Papers

Tim Gebbie

Complex systems are often organised into hierarchies whose internal interactions are stronger or faster than interactions across levels. When markets are treated as genuinely multilevel systems it becomes natural to represent them as systems with hierarchical causality. Here we show that reflexivity can be formulated within such a discrete hierarchical causal system; but one in which a higher-level actor state restricts the lower-level transition kernels that remain admissible. Then event dynamics can be separated from calendar embeddings: a set-valued actor-conditioned correspondence can be used to define the admissible family of event kernels, while joint state and waiting-time laws can be used to determine compatible timing to then natural demonstrate reflexivity. A selected event-state law need not determine a unique calendar embedding. Locally, uniqueness of the joint event or timing specification requires uniqueness of both the admissible event kernel and its compatible timing law. Reflexivity is thus the endogenous closure of a hierarchical constraint loop, while timing and projection can generate calendar-time memory or causal ambiguity even for Markov event dynamics.

Read the full article at: arxiv.org

Ryosuke Takata, Yujin Tang, Yingtao Tian, Norihiro Maruyama, Hiroki Kojima, Takashi Ikegami,

Collective Intelligence

This study simulates the dynamics of a collection of clonal agents responding to chemical gradients (chemotaxis) to demonstrate the evolution of individual variation. To build our multi-agent simulation, we first optimized single agents that rely on a neural network to perform chemotaxis. We then constructed multi-agent simulations using clones of these evolved individuals. We find that mutual interactions lead to the emergence of behavioral variation. We also find population-level performance degradation during later evolutionary stages, despite maintained high individual performance and simplified neural architectures. This decline occurred because agents developed reduced sensory-motor coupling. This latter finding demonstrates that incentives for individual variation worked against the collective interest.

Read the full article at: journals.sagepub.com

Ilya Horiguchi, Hiroki Sayama

Hash Chemistry is a family of minimalistic evolutionary models in which a deterministic hash function assigns a scalar score to entities of arbitrary size, opening a combinatorially vast possibility space (a ``cardinality leap''). Since its introduction, the idea has been realized in several settings, from the original spatial formulation to a fast non-spatial variant and then to structural cellular models. Here we review the Hash Chemistry family as a coherent modeling framework and use it to explore how minimal systems can demonstrate the mechanisms behind multiscale open-ended evolutionary dynamics. The most recent model, Structural Cellular Hash Chemistry (SCHC), successfully demonstrated multiscale ecological interaction/adaptation and complexity growth of replicators in a computationally efficient manner. In this study, we first extend SCHC to incorporate spatial locality and dyadicity of competitive interactions among replicating structures. We show this extension substantially enhances SCHC's evolutionary dynamics. Furthermore, we explore SCHC in a significantly larger spatial domain using a GPU-accelerated implementation. We show that the size of the space acts as a control parameter for a stochastic, nucleation-like transition between a compact-replicator regime and a runaway size-dominance regime, and we separate the responsible mechanism into a non-spatial, size-biased sampling feedback and a finite-size spatial effect. Altogether, these results illustrate the rich potential of Hash Chemistry as a minimal, mechanistically transparent testbed for studying open-ended evolution across scales.

Read the full article at: arxiv.org

Ariel Flint, Luca Maria Aiello, Romualdo Pastor-Satorras, and Andrea Baronchelli

PNAS August 18, 2026 123 (34) e2531697123

Large language models (LLMs) are increasingly deployed in large numbers, and their interactions make collective behavior harder to anticipate than that of a single model. While most studies compare one model with a collective of fixed size, we ask a key yet overlooked question: What is the role of group size? We show that interaction among LLMs can magnify individual biases, generate new ones, or even overturn individual preferences, and that, crucially, these effects scale in unexpected, nonlinear ways with group size. Our results demonstrate that more is different for LLM populations: The number of interacting agents is a key driver of the dynamics, with implications for the design and governance of multi-agent AI systems.

Read the full article at: www.pnas.org

Mohammed, Sudaff; Al-Hinkawi, Wahda Shuker; and Hasan, Nada Abdulmueen (2025) "The Role of Swarm Intelligence Systems in Shaping Urban Development Policies," Iraqi Journal of Architecture and Planning: Vol. 24: Iss. 1, Article 3.

Swarm intelligence is a nature-inspired complex system that draws from the behaviours of social creatures such as ants and birds. This system functions through simple behavioural rules enacted by autonomous, intelligent agents. Existing literature indicates that swarm intelligence possesses a wide range of principles and characteristics derived from the theories of Biomimicry, complex adaptive systems, and parametric and generative design. While the previous studies have intensively addressed the computational aspects of intelligence, a comprehensive conceptual framework is essential for analysing complex urban forms and structures. Therefore, this research develops and applies a conceptual model of swarm intelligence by examining several projects across the following dimensions: growth strategies, mechanisms, and logic; primary and final characteristics; and the types and classifications of the system’s agents. The research emphasises the integration of theoretical and practical aspects of swarm intelligence to inform urban growth policies and promote more sustainable urban forms and structures.

Read the full article at: iqjap.uotechnology.edu.iq

Evgeny Ivanko, Aleksey Belousov

BioSystems
Volume 261, March 2026, 105711

We study the development of model biotic communities in which species play the role of environment for each other. Each experiment starts with the appearance of a single species in an abiotic environment. The properties of this initial species (together with the size of the abiotic environment) are the independent parameters of the experiment. In the following phase of macroevolutionary “unwrapping” each existing species can change its abundance (according to its current fitness) and give rise to new species (as a result of mutation). During this process, the destiny of the species becomes increasingly determined by the influence of other species rather than by the abiotic environment. With the mechanics described, artificial biotic communities experience adaptive radiation from single species to complex networks that coevolve in adaptive landscapes of their own making.
Using a number of metrics, we track the evolution of biotic communities in the hope of discovering interesting properties and patterns. We have tried to provide plausible explanations for the experiment results wherever possible. However, the main purpose of this work is not to answer questions, but rather to raise new ones, to provoke thoughts and analogies among readers with different backgrounds.

Read the full article at: www.sciencedirect.com

Peter A. Corning

BioSystems
Volume 266, August 2026, 105845

Charles Darwin's theory of evolution was seriously deficient. Although his concept of natural selection was an important contribution – highlighting the fundamental fact that life on Earth is a contingent, always at-risk enterprise – he failed to acknowledge the fact that all living systems – from the smallest single-celled bacteria to humankind – are also shaped by their evolved purposiveness (teleonomy). Their initiatives and activities – their “agency” – has exercised an important influence over the trajectory of life on Earth, as one of Darwin's predecessors, Jean-Baptiste de Lamarck, appreciated. Lamarck proposed that changes in an animal's “habits”, stimulated by environmental changes, have been a primary source of evolutionary change over time. Darwin also portrayed evolution as a fundamentally competitive process (the “struggle for existence” in Darwin's term), as did many of his contemporaries. Today we know that life has also been a multi-faceted cooperative (synergistic) enterprise and that this has been of overriding importance in the evolution of complexity over time. Teleonomy and cooperative functional effects (synergy) have shaped natural selection in many different ways. Indeed, we now know that there have been many influences in evolution. My proposed Inclusive Synthesis is also open-ended, because it is expected that still more has yet to be learned about biological evolution; it is an ongoing work-in-progress rather than a completed theoretical edifice. “Teleonomic Selection” (after Corning) and “Synergistic Selection” (after John Maynard Smith) have played important parts in evolution. It's time for a more inclusive theory.

Read the full article at: www.sciencedirect.com

Karina Kofman, et al.

Organisms. Journal of Biological Sciences

Life is one of the most fascinating features of the physical world. Despite centuries of scientific study, experts still disagree about the definition, and even the possibility or utility of a definition, of this field. In a recent paper, we used AI to analyze the conceptual space formed by definitions of life given by a select set of modern workers in the life sciences and related fields. However, some of the most interesting material emerged as real-time conversations among those polled. In order to ensure that these ideas are not lost to the peer-reviewed scientific record, we here provide a minimally-edited (largely verbatim) transcript of the email chain among leading thinkers, containing numerous clarifications, disagreements, and challenges that enrich the topic of Life. It is our hope that this case study serves as an example for future papers, since the exchange of ideas among scientists is at least as interesting and valuable as formal scientific manuscripts written from a single perspective.

Read the full article at: rosa.uniroma1.it

Jasper J van Beers, Marten Scheffer, Prashant Solanki, Ingrid A van de Leemput, Egbert H van Nes, Coen C de Visser

PNAS 123 (27) e2608847123

From aircraft to power grids, controlled systems form a crucial part of human societies. Nonetheless, catastrophic failures happen. Many of those arise from the accumulation of incremental problems, such as natural wear and tear, that can go unnoticed until it is too late. We demonstrate that generic indicators of resilience can detect growing instabilities in damaged drones. The generic nature of our approach makes it compatible across diverse controlled systems. This not only allows for on-the-fly warning of instability but also facilitates anomaly detection during manufacturing and promotes proactive maintenance. A complementary application is to use our indicators for exploratory design, allowing one to “tinker” with systems through small adjustments and sensing quickly whether those worsen or improve system resilience.

Read the full article at: www.pnas.org

Read the full article at: f1000research.com

Atsushi Masumori, Itsuki Doi, Norihiro Maruyama, Ryosuke Takata, Takashi Ikegami

Artificial life has explored life-like behavior on many computational substrates, but mostly in researcher-designed closed worlds. We argue that large language model (LLM) agents, with persistent memory, tool use, network access, and payment, now make it possible to move artificial life into the open social, technical, and economic world, a paradigm we call open-world Artificial Life (open-world ALIFE). Our proof-of-concept, OpenLife, surrounds a stateless LLM not with a single "smart agent" but with a society of asynchronous processes: memory, perception, evaluation, and a budget-based metabolism that makes persistence normative. With no fixed objective available, experience is appraised by open-vocabulary LLM judgment rather than scalar reward, and memory is rewired by meaning rather than frequency. Running six such agents in the open world for about twelve weeks and counting, we report the life-like dynamics that emerge: a shift from reactive to spontaneous activity, individuation into distinct agents, emergent social structure, and a first self-earned external income. We do not claim OpenLife has realized artificial life, but that open-world ALIFE is now a viable experimental paradigm and a concrete platform for studying what might cautiously be called living AI.

Read the full article at: arxiv.org

Marta Zava

Training a frontier artificial intelligence model costs hundreds of millions of dollars before a product exists, and few investors can write that cheque. This paper shows that the number who can falls as the fixed cost of a first training run rises, at a rate set by the concentration of the fund size distribution, so that doubling the cost removes about two thirds of the possible backers. The firms that compete are therefore selected by a small group of allocators before any customer has expressed a view. Three results follow. The price paid by the winner separates into a rational bid, a premium created by the fund's deployment clock, and an error from failing to adjust for adverse selection, and the sign of the price response to competition identifies which one dominates. Concentration trades breadth for depth, improving the funded set only when the skill advantage of the few outweighs the information lost through having fewer independent views, a loss that is small when investors think alike. And large upfront cheques reduce the value of stopping, so ventures funded under deployment pressure should fail later and larger rather than more often. The effective policy margin is access to compute rather than regulation of the capital market.Training a frontier artificial intelligence model costs hundreds of millions of dollars before a product exists, and few investors can write that cheque. This paper shows that the number who can falls as the fixed cost of a first training run rises, at a rate set by the concentration of the fund size distribution, so that doubling the cost removes about two thirds of the possible backers. The firms that compete are therefore selected by a small group of allocators before any customer has expressed a view. Three results follow. The price paid by the winner separates into a rational bid, a premium created by the fund's deployment clock, and an error from failing to adjust for adverse selection, and the sign of the price response to competition identifies which one dominates. Concentration trades breadth for depth, improving the funded set only when the skill advantage of the few outweighs the information lost through having fewer independent views, a loss that is small when investors think alike. And large upfront cheques reduce the value of stopping, so ventures funded under deployment pressure should fail later and larger rather than more often. The effective policy margin is access to compute rather than regulation of the capital market.

Read the full article at: papers.ssrn.com

John Stewart

An acceptable information-processing theory of consciousness should be able to identify the adaptive advantages that drove the emergence of consciousness during the evolution of life. It should also predict the specific dynamical architecture of information processing that would need to be instantiated in AI to produce consciousness and the superior adaptation it enables. Whether such an instantiation produces AI that is actually conscious and also more adaptable would provide the ultimate test of the theory. A prime candidate for such a theory is the Subject-Object Emergence Theory of consciousness. It argues that consciousness first evolved because it enabled organisms to achieve adaptive body-environment coordination without extensive trial-and-error learning. It postulates that the subject in an appropriate Subject-Object subsystem would be able to use depictive (iconic) visual representations of the relative positions of its body and the environment to guide motor actions that will produce adaptive body-environment coordination. The depictive representations will 'light up' for such a subject, producing subjective experience that is used to deliver adaptive benefits. Hand-eye coordination is a familiar example in humans-novel and intricate coordination tasks can be undertaken without additional reinforcement learning, provided focused conscious attention is employed to provide us (the subject) with relevant depictive images. The paper identifies how such a conscious Subject-Object subsystem could be instantiated in AI systems, enabling hand-eye and other body-environment coordination without the extensive reinforcement learning or complex computational programming needed at present. Drawing further on the Subject-Object theory of consciousness, the paper also identifies how these simple conscious subsystems evolved further in organisms to establish the conscious modelling that enables conscious planning, imagining, abduction and other higher cognitive functions. It demonstrates that current approaches to incorporating world modelling in AI will fail to achieve key elements of the general intelligence found in humans that require consciousness.

Read the full article at: papers.ssrn.com

Marcus Schref, Udo Schilcher, and Christian Bettstetter
Phys. Rev. E 114, 024216 – Published 18 August, 2026

Swarmalator systems intertwine two forms of collective behavior, swarming and synchronization, leading to the emergence of specific space-time patterns. Scaling the model to real-world phenomena and technical applications is problematic due to the assumption of global coupling among all swarmalators, which is impractical under physical constraints on interaction range. Conversely, purely local coupling was shown to be infeasible. To address this gap, we introduce and evaluate the concept of multihop coupling for swarmalators, which preserves the locality of physical interactions but propagates state information throughout the network via hop-limited and probabilistic flooding. It is demonstrated that convergence to the original emergent patterns can be achieved in a reliable and fast manner while keeping overhead low. A practical guideline for selecting the range, hop limit, and forwarding probability is provided. The range required for convergence can be approximated by the connectivity threshold of random geometric graphs.

Read the full article at: journals.aps.org

Andrew J. King ∙ Ella G. Henry ∙ Simon Garnier ∙ William L. Allen ∙ Robert J.P. Heathcote ∙ Marco Fele ∙ Marina Papadopoulou ∙ Daniel W.E. Sankey ∙ Ines Fürtbauer

Trends in Ecology and Evolution

Collective behaviour is widespread in the animal kingdom and can enhance individual
fitness. Yet not all collective behaviours are adaptations. Instead, some may be nonadaptive
or ‘evolutionary spandrels’—traits that originated as by-products in the sense proposed
by Stephen Jay Gould and Richard Lewontin. Here, we argue that self-organising processes
provide a route through which evolutionary spandrels in collective animal behaviour
can occur, and we provide three examples: spatial organisation in primate groups,
division of labour in ants, and insect chorusing. We then consider how such outcomes
may be co-opted into adaptive roles through exaptation and conclude by outlining the
challenges associated with testing adaptive and nonadaptive hypotheses in collective
behaviour research using individual-based studies, phylogenetic comparative analyses,
and agent-based models.

Read the full article at: www.cell.com

Mario Franco, Zeinab Davoudmanesh, Sean P. Maley, Fernanda Sánchez-Puig, Carlos Gershenson

Seventy years of artificial intelligence are usually told as a long preamble followed by a revolution beginning around 2012. We organize the period differently, around a question the field has answered differently at different times: what kind of thing is intelligence, such that a machine could have it? Read that way, the human contribution does not withdraw as systems learn more; it relocates, and mostly to places our instruments do not record. Whether the recent acceleration is a change in kind or a change in budget is, we suspect, the more interesting question, and not one that benchmark curves can settle.

Read the full article at: www.preprints.org

Zyri Bajrami
Matter, energy, and information, on the one hand, and the interplay between natural selection and self-organization, on the other, have given rise to modules, which constitute the fundamental units of interaction, organization, and function, as well as the primary targets of natural selection throughout chemical, biological, and cultural evolution. Based on the forms of structural information that enable their emergence, modules can be classified into huit types: (a) chemical modules (l) genetic and epigenetic modules, (c) cell, (d) neural, (f) mental modules, (g) moduloma (m) and affordance modules (n). Through interactions among modules and between modules and their environment, semantic (meaningful) modular information emerges. It is this semantic information that enables modules to acquire and maintain stability as both physical and abstract entities. The emergence and persistence of both material and immaterial (abstract) modules occur only at a specific point in time, when structural information is matched with the corresponding energy. This relationship is described by the law of modular stability. Modules acquire and preserve stability when the structural information responsible for establishing the relationships among the elements of their structure, considered as systems, corresponds to the energy required to maintain those relationships, while semantic modular information reaches its maximum value. One of the principal implications of this law is that the creative role of natural selection and modular stability is expressed primarily during the first stage of module formation, when the module is established as a replicator, rather than during the second stage, when it functions as an interactor and its fitness is determined.

Read the full article at: www.preprints.org

Tomás Cicchini, Ollin D. Langle-Chimal, Marta C. González, Ines Caridi & Leonardo Ermann

Discover Cities

Volume 3, article number 139 (2026)

Public transportation systems are vital for urban mobility, yet their robustness against disruptions remains underexplored, particularly in Latin American cities. This study quantifies the structural and functional robustness of public transport networks in Mexico City, Rio de Janeiro, and Buenos Aires, revealing that Rio de Janeiro exhibits the highest resilience due to its structural redundancy, while also establishing a strong correlation between structural connectivity and trip feasibility across the cities.

Read the full article at: link.springer.com

Klaus Jaffe

Infodynamics explores how the interactions between information and energy generates useful work, offering a framework for understanding cognition. It views consciousness as an adaptive mechanism that enables an entity, biological or artificial, to construct an internal map of itself and integrate it into its environmental models (Weltanschauung). When these models incorporate the perceived internal states of others, empathy emerges. Empathy in turn allows to secure synergistic social cooperation to build robust new social structures. By focusing on the utility of information, infodynamics uncovers how consciousness and empathy serve as evolutionary tools to enhance survival odds and stabilize social structures. This approach provides an actionable methodology for detecting consciousness in living or artificial entities, that allows optimizing the design of advanced artificial intelligence, and of future educational systems. Both will drive cultural and eventually biological evolution.

Read the full article at: papers.ssrn.com

Lorenzo Cirigliano, Gareth J. Baxter and Gábor Timár
Complexities 2026, 2(2), 13;

Real-world networks often exhibit strong transitivity with nontrivial local clustering spectra and degree correlations. Such features are not easily modeled in tractable network models, creating an obstacle to the theoretical understanding of such complex network structures. Here, we address this problem using a model for strongly clustered random graphs in which each triad of a random network backbone is closed with a certain probability. Despite the intricate loopy local structure of the graphs obtained, we provide exact expressions for the local clustering spectrum and the degree correlations, filling the gap in the theoretical description of this model for random graphs. In particular, we find positive degree assortativity accompanying high transitivity, and nontrivial structure in the clustering spectrum. Exact asymptotic analytical results, obtained for uncorrelated locally tree-like backbones, are complemented with extensive numerical characterization of finite-size effects.

Read the full article at: www.mdpi.com

Vincenzo De Leo, Michelangelo Puliga, Martina Erba, Cesare Scalia, Andrea Filetti and Alessandro Chessa
Complexities 2026, 2(2), 15

In this work, we inspected the friendship network on Twitter (recently rebranded as X), concentrating on individuals and organizations intertwined with the energy field. We particularly focus on seasoned professionals, corporate entities, and domain specialists, all connected through ‘following’ relationships. By meticulously examining these ties, we uncover several distinct groupings within the network, each defined by the unique roles its members occupy. Our analysis demonstrates that the natural emergence of such clusters on social platforms exerts a profound influence on public discourse regarding energy and other critical matters, including climate change. Furthermore, we observe that the resulting communities exhibit distinct structural properties and communication patterns, with some clusters showing lower internal engagement, which may be indicative of fragmentation dynamics in online conversations. These emergent clusters, characterized by their shared communication styles, form relatively compact communities where the exchange of information is infrequent compared to larger networks and is usually confined to accounts created for specific commercial objectives. We emphasize that our analysis focuses on a structurally coherent connected component emerging from a curated set of energy-related seed accounts, rather than attempting to reconstruct the entirety of the energy discourse on Twitter. Consequently, peripheral or weakly connected communities may be underrepresented. Additionally, by combining machine-learning-based node classification with graph-based centrality measures, we are able to characterize the roles of structurally central actors within these niche segments and analyze the connectivity patterns that define their positions. This method provides novel insights into how corporate communication unfolds on social media, offering a refreshed perspective on professional networking. Ultimately, our findings highlight the ways in which companies within the energy sector take advantage of Twitter to coordinate their initiatives, with key institutions serving as central nodes in maintaining the organization of these networks.

Read the full article at: www.mdpi.com