AI's first cut to knowledge work is pay, not layoffs.你的工资给‘稀缺’定价，不是给‘能力’定价；当 AI 达到平均水平、能以算力成本复制，稀缺消失，工资随之下跌。

New research from DeepMind cofounder Shane Legg and colleagues reaches a counter-intuitive result: in a 14-person model economy, when AI merely matches the human average, specialist pay falls from 4.15 to 2.90 (about 30%) with no one laid off.

The mechanism is hard. Your wage was never pricing your skill; it was pricing your scarcity. A skill pays because few people have it and it cannot be copied. Once AI reaches the average, it can be reproduced at compute cost, and the scarcity ends at once; the price drops even if your ability has not. The sharper step in the paper: when AI surpasses every human expert, top programmer pay falls below a tenth of its peak, and the only role left with bargaining power is the coordinator who directs the AI.

What it means for you: don't treat your wage as a certificate of ability. Ask one question: is the money I earn protected because no one can copy me, or only because no one has copied me yet? If it is the second, the real leverage is to move toward the coordinator end: do the judgment, the allocation and the owning-of-outcomes, not the part that can be reproduced.

This call is wrong if independent labor data show that, in the jobs most exposed to rising AI capability, experts' real wages do not fall.