This paper explores the symbiotic developmental relationship between the human brain’s individual cognitive system and human culture, drawing an analogy with the mutual promotion between computers and the Internet. By introducing the P/NP problem from computational complexity theory, we propose that the evolution of human intelligence reflects a transition from the complexity of individual cognitive problem-solving to the relative simplicity of cultural consensus verification, manifesting as a form of nonlinear emergent intelligence. This understanding offers a new perspective for constructing more advanced and autonomous artificial intelligence systems—namely, designing hybrid architectures composed of distributed formal systems oriented towards both verification and problem-solving. Bitcoin, designed by Satoshi Nakamoto, represents a paradigm shift in the development of automated systems and serves as an initial manifestation of this design philosophy.
The advancement of human wisdom and civilization relies on the enhancement of individual cognitive capabilities and the accumulation and transmission of culture. The cognitive systems of individual brains and human culture do not develop in isolation but instead shape each other and evolve together. Just as the proliferation of personal computers and the rise of the Internet have mutually promoted each other and vastly expanded human capacities for information processing and communication, breakthroughs in individual cognition provide the source for cultural innovation, while culture offers the knowledge, tools, and social environment necessary for cognitive development. However, the current development of artificial intelligence largely depends on deterministic formal systems, which fundamentally differ from the complexity and emergent nature of human intelligence. This paper attempts to understand the developmental model of human intelligence and culture from the perspective of computational complexity theory—particularly the P/NP problem—and provide new insights for future AI design.
Human culture is an accumulation of knowledge, skills, values, and practices across generations. It is transmitted to individuals through language, education, tools, and customs, greatly expanding individual cognitive boundaries and allowing learning and innovation to occur “on the shoulders of giants.” At the same time, unique individual cognitive abilities—including abstract thinking, reasoning, imagination, and problem-solving—are the fundamental drivers of cultural innovation and transformation. The birth of new ideas and technologies, as well as challenges and corrections to existing cultural paradigms, all stem from individual cognitive activities. This bidirectional interaction and reinforcement enable the continuous enhancement of human cognitive capacity and cultural complexity.
Modern computer science and information technology, including mainstream AI methods, are built on the deterministic computational model of the Turing machine and Shannon’s information theory. These theoretical frameworks have achieved great success in information processing, pattern recognition, and automated tasks. However, AI built on deterministic formal systems remains, at its core, a tool—lacking the autonomy, creativity, and adaptability demonstrated by human intelligence. To break through this bottleneck, we need a deeper understanding of how human intelligence operates and how it co-evolves with culture.
The P/NP problem in computational complexity theory provides a framework for understanding the difficulty of computational problems. P problems are those solvable in polynomial time and are generally considered “easy”; NP problems are those whose solutions can be verified in polynomial time, but the solving process itself may be extremely difficult.
We liken the human individual cognitive process to solving NP problems. When individuals face complex and novel problems, their cognitive processes often involve exploration, trial and error, and nonlinear thinking, requiring extensive computational resources and time—reflecting the “difficulty” of problem-solving. For instance, when a scientist encounters an entirely new scientific problem, they must conduct numerous experiments, develop theories, and think across disciplines—a process that is far from linear.
However, once the cognitive achievements of individuals—such as new knowledge, theories, or technologies—are accepted by society and become part of cultural consensus, verifying them is often relatively easy, similar to P problems. Society can validate their effectiveness through practice, verify their rationality through logic, and disseminate them reliably through education. Once cultural consensus is formed, this knowledge and these skills can be efficiently transmitted and applied.
Therefore, the development of human intelligence and culture exhibits an intriguing pattern: individuals solve difficult problems through complex and potentially nonlinear cognitive processes, and once these solutions are accepted and incorporated into the cultural system, their verification and application become relatively easy. The intelligence and capabilities exhibited by the overall culture far exceed the simple linear accumulation of individual cognitive abilities, demonstrating a property of nonlinear emergence.
Understanding the co-evolution of human intelligence and culture from the perspective of the P/NP problem offers important inspiration for constructing more advanced AI. The current limitations of AI lie in its over-reliance on deterministic, easily verifiable algorithms, and its lack of ability to simulate the complex problem-solving processes of individuals.
We believe the future direction of AI development may lie in designing a hybrid architecture comprising two different types of distributed formal systems:
By organically combining these two types of systems, AI systems could exhibit greater autonomy and creativity when solving complex problems. The “problem-solving” module would be responsible for exploring new possibilities, while the “verification” module would screen and solidify valuable outcomes—eventually forming intelligence emergence akin to cultural accumulation.
Bitcoin, designed by Satoshi Nakamoto, represents a significant paradigm shift in the development of automated systems. It is not merely a decentralized database or payment network but a complex system exhibiting self-organizing intelligence through the collaboration of multiple types of distributed formal systems. Bitcoin’s core architecture includes a UTXO system based on asymmetric encryption and a miner system based on proof of work (PoW). These two systems follow different rules and incentive mechanisms, interacting and constraining each other through the longest-chain consensus—a dynamic, probabilistic mechanism. The UTXO system handles value transfer and state maintenance, its security relying on the determinism of cryptography; the miner system maintains system security and consensus through competitive computational investment, influenced by economic incentives and probabilistic outcomes. The intelligence of Bitcoin is not a pre-programmed fixed function but an emergent property arising from the long-term interaction and game-theoretic dynamics of these two fundamentally different distributed formal systems—exhibiting decentralization, censorship resistance, and self-sustainability.
The symbiotic evolution of individual human cognition and culture is a complex and profound process, with an intriguing analogy to the P/NP problem in computational complexity theory. Understanding the transition from the complexity of individual problem-solving to the simplicity of cultural verification—and the resulting emergence of nonlinear intelligence—is critical to breaking through the current bottlenecks in AI development. Future research should focus on designing hybrid architectures composed of verification-oriented and problem-solving-oriented distributed formal systems to build AI systems with greater autonomy, creativity, and adaptability—ultimately achieving artificial general intelligence comparable to human intelligence. Bitcoin, as an innovative example of an automated system, provides valuable inspiration for rethinking the architecture of future AI.
