Wet-State Human Agency
Wet-state human agency is the E45 claim that AI makes the emotional, bodily, desirous, and relational parts of the person more important, not less. In E45 孟岩对话李继刚：人何以自处, Li Jigang / 李继刚 borrows the contrast between dry and wet states: society often strips away feelings, moods, and inner movement as noise, while AI may take over more dry knowledge work and force people to ask what remains human.
Li’s answer splits human ability into body power, brain power, and heart power. Industrial technology amplified body power, AI is absorbing much brain power, and the remaining human center is heart power: intention, emotion, will, desire, aesthetic weight, value choice, and real connection. Meng Yan / 孟岩 extends this with the idea that after body labor moved to machines, people built gyms; after brain work moves to models, people may need practices that keep the mind and will alive.
This concept is not anti-rational. It says explicit cognition is only one layer of agency. A person still needs to choose what deserves delegation, what world they want, what kind of relationship they maintain with others, and when to leave the high-flow AI environment for the body.
Key Claims
- AI makes dry cognition cheaper, so the scarce human layer shifts toward heart, will, taste, and embodied life.
- Emotion and mood can be information rather than residue to be removed from thought.
- Human agency weakens when the person becomes a passive relay between model output and the next task.
- Offline presence and small-scale human connection become more valuable when AI language and AI workflows dominate daily cognition.
- Body pacing, sleep, meals, movement, and friendship are part of AI-era self-governance.
Connections
- Human Agency Under AI — broader agency frame extended by E45.
- Subjectivity As AI Asset — taste, values, and self-understanding become AI-era inputs.
- Human-Machine Amplification — technology amplifies the user’s state, including weak or strong will.
- Embodied Judgment and Trained Intuition — body and emotion as parts of judgment.
- AI Use Pacing and Attention Industrialization — practical risks when tools and feeds capture volition.
- Feed Curation — input selection as a way to protect wet-state agency.