Research on information diffusion suggests that audience engagement often depends on a balance between novelty and familiarity, but it remains unclear whether narratives follow a similar principle when organizing information for audiences. To examine this possibility, this study analyses 2578 TED talks and models narrative semantic progression along three dimensions: speed, breadth, and circuitousness. The results reveal that semantic breadth and circuitousness are associated with viewership in a nonlinear manner, with moderate levels linked to higher engagement, while both low and high extremes are less effective. In contrast, semantic speed is predominantly negatively associated with viewership. The interaction between semantic speed and semantic breadth further indicates that breadth influences the level of speed at which predicted viewership reaches its maximum. These findings indicate that narratives associated with higher engagement tend to be characterized not by maximum novelty or familiarity, but by a balance between the two. Analyses of podcast data reveal broadly consistent patterns, extending the findings beyond the highly curated context of TED Talks. This study advances a nonlinear perspective on narrative effectiveness and demonstrates the value of Transformer-based embeddings, including large language model embeddings and widely used sentence embedding models, for capturing how meaning unfolds across narrative sequences.
Although artificial intelligence (AI) has become increasingly smart, its wisdom has not kept pace. In this opinion article, we examine what is known about human wisdom and sketch a vision of its AI counterpart. We introduce human wisdom as strategies for solving intractable problems-those outside the scope of analytic techniques-including both 'object-level' strategies, such as heuristics (for managing problems), and 'metacognitive' strategies, such as intellectual humility, perspective-taking, or context adaptability (for managing object-level task fit). We argue that AI systems particularly struggle with this type of metacognition. Wise metacognition would lead to AI that is more robust to novel environments, explainable to users, cooperative with others, and safer by risking fewer misaligned goals with human users. We discuss how wise AI might be benchmarked, trained, and implemented.
Higher-order interactions that nonlinearly couple more than two nodes are important in many networked systems, and their effects on collective dynamics are increasingly being studied. Here, we provide an overview of this rapidly growing field and of the techniques that can be used to describe and analyse them. We focus in particular on new phenomena and challenges that emerge when non-pairwise interactions are considered. We conclude by discussing open questions and promising future directions on the collective dynamics of higher-order networks. This Review surveys how higher-order interactions, which link more than two units at a time, reshape collective dynamics in complex systems. New synchronization phenomena, analytical frameworks and emerging methods to reduce or infer higher-order structure from data, are highlighted.
Emergence is a concept in complexity science that describes how many-body systems manifest novel higher-level properties, properties that can be described by replacing high-dimensional mechanisms with lower-dimensional effective variables and theories. This is captured by the idea 'More is Different'. Intelligence is a consummate emergent property manifesting increasingly efficient-cheaper and faster-uses of emergent capabilities to solve problems. This is captured by the idea 'Less is More'. In this paper, we first examine claims that Large Language Models exhibit emergent capabilities, reviewing several approaches to quantifying emergence, and secondly ask whether LLMs possess emergent intelligence. This article is part of the theme issue 'World models in natural and artificial intelligence'.
Large language models have achieved remarkable capabilities across domains, yet mechanisms underlying sophisticated reasoning remain elusive. Recent reasoning models outperform comparable instruction-tuned models on complex cognitive tasks, attributed to extended computation through longer chains of thought. Here we show that enhanced reasoning emerges not from extended computation alone, but from simulating multi-agent-like interactions -- a society of thought -- which enables diversification and debate among internal cognitive perspectives characterized by distinct personality traits and domain expertise. Through quantitative analysis and mechanistic interpretability methods applied to reasoning traces, we find that reasoning models like DeepSeek-R1 and QwQ-32B exhibit much greater perspective diversity than instruction-tuned models, activating broader conflict between heterogeneous personality- and expertise-related features during reasoning. This multi-agent structure manifests in conversational behaviors, including question-answering, perspective shifts, and the reconciliation of conflicting views, and in socio-emotional roles that characterize sharp back-and-forth conversations, together accounting for the accuracy advantage in reasoning tasks. Controlled reinforcement learning experiments reveal that base models increase conversational behaviors when rewarded solely for reasoning accuracy, and fine-tuning models with conversational scaffolding accelerates reasoning improvement over base models. These findings indicate that the social organization of thought enables effective exploration of solution spaces. We suggest that reasoning models establish a computational parallel to collective intelligence in human groups, where diversity enables superior problem-solving when systematically structured, which suggests new opportunities for agent organization to harness the wisdom of crowds.