Meng Yan says the rise of AI agents is widening the gap between AI capability and social resources
Chinese commentator Meng Yan argues that rapid gains in AI performance do not automatically translate into economic rewards or social mobility, because modern organizations are built as dense collaboration networks where most roles require only baseline competence, not ever-rising individual output. In his framing, extra productivity often spills over without being absorbed by the system. That is why many people who improve their efficiency with AI still fail to see corresponding gains in income, status, or access to resources.
He says the mismatch is becoming sharper in the AI agent era. The people pushing hardest into tools such as Agent frameworks, MCP, Skills, workflow systems, APIs, Git, IDEs, and coding agents are often programmers, designers, researchers, and analysts — secondary nodes in the social and commercial network. Meanwhile, entrepreneurs, executives, investors, and institutional leaders control clients, capital, hiring power, and connections, yet many of them remain far from the front edge of agent-based AI because the new wave is increasingly programming-driven.
Meng also points to an additional constraint in mainland China: users cannot directly and reliably access the OpenAI and Anthropic ecosystems, leaving them cut off not only from the models themselves but also from the fast-growing surrounding stack of agent tools, coding workflows, and developer communities. He argues that AI’s real social impact will unfold in three stages: changing nodes, changing connections, and finally changing network topology.
