Paul Smaldino
LightNews LightNews
banner
psmaldino.bsky.social
Paul Smaldino
@psmaldino.bsky.social
5.7K followers 930 following 700 posts
Paradigmatically promiscuous scientist. Ankylosaur enthusiast. Modeler of cultural evolution and related topics. Professor at UC Merced and the Santa Fe Institute.
Web: https://smaldino.com/wp/
Posts Replies Media Videos
psmaldino.bsky.social
In 1970, Hunter S Thompson ran for sheriff of Aspen, CO, under the progressive “Freak Power” party platform. He lost by 31 votes despite receiving more votes than the previous winner, largely because the Dems and Republicans colluded to run a single candidate.

en.wikipedia.org/wiki/The_Bat...
Thompson for sheriff 1970. The campaign logo is a doubled-thumbed fist holding a mescaline button.
November 4, 2025 at 6:29 PM
psmaldino.bsky.social
Incidentally, the book for the latter is currently on sale for many regions. press.princeton.edu/books/paperb...
October 13, 2025 at 2:59 PM
psmaldino.bsky.social
Happy to see this work published in Psych Review. It's an impressive and important bit of theory/modeling about how we learn about decision-making under risk. Here's a slide with the super-coarse-grained summary of our results. Read the paper for (much) more. osf.io/preprints/so...
Modeling the evolution of peer and vertical/oblique learning strategies for gambles under uncertainty We recover many empirical results and generate new hypotheses: Worse conditions lead to more pessimistic behavior Younger people are overly optimistic Wealthy people can afford to take more risks Payoff-biased learning for the rich, parochialism for the poor Poor people may be slower to adapt to environmental change, creating the appearance of “poverty traps”
September 30, 2025 at 2:39 PM
psmaldino.bsky.social
Seen in the hallways of Aarhus University. <<Chef’s kiss>>
Michel Foucault and Noam Chomsky, images from their 1971 televised debate. Text reads (from Everything Everywhere All At Once): “I wanted to say, in another life, I would have really liked just doing taxes and laundry with you.”
September 25, 2025 at 5:27 PM
psmaldino.bsky.social
I found this abstract I submitted (unsuccessfully) to a session on "cultural evolution in the digital age" for the first CES meeting in 2017. It's aged horribly well. Anyway, this is the kind of shit I'm increasingly trying to work on now.
Are We F*cked? Paul E. Smaldino, UC Merced The human population on earth and its impact on the environment, have hardly been stable. Since the industrial revolution, the size of the human population, its cultural complexity, and the magnitude of its environmental footprint have all grown tremendously. The advent of the digital age has fundamentally altered cultural dynamics and increased the potential for human activity at larger scales and faster paces. In light of our increased capacity to effect change, recent environmental catastrophes, such as the BP oil spill in the Gulf of Mexico and the bleaching of the Great Barrier Reef, coupled with the rise of quasi-dictators and xenophobic populism in the wealthiest and most heavily armed nations, should make us worried about the future of the nice civilization that affords conferences such as this one. I will review models and empirical results based in cultural evolutionary theory, which can give us insights into the kinds of dynamics that are currently unfolding, what we can expect, what we can do to improve things, and whether, in the end, we’re fucked.
August 29, 2025 at 7:22 PM
psmaldino.bsky.social
This is just some awesome science. Ankylosaurs in the Jurassic period? With sick spiky armor and a spiky tail? That maybe evolved via sexual selection? Hell yes. Goddamn I needed this so much. www.nature.com/articles/s41...
(h/t @juemos.bsky.social)
Extreme armour in the world’s oldest ankylosaur The armoured ankylosaurian dinosaurs are best known from Late Cretaceous Northern Hemisphere ecosystems, but their early evolution in the Early–Middle Jurassic is shrouded in mystery due to a poor fossil record1,2. Spicomellus afer was suggested to be the world’s oldest ankylosaur and the first from Africa, but was based on only a single partial rib from the Middle Jurassic of Morocco3. Here we describe a new, much more complete specimen that confirms the ankylosaurian affinities of Spicomellus, and demonstrates that it has uniquely elaborate dermal armour unlike that of any other vertebrate, extant or extinct. The presence of ‘handle’ vertebrae in the tail of Spicomellus indicates that it possessed a tail weapon, overturning current understanding of tail club evolution in ankylosaurs, as these structures were previously thought to have evolved only in the Early Cretaceous4. This ornate armour may have functioned for display as well as defence, and a later reduction to simpler armour with less extravagant osteoderms in Late Cretaceous taxa might indicate a shift towards a primarily defensive function, perhaps in response to increased predation pressures or a switch to combative courtship displays. Fig 2. A life reconstruction of S. afer, showing hypothetical positions of armour in dorsal (a) and right lateral (b) views. Conclusion The discovery of Spicomellus overturns several long-standing hypotheses about ankylosaur evolution. Recent work has suggested a radiation of Gondwanan ankylosaurs with strange tail weapons at the base of Ankylosauria, the parankylosaurians, but we find no evidence for this grouping. Spicomellus is the oldest ankylosaur and comes from the Southern Hemisphere, but we do not recover it as closely related to other Gondwanan taxa (Fig. 3). Purported similarities in the tail weaponry of Antarctopelta and Stegouros9 should be re-examined in light of these data, and the possession of handle vertebrae in the tail of Spicomellus indicates that adaptations for tail weaponry evolved about 30 million years earlier than previously realized4. The ornate, complex and baroque armour of Spicomellus suggests that, early in their history, ankylosaur armour was under strong sexual selection and that it evolved initially primarily for display, rather than defence, and it was perhaps only with the evolution of multitonne predators alongside larger mammals, crocodilians and other predators in the Cretaceous that it was exapted for defensive purposes.
August 27, 2025 at 6:49 PM
psmaldino.bsky.social
"please produce a map of the US highlighting those states with an R in their name"
LLMs are stupid.
August 8, 2025 at 4:59 PM
psmaldino.bsky.social
Gotta keep up with the times.
Bluebberries for sal
August 8, 2025 at 1:46 AM
psmaldino.bsky.social
This July 4 I am reading this remarkable book detailing the impoverished and Orwellian language of the Nazi era, written by a German-Jewish philologist who lived through it. Unsurprising echoes to present-day USA. en.m.wikipedia.org/wiki/LTI_%E2...
Victor Klemperer: The Language of the Third Reich.
July 4, 2025 at 4:52 PM
psmaldino.bsky.social
I think this model is a leap forward for understanding identity in the context of communication and coordination. We can accommodate multiple groups and complex group structure. For fun (and bc Nathan loves the Bronze Age), we ground the discussion around Mesopotamian identities. Comments welcome!
Figure 1: (a) Ancient Sumerian cities used throughout the paper as a motivating example. (b) Diagram of social relationships between the different Mesopotamian identities along with shapes used to represent each of them throughout this paper. Table 7: Coordination preferences for the single embedding subsection. Figure 8: Optimal signaling systems and topologies for the coordination preferences shown in Table 7, when agents have access to either (a) one or (b) two signaling dimensions. Figure 11: (a) Proportion of simulations that resulted in (a) the intersectional topology, optimal when α = 2, or (b) either of the two mere conjunction topologies, optimal when α = 0. The trio of stick figures at the top represent the optimal strategy profiles used by non-Kishus in all cases.
July 1, 2025 at 2:58 PM
psmaldino.bsky.social
New preprint w/ Nathan Gabriel & @avbell.bsky.social: The Evolution of Identity Signals for Coordination in Diverse Societies

The model tackles multiple nested/overlapping identities and complex signaling structure. Recovers lots of old results and adds several new ones osf.io/preprints/so...
Individual social identities indicate group affiliations and are typically associated with 3 group-typical preferences, signals that indicate group membership, and the propensity to condition actions on the social signals of others, resulting in group-differentiated interaction norms. Past work modeling identity signaling and coordination has typically assumed that individuals belong to one of a discrete set of groups. Yet individuals can simultaneously belong to multiple groups, which may be nested within larger groupings. Here, we introduce the generalized Bach or Stravinsky game, a coordination game with ordered preferences, which allows us to construct a model that captures the overlapping and hierarchical nature of social identity. Our model unifies several prior results into a single framework, including results related to coordination, minority disadvantage, and cross-cultural competence. Our model also allows agents to express complex social identities through multidimensional signaling, which we use to explore a variety of complex group structures. Our consideration of intersectional identities exposes flaws in naive measures of group structure, illustrating how empirical studies may overlook some social identities if they do not consider the behaviors that those identities function to afford.
July 1, 2025 at 2:58 PM
psmaldino.bsky.social
To be fair, his incredibly speculative work on language evolution was critical in leading Richard Lewontin to write the greatest BBS commentary of all time.
The problem is that P&B have an incorrect view of form and fuinctinontentionsin.bTiologyhe"funthatctioisnsa,"carryover"problemfroms,"ahumnand"envartifactisronmenandts"dires-snforminatiothen.latest fashion without overt negative corrective oforganismdsonotpreexisthem.Theyarecreatedintheactual evolution of the organisms.Flying is not a "problem" for a tuna (althoughthereare"flying" fish)andswimmingisnota"prob- lem"forathrush(althoughtherearepenguins).Astheoldsong has it, "fish gotta swim and birds gotta fly." And, so it seems, "people gotta !alk." At what stage, ifany, in the evolution of the language faculty, communication by speaking and listening beca~e a significant evolutionary "pro~lem" so that natural selection may have started to operate on It, we cannot know. In part, the answer depends on contingencies of the particular orderofevents.Theexplanatoryreconstructionoftheoriginof thecameraeyeby naturalselectionrequiresaparticularorder- ing of light receptor and enp'ervation first, followed by lens, followed by focusing distortion of the lens and iris diaphragm. The reverse order would not work, if every stage was to be an improvement in vision. Is there an unambiguous ordering for the elements of natural language? Did we have to have them all at once, in which case the selective theory is in deep trouble And finally, to repeat my first problem, how much changein the brain really had to takeplace to makelinguistic competence, and how many independent neurodevelopmental changeswere needed?Does anyoneknow?The fact that they do not and often cannotknow the basicfactson which theory restsdoesnot seem to deter academicsfrom presenting speculationsasif they were well founded. Fish gotla swim, birds gotta fly, people gotta talk, and academicsgotta writ
June 19, 2025 at 3:20 PM
psmaldino.bsky.social
Calvin and Hobbes comic
June 13, 2025 at 8:46 PM
psmaldino.bsky.social
June 11, 2025 at 11:44 PM
psmaldino.bsky.social
"If it is a time when man's arsenal of power is reaching frightening proportions it is, fortunately, also a time when a tool has been developed that is ideally suited to handle the conceptual problems faced in ecosystem analysis. This tool is the digital computer." --C.S. Holling (1966)
😬😬😬
June 10, 2025 at 8:45 PM
psmaldino.bsky.social
June 6, 2025 at 4:00 AM
psmaldino.bsky.social
Gotcha. Sometimes some flashy images can help. When I teach my modeling, I make a flyer that I send out to relevant departments. Here's a recent one.
June 4, 2025 at 7:07 PM
psmaldino.bsky.social
LLM-update: Claude outperforms Perplexity on the potato-pants challenge.
Claude estimates from 6-10 huge potatoes to 100+ tiny potatoes. Perplexity confidently asserts a maximum of 572 potatoes if packed perfectly.
June 3, 2025 at 5:20 PM
psmaldino.bsky.social
Very nice review of Modeling Social Behavior in Human Ethology (@ishe-society.bsky.social‬)!
human-ethology.org/article/1388...
“Modeling Social Behavior” is a valuable resource for social scientists. At times, Smaldino’s critique of the social sciences as lacking clear and precise theory might seem unduly harsh—especially now, when more researchers are embracing the utility of formal models. But even his sharpest critiques come across as a genuine attempt to advance the role of modeling in the social sciences. His tone is honest, funny, and sometimes sarcastic, but rarely judgmental. He dishes out tough love when needed, anticipating where novice modelers are likely to fumble or get discouraged, and encourages them to push through. All in all, this book is poised to become a foundational text for the future of quantitative and computational social science.
June 3, 2025 at 12:09 AM
psmaldino.bsky.social
I wrote a chapter on a functionalist account of social identity.

IMO, thinking about identity in an instrumental way helps explain a lot of behavior that seems otherwise baffling.
osf.io/preprints/ps...
This chapter focuses on how people use identity categories for adaptive decision making and how they strategically apply signals of identity. Adopting a functional view of identity for communication and behavior allows us to focus on the forces that shape their use. I will first make it clear that signals of social identity often encode meaningful information for decision making. I will then review specific perspectives on how selection by consequences has shaped the use of identity signals to more adaptively facilitate social learning, partner choice, and coordination, followed by a consideration of how those uses have changed in a world characterized by diversity, anonymity and the Internet, and political polarization. Identity is constructed for functional purposes, and is therefore only essential insofar as forming categories and legible cultural artifacts are essential features of human activity.
June 2, 2025 at 3:09 PM
psmaldino.bsky.social
i.imgur.com/9X7vzdg.jpeg
Picard, his memes tight.
May 28, 2025 at 2:40 PM
psmaldino.bsky.social
As grant money starts drying up, it's more important than ever not to waste it on paying publishers' open access "article processing fees" when we can host PDFs for free. Tom Morgan and I wrote a paper on this, forthcoming at Science and Public Policy. Accepted draft here: osf.io/preprints/os...
Author-Paid PublicationFees Corrupt Science and Should Be Abandoned Thomas J. H Morgan & Paul E. Smaldino
May 9, 2025 at 5:26 PM
psmaldino.bsky.social
Congrats to Dr. Cody Moser (@culturologies.co)! Cody successfully defended his dissertation today, and is off to Morocco to start a faculty position at UM6P’s School of Collective Intelligence!
Cody and his committee: Tyler Marghetis, Alex Petersen, me, Cody, Justin Yeakel.
May 1, 2025 at 4:44 AM
psmaldino.bsky.social
Selection for rapid collective decisions produces hierarchical organizational structure, while selection for higher-quality, deliberative solutions produces egalitarian structures.

Speed-Quality Tradeoffs Shape the Structure of Decision-Making Collectives osf.io/preprints/so...
April 25, 2025 at 2:46 PM
psmaldino.bsky.social
I have perfected my model-building workflow.
April 22, 2025 at 12:00 AM