One of the most influential scientific and philosophical viewpoints is “More is Different”, introduced in 1972 by Nobel Prize–winning physicist Philip W. Anderson, highlighting the limitations of the reductionist approach. The emergent properties cannot be derived from the fundamental laws that govern their elementary particles. The generalization of this approach suggests a hierarchical structure of science, where explainable properties of small-scale systems cannot necessarily predict the emerging phenomena on larger scales of similar systems. Its interdisciplinary perspective covers chemistry, molecular biology, cell biology, and social sciences besides physics.
Anderson’s pioneering philosophical viewpoint was announced where machine learning applications based on complex architectures and datasets were nonexistent. Machine learning and artificial intelligence (AI), which perform tasks typically associated with human intelligence, have become a part of daily life only in the last two decades.
The relevance of the original "More is Different" viewpoint to AI models can now be quantitatively examined. Prof. Ido Kanter, from the Department of Physics and Gonda (Goldschmied) Multidisciplinary Brain Research Center at Bar-Ilan University, explores this question in a paper just published in Physica A . The main finding is that physics represents “More is the Same” from an information viewpoint, whereas AI embodies "More is Different", as a consequence of learning and cooperation among the nodal architecture.
The research shows that as AI models learn, their internal units -- known as nodes -- begin to specialize. Rather than performing identical functions, different nodes take on distinct roles, such as recognizing specific patterns or linguistic features. This division of labor allows the system to become more effective, suggesting that AI’s strength lies not only in its size but in the coordinated interaction among specialized components. "Even a single node within a language model can contain meaningful information about the model’s overall task," said Prof. Kanter and added, " When multiple nodes operate together, their combined capabilities exceed the sum of their individual contributions, demonstrating emergent intelligence in action -- More is Different."
The research also identifies a key distinction between AI systems and many physical systems. In physics, individual components often reflect the same information about the system as many components, an idea that can be described as “More is the Same,” Adding information of more components does not necessarily increase the total information about the system’s state.
The findings may have implications for neuroscience. Drawing on experimental evidence on dendritic learning as an alternative mechanism to synaptic plasticity, Prof. Kanter suggests that the brain may rely on neurons that are more specialized and information-rich than previously assumed.
The study points to a broader conclusion: intelligence in AI may emerge not simply from scale, but from the ability of individual components to specialize, share information, and work together. Sometimes, understanding the future of artificial intelligence begins with a foundational question from physics.
More is Different in AI—More is the Same in Physics
2-Apr-2026