王峰 AI × 经营 约一次诊断

总目 / 文章 / WF-0482

文章 · 存档

我对着AI发了火,然后照了照镜子

I Yelled at My AI — Then Took a Look in the Mirror

Hui Jin

Lately, I've noticed something — working at the speed of AI comes with a cost: I'm getting more impatient. A lot more.

Impatient with slow builds. Impatient with unclear prompts. Impatient with, well, my family — the very reason I'm doing this in the first place.

And while I'd never treat a real teammate with the kind of clipped frustration I sometimes aim at my AI tools… it's starting to seep out. Into how I context switch. Into how I handle setbacks. Into how I show up when things get messy — which, let's be honest, is most of the time.

And ironically, that impatience slows me down. It shortens my fuse, clouds my thinking, and chips away at the clarity I need to actually move fast. The very thing I pretend to care most about — speed — became the victim of the mindset I've let creep in.

That's when I realized: speed had warped my expectations. And unless I wanted to become the startup version of someone yelling at their microwaves for taking too long, I needed to ground myself.

So I built something simple.

Because taking the work seriously is essential.

But taking myself too seriously? That's the fastest path to burnout, bad decisions, and — let's be honest — being harder to live with.

Here's what I'm trying to remember:

1. I Am Not the Code

If something breaks, stalls, or ships with a bug, it's not a referendum on my intelligence, future, or identity. It's just broken code. Fix it. Move on.

Maybe even put my frustration in the comment — so I can laugh at myself six months later.

2. Give My Genius a Dumb Name

Call my internal tool gremlinWrangler() or duct_tape_alpha.

Humor breaks tension. Nothing I build is too sacred to laugh at.

3. Narrate Like a Nature Documentary

Here we see the overcaffeinated founder in his natural environment… pacing while talking to an AI cofounder and wondering why his build step failed. Again.

It sounds ridiculous. That's the point. Zooming out gives me perspective. And better decisions live at altitude.

4. Ask:“Will Future Me Even Care?”

Most of what I get spun up about won't make the highlight reel. But losing my ability to stay calm, present, and reflective?

That's the kind of thing Future Me would care a lot about.

AI Makes Me Fast. But I Still Have to Stay Human.

I want to stay human because that's the point. Because presence, patience, and kindness — with my family, myself, and everyone around us — are the reasons I'm building anything at all.

AI may be my cofounder — tireless, brilliant, maddeningly literal. But I'm still the one steering.

And if I do this right, maybe I get to help shape not just what we build, but how we relate to this technology — and to each other.

So yeah. I'm keeping my standards high. But I'm also keeping a clown nose on my desk.

Just in case.

我对着AI发了火,然后照了照镜子

作者:Hui Jin

AI让我如虎添翼。

可后来,它也开始让我“碎了一地”——而这种崩溃,其实早就有迹可循。

于是我给自己做了一个“”,提醒自己脚踏实地,别变成一个混蛋。

最近我发现了一件事:用AI的速度在工作,是有代价的。

我变得越来越没有耐心,真的——越来越严重。

对开发进展慢的不耐烦、对提示词不清晰的不耐烦、甚至对我家人的不耐烦……而他们才是我努力的初衷啊。

我永远不会像那样对真人团队成员发火,但对AI的那股带刺的情绪已经开始溢出——

体现在我频繁切换上下文时的急躁,

体现在遇到问题时的处理方式,

体现在一切变得混乱时我的状态 —— 说实话,我大多数时候都很混乱。

讽刺的是,这种急躁让我变得更慢了。

让我更容易爆炸、更难专注、更容易失去判断力——

我最在意的“效率”,恰恰成了我这种心态的受害者。

这时我才意识到:“追求速度”正在扭曲我对我自己的期望。

我不想成为那种“对着微波炉大吼‘怎么还没热好’”的创业版疯子,

所以我必须重新找回自我。

于是我做了个很简单的东西:

因为认真对待事业当然重要,

但如果太把自己当回事儿,那就是走向崩溃、做错决策、变成混蛋的最快路径。

以下是我时常提醒自己的几点:

1. 代码不代表我

如果出了bug、卡住了、发布有问题,那只是代码有问题,

不是我笨,不代表我未来完蛋,也不代表我不行。本来代码大多不是从网上抄的就是AI写的,说实话,从用Cursor的第二天起,我就放弃了逐行审阅AI代码的计划。

修就完了,继续前进。

还是愤怒的话,可以把自己的抱怨和脏话写进注释里,半年后看到还能笑一笑。

2. 给“天才创意”起个蠢名字

把我内部的工具命名成 gremlinWrangler() 或者 duct_tape_alpha。

懂的都懂。

幽默能化解张力。

没有任何一个项目神圣到不可以被调侃。

3. 把自己当成自然纪录片的主角来解说

“现在画面中出现的是一位摄入过量咖啡因的创始人,在他自然的栖息地里来回踱步,嘴里对着AI合伙人念叨:‘这构建怎么又失败了?’”

听起来很蠢吧?这正是重点。

从上帝视角看问题,会让我更有判断力。

而更好的决策,往往就藏在这种“俯瞰视角”中。

4. 问自己一句:“未来的我会在乎这事吗?”

大部分让我抓狂的事,将来连我的回忆集锦都进不去。

但如果我因此失去了平静、专注、自省的能力——

未来的我,真的会很在意。

AI让我更快,但我必须保有人性

我之所以想保有人性,是因为——没有因为,人性就是一切。

能活在当下、有耐心、保持善意——

对家人、对自己、对周围的每一个人。

这些,才是我做一切事情的理由。

AI可以是我的合伙人,它不知疲倦、才华横溢、但死板到令人抓狂。

可舵,还得我来掌。

如果我做得够好,

也许我不仅能影响我们“造出什么”,

还能影响我们和AI如何相处共存,以及——我们彼此如何相处。

所以是的,我依然对自己要求很高。

但我也会在办公桌上放一个小丑鼻子。

以备不时之需。