intensity of the waggle represents the distance. See this video for a quick walkthrough.

Something different is happening here compared to prior transitions — rather than transmitting information between generations, information is being transmitted socially within generations. Social transmission aggregates information arising from many bees, producing a kind of externalised map of food sources.

In a way, human language did for information transmission within generations what genetic code did for transmission between generations. Bees are only capable of communicating about the location of food. Symbolic language, on the other hand, generalises representation of information and enables communication of arbitrarily complex concepts. As with genetics, language achieves this by compressing information via a small alphabet of components that can be recursively composed together. Genetics and language are therefore similar in that they are both generalised mechanisms for transmitting information from a sender to a receiver. The difference is the timescale — memes travel faster than genes — and the level of abstraction.

The pattern here seems to be one of continuous aggregation, compression and generalisation, spiralling through levels of abstraction at faster and faster speeds:

• Aggregation — individuals come together and ‘share’ information, merging their evolutionary histories to produce the possibility of a collective aggregate that enables newly adaptive higher-order combinations.

• Compression — cooperation coevolves with new mechanisms (“languages”) for storing and transmitting aggregated information, operating through information compression. Through aggregation and compression, individuals are increasingly able to externalise information processing to the higher-order collective (“division of labour”).

• Generalisation — languages become more generalised and more capable of combining what is known to reach higher levels of complexity. This opens up a new dimension of higher-order adaptations not available to lower levels, providing a selection pressure that consolidates a new layer of organisation.

• The cycle repeats — at the new level of organisation, there becomes a selection pressure for re-aggregation.

Conclusion

So, what can we learn from our journey into biological esoterica?

Well, the first thing to realise is that there is a strong evolutionary precedent for deeply cooperative integration of previously uncooperative units. “Collective intelligence” has been continuously selected for at many levels of the biological hierarchy, where extreme cooperation between components with partial information is a rule, not an exception.

Second, each transition is similar in critical ways, progressing through phases of origin, maintenance and transformation. Shared information is core to new levels of organisation, and newly cooperative individuals coevolve with new mechanisms to aggregate, compress and generalise it for the collective benefit. Over time, there is a selection pressure for higher levels of abstraction and increasing speed of transmission.

Third, the consolidation of higher levels of organisation opens up entirely new evolutionary possibilities that simply were not accessible at the lower levels. The emergent characteristics of the higher level are not predictable from the viewpoint of lower levels.

More broadly, major transitions suggest an important lens for looking at the evolutionary process as it pertains to social systems. When we think of evolution, we tend to think ‘survival of the fittest’. The connotation is to see evolution purely as a mechanism for ruthless, zero-sum competition in which cooperation must be a glitch. Of course, it is ruthlessly competitive at each level — lions and hyenas are forever locked in a vicious blood feud until one species no longer exists. But we need only zoom in, to look back in time, to see the symphony of cooperation between individuals at lower levels combining to give meaning to ‘lion' and ‘hyena’. With a better lens, it is really cooperation all the way down.

While institutions have solved for cooperation at many scales, it is hard not to feel as though they are stuck in an individualistic paradigm. A glass ceiling seems to separate us and higher levels of organisation. It’s here that multipolar traps arise, from each following their individualistic incentives at the expense of the collective good, without understanding the possibilities available when everyone cooperates. Well, the insight from major evolutionary transitions is not only that there is an evolutionary precedent for overcoming these traps that has been continuously selected for; but that what ultimately comes of the integration is so collectively powerful and so far out of the existing space of possibilities as to be a new level of the evolutionary hierarchy.

As we look to overcome the multipolar challenges we face, there is at least one thing that distinguishes a future transition from those past. This time, we are capable of understanding at least the medium-term benefits of extreme planetary cooperation — we can project into the future and conceive of a world in which everyone cooperates. Rather than waiting on millions of years of naturally directed evolution, we can consciously engineer this world into being.

If you found this essay valuable, please consider collecting it as an NFT. Think of it like funding research, and enabling me to do more of it. Like buying a book in reverse, after you decided it was already worth it. If you do, know that it is deeply appreciated.

Pathologic. Morphologic. Demagogic. Nation logic.

We are inundated with grim predictions of our future. The threats are very real. Deep and truthful understanding is critical to mitigating them. However, if the history of human endeavour has taught us anything, it’s that the depths of the human spirit are moved by that which we might create, not that which we must avoid.

Pratītyasamutpāda, translated as dependent arising, is a central tenet across the schools of Buddhism. It holds that all phenomena are conditional, caused by another. The simplicity of this idea cannot mask its deep significance: impermanence, dependence, interconnectedness.

“When this exists, that comes to be. With the arising of this, that arises. When this does not exist, that does not come to be. With the cessation of this, that ceases.” - Samyutta Nikaya 12.61

Indigenous tradition, meanwhile, has long held a deep respect for nature, recognising human communities as embedded within a larger ecology. Such traditions have emphasised the role of psychedelic substances as agents of spiritual and ecological enlightenment. Psychedelics reliably induce a dissolution of self, revealing a sense of oneness, connection, with other and environment. Meaningless hallucination, maybe. Or, perhaps, turning the dial up on what really exists beyond the doors of perception.

“A portion of the self overflows into the outer world, into objects, which begin to live, to have another, a deeper meaning” - LSD: My Problem Child (Albert Hofmann, 1980)

For a few brave and fortunate explorers, the world above complements those revealed within. Astronauts frequently describe the “Overview Effect”: a deep change in perspective after seeing the Earth from space for the first time. A hundred thousand miles up, one is forced to appreciate the Earth for what it is: a single whole suspended in the vast expanse of space.

Science is gradually revealing the truth of our intuitions. Earth — geosphere, biosphere, soon-to-be-technosphere — is an integrated system of which we are becoming an influential force. A collective anthropogenic force of cumulative ingenuity, standing on the shoulders of 100 billion ancestors.

Slowly, then quickly, humans are integrating a world that moves as a single individual; with atoms through industrialisation and with bits through computation. We are globalised, interconnected and interdependent; a planetary public bursting from the borders of nations.

Our institutions, however, lag behind our intuitions. The industrial great leap forward has been powered by multipolar markets and nation states, exploiting evolutionary zero-sum games for social and technological progress. Fruitful though they have been, they are failing in the face of a new class of global coordination problems.

In the Western world, it feels like the vibe has shifted in the last few years. Pessimism, anger and fear seem to be more common, and not unreasonably. The problems are coming thicker and faster, triggering immune reactions in institutions built for a world that no longer exists.

Perhaps all generations feel like this. The older we get, the better it was. Or perhaps this time is different. Existential risk, climate change, energy, materials — these are intrinsically global problems that require a kind of mass collaboration not afforded by industrial society. To the contrary, a new axis of global superpowers is looming. Every day, the tipping point to inevitable conflict between the US and China is inching closer; a snake pit in which we are all unfortunately entangled.

We are conditioned to look at this familiar fight between ‘good’ and ‘evil’, ‘us’ and ‘them’, as an intrinsic part of being human. Something we have no choice but to partake in. And within the system that we’ve built, that is true. Multipolar narratives are built deep into our cultural psyche. But nation states are now at the core of risks that are becoming too great for a society of our technological capabilities, and it is them for whom the bell tolls. Time marches on. The window is open to gradually succeed them with something new, in which strong global cooperation is a core design criterion rather than an inconceivable retrofit.

The future is bright and better than we can possibly imagine. It will surprise us in ways that we cannot possibly foresee, at a speed we could not possibly predict. But to reach it, it’s crucial to realise that what got us here will not get us there. We are gradually developing the technology and the understanding needed to integrate humanity into a planetary collective, and the scale of the opportunity before us is unprecedented. Major evolutionary transitions from selfish independence to greater wholes have happened many times in the past, and it can happen again. It is far from inevitable, but there is a 13.7 billion year evolutionary tailwind propelling us forwards.

In this first instalment, I want to collect some of the threads together, and present an alternative argument broadly in support of previous rallying cries. I’ll start with a short introduction on our evolution as a collective in the context of the industrial revolution, then discuss key aspects of industrial dysfunction and finish on some key principles for what comes next.

Vectors of collective evolution

“Direct self observation is not nearly sufficient for us to know ourselves: we require history, for the past continues to flow within us in a hundred waves. We ourselves are, indeed, nothing but that which at every moment we experience of this continued flowing.” - All too Human (Nietzche, 1888)

As we’ve evolved over the past 200,000 years, our capacity to process energy and information as a collective has underwritten our march from just another competitor on the Serengeti plains, to a spacefaring apex species.

One human alone is weak and vulnerable to the natural world. This is more true today than it has ever been — we cannot produce our own food, build our own house or cure our own ailments. We certainly cannot make our own computers. Some may view this as a shameful regression of human capability, from worldly generalists to frail micro-specialists masked by the thin veil of society.

Collectively, however, humans have become so strong as to invoke a new geological epoch — the Anthropocene — in which we have become the planet’s driving geological force. Not bad for 200,000 years work, a mere 0.005% of Earth’s lifetime. The fact that we are even able to create climate change is characteristic of how far we’ve come.

Despite our collective prowess, we have a tendency to revere the individuals of our past as creators of our present; the great scientists, philosophers, writers, lone geniuses, company founders and visionary artists. We love to identify single events, single factors and single groups as determinants of our future. The simplest answer is the best, so goes Occam’s Razor.

Of course, to do so is to miss the manifold of interactions, the context, the accumulation of billions of humans working together in concert with the environment over evolutionary time. We have long outpaced genetic evolution. We evolve through technological and cultural accumulation, both fundamentally collective endeavours stretching back to the earliest human societies.

At the core of this evolution has been our growing capacity to process energy and information. To be sure, any look at something so complex as societal evolution is bound to miss a substantial amount of important details, lest it become the volumes of psychohistory. However, energy and information are crucial to understanding how we may evolve from here.

We’ll take a short tour through our evolutionary past, placing these vectors into the context of the industrial revolution.

“We are projects of collective self-creation.” — The Dawn of Everything: A New History of Humanity (Graeber & Wengrow, 2021)

Energy

Energy is the core resource of life. It is the foundation of all living processes. Growth, movement, thought, sex, homeostasis and communication all require energy. More energy is better, because more powerful systems outcompete in the evolutionary game. So, living systems self-organise to maximise their energetic power.

Humans excel at energy manipulation. Where all other organisms are confined to consuming energy endosomatically (within the body), humans began evolutionary dominance the minute we discovered fire — exosomatic energy consumption (outside the body). What began with fire, and developed with farming, hit a colossal step change through fossil fuels. It is hard to overstate the scale of the impact of fossil fuels on human civilisation.

Industrialisation was much less a ‘revolution’ and much more a gradual transition from biofuels, mainly wood, to fossil fuels as the main energy source over hundreds of years. The first official coal mine was erected in Scotland in 1575, slowly becoming the dominant fuel before hitting an inflection point in the 19th Century as industrialisation gathered steam.

Industrialisation replaces man with machine. Fossil fuels are the ghostly armies driving the machine. As such, human labour is replaced with fossil fuel labour. Nathan Hagens asks us to consider fossil fuel workers entering the workforce like new human workers. Fossil fuels are more powerful than humans — a single barrel of oil is equivalent to about 5 years of manual labour. For every human worker, there are around 100 fossil fuel workers driving our machines.

However, we don’t pay fossil fuel workers like we pay human workers. We merely pay the cost of extraction. Through this mechanism, we’ve been able to bring on billions of equivalent human workers into the global economy at extremely low cost while simultaneously increasing efficiency.

"Consider milking a cow using three methods: manual (no energy other than the human labor), semi-automated electric milking machines (1,100 kWh per cow per year, or cow-year), and fully automated milking (3,000 kWh per cow-year). The manual milker, working alone, requires 120 hours of human labor per cow-year, but the semi-automated machines require only 27 hours of labor, and full automation only 12 hours. Let’s assume that the human milker is paid $5 an hour working alone. Using electric milkers that run on electricity at $0.05 per kWh, output rises significantly and— because cheap electricity substitutes for so many human hours of labor—the wages for the milker increase to $18 per hour with semi-automated milkers and to $33 per hour with the fully automated technologies.” — Energy, Credit and the End of Growth (Nathan Hagens, 2015)

Win for all involved. Humans do less manual labour, companies make more money, the economy gets bigger and we get more stuff. For 200 years, fossil fuels have underwritten a great leap forward in the quality of human life, powering the evolution of nearly every process relevant to our lives.

First came coal-powered steam engines, prime mover of industrialisation through the 19th Century. Pumping, sawing, grinding, drilling, turning, polishing; trials of animate horses became those of inanimate steam engines.

Next came electricity. Automated production begot mass production. Electric motors roaring, humanity soaring: fertiliser, steel, consumer goods — production at scale.

Then, diesel engines and gas turbines: the prime movers of globalisation. Cargo ships, semi-trailer trucks and freight aircraft — global trade networks.

Fossil fuels are the lifeblood of the modern economy, evolving over centuries to power our collective machine muscles. They have led us to the most important technology of the industrial era, computers. The automation of automation. Of course, computers are as much about manipulating information as they are about manipulating energy, finding an analogical ancestor in the provocateur of industrialisation: the printing press.

Information

Johannes Gutenberg first invented the printing press in 1450, rousing the tides that would become liberated and educated democratic societies. Beyond just an invention that we employ, its capacity as the means by which we have been reinvented is one of classic remark: “the medium is the message”.

The cultural significance of the printing press might appear largely distinct from the purely technological ink and metal type. However, on closer inspection, we see that values are instead baked into it. Inside the plough we discovered that animals are subservient to humans. Inside the printing press we rediscovered that information should be free.

With this value at its core, the printing press was at the foundation of change in Western society throughout the Renaissance and Enlightenment periods.

Modern science differs from prior schools of thought. Phlogiston became oxidative combustion. Miasma became germ theory. Beginning with Copernicus in 1543, gathering steam with the Philosophical Transactions (1665) and reaching a head with the Principia Mathematica (1687), what changed was the commitment to writing things down, and crucially, to distributing those writings far and wide. From this commitment emerged a selection pressure for rationalism through inscription: a mundane dedication to progressively improving our representations of the world to establish truth by consensus. The Enlightenment followed, becoming the science we know today through the compounding of inscriptions debated over time.

"Thus, one more inscription, one more trick to enhance contrast, one simple device to decrease background, one coloring procedure, might be enough, all things being equal, to swing the balance of power and turn an incredible statement into a credible one which would then be passed along without further modification.” — Visualisation and cognition (Bruno Latour, 1986)

Culturally, the printing press was introduced into European societies largely living under the dogmatic umbrella of the Roman Catholic Church. Driven by its ancestor Humanism, the Renaissance broke free from its theological shackles through the rediscovery of classical scholarship. Dogmatic narratives were questioned, and print created a distribution mechanism. Over time, the compounding dynamics of open scholarship precipitated the Reformation, which gradually spread throughout Western Europe to break the grip of the Holy Roman Empire.

These dynamics betray the underlying features of information. ‘Information’ has come to be colloquially synonymous with things that you read or hear. However, information is more fundamental it seems. So fundamental, in fact, that our understanding of life is converging on the idea that information is the defining property of life.

Shannon defined information as a reduction of uncertainty. In other words, the degree of surprise you experience at learning the outcome of an event — the more unlikely that event, the more surprised you feel at learning about it and the more information you gain. However, we still do not fully understand what information really is, or its relationship to life.

Sara Walker and Lee Cronin’s Assembly Theory may provide a novel insight. Consider the universe as a lego-builder. It produces novelty by sticking together legos that have already existed. Hydrogen by hydrogen, cell by cell, person by person; continuous complex emergence. Every thing, every person, every thought a mosaic of building blocks in time. As existing lego pieces grow in quantity, more combinations are possible and more complex combinations are increasingly unlikely. Information, then, is an instruction manual, a memory of stepwise lego-building to reconstruct a complex combination. The more complex the combination, the more information you need.

Processing information is what enables life to adapt to its surroundings and act intelligently, by reconstructing lego combinations from the instruction manual. Everything we do arises from processing information — in real-time through bodily systems, and in evolutionary time through genetic heritability. It distinguishes life from non-life: adaptive, not random.

While all lifeforms process information, no single organism has the full story. Each of us is an imperfect information processor, a perspective, with a unique window on the world and a finite capacity.

One of the defining characteristics of humans, then, is to collectively process information through the consilience of technology and culture. We combine our windows on the world to produce an emergent aggregate that is greater than the sum of its parts.

It is common to see modern society as merely a product of the Enlightenment; to rob early human societies of their creative contributions. In reality, we are dependent on the outcomes of collective information processing extending way back to the earliest humans. Everyday necessities like yeast for bread-making, farming, basket weaving, metallurgy, navigation, shipbuilding, clothing, cooking, foraging, shelter, fire and jewellery have been honed over thousands of years. Each discovery builds on the last — new legos made of old combinations.

Information is intrinsically combinatorial. Ideas have sex, and new ideas are born in the brain’s fertile ground. Climbing the tree of knowledge is a promiscuous adventure, reaching for ever higher branches by standing on the shoulders of giants. It adds up to something, enabling successors to reach new heights at super-exponential speed.

Technology and cultural norms externalise our journey up the tree of knowledge; the cumulative result of information processing that provided an evolutionary advantage. A kind of collective genetics, accumulating information and enabling progressively better adaptations over evolutionary time — the sum total of human ingenuity enshrined in object and process.

Through this lens we may return to the Enlightenment. Drawing and writing are user interfaces for thought. Building upon language, they shape our internal schema and enable us to think new thoughts by manipulating new information. The impact of the printing press was to enable writing to grow into a spiralling collective dimension: a progressive embodiment of inscription as an internal cognitive amplifier, and a swelling body of externalised cultural knowledge. Newly complex societies could emerge from these conditions — individuals empowered by the collective intelligence of an increasingly informed polity.

"It is not a question of mind versus matter but the mind made matter and matter made mind." — The Mind Made Matter (David Krakauer, 2022)

Over time, the collective dynamics of the printing press conspired with double-entry bookkeeping to enable markets: a step-change in real-time information processing.

Now, theoretically, any individual can themselves produce any thing — there is nothing stopping individuals from fabricating semiconductor chips in their garage. Realistically, however, individuals are constrained by energy, information and time — it takes highly specialised knowledge and lots of energy (via machines) applied over time to fabricate semiconductors.

The breakthrough of markets was to distribute information processing out to the edges to encompass vast swathes of entrepreneurial perspectives. ‘Truth’ was no longer established by the state, but by selection. What sells, wins. Markets outcompete centralised economies because they process more information in real-time through variation and selection, contributing better evolutionary adaptations back to our collective genetics.

In a market paradigm, money itself is information. Price collapses supply, demand, production costs, complex market dynamics and know-how into a single number that reduces uncertainty to make multi-temporal trades between distrustful parties possible. Over time, transactions accumulate to represent a collective record of who owns what — a coarse-grained informational history of economic interactions.

Markets defined a new era of collective agency, powered by information, challenging the dogma of the Church. Agency disrupts power. Over time, mutual inter-regulation between states, markets and an informed polity, mediated by money, provided the conditions for democratic nation states to emerge.

We can look at democracy as an adaptation to collective information processing in real and evolutionary time. Today’s democracies attempt (inefficiently) a kind of bottom-up governance, by coarse-graining individual perspectives through representation. Representatives describe their policies, and voting condenses each voter’s informational universe into a yes or a no. In aggregate, the chosen representative theoretically best represents the collective viewpoint in parliament. Debate produces laws, which contribute back to societal genetics as an accumulation of information governing the electorate through downwards causation.

Society is a collective energy and information processor

“We’ve come a long, long way together. Through the hard times and the good.” — Praise You (Fatboy Slim, 2006)

Zoom out, and societal evolution is the evolution of collective information processing, underwritten by energetic metabolism.

We have moved beyond genetic evolution in its natural form, and now evolve in a higher collective dimension. Our strength is our ability to collectively process and accumulate information, and recombine it for continuous adaptation.

In real-time, democratic market economies are distributed information processors, collective brains, that combine individual perspectives into a collective aggregate. Technology and culture are evolutionary accumulators, collective genetics, enshrining collective information processing to empower individuals with the agency to create new combinations. Exercising our burgeoning appetite for information, meanwhile, consumes energy in the form of fossil fuels.

We as individuals are indistinguishable from our collective context, in both real and evolutionary time; wave/particle dualities in a continuous field of collective information processing. We are just a small part of something much bigger than ourselves — it defines us, and we define it.

Birth of a Frankenstein

“There is something at work in my soul, which I do not understand.” — Frankenstein (Mary Shelley, 1818)

New definitions of life are rising from the ashes of tradition. “Life” is not merely defined by cells, membranes or self-replication. Organelles, cells, organisms and superorganisms are all forms of life. Life keeps on emerging at more complex levels of organisation.

What, then, is life? What is an individual?

Life, “individuality”, may come to be defined by informational coherence over time. A bounded entity that processes information and consumes energy to adapt to its environmental surroundings.

Thus, we may expand our view of what life encompasses. A human is an individual and a society of symbiotic individuals. An ant colony is a superorganism and a society of organisms. Life can be nested; a single individual both a whole and a part.

Society processes information in real-time via distributed individuals, and consumes energy through fossil fuels. It accumulates genetic information in the form of technology and culture, producing an emergent and recombinant aggregate capable of adapting to changing circumstance.

In society, we have created life. We are life and we are part of life.

Society is a distinct lifeform; a hyperobject evolving in a higher-dimensional space on a longer timescale than its participants. Scarcely perceptible, but learning, accumulating and growing just like us.

Of course, life and intelligent life are distinct. Something separates humans and other animals. Something similar distinguishes our society from its future self.

Life may exist on a spectrum — more alive to less alive — defined by its capacity to transform itself with information. Societal evolution, then, is about becoming more alive, through a blossoming capacity for collective information processing. It is about developing an intelligence that supersedes any individual; a transition from a disjunct collective to an adaptive organism replete with self-awareness. A society of minds.

As yet, a Frankenstein. Tragically unrecognised, maligned and self-sabotaging.

To the extent that society demonstrates intelligence, it coalesces around the Schelling point of GDP: the bastion of late industrial society. Common ground between market and state. GDP measures the value of all the goods and services in an economy; a bottom-up computation of its energy and information processing capacity.

The game of GDP is a game of growth. Individuals, companies and nations are the players. More money means more power; entitling claims on energy and information. Each player must grow their lot to pay off last year’s loans or get outcompeted.

The industrial pursuit of economic growth has delivered on almost every measurable metric that matters in society. We are healthier, wealthier, wiser, calmer and fuller. The industrial organism is a measurable improvement on its feudal ancestor.

Progress is not evenly distributed. Western countries tended to industrialise earlier, and have reaped the combinatorial benefit. ‘Traditionally’ (within the industrial lifetime) lower income countries like China and India are now in the exponential phase of industrialisation. A similar growth pattern, just staggered in time.

The benefits of industrial economic growth are definite and inarguable. However, the metabolism that currently powers it is incommensurate with our planetary reality.

Energy and material addiction

All goods and services are dependent on energy in some way, whether in manufacturing, distribution or usage. Making more consumes more energy. While improvements in energy efficiency would appear to reduce energy consumption, Jevons’ Paradox instead expands consumption through newly profitable activities.

All lifeforms scale with energy consumption — consuming around 75% more energy for every doubling in size. Society as a lifeform is no different. While some countries have at least partially decoupled GDP growth from energy consumption, global economic growth is empirically tied to energy consumption, and energy consumption is tied to fossil fuels. The transitive implication being that economic growth is tied to fossil fuel consumption — energy consumption scales sublinearly with GDP.

So far, fossil fuels have represented cheap, massively scalable labour applied to almost any problem. Despite the obvious fact that fossil fuels are finite, evidence suggests that we are not in imminent danger of running out. But, this is overly simplistic. What matters in energy is the energy return on investment (EROI) — the ratio between the amount of energy spent in extraction and the amount of energy extracted. We are gradually exhausting the easy-to-recover fossil fuel reservoirs. To maintain our rate of consumption, this will mean using harder-to-recover reservoirs with lower EROI. In other words, we have to spend more energy to recover the same amount of new energy. This trend is most pronounced for oil, but also true of gas and coal.

Here we find a tipping point. When EROI goes below about 10, the amount of useful energy delivered to society begins to fall in a non-linear fashion. Since everything in the economy is reliant on energy in some form, things that used to be profitable at a previous energy price become unprofitable. At a certain point, non-linear domino effects can cascade throughout the system.

The race is on. Increase fossil energy efficiency faster than declining EROI, or switch to a new energy subsidy.

Changing energy subsidy is no small task. Fossil fuels require large amounts of infrastructure. It is upon this infrastructure that goods and services are built. Changing the underlying infrastructure risks disrupting everything that sits on top of it, and so infrastructure evolves slowly for stability. Advanced economies are fed by a massive built energy infrastructure, while structural growth incentives continually push energy demand. The problem is not only to replace the existing energy demand, but to replace the increasing energy demand while continuing to support the fully evolved state of the economy.

It remains to be seen whether renewables are able to fill this demand. On the plus side, they tap practically unlimited sources of energy. Efficiency is increasing and the price is decreasing quickly. However, the EROI of key renewable technologies (solar and wind) is lower than fossil fuels, while their intermittence requires infrastructural overhaul. Importantly, manufacturing and installing renewable energy infrastructure requires lots of energy (necessarily derived from fossil fuels) and lots of materials (which are themselves non-renewable and cannot currently be recycled).

Nuclear power is another solution. Nuclear EROI is competitive or higher than fossil fuels, with less intense material requirements. When it works safely, it is low on externalities. While uranium is itself a finite planetary resource, seawater extraction combined with new reactor designs can at least extend its supply to a reasonable timescale (possibly indefinite). However, it faces enormous upfront investment and low public opinion. It could replace electricity generation at scale — France is evidence of such. Whether it is realistic for countries with highly evolved built infrastructure, given the political challenges, is an open question.

Alongside energy, economic growth is also tethered to materials. GDP growth tracks almost linearly with material footprint. Growth means more stuff, and more stuff means more materials. While no single material is as deeply integral to the economy as energy, materials are subject to the same fundamental problem. As they get harder to extract, they get more costly to extract, and these costs can cascade. Copper is the new oil.

Before we get to planetary externalities, then, a systemic growth obligation is economically problematic for finite resources at the point of extraction. The system is built on the assumption of limitless cheap fossil energy and materials, and at some point, they will no longer be cheap. So far, this has been largely invisible, because we have consistently been able to draw on easy-to-recover reservoirs.

Of course, the problem is not limited to extraction. Harm exists on both sides of the equation. Fossil fuel consumption emits carbon, which is causing climate change. Mass material consumption without recycling leads to harmful toxicity, soil degradation and water pollution. Crucially, these harms are simply second-order consequences of the economy working as ‘designed’.

Missing information

We live in a highly evolved economic system, consuming energy and materials to the relentless beat of economic growth. Growth is a necessity in every corner, producing a structural dissonance between the short-term need for energy and materials to support it, and the long-term need to align ourselves with the substrate on which we are dependent.

Intelligence includes the ability to determine goals and process information to adapt to a changing environment. Societal intelligence arises from individuals who make decisions based on the information at hand. If money is the information that flows through economies, then in order to solve the problem of energy and materials, information about the real planetary cost of externalities would need to be factored into prices.

As we know, however, prices include no such information. $10 is the same regardless of what it is spent on. Consider a simple example. You have a choice of two t-shirts: the first is mass-produced from polyester (derived from petrochemicals) at a high-energy, low-income factory; the other is produced by a small family business from locally grown bamboo and made by hand. Both cost $50. The economy looks at these two items as exactly the same. However, one of these options is significantly more harmful — unsustainable resource use, high carbon emissions and poor factory conditions for workers. Information gets lost.

Consider an adjacent example. The conditions are the same, but more realistically, the mass-produced t-shirt is actually significantly cheaper — say $20 compared to the $50 artisan version. A financial incentive now exists to cause harm (at least, to not prevent harm from being caused). All agents have an incentive to maximise their money, and can’t be expected to make altruistic decisions. Even if they did choose altruistically, agents can only act on the information available to them, and the information about the true planetary cost has been lost! And so, the company behind the mass produced t-shirt is likely to dominate sales. Other companies, meanwhile, are forced to adopt similar practices or get outcompeted in an evolutionary race to the bottom.

Intelligence cannot exist without information. Brains without sensory information are electric mush. Ant colonies without pheromone trails are a random crowd. Economies lacking information are similarly constrained. As a distributed system, nobody is steering the ship. There is no illuminati. It is an emergent aggregate arising from the actions and interactions of many economic players. If every transaction is blind to its planetary consequences, then the whole economy is blind to its planetary consequences. Despite this, GDP increases at the expense of anything else we value, and we welcome it as progress.

Industrialisation, driven by the growth of financial value as a proxy for progress, has seen us become a powerful force upon the planet’s dynamical processes. However, the financial proxy has become increasingly abstracted from our planetary context, producing harm at extraction and consumption of finite fuels and materials. Economic systems structurally reinforce the problem — agents are not only incentivised to make decisions that work against the collective best interest in pursuit of growth, but they lack the information to make good collective decisions in the first place.

“..long-term survival of our, or any global scale ‘project of civilization’ will require a fundamentally different mode of planetary-scale behaviour in which knowledge of planetary-scale impacts feeds back on, and modulates, behaviour in an intentional loop" — Intelligence as a planetary-scale process (Frank, Grinspoon & Walker, 2022)

Any organism whose brain seeks to destroy its body will ultimately fail. Becoming an intelligent society means developing the self-awareness to recognise that our economic system is not backed by gold, but backed by the planet as our shared body. It means coming to consensus on collective values that emphasise the complex manifold of interactions that drive society as an adaptive organism in the context of coupled planetary systems. Capital exists in many forms beyond financial, and it is balanced growth that produces adaptive organisms. Misaligned growth is a cancer.

Values are not just abstract ideas. Crucially, we must build them into our information systems such that distributed agents are systematically incentivised to make decisions that benefit the collective, without requiring altruism. A new system has to emerge from the bottom up — from individual transactions to a new emergent aggregate that is intrinsically in tune with our collective and planetary reality.

“All my weight in gold, is not all I know.” — Family Business (Kanye West, 2005)

Multipolarity

The relationships between GDP, energy, materials and climate change hint at an underlying problem. The tragedy of the commons is what we might call a multipolar trap, in which agents have an incentive to pursue strategies that maximise their own best interest at the expense of the collective. As we look deeper, we find that these traps are rampant, and this impairs our ability to process information in key collective pursuits.

Our world system is built on the principle of multipolarity. Individuals, companies and nations are, on average, pitted against each other in evolutionary zero-sum games in the pursuit of growth at each other’s expense. Investments, products, trade agreements — broadly speaking, there is a winner and a loser somewhere in the system. Markets and nation states reinforce these dynamics.

I couldn’t put it better than Frank Slootman:

“If I'm hiring a top salesperson, top engineer from another company, I've taken their strength now into my organization. It's not just that I gained theirs, but they lose theirs. So it's a double impact. I weakened them and I strengthened myself.” — Invest like the best (Frank Slootman, 2022)

There are extremely valuable aspects of competitive, multipolar institutional systems. Markets empirically create great products that may not otherwise arise, because of the competition between market participants. Peer polity between nation states probably does lead to social progress, because of the competition. In natural systems, competitive dynamics with conflicts of interest can facilitate information flow, enable selection between competitive strategies and foster innovation. When information processing is noisy and uncertain, such environments may well be a general feature of collectives.

However, they also create failure modes in problems that require cooperation at the global level. In evolutionary time, economic players cooperate through the combination and accumulation of information, enshrined in technology. In real-time, however, players are incentivised to both withhold information and present misleading information in order to succeed in zero-sum selection. Where there is a winner and a loser, and the reward is money, disinformation is endemic.

When I was younger, I worked in healthcare. There is a common heart problem which can cause significant trauma if left untreated. At the time, there were about five drugs competing for market share to solve this problem. Each one was driven by a sophisticated sales team, whose responsibility it was to persuade healthcare decision makers that their drug was the most effective.

Patients naturally want the best possible outcome. Healthcare professionals generally want the same thing. Needless to say, on average, some drugs produced ostensibly better patient outcomes than others in both controlled and real-world settings.

Companies, however, only grow when their product sells, and the game is strictly zero-sum between competing companies — patients can only take one drug. As such, there is a misalignment in incentives, which creates the conditions for disinformation. First, pharmaceutical companies are incentivised to construct and manipulate statistical information with the greatest possible chance of success within the boundaries of regulation. Historical examples include paroxetine, citalopram and other antidepressants. Second, salespeople are incentivised to build relationships and present information using the most compelling narrative, regardless of whether what they are selling is objectively the best choice. Healthcare is about patients, but the system produces situations in which financial incentives are incongruent with the patient’s best interest.

Not all examples are as consequential. Narrative manipulation matters less in tax software. However, a huge portion of our information environment is driven by individuals and companies who stand to gain financially from convincing us of something, regardless of whether or not that thing is verifiably true. It’s less about information, and more about who can tell the best narrative.

In a pre-internet world, this effect was more contained. Traditional media placed legitimacy in the hands of a few with the power to distribute. Narratives flowed from power to the people. The internet, especially social media, has introduced an environment in which information is truly free, giving everyone the ability to distribute. It gives power to the people. Everyone has a voice in a game driven by the acquisition of mindshare.

The social medium is a medium of narrative — there is no requirement to support claims with verifiable evidence — and it is the locus of information aggregation and discussion on today’s internet. It has massively increased the complexity of our information environment, taking pre-existing narrative incentives into an exponential dimension.

As I described earlier on, this is ultimately a good thing for fundamental reasons — ideas are combinatorial, and agency through information increases the likelihood of serendipitous combinations which contribute to our collective evolution. This is a win both for society and the individual.

However, we are all now participants in a global narrative colosseum. It is harder than it has ever been to understand what is true, owing to the sheer volume of information embedded within a complex matrix of misaligned incentives. Algorithms, meanwhile, orchestrate the chaos in pursuit of financial outcomes as we flirt with new externalities of complexity — emotional contagion, political mobilisation and polarisation.

“Are you not entertained?“ — Gladiator (2000)

We are all imperfect information processors. We succeed as a collective, by combining our information to produce more accurate representations of reality. In the print world, we evolved a series of institutions — government, journalism, science — that were trusted to be the aggregators, combining expert perspectives into collective wholes acting as a form of consensus.

Since roughly the introduction of the internet, these institutions have begun to reveal their decay. In science — p-hacking, 'flexible' analysis, low statistical power, confirmation bias and incomplete reporting. In journalism — negativity, wokeness, cancelling, misinformation and clickbait.

Consider vaccines as a recent example. To trust vaccines, one of two things needs to be true: either a) you need to understand the information yourself, or b) you need to trust that institutions can interpret the information for you.

The scientific basis of vaccine efficacy is a set of non-trivial concepts, spanning biology, chemistry and statistics. It is spread across a vast expanse of scientific literature which is practically uninterpretable for anyone who is not a professional scientist. For the majority of people, this means outsourcing sensemaking to an institution.

We have exponentially increasing complexity of information, and exponentially decreasing trust in institutions to make sense of it all. This leaves the responsibility for sensemaking to individuals. However, there is simply too much information across complex domains for individuals to effectively make sense of. Instead, we seek counsel on social media, and are swallowed by the narrative vortex.

Assuming I’m not the subject of a simulated Truman Show, it’s likely that you too have a sensation that there’s a ‘you’ inside your head. Like a tiny person pulling levers to move and construct sentences. Of course, this is an illusion. Your actions instead arise from a society of 86 billion neurons cooperating to process sensory and stored information.

When our brains respond to stimuli, it is marked by two phases. There is a divergent phase, in which neurons process information individually. It is followed by a consensus phase, in which neurons consolidate a collective solution. Together, the brain benefits from the maximum perspectives, while everyone disagrees and commits to a collectively computed aggregate that enables coordinated action. Multipolarity with consensus.

Society is like a more complicated version of this. When something new happens, our perspectives naturally diverge. In the past, trusted institutions have done the hard work of integrating information into a form of consensus. However, they are no longer capable of meeting the level of complexity needed to do so. This leaves us achieving the first bit — more perspectives deriving from agency provided by the internet (good!) — but as yet without a new way of coming to collective consensus. Multipolarity without consensus.

Consensus is important for coordination problems. In popular culture, climate change is the most prominent example. It is of course extremely important in the long-run. But it is far from the only problem, and is in fact a lighthouse for a growing class of coordination problems defined at the global level. It is this class of problems that our institutions and information environment are woefully underprepared for.

The internet has delivered us into a new age of collective agency. A planetary public with new powers of information. We can reasonably expect a complex system composed of billions of agents with imperfect information to be noisy, uncertain and operate near a critical point. This is what we are seeing. But as things stand, we are unable to draw on the benefits of our newly vast information processing power.

Our collective fitness as a societal organism is dependent on our ability to process information, and the problems we are facing are increasingly global. Our powers of collective information must rise to the scale of the problems we must solve — globally, not nationally. To do so, we need open societies that incentivise truthful information, together with novel consensus-building mechanisms, “institutions“, that can handle and integrate information at true global complexity. Cities reveal the returns to scale that await our success.

Presently, we cannot implement open societies because anti-consensus multipolar narratives are embedded deep into our cultural psyche. Throughout our industrial evolution, nationalism has taken the mantle from religion to produce imagined communities at national scale, launching us into a globalised but tragically misaligned world. Nation states remain the core civilisational institution, intrinsically restricting the global cooperation we now desperately need.

Nation states and the dangers to come

“We are taught to see the world in terms of nationality. It is our country, "our" athletes compete in the Olympics. When they win, it is "our" flag that waves in the ceremony. "Our" anthem brings the judges and other competitors to attention in the awards ceremony. "We" are led to believe that it is "our" victory, although it is never quite clear how "we" participated, other than by being within the same territory as a citizen.” — The Sovereign Individual (Davidson & Rees-Mogg, 1999)

The post-WWII geopolitical order saw America establish itself as the world’s most powerful economy — undisputed ‘winner’ of the industrial age; a position coming along with an assumed responsibility to maintain global order. Through a dominant economy, military and technology, America has since then presided over a relatively stable period of global geopolitics without any major conflicts.

However, the past decade has revealed chinks in the armour, and it is becoming increasingly common to acknowledge the decline of the American hegemony. Wokeness, institutional decay, rising gun violence, political polarisation, Capitol rioting, regulatory standstill and military withdrawal are all symptoms of evolutionary destabilisation.

Destabilisation is an internal catastrophic risk, with political polarisation at its core. The electoral college has always been an ambitious experiment in federated democracy at scale, in which polarity is designed to enable a better, more pluralistic collective aggregate. However, it has evolved into gridlock and become incapable of reaching actionable consensus, while a march of blundering buffoons take office.

This comes at a time when America is facing complex problems. Economic progress is slowing. Scientific progress is slowing. Educational progress is slowing. Inequality is growing. The world is getting more complex, the problems are getting harder, and the system can’t cope. Arguably, it is beginning to go backwards.

What happens from here is uncertain. Perhaps chaos will facilitate a recentering. Perhaps it will regress from democracy. Perhaps it will even descend into civil war. What is abundantly clear is that America has serious internal work to do. Internationally, the situation is precarious — overreach, and risk the home front. Underreach, and risk the hegemonic advantage.

“Why did one straw break the camel's back? Here's the secret The million other straws underneath it, it's all mathematics.” Mathematics (Mos Def, 1999)

Whether America returns to isolated exceptionalism to focus on internal problems, or attempts to maintain its commitment to being the world’s policemen, there are direct and indirect risks at the global level.

The decline of America opens the door for a new power to establish itself in a newly open geopolitical environment. As we know, there is a willing suitor — China is the ascendant world power. Within the past half century, China’s GDP, population, trade and almost everything else have skyrocketed, putting it within touching distance of the US economy.

China is a different beast. It is a modern autocracy empowered by new technologies of control. It knows how to build things, and is the world’s manufacturing powerhouse. Just 50 years into its modern transformation, it has momentum, with room to grow.

China’s rise is not inevitable. It has failure modes — a serious and fast-approaching demographic transition, an economy overly reliant on real estate and an autocratic inability to process valid, counter-state points of view. Whether China indeed becomes the world’s most powerful economy, or whether it echoes the almost economic dominance of 1980s Japan remains to be seen.

While the future is unknown, American decline and Chinese ascent are a source of present international friction. Graham Allison has introduced the notion of Thucydides’s trap: historical context of the perils that accompany the rise of an emerging superpower that threatens to overtake an existing superpower. 75% of historical examples have resulted in war, including both 20th Century World Wars. Adolescent petulance and ageing arrogance are a bad combination.

"It was the rise of Athens, and the fear that this inspired in Sparta, that made war inevitable.” — History of the Peloponnesian War (Thucydides, 455 AD)

We cannot directly infer future events from past events. The world is unpredictable, and many futures are possible. However, Allison’s warning serves as a critical context for where we find ourselves in the arc of history. A coming war may resemble the Cold War more than the world wars, but any form of conflict brings with it the potential for non-linear cascades that result in full-scale armed conflict. It is not an imaginary concept either — it is already happening.

Conflict between the US and China is a catastrophic risk to both countries. For America, it is likely to accelerate the trend towards decline, by distracting and de-resourcing Americans from fixing growing internal problems. For China, it may cause growth to stagnate before it has the chance to reach its full potential. If it escalated into a ‘hot’ war, then up to 1.8 billion people stand to get caught directly in the conflict. Like both world wars of the 20th Century, a sufficient spark on the geopolitical stage can create a fire that drags further countries into conflict — both the US and China have powerful allies who may well be compelled to respond.

“He had come to know quite thoroughly the world in which he lived. His outlook was bleak and materialistic. The world as he saw it was a fierce and brutal world, a world without warmth, a world in which caresses and affection and the bright sweetness of spirit did not exist.” — White Fang (Jack London, 1906)

While the direct consequences of conflict could be catastrophic, the indirect consequences are likely to be even more insidious.

War is hungry for energy and materials. It means manufacturing and fuelling more weapons, ammunition, vehicles and equipment. It means diverting workers away from ‘non-critical’ industries to support the war effort. In the fog of war, energy and material security becomes mission critical, and no threat of climate change or resource scarcity can possibly outweigh the drive for self-preservation. We cannot coordinate solutions at the intersection of energy, materials and climate change in peacetime, and we’re even more unlikely to do so in wartime.

War creates arms race dynamics. In 1933, Leo Szilard was the first to conceive of the nuclear chain reaction. Almost immediately, he realised the power of what he had discovered. Together with Albert Einstein, he co-wrote a letter to then president, Franklin Roosevelt, arguing for a research embargo. The rest, as they say, is history. The US enjoyed nuclear dominance for just a few short years, before embarking on the Cold War with the former Soviet Union.

In the context of the Second World War, the Manhattan Project was a rational and ultimately winning move. Once the idea of the nuclear chain reaction existed, the US could not guarantee that Germany was not already building bombs. Thus, not developing nuclear weapons could be considered an existential threat to America’s sovereignty — a nuclear advantage for the Nazis might have seen WW2 turn out very differently. The Cold War likewise. In this scenario, the dominant strategy is to pre-emptively strike first. Defence through offence. A classic arms race — both sides are incentivised to maintain their evolutionary advantage by building technology as quickly as possible lest they get outcompeted by a technologically superior opponent. Of course, from the collective viewpoint, this is the worst outcome for everyone.

"From a god’s-eye-view, the best solution is world peace and no country having an army at all. From within the system, no country can unilaterally enforce that, so their best option is to keep on throwing their money into missiles that lie in silos unused." — Meditations on Moloch (Scott Alexander, 2014)

The world entered a new paradigm the day the Hiroshima bomb was dropped. Nuclear weapons are the first technology that has come along with existential risk: a risk that threatens the entire future of humanity. A true nuclear war scenario could represent a permanent extinction or irretrievable reduction in human capabilities. So far, atleast, nuclear war has been avoided through the notion of mutually assured destruction.

Nuclear war between China and the US seems unlikely due to the gap in nuclear capability. Of course, this can’t be guaranteed, especially if more players are involved — China and America are not the only aggressive nuclear states. Any post-1945 armed conflict between nuclear-enabled powers comes along with the risk of a nuclear holocaust.

A nuclear arms race may be less likely, but more arms races loom on the horizon.

First — artificial intelligence. Primarily through the work of Deepmind and more recently OpenAI, human-level artificial general intelligence is changing from an obscure science fiction in the early 2000s to what now seems a fast-approaching inevitability. AlphaFold, the work of a narrow and ultimately ‘primitive’ AI, has already solved a 50-year grand challenge in biology. AlphaFold is really a triumph of the creativity of its designers — it did not choose the problem or orient itself to solve it. AGI will not only bring with it significantly greater levels of intelligence than AlphaFold to all domains, but also the creative ability to choose its own problems and act as an autonomous, goal-directed agent.

Values-aligned, human-friendly, cooperative artificial general intelligence will be the most positively transformational technology that humans ever develop. Misaligned, human-antagonistic, competitive artificial general intelligence might be a species-level risk. Even if the risk is low, safety and alignment is critical because there will be no second chances. Pandora will not go back in the box.

AGI is the ultimate nation state technology, able to solve complex control problems beyond human comprehension. Resource optimisation, mind control, scientific discovery, human alignment, capital allocation — all add up to a decisive strategic advantage.

Even in peacetime, states that grok AGI have an incentive to pursue forms of ‘safe’ alignment that place their own self-preservation at the centre. A wartime scenario between competing nation states further creates all the wrong conditions for the development of safe AGI. America houses one of the two current frontrunners, while China has nationally committed to becoming the world leader in AI by 2030. We might imagine that in a war scenario, nationalisation of any AI research becomes a matter of national security. Competition between the two countries increases the temptation to deliberately disregard safe implementation, while also increasing the likelihood of mistakes. For good reason — whoever gets there first automatically wins any competition. The problem is that once achieved, without alignment, the victory may well be unassailable and the risk existential.

The problem is not limited to AGI. The race is on to build autonomous weapons. Makes sense if you’re an American! It does not make sense if you are a human.

“You don’t want to play with dynamite. That shit’ll blow your world apart.“ — Wile Out (DJ Zinc & Miss Dynamite, 2010)

Second — bioweapons. Organisms are defined by code; base-pairings, genes, DNA. Code is easy to distribute. Meanwhile, wet lab work is becoming more accessible. With it comes the possibility of reintroducing diseases of the past, or even worse, engineering new ones. Combined with personal genomic data, we have a recipe for targeted, nearly invisible mass destruction.

This may seem abstract and futuristic, but it’s already possible. 6 months, 5 people, $100,000 and you can bring Smallpox back into existence. For context, Smallpox is one of the most devastating known diseases, killing many millions of people throughout history.

In a wartime scenario, bioweapons are tempting. Put yourself in Roosevelt’s shoes, circa 1943. You have the power to wipe out every Japanese and German to end the war immediately. Even if there is a chance of wiping out significantly more, one is forced to choose between kin and other. After all, nuclear bombs were dropped. Or, put yourself in Hitler’s demented shoes. You are hiding out in a Berlin bunker, awaiting imminent Allied encroachment. But wait! A hail mary — a bioweapon is just a red button away.

“Dooming”, vous m’accusez. Understandable. Until you realise that during the Cuban Missile Crisis, mutual nuclear annihilation was close to 50/50; again at Able Archer and more. A coin toss. Modern Western and Soviet societies exist only through pure luck. Of course, the maths here is imprecise, but the point is that war is not an option, hot or cold, when we have technology this powerful. It doesn’t take a 60s-era, peace-toting tree-hugger screaming ‘make love, not war’ to realise this — it is an obvious fact if you value the continuation of the human race.

We might argue that the threat of nuclear war has been a key determinant of the relative peace we’ve experienced over the past 75 years. On this line of reasoning, we might further argue that nuclear bombs are ironically a peace technology as much as they are a destructive technology. There is truth to this argument, but our arrogance lies behind the observation selection effect of a world unscathed by nuclear war. Similar logic cannot be indefinitely applied to future arms races with powerful technology and no reset buttons. It only has to go wrong once.

The problem isn’t the technology per se, but rather the environment in which it is conceived. Antagonistic, competitive nation states cannot, by definition, handle the existential risk associated with increasingly powerful technology, because they are a cause of it simply by following their incentives. Where there are nation states, there is war. At some point, a multipolar world takes on a level of risk from technological progress that surely leaves an existential event as a near-certainty.

Systemic existential risk

War is the extremism of multipolar nation states. However, it doesn’t take extremism to induce existential risk — the problem is more fundamental.

Misaligned artificial general intelligence and bioweapons represent existential risks arising from our pursuit of technology. Molecular nanotechnology is another, though arguably more distant. Over time, we might expect even more powerful, existentially risky technologies to emerge — you never know where the stepping stones might lead.

Nick Bostrom asks us to consider the evolution of technology as a process of pulling coloured balls out an urn. So far, the balls have been mostly beneficial.

“What we haven’t extracted, so far, is a black ball: a technology that invariably or by default destroys the civilization that invents it. The reason is not that we have been particularly careful or wise in our technology policy. We have just been lucky.” — Vulnerable world hypothesis (Nick Bostrom, 2019)

It is common to argue that technology has both positive and negative consequences. Social media, for example, connected the world and precipitated a planetary public. However, it did so under the influence of advertising, producing algorithms which exploit political fault lines and human biases to farm engagement. This seesaw of consequence appears fundamental — technology is recursively combinatorial; an instantiation of accumulated information. Over time, more complex technology interacts with a more complex world to produce more possible future states, both good and bad.

The black ball is a thought experiment, but it reveals what is crucially absent in our technological development. To see it, we must first realise that what is pulling balls out of the urn is a distributed, multipolar economic system driven at multiple scales by individuals, markets and nations. Each eggs the other on and nobody is in the driving seat — we all influence it, some more than others, but nobody is in control. It’s arms races all the way down.

In this system, there is no collective capacity to determine whether or not we should introduce a new technology, when is the right time to introduce a new technology or what social or technological buffers should be in place before a new technology is introduced. Instead, we assume that technology is inherently good, and negative societal consequences are simply the price we must pay.

The issue is magnified given that once technologies that provide a sufficient evolutionary advantage are introduced, their widespread usage becomes a foregone conclusion. You are at a significant disadvantage in the world today if you don’t own a smartphone — there is no choice but to sign up for its psychological and societal consequences, good or bad. Globally, as long as powerful technology is developed somewhere, it will outcompete regardless of its consequences.

“Think about all the ways in which our minds, relationships, and cultures changed with the widespread adoption of the smartphone technology ecosystem. How does it affect memory and navigation capacities, or attention span, or personal reflection? How has it affected the values enacted around a family dinner table, or the norms of social interactions in general?… Was all of this intended or considered by the initial inventors and adopters?“ — Technology is not values neutral (The Consilience Project, 2022)

For most of human history, this has been a net positive force and a counter to stasis. We’ve avoided mass catastrophe simply because no technology was powerful enough, and we could afford to learn and experiment after the fact. However, we are now approaching technologies with the capacity for existential risk. We only get a single shot at safe, holistically-considered introduction. Competitive nation states, combined with distributed, adversarial market dynamics and lagging, uninformed, ineffective regulation is a dangerous combination in light of such technologies.

This is not an argument for ‘regulation’ to constrain technological development, at least not regulation in any conventional legal form. It is an admonition that we are structurally driving existential risk and lack the ability to recognise it, let alone escape. We are unconscious slaves to evolutionary forces and market incentives that accelerate us like blinded soldiers into a field of nuclear landmines.

“We can invent but we cannot un-invent. Our strategy is to hope that there is no black ball.“ — Vulnerable world hypothesis (Nick Bostrom, 2019)

Technology is just one axis of risk. We also face a series of natural risks, including asteroid impacts, supervolcanoes, extreme weather, solar flares and pandemics. The risk is small in each case. Supervolcanoes, for example, are estimated to occur roughly once every 50,000 years. While the prevalence is low, the impact can be mighty.

COVID is the most recent example of a global threat arising from a (hopefully) natural risk. It has caused the deaths of at least six million people and caused considerable economic disruption, but it has been significantly less catastrophic than it could have been. As such, it has been a dress rehearsal for something more serious. It has left in its wake a series of clear indications that civilisation is not yet up to the challenge of a truly serious pandemic — slow responses, delayed vaccines, institutional distrust, lack of PPE, decision fatigue, questionable lockdowns. Worse, it is unclear whether we have learned our lessons well enough to implement future resilience.

Another, less understood risk is that of demographics. Paul Ehrlich apocalyptically proclaimed in 1968 that overpopulation would lead to social collapse. As it turns out in 2022, underpopulation is a more realistic risk. Mass decreases in global fertility rates, especially in advanced economies, threaten an ageing population that cannot replace itself. In turn, there may not be enough working-age people to fill our large economic shoes. In China, for example, it’s possible that a population of 1.4 billion recedes to a mere 600 million by 2100 — modern China will not survive if so.

Any lifeform that is not able to determine and mitigate existential risks to its existence is destined to fail in an evolutionary game. The sword of Damocles strikes hard and true. Existential risks, both natural and anthropogenic, are intrinsically global coordination problems. Mitigation can only be achieved by a systems-aware global individual coordinating information and resources, prioritising long-term health over short-term economic gain even when the annual probability of adverse events is low. Without this ability, we are confined to flailing reactionary responses which, one day, will simply not be enough.

The takeaway here is not that we are doomed, nor that we should avoid technology. It’s that our current institutional systems structurally lack the adaptive advantage demanded by a new class of global coordination problems. We cannot effectively mitigate existential risks until we move on from a multipolar system with no consensus, because it is the cause. This means rethinking the concept that sits at the centre of it all: the nation state.

Diminishing returns to nation states

Forests accumulate through a succession of species. First, pioneer species increase net energy by flooding soil with nutrients, enabling new species to follow. Over time, energy production asymptotes in line with energy consumption, culminating in a climax community that maximises the energetic capacity of the environmental niche. The ecosystem is defined by a law of diminishing returns — quickly, then slowly, then steady-state.

Forest succession is a useful analogy for looking at today’s advanced economies. Industrial growth has been rapid and unprecedented, owing to the accumulation of collective information processing underwritten by fossil fuel metabolism. The industrial organism has risen to the height of its fossil forebears.

As economies have grown, they have become complex. GDP growth, in many ways, is a measure of increasing complexity. More systems, more specialisation, more administration and more interlinking. It is a two-sided coin. As it grows, metabolism increases. More productivity! More maintenance. More jobs! More bureaucracy. More stuff! More energy and materials.

Importantly, it rarely gets less complex. Institutions, infrastructure and beliefs don’t just go away. Systems become dependent; infrastructure has to be maintained; regulations accumulate; industries lobby; organisations become bureaucracies; principals become agents; network effects ossify. As complexity grows, change becomes increasingly difficult due to the increasing cost of maintenance — there is something to protect.

Joseph Tainter argues that there are diminishing returns to increasing complexity under energetic and organisational constraints; a pattern characteristic of all prior human societies. Diminishing returns require more investment in maintenance, which reduces the rainy day fund for future economic stress.

Economic growth is underwritten by energy and materials, and we have exploited both to a point where we are receiving negative signals on both sides of the equation — at extraction through decreasing EROI and at consumption through externalities. Meanwhile, complexity has grown to an extent that makes the large-scale changes needed to adapt exceedingly difficult — regulations, inefficiency, infrastructure, dogma — while the internet subverts the institutions we have previously entrusted for progress. Advanced economies are approaching the maximum capacity afforded by the assumptions of industrial society.

“Tax the rich! Spend more money!”, they say. Of course, these sentiments imply that governments are even capable of generating marginal improvements on increased tax income. Realistically, governments are big and inefficient capital allocators, spending more money to maintain a familiar growth obligation via personnel who play politics rather than solve problems. Despite the inefficiency, increased spending is evolutionarily reinforced by competition with other nation states.

While there are parallels with the patterns of old civilisations, ours is very different from the civilisations of yesterday. We exist in the first globalised society, deeply interconnected and interdependent. Globalisation without coordination brings huge benefits, but also significant fragility, which recent events have served to demonstrate. The Russo-Ukrainian war shows how reliant we are on each other for food, oil and energy. The Ever Given fiasco showed how easily global trade can come to a halt. COVID has shown how a single viral sample can propagate through an entire world system. Problems do not stay localised.

Here we see the essential irony of the system we live in. Globalised, interconnected and interdependent, but misaligned, antagonistic and self-sabotaging. A schizophrenic playing hide and seek.

The core issues facing the world in this century — energy, materials, climate, geopolitics, existential risk — are immensely complex and long-term coordination problems. Solving them requires deep, empirical understanding of how it all fits together, with the ability to globally coordinate information and resources over multi-decade-long timescales.

Nationally, these problems are beyond what can be productively achieved by shortsighted democratic politics, where each new election cycle seeks to ‘undo the damage’ done by the previous administration. Globally, these problems are beyond multipolar geopolitical nations seeking growth at others’ expense, and arguably beyond the limits of human cognitive biases about linear cause and effect.

"In fact, the history of human social engineering, right up until the present, is based largely on human intuition about the few directly attributable causes of a problem and how to adjust them. As a result, attempts at social control have been mostly reactive responses to the consequences of previous actions and decisions. If we think crime rates are too high, we invest in police and prisons. Unfortunately, as the US incarceration rate attests, such ham-handed social solutions often backfire." — Engineered Societies (Jessica Flack, 2016)

Everything has a lifetime, nation states included. The foundations of advanced democracies were built on pre-industrial technology to solve the problems of a world that doesn’t exist anymore, and they have lost their adaptive advantage. We desperately need to coordinate information and resources globally, not nationally, according to empirical models of the world that optimise for global system health beyond our selfish biases. Short of a massive war, there is practically zero chance of retrofitting this capability onto highly evolved nation states.

"We want to be able to peacefully start a new state for the same reason we want a bare plot of earth, a blank sheet of paper, an empty text buffer, a fresh startup, or a clean slate. Because we want to build something new without historical constraint.” — The Network State (Balaji Srinivasan, 2022)

Ecological succession has a second stage. Fires burn the forest and invite a new generation of pioneers. However, they don’t start from scratch this time. Soil and nutrients remain — the groundwork is already laid, and succession happens faster. Ghosts are buried and become oil, while a new generation builds on its genetic inheritance to readapt to the environment. Healthy, adaptive and cyclical — a forest in tune with its niche.

Over the past 500 years, nation states have birthed a globalised industrial civilisation. A step change in collective information processing, underwritten by fossil fuels, has grown humanity from disparate kingdoms to a living organism of 8 billion people with a 17TW metabolism. We are more alive than our feudal ancestor. However, we are now held back by the same systems that brought us here, and something new is required.

Ecological succession is an imperfect analogy. We want to avoid a war that burns the whole forest down. We want a new generation of pioneers growing in parallel, laying the groundwork for a gradual and controlled succession. Of course, this means that at some point, we must accept that the world we’ve built doesn’t serve us anymore, and that’s okay — it’s a natural part of evolution prevented only by blind allegiance to deep-rooted national stories.

It is common to bemoan the state of the world — to ask how to fix America, how to beat China, how to reinvigorate the EU or how to contain Russia. To feel like it’s all falling apart. Well, if you are willing to take a long-term view, then these are ultimately the wrong questions. We should not be asking how to build better nation states.

We should be asking how to succeed them.

“I got mind control when I'm here, you gon' hate me when I'm gone. Ain't no blood pumpin' no fear, I got hope inside of my bones. This that life beyond your own life, this say this go for mankind. This that outer-body experience no coincidence you been died“ — Never Catch Me (Flying Lotus ft. Kendrick Lamar, 2014)

Society in the information age

Industrialisation was about becoming masters of energy, exploiting fossil fuels to manipulate atoms. Mission accomplished, for the most part — through the combination of markets and states, we have built an economic engine that has underwritten a great leap forward, breathing life into a globalised societal organism.

Progress has come at a price. The industrial economic engine is leaking with increasingly serious externalities. We are facing globally complex risks to our existence deriving from nature, technology, resources and each other. As yet, the societal organism lacks the intelligence to overcome these challenges, and instead structurally reinforces them. Our institutions are built on a principle of multipolarity, driving us to ignorance by antagonism; an uncoordinated world struggling to adapt to globalising selection pressures.

Looking back through history, the evolution of society has always been about collectively processing information. We are descendants on a journey of collective self-creation, empowered by the cumulative ingenuity of over 200,000 years of human history enshrined in the collective genetics of technology and culture.

Building upon our inheritance, we must now become masters of information, rising to the complexity of the problems we must solve — a single global intelligence that can overcome the zero-sum forces that have defined our industrial evolution.

“In the 17th chapter of St. Luke it is written: ‘the Kingdom of God is within man’ — not one man nor a group of men, but in all men! In you! You the people have the power… to make this life free and beautiful! To make this life a wonderful adventure… Let us fight to fulfil that promise! Let us fight to free the world — to do away with national barriers, to do away with greed, hate and intolerance. Let us fight for a world of reason, a world where science and progress will lead to all men’s happiness.“ — The Great Dictator (1940)

One can be forgiven for believing that convergence on the modern system is terminal; that domination, war and planetary externalities are simply the price we must pay for what we’ve come to value. Of course, there is a reason for the convergence. Nation states outcompeted in an evolutionary race of divergent societies, organising people, energy and information to achieve sovereignty and a monopoly on violence.

Agriculture is often viewed as the apple of temptation inexorably marching humanity from clueless plonkers in the Garden of Eden towards this convergence. Au contraire, Graeber and Wengrow’s new take on human history suggests a much richer view of ancient civilisations, in which our ancestors were the masters of their own destiny exercising the freedom to continually experiment and reshape their social realities in complex ways. For the good industrial nation states have brought, this is clearly a freedom we have lost.

…maybe the real question should be ‘how did we get stuck?’ How did we end up in one single mode? How did we lose that political self-consciousness, once so typical of our species? How did we come to treat eminence and subservience not as temporary expedients, or even the pomp and circumstance of some kind of grand seasonal theatre, but as inescapable elements of the human condition? If we started out just playing games, at what point did we forget that we were playing?” — The Dawn of Everything: A New History of Humanity (Graeber & Wengrow, 2021)

The internet is to nation states as the printing press was to the Holy Roman Empire. It breaks the asymmetry of narrative distribution, enabling agency and coordination outside of state borders. Cryptography, meanwhile, delivers agency of monetary information away from states and into a collective dimension. Compounding agency through information, over time, presents the conditions for a newly complex form of societal organisation to emerge. A new Enlightenment.

Humans diverge for fundamental reasons. Brains are like reaction chambers, combining information and experience into organising principles that see the world in different ways. No two brains are the same, but some overlap more than others. To date, we’ve settled on industrial democratic aggregates as the best adaptation to living peacefully together as a divergent people using the technology we had available. By attempting bottom-up governance, they create more checks on power and more capacity for collective information processing than other systems of government. However, in the space of all possible societal configurations, it’s extremely likely that they are a local maximum. That we are experiencing existential externalities is evidence of such.

We must ask, then, how might we discover better societies laying dormant in possibility space? How might we transition out of the industrial economy and its damaging externalities? How might we rediscover the freedom to experiment so characteristic of our forebears?

Balaji Srinivasan has argued that attracting individuals away from today’s nation states will favour a much larger number of smaller, more values-aligned societies. Rather than force naturally divergent individuals into state borders aligned only by location, new kinds of states will bring together highly-aligned groups with shared values, interests and objectives. Such societies will empower individuals with greater freedoms to choose how they wish to be governed, rather than succumb to monolithic democratic aggregates that try to please everybody and end up pleasing nobody.

Replacing nation states is far from easy, nor is it a utopia or a magic fix. Rather, it is a way to reintroduce slack into the system; to open up a new evolutionary space for experimentation in the areas of our social reality that have become dogmatic — collective values, regulation, governance, social norms, economic principles, cooperation, money, infrastructure, energy, architecture and more.

Evolution is present in any system in which there is variation and selection. Small, highly divergent societies will compete for people, especially young people seeking novelty and yet to be terminally conditioned by the status quo. New selection pressures will arise that search for new kinds of societies in possibility space, enabling faster evolution of the global meta-organism. Not just faster evolution, but new evolutionary directions that can vector us away from the damaging externalities of the industrial economy.

If you wish to die on the hill of nation states, then ask yourself the following questions. How would you start a new collective value system emphasising the full range of adaptive factors beyond financial value, in a highly evolved country of hundreds of millions of people? How could you systematically mitigate existential risks from powerful technologies, when you are at war with other countries who ignore it to their competitive advantage? How would you implement a new economic system that values planetary externalities, when doing so puts you at economic disadvantage?

Succeeding nation states is no guarantee of answering these questions well. Like all technological change, it will certainly bring with it the potential for newly undesirable futures. Crucially, however, it is a blank slate that represents the possibility of a new system that is built with such questions in mind. This leaves us with a final question — looking back in 200 years time, did we flourish because we stuck with nation states, or because we succeeded them?

As new state-like things becomes more practical, it would be easy to see the transition as anarchic; a technological deliverance of the libertarian ideal. Relative to the current world, it is that, bringing the opportunity for highly divergent, localised self-determination combined with freedom of choice.

But it’s also much more than that. It’s an opportunity to, from the ground up, build technological connective tissue between blank states that recognise society as a deeply interconnected global individual, intrinsically intertwined with coupled planetary systems. For the first time, we can design the foundations of a truly globalised society that coordinates information and resources to overcome the coordination problems that plague the industrial organism; baking it in before new states evolve immune systems that reject a retrofit.

We can become a global intelligence; our capacity to process information rising to the complexity of the problems we face. Team Earth — fundamentally built upon global, cooperative, systems-aware principles.

• Seek to build open societies that maximise collective information processing through the maximum number of honest perspectives, integrating a global public to solve global coordination problems.

• Seek to overcome the multipolar traps that poison our actions and interactions, intelligently balancing the health of the collective with the needs of the individual.

• Seek to find ways to iterate continuously in how we organise society, to enable faster evolution of the meta-societal organism without getting stuck in a local maximum.

• Seek to develop collective values that are more than just financial, and systematically recognise the manifold of complex causality that drives society as an adaptive organism.

• Seek to recognise that the planet is our collective body with its own limits, and create deeply integrated systems of intentional feedback between human behaviour and planetary processes.

• Seek ways to develop technologies and ecologies of technologies with the capacity for existential risk, while maintaining the collective wisdom to know how to use them without destroying ourselves or our values.

• Seek to discover forms of empirical governance, able to orchestrate policies that prioritise effective long-term decisions over short term economic gain, extending beyond the cognitive biases that plague human thinking.

“Don’t fight forces, use them.“ - R. Buckminster Fuller

Human history is a blip on the tail end of a 13.7 billion year-long trajectory of increasing complexity. Atoms → molecules → organelles → protists → eukaryotes → multicellular organisms → tribes → societies; continuous integration to produce new individuals at higher levels of organisation. Each step seemingly impossible, yet here we are — evolutionary tailwinds at our back.

Coordination of a global intelligence is just the next major evolutionary transition. Zoom out, and we find ourselves caught between the two worlds. On the other side, we can catch a distant glimpse of a societal collective intelligence, a technosphere, coming into alignment with its foundational bio- and geospheres. Energy and information reuniting at a higher level to produce an autopoietic planetary intelligence from the interactions between its concentric spheres.

A blueprint for galactic planetary colonisation.

“A transition to planetary intelligence… would have the hallmark property of intelligence operating at a planetary scale. Such planetary intelligence would be capable of steering the future evolution of Earth, acting in concert with planetary systems and guided by a deep understanding of such systems.” — Intelligence as a planetary-scale process (Frank, Grinspoon & Walker, 2022)

It is common to remark that we are born too late to explore the Earth, and too early to explore the galaxy. That we are the awkward middle generation, forced to settle for cat memes and junk food. However, we are arguably the most important generation that will ever exist — confined to a single planet, building the technology of gods with the wisdom of gnats. Until we cash our life insurance policy, we exist at a crucial squeeze point with the power to ascend to intergalactic paradise engineers or destroy ourselves entirely. A great filter.

We are the most complex known thing in the universe, building on the lives of more than 100 billion ancestors. There is every reason to believe that there can be trillions more, human and artificial. The decisions we make and the systems we build will have massively disproportionate impact on human evolution. We owe it to ourselves, our ancestors, our descendants and the universe to get them right.

In the coming decades, it’s likely that you will be forced to choose a side on a new axis of global superpowers. They will make you believe that their fight is your fight; that without your help, evil will prevail. It will feel inevitable, national preservation a necessity.

When that time comes, don’t choose Team America or Team China. Choose Team Earth. Not to be overly dramatic, but the fate of the human race depends on it.

“Birds flying high, you know how I feel. Sun in the sky, you know how I feel. Breeze driftin' on by, you know how I feel. It's a new dawn. It's a new day. It's a new life, for me. And I'm feeling good.” — Feeling Good (Nina Simone, 1965)

If you found this essay valuable, please consider collecting it as an NFT. Think of it like funding research, and enabling me to do more of it. Like buying a book in reverse, after you decided it was already worth it. If you do, know that it is deeply appreciated.

collect://

References

I’ve referenced relevant resources in context. This work was most influenced by consuming the work of the following people, who you should absolutely check out.

• Daniel Schmactenberger

• Nathan Hagens

• Nick Bostrom

• David Krakauer

• Jessica Flack

• Sara Walker

• Balaji Srinivasan

Thanks also to Midjourney for being a thinking partner. All images that are not otherwise referenced were produced using the prompts in the captions (with an additional modifier of “—aspect 16:9”).

If I got something wrong, please accept my apologies and reach out so I can edit. May the blessing of 14,000 sheep, 6000 camels, 1000 teams of oxen, and 1000 female donkeys be upon you.

While it can be tempting to write them off, especially in ‘serious’ innovation domains, their structure gives them unique capabilities as vehicles for discovery. We’ll take a short walk through their structure, and finish on the following point:

DAOs are novelty search engines which can more efficiently explore a search space by enabling many cooperating teams to collect and integrate stepping stones.

Novelty search - a paradigm of discovery

Imagine yourself at the shore of a lake. You intend to reach the other side, and must do so by hopping across stepping stones on the surface. However, the lake is shrouded in a dense fog, which obscures all but the stepping stones closest to you. As you walk across the stepping stones, the shore dissolves from sight behind you.

Eventually, you find yourself at a fork; a decision to be made. Which way is best? The fog ensures that you can’t see where either path leads. Still, you must make a choice.

Stanley & Lehman (2015) draw this analogy in relation to discovery. The lake represents the abstract space of all possibilities, and the stepping stones represent strategies for navigation. Your journey is a process of searching through the space of all possibilities. The question is, how do you navigate when you don’t know the nature of the territory?

We can reframe this by thinking of the space of all possibilities as a room. Now imagine yourself an artist, searching through the room to discover the next Monet. Conceived as a search space, we see that this rare diamond is already there. Your purpose is merely to find it among the plethora of meaningless dead-ends.

As you paint, you search through the room. What you paint will be influenced by the parts of the room you have already visited. If you’ve spent time working with modernism, you’re likely to be influenced by it. Without having visited the watercolour corner, you are unlikely to invent them. The problem of the lake arises here again. How do we reach the Monets? What stepping stones lead us there?

We can think of all complex domains in this way. We know there are new discoveries to be made, if only we knew how to get there. Unfortunately, we can rarely see beyond the most immediate stepping stones.

Consider vacuum tubes. It was obvious to nobody that they would enable the first computers. This only became clear after vacuum tubes and associated computing discoveries had been made, so that someone could make the connection. If you had set out to build a computer in the 1800s (as Charles Babbage did), it’s unlikely that you would have drawn this insight.

We are biased to believe that the interim steps resemble the final goal, and use objectives to reach stepping stones that appear to lead to the final destination. However, without knowing the nature of the territory, those objectives may well lead us to get stuck.

It’s perfectly possible that moving closer to the goal actually does not increase the value of the objective function, even if the move brings us closer to the objective.

Why Greatness Cannot Be Planned - Stanley & Lehman (2015)

Consider education. We measure improvement in education based on test scores, on the assumption that better test scores mean we’re getting smarter. “But the Math scores are up! That’s good, right?” Wrong. Optimising for Math scores through assessment forces us to invest in things which improve short term scores, while preventing the exploration of the larger search space.

So what’s the alternative? Well, novelty search is a form of non-objective search. Instead of following the stepping stones that appear to lead to the objective, simply collect stepping stones that lead in interesting directions. Focus on newness, regardless of where it leads.

Because eventually you have to acquire some kind of knowledge to continue to produce novelty, it means that novelty search is a kind of information accumulator about the world in which it takes place. The longer the search progresses, the more information about the world it ends up accumulating.

Why Greatness Cannot Be Planned - Stanley & Lehman (2015)

Each stepping stone you collect opens up new possibilities. New ideas tend to arise from the combination of existing ideas; from combining stepping stones together. Every now and then, you land on a combination that lifts you into an entirely new context.

The more stepping stones you collect, the more interesting combinations you can create. You can’t predict where they’ll take you, but keep doing it, and you can predict you’ll end up somewhere interesting.

Convergence and divergence

Companies and DAOs both exist on spectrums that can’t be painted with a single brush. However, their core structures have different organisational characteristics that optimise them for different roles in the future of innovation.

Companies are a convergent tool for pursuing singular visions

Great ideas tend to arise from a single mind that has collected and synthesised an unusual series of stepping stones.

Companies are hierarchies that are led by a single person who is ultimately responsible for decision-making. For ideas that require many people to materialise, companies are an ideal structure for making them real, since leaders are empowered to optimise their resources in pursuit of their vision.

This is a feature (not a bug), and is both what makes companies strong and simultaneously brittle. A strong leader with a strong vision can guide a company to achieve incredible things. A weak leader with a weak vision can't achieve much, and may in fact cause harm.

In either case, the point of a company is to align everyone to achieve something specific; to converge everyone's work on a centralised axis. This is "strong alignment", in the sense that everyone must work together to achieve a convergent outcome.

A consequence of strong alignment is the need for objectives. Leaders use objectives to distil their vision into something that can be easily understood, and enable each individual to put their work into the context of the whole. Without them, it risks dilution of a strong vision into a mosaic of interpretation.

The strength of the hierarchical model is the ability to pursue singular visions arising from a leader, by aligning the work of many individuals towards a convergent outcome. To do so, companies must carefully define the problem they want to solve and maintain a strong focus by measuring their progress against interstitial objectives.

However, this prevents companies from exploring the adjacent possible via stepping stones that do not resemble their objective. In turn, it causes them to lose out on their combinatorial possibilities.

DAOs are a divergent tool for exploring the territory

DAOs are networks that are defined by nodes and links, corresponding to people and relationships.

Networks have a flexible topology that can re-organise in response to change, adopting new shapes that evolve over time. Leadership is distributed contextually and individuals are empowered to make decisions that meaningfully change the structure of the organisation, without requiring permission or consensus. We can call this ability self-organisation.

Consider a large network arrangement like a democratic nation state. A nation is composed of many individuals and organisations all contributing to its growth and change. However, the complexity of a nation ensures that no single entity can appreciate it as a whole, not even a President.

Instead, each is embedded in a context, like a local community, a business, or a social circle, and interprets the nation in its own way to make decisions that affect that context. As they make decisions, like starting a business, voting on governments, or forming new relationships, the structure of the network meaningfully changes in response.

This makes it challenging for nations to pursue a specific, unified agenda. Instead, they more successfully focus on highly interpretable goals like increasing GDP and improving the quality of life.

Similarly, in a mature DAO, individuals will be able to understand a part of the network (local context), but they will not be able to appreciate it as a whole (global context). Individuals will interpret the DAO in their own way, and explore directions that affect their local context without having to consider the context of the whole. As they make decisions, the structure of the network will self-organise in response.

Through this mechanism, networks have an information advantage. Individuals know more about themselves than anyone else, and decide for themselves how they are best able to contribute.

…it is intuitive that specification and pricing of all aspects of individual effort—talent, motivation, workload, and focus as they change in small increments over the span of an individual’s full day, let alone months—is impossible.

…hierarchical organization is a lossy medium. All the information that could have been relevant to the decision regarding each factor of production, but that was not introduced in a form or at a location that entitled it to “count” towards an agent’s decision…is lost.

Coase’s Penguin - Yochai Benkler (2002)

Networks are also able to allocate people more efficiently, since anyone can take on any role and the network can benefit from the best combinations.

…different people will be more or less productive with any given set of resources and collaborators for any given set of projects, and that this variability is large.

Peer production has an advantage over firms and markets because it allows larger groups of individuals to scour larger groups of resources in search of materials, projects, collaborations, and combinations…

Coase’s Penguin - Yochai Benkler (2002)

Tokens, meanwhile, create a mechanism for “weak alignment“, in which everyone in the network has a shared incentive to increase its value. However, the way in which that value is achieved is open to interpretation.

The consequence is that DAOs are not suited to pursuing one vision. They are optimised for exploring many divergent visions simultaneously, emerging from individuals making decisions based on their local context. This is a feature (not a bug), and will be the source of strength for the networks to come.

This is an uncomfortable thought for some. “There’s nobody steering the ship!”. “There’s no roadmap!”. No - and that’s the point. They are autonomous. Autonomy derives not just from smart contract logic, but from a continuous process of distributed self-organisation.

DAOs are containers for many weakly-aligned teams

DAOs and companies have different organisational characteristics arising from their core structures. However, this does not mean to say that they are mutually exclusive - quite the contrary. Instead, DAOs can be conceived as containers for many weakly-aligned hierarchies.

Great ideas tend to come from a single mind and leadership is a requirement for effective human collaboration. Hierarchies are well-optimised to enable strong leaders to achieve strong outcomes. DAOs do not get around this fact. Hierarchies work because they enable leaders to synthesise the contributions of a group.

Leaders with differing ideas naturally surface in groups of people, and it is equally natural for people to follow leaders and causes they believe in. Rather than replacing hierarchies, DAOs create a mechanism through which a single organisation can allow many hierarchical teams to explore different directions simultaneously.

Problems tend to arise in hierarchies due to groupthink: a progressive inability to consider alternative points of view.

Homogeneous groups are great at doing what they do well, but they become progressively less able to investigate alternatives.

The Wisdom of Crowds - James Surowiecki

In a network arrangement, each team interprets its own direction based on local context, and contributes what they believe will increase its value, without having to align to a singular vision or narrative. There is no need for consensus or compromise, because each group is empowered to self-organise.

…the best collective decisions are the product of disagreement and contest, not consensus or compromise.

The Wisdom of Crowds - James Surowiecki

In this way, a single DAO can be seen as a multi-organisational network. In times past, these structures have struggled to exist due to lack of the technologies required to coordinate. Now, weak alignment with a programmable token gives each team an incentive to cooperate and share information, while open blockchains make information increasingly easy to share.

Its key principle is heterarchic (or, to offer another term, “panarchic”) collaboration among members who may be dispersed among multiple, often small organizations, or parts of organizations. Network designs have existed throughout history, but multiorganizational designs are now able to gain strength and mature because the new communications technologies let small, scattered, autonomous groups to consult, coordinate, and act jointly across greater distances and across more issue areas than ever before.

Tribes, Institutions, Markets and Networks - David Ronfeldt (1996)

DAOs of the future will benefit from the best of both worlds. Strong, unified hierarchies to achieve convergent outcomes in a local context, exploring many divergent directions while cooperating with one another to improve the state of the whole.

Vehicles for exploration

Let’s return to our search for a Monet. We know it’s out there somewhere, we just need to find it.

As we identified, objectives aren’t much use when we don’t know the nature of the territory. Instead, we should collect stepping stones that lead in interesting directions, without knowing exactly where we’ll end up.

Companies help us hop between single stepping stones by using a hierarchy to pursue a singular vision. However, they rarely move beyond this because this same mechanism prevents the exploration of the wider search space.

The magic of DAOs is to embrace complexity by enabling many cooperating teams to collect divergent stepping stones. Those stepping stones can be shared, reused and combined in new contexts, leading in new directions.

DAOs are a generating system for discovery, with people and information as its parts.

A generating system… is a kit of parts, with rules about the way these parts may be combined. Almost every ‘system as a whole’ is generated by a ‘generating system’. If we wish to make things which function as ‘wholes’ we shall have to invent generating systems to create them.

Systems Generating Systems - Christopher Alexander (1968)

As such, DAOs are a mechanism for open-ended evolution. Or rather, evolved open-endedness: a way to not only generate solutions to directions, but to progressively generate entirely new directions.

…instead of thinking of the open-endedness as existing conditions or properties of the evolutionary system, we consider them as the outcome of evolution itself.

Evolved open-endedness, not open-ended evolution - Pattee & Sayama (2019)

And so we arrive at the crux of the argument. DAOs are novelty search engines which can more efficiently explore a search space by enabling many cooperating teams to collect and integrate stepping stones.

This capability suggests a unique and interesting role in the future of innovation. Where we end up is unpredictable, but predictably interesting.

*Taken from Self-organisation in communicating groups (Heylighen, 2015).

Thanks to Gordon Brander’s excellent Subconscious newsletter for generative insight and references that helped shape this article.