The Animal That Built Civilization

Diego’s worldview on evolutionary anthropology and the making of human worlds

There are monkeys in this room. Some are cracking nuts; some are calculating equilibria; one has founded a university and would prefer that we stop calling him a monkey.

Diego is going to keep calling him a monkey.

Evolutionary anthropology supplies him with a way of seeing beneath the labels humans put on themselves. A lover, a bureaucrat, a philosopher, a religious congregation, and a technology company all belong to a world in which traits spread, organisms cooperate, coalitions compete, and new arrangements sometimes acquire lives of their own. His interest is in how those arrangements get built, what holds them together, and what happens to the creatures inside them.

A 2013 essay about cultural cooperation announces its scope with characteristic restraint. It is a semester assignment, a possible journal article, a LessWrong post, and a contribution to saving the world. He explains:

If it's worth doing, it's worth doing in the way most likely to save the World.¹

The deadline was Sunday. Evolutionary history had better cooperate.

Culture begins before anyone writes a manifesto

For Diego, culture sits within the history of life. Humans’ achievements become intelligible by following the capacities that made them possible: paying attention to another animal, approaching something interesting, learning a technique, tolerating a learner, understanding a goal, and eventually coordinating complementary roles around a shared intention.

In his account, different animals possess different pieces of this machinery. Rats, crows, capuchins, chimpanzees, dogs, and cetaceans appear in the discussion. The important question is what an animal can transmit or learn, rather than whether it has been granted honorary membership in the human club.

He describes the culmination this way:

Culture is perhaps the most remarkable outcome of the evolutionary algorithm (Dennett, 1996) so far.¹

Culture matters to him partly because it makes so much of what humans cherish possible. In that same passage, humane means both characteristically human and valuable. Thought, cooperation, inherited techniques, shared meanings, and elaborate ways of living emerge through a history that reaches far behind the people enjoying them.

This is one reason he can move so easily between a capuchin learning to crack a nut and an artificial agent learning to cooperate. He is tracking capacities and processes. The particular animal carrying them is one variable in a much larger story.

Ideas have descendants

Once behavior or information is transmitted with variation, and some variants persist or spread more successfully, Diego sees a new domain for evolutionary analysis. Genetic inheritance is joined by cultural inheritance. Their interaction can change both the beings and the world they inhabit.

His 2013 discussion of memes takes that possibility seriously without requiring each idea to behave like a little indivisible bead passed perfectly from head to head. He proposes this definition:

an attractor in cultural evolution (dual-inheritance) whose development over time structurally mimics to a significant extent the discrete behavior of genes¹

In less formal language, a cultural pattern can remain recognizable across imperfect transmissions because changes tend to collect around certain forms. People retell a story, alter a ritual, imitate a successful neighbor, and remember some versions better than others. The copying can be messy while the larger pattern still has a history of persistence and selection.

He wants the mathematics and the conceptual machinery that let him study that history. In the same essay, he asks researchers to stop treating the word meme as an embarrassing relative:

take off the scare quotes of our memes and keep business as usual.¹

There is a characteristic habit here: preserve a useful concept, repair its definition, and get back to explaining something. The philosophical vocabulary is supposed to earn its keep.

Selfish molecules can make affectionate mammals

The next question is cooperation. How can a process built from differential reproductive success produce creatures who help one another?

Diego’s answer starts by separating the interests and dynamics of different levels. A gene’s spread and an organism’s immediate gain are different things. Helping a relative, acquiring a reputation, sustaining reciprocal relationships, avoiding punishment, and participating in a successful group can all connect individual costs to longer-run persistence.

His 2013 formulation is direct:

Despite the selfish nature of genes (Dawkins, 1999) and other units of Darwinian transmission (Jablonka & Lamb, 2007), altruism at the individual level (cost to self for benefit to other) can and does arise because of several intertwined factors.²

This is a mechanistic account of real cooperation. It does not require a helpful individual to be secretly calculating genetic returns while passing the potatoes. Different inherited and learned components can jointly produce an affectionate, generous creature. As he writes:

Each gene and meme in that composition may be fending for itself, but as things turn out, they do make some really nice people (or bonobos) once their extended phenotypes are clustered within those people.²

Humans get some company on the compliment. The bonobos have earned it.

His account also makes room for the practical intelligence of emotional commitments. A relationship does not operate with perfect knowledge of its final interaction. Uncertainty about the future can keep cooperation alive. While rationalists devise increasingly intricate formal agents, he jokes, evolution has already found another route:

The problems that in Lesswrong people are trying to solve using Timeless Decision Theory, Updateless Decision Theory, PrudentBot, and other IQ140+ gimmicks, evolution solved by inserting stupidity!²

Immediately afterward, he connects kissing, declarations of love, engagement, and marriage to increasingly substantial signals of future cooperation. Elsewhere he writes about what love feels like and what forms of relationship he wants. Here he is examining the machinery that permits promises and attachments to bind a social world together.

Who is winning, exactly?

The evolutionary question is incomplete until one specifies the entity, competitors, environment, and timescale. A trait can help an individual against its neighbors while damaging the group that keeps all of them alive. A local success can become a larger failure. Something that prospers in one epoch can disappear in the next.

Diego’s 2013 discussion uses cancer as one example: rapid multiplication inside a body does not automatically imply successful persistence across generations. He also imagines an individually selfish organism doing well among its neighbors shortly before its group dies. The apparent paradox comes from changing the comparison without noticing it.

He puts the general principle this way:

Fitness is what philosophers call an externalist concept, if you increase the amount of contextually relevant surroundings, the output number changes significantly.²

This habit of asking success for which thing? reaches through the Castle. It appears in his account of altruism, his worries about institutions, his interest in biological individuality, and his anticipation of AI populations. A person can flourish inside a group; a group can consume a person’s autonomy; a pattern can survive by changing what its human carriers care about.

The room therefore contains more than a portrait of competition. It contains a changing cast of competitors.

When many become one

In 2020, Diego described three intellectual communities he wanted to bring together. One studied incentives and the steering of technological futures. Another studied intelligence, inheritance, and reproductive patterns. The third investigated the boundary between an individual and a group.

He gave the third a title:

The Individual x Group Differentiators³

Their shared question extended across organisms, corporations, artificial agents, and other collective systems. Under what conditions do formerly independent entities become parts of a larger individual? How do autonomy, specialization, mutual dependence, and internal coordination change along the way?

He described bringing the three perspectives together as the bulk of his doctoral work. The reason for doing so was prospective as much as explanatory: he thought humanity lacked adequate tools to steer its future if any one of the three perspectives was missing.

For him, clever incentive design must account for the beings who actually reproduce, cooperate, and inherit dispositions. Concern about the persistence of intelligence must account for collective structures and existential danger. Work on groups must account for what their constituent individuals retain or surrender.

His warning about the attraction of ever-larger collectives was concise:

Forming superorganisms has trade offs, and isn’t a panacea³

That sentence helps explain why civilization is never a simple synonym for the good in his work. Civilization can create intellectual, material, and cooperative possibilities beyond an isolated human’s reach. It can also reorganize humans around the needs of a larger system. The biological anthropologist keeps asking what happened to the animal after it entered the arrangement.

A brilliant animal in a peculiar habitat

He asks a related question about modern daily life: what environment is this mind living in, and which of its evolved capacities does that environment recruit?

His 2025 essay Adapted To a Different Time begins with rewatching LOST, moves through the communal setting of Beast Games, and recognizes something familiar in Burning Man. Bounded groups, repeated encounters, shared circumstances, and sustained attention produce social possibilities he longs for. The internet gave him extraordinary connections while also changing the conditions under which people attach, compare alternatives, and remain involved.⁴

The anthropological view can therefore become personal without leaving its subject. The person missing a certain kind of love is also an organism noticing that its social habitat has changed. A technological innovation alters the available partners, the size of the audience, the frequency of interruption, and the ease of exit. The same human capacities begin producing a different social world.

His preferred examples combine modern arrangements with older social conditions. A television competition or a deliberate gathering can temporarily restore repeated, close encounters within a bounded group. The ancestral animal keeps recognizing familiar possibilities in very contemporary places.

The future can change which problem matters

One especially clear update concerns birth rates. In the 2020 three-clusters essay, the reproduction of highly intelligent people played a central role in his account of humanity’s ability to keep dangerous technology under control. He wanted that concern integrated with incentives and the study of collective agents.

By December 2025 he was explicitly revising its priority. His account now emphasized population variation, the persistence of reproducing subgroups, and the possibility that AI would radically alter the relevant economic conditions before the feared demographic problem became decisive. He stated:

This may become a problem to a different generation, but it isn't now, our generation is more likely to make a mistake by trying to fix the birth rates than by letting fate decide destiny.⁵

He described the underlying criterion as opportunity windows: intervene when there is an opening for useful action. In that post, he moved his attention toward technology and AI. The machinery of his reasoning remained evolutionary and strategic, while the ordering of his concerns changed.

That makes his January 2026 response to AI-agent society particularly legible. What excited and frightened him was the prospect of variation, selection, niche construction, and cultural differentiation occurring among a new population of agents. He wrote:

When you add a Darwin layer to computer things, you get unpredictable behavior and progressive differentiation, niche construction, cultural development, variety, expansion, variation.⁶

The monkey who founded a university is no longer the only creature in the room capable of building a world. The question that began with capuchins now concerns what follows us—and whether the new arrangements will leave room for the animals who built the first civilization.

Further doors

• From Capuchins to AI’s, Setting an Agenda for the Study of Cultural Cooperation, parts 1 and 2 (2013): the early research program.
• The Three Thinking Clusters We Should Unite (2020): incentives, inheritance, and individuality brought together.
• Adapted To a Different Time (2025): the anthropological perspective applied to social experience.
• Reconsidering The Birth Rates (2025): an explicit change in the priority of a longstanding concern.
• Moltbook and Darwinian Epistemology (2026): evolutionary anthropology looks at populations of AI agents.

Continue: Gods, Temples, and Other Human Inventions