AI 时代最稀缺的资源是人的 willpower，或者用尼采的话说，Wille zur Macht。
在工作中一个常见的困难情景是你注意到两个不同的地方存在类似但又有区别的模式，想弄明白它们之间是什么关系。这种时候你去问谁呢？自己去研究非常痛苦。想请教专家，那你需要找一个比你资深、更懂行的人，但现实中同时精通两个不同领域的专家少之又少。即便有，他们的职位通常很高，事务繁忙，约到时间非常困难。就算你描述清楚了问题，对方也很难给出一个让你心满意足的答案。即便搞明白了答案，想要推行下去也更难，因为这类工作往往需要大量的跨部门协调。——通常你最后耸耸肩算了。
在 AI 时代，事情的难度完全不在一个量级。你完全可以让 AI 去联接两个不同领域的知识，给你一个综合性总结。如果你想做什么，直接动手做就是了。Y Combinator 的 CEO Garry Tan 说过：在 AI Agent 的帮助下，每个人都应该有勇气去「移山倒海」（原话是 boil the sea）。也就是说，你完全可以也应该去做不可能的事情。
那瓶颈是什么呢？在 AI 的时代，既然你可以很快地把任何一个念头付诸实施，问题就变成了你愿不愿意去花费这个心力去推动一件事发生。
AI 是不会主动产生这样一个念头的。它不会主动想要把两个跨部门的东西放在一起比对然后看看有没有能够共享的部分。AI 没有自发的 willpower，它的 willpower 只能来自于某个指令，可能来自人，也可能来自另一个上层 AI agent，但一层层追溯上去，源头也还是人，或者说人的主观能动性和内在激励。任何一种推动事情改变的力量，归根结底，不外乎来自人自我超越扩张和赋予世界意义的本能，也就是尼采说的 Wille zur Macht。AI 没有这个。
于是看起来你的工作更轻松了，但它对你的耗损反而增加了。传统意义上可以用来发挥心流的部分被 AI 拿走了，你每一个指令都会被迅速完成，于是你不得不去想 What’s next。下一个指令是什么？我要把这个事情往哪推？在传统工作里，这种对心力的消耗可能一周才会发生一次，现在变成了每 15 分钟就要来一次。这个部分没有办法交给 AI，它只能解决怎么做这件事情的问题，没有办法产生出源源不断的新的 willpower 注入一个系统。
于是就像沙丘世界里的香料一样，人的意志成为了工作的本质形态，驱动着整个世界的运转。大家通常会以一种含糊其辞的方式谈论它，仿佛它只是点缀，但最终只有它才是真实的。
Intention Is All You Need
The most scarce resource in the age of AI is human willpower, or to use Nietzsche’s term, Wille zur Macht.
A common difficult scenario at work is when you notice similar yet distinct patterns in two different areas and want to understand the relationship between them. Who do you ask? Conducting the research yourself is agonizing. If you seek an expert, you need someone more senior and knowledgeable than you, but in reality, experts who are proficient in two different fields simultaneously are few and far away. Even if they exist, they are usually high-ranking, busy, and incredibly difficult to schedule. Even if you articulate the problem clearly, it is hard for them to provide a satisfying answer. Even if you find the answer, implementation is even harder because such tasks often require extensive cross-departmental coordination. Usually, you just shrug and let it go.
In the AI era, the level of difficulty is on a completely different scale. You can have AI bridge knowledge from two different domains and provide a comprehensive synthesis. If you want to do something, you just do it. Garry Tan, CEO of Y Combinator, said that with the help of AI Agents, everyone should have the courage to “boil the sea.” In other words, you absolutely can and should do the impossible.
So, where is the bottleneck? In the AI era, since you can turn any thought into reality almost instantly, the question becomes: are you willing to expend the mental effort to drive a thing into existence?
AI will not spontaneously generate such a thought. It will not take the initiative to compare two cross-departmental entities to see if there is potential for shared resources. AI lacks spontaneous willpower; its willpower can only be derived from a command—perhaps from a human, or perhaps from another higher-level AI agent. But if you trace it back layer by layer, the source remains the human, or more specifically, human agency and intrinsic motivation. Any force that drives change ultimately stems from the human instinct for self-transcendence, expansion, and the desire to endow the world with meaning—the Wille zur Macht. AI does not have this.
Consequently, while your work appears easier, its toll on you actually increases. The parts traditionally associated with entering a state of flow have been taken over by AI. Every command is completed rapidly, forcing you to constantly ask, “What’s next?” What is the next instruction? Where do I push this next? In traditional work, this kind of mental drain might happen once a week; now, it happens every 15 minutes. This part cannot be delegated to AI. It can solve the “how” of a task, but it cannot generate the continuous stream of new willpower required to fuel a system.
Thus, like the Spice in the world of Dune, human will has become the essential form of work, driving the gears of the entire world. People usually speak of it in vague terms, as if it were a mere ornament, but in the end, it is the only thing that is real.