AI News HubLIVE
In-site rewrite1 min read

"Five Types of People AI Cannot Replace; Being Second Is the Safest for Companies" | Kunlun Wanwei's Fang Han at AIGC2026

At the 2026 China AIGC Industry Summit, Kunlun Wanwei CEO Fang Han pointed out that token consumption has become a measure of strength in the AI era, but experience is no longer a moat. He identified five types of people AI cannot replace and advised companies to aim for second place in AI implementation to minimize risk.

Source量子位Author: 一水

At the 2026 China AIGC Industry Summit hosted by QbitAI, Kunlun Wanwei's CEO Fang Han delivered a keynote speech that challenged common anxieties about AI. He argued that while token consumption—the amount of data processed by AI models—is becoming a critical indicator of strength in the AI era, experience is no longer a reliable defense.

Fang highlighted that ordinary employees consume millions of tokens monthly, while heavy users like the creator of "Crayfish" burn 600 billion tokens per month at OpenAI. This disparity underscores token usage as the "electricity consumption" of the AI era. However, he cautioned against alarmism: AI mainly threatens jobs that are closed-loop and fault-tolerant, such as coding, middle management, and junior apprentices. High-judgment, responsibility-heavy roles like surgery and on-site construction remain secure.

Drawing from his experience through the internet and mobile internet cycles, Fang advised against both short-term overestimation and long-term underestimation of AI. He dismissed predictions about AGI's imminent arrival from company executives as self-serving, noting that only DeepMind's Hassabis gave a distant date of 2034.

Fang identified five irreplaceable human roles: storytellers (e.g., writers in short dramas), idea creators (now empowered by AI tools), beauty definers (creating new aesthetics), system builders (like the Crayfish founder who built a successful product), and paradigm shapers (like Einstein or Newton).

For businesses, Fang recommended a pragmatic approach: most industries should aim to be the second mover in AI adoption, as first movers bear excessive exploration costs and third movers gain no benefits. IT is the exception, where only the top survives. He cited Anker's cautious, business-driven AI adoption as a model. Kunlun Wanwei itself provides every employee with a monthly token subsidy of 700 RMB and mandates AI coding usage since 2024.

Fang concluded that AI resets the playing field for everyone, breaking the ceiling for young people. He urged attendees to seize the opportunity rather than succumb to fear.