Where Are the Vibecoded Photoshops?
Where Are the Vibecoded Photoshops?
“Vibecoded” 的 Photoshop 在哪里？
If vibecoding is what people say it is, the world should be drowning in vibecoded artifacts right now. Two years of access. Millions of people with the tools. The barrier supposedly fell, right? RIGHT? So where is everything? 如果“Vibecoding”（凭感觉编程）真如人们所说的那样，那么现在全世界应该已经被这类产物淹没了。两年的使用权限，数以百万计的人拥有这些工具。门槛理应已经消失了，对吧？真的吗？那么，东西都在哪儿呢？
Show me the evidence. Where is the vibecoded Photoshop. The vibecoded Excel. The vibecoded Maya. The vibecoded Blender. The vibecoded compiler that compiles itself. The vibecoded database, the vibecoded OS, the vibecoded anything-that-requires-architectural-judgment-to-hold-together. Huh? I am not asking for slop. Slop exists, slop is easy, slop is not the question. I am asking for the coherent, complex, non-trivial things that vibecoding allegedly makes accessible to anyone who can prompt. Silence. Every time. The category is empty. And the accusers never want to address that, because addressing it means admitting the accusation doesn’t hold up. 给我证据。Vibecoded 的 Photoshop 在哪？Vibecoded 的 Excel、Maya、Blender 在哪？能自举的 Vibecoded 编译器在哪？Vibecoded 的数据库、操作系统，或者任何需要架构判断力才能维持运作的 Vibecoded 产品在哪？嗯？我指的不是那些垃圾内容。垃圾内容确实存在，也很容易产生，但那不是问题的关键。我想要的是那些连贯、复杂、非琐碎的成果——据称只要会写提示词（Prompt）的人都能通过 Vibecoding 实现的东西。一片寂静。每次都是这样。这个类别是空的。而那些指责者从不愿面对这一点，因为面对它就意味着承认他们的指责根本站不住脚。
Accusation without evidence
无证据的指责
There are no vibecoded Photoshops because vibecoding does not do what the rhetoric claims it does. The accusation itself is the vibe. The accuser feels that a thing must have been easy because it was made with AI. They post that feeling as if it were a finding. The finding never has to be checked, because nobody else checks it either. The accusation travels because it feels right, not because it is right. Big difference. 根本不存在什么 Vibecoded 的 Photoshop，因为 Vibecoding 并没有实现其修辞中所宣称的功能。指责本身就是一种“感觉”（vibe）。指责者觉得某样东西既然是用 AI 做的，那一定很容易。他们把这种感觉当作一种“发现”发布出来。这种发现从不需要被核实，因为也没人去核实。这种指责之所以传播，是因为它“感觉上是对的”，而不是因为它真的正确。这两者有天壤之别。
The accusation that someone produced unverified output is itself being produced as unverified output. Or in other words.. The thing they accuse vibecoders of is the very thing they are doing: No definition. No test. No falsification. Just a claim, shipped fast, never checked. The accusation is the real vibecoded content. The accuser is the vibecoder. They accuse others of the very crime they are committing themselves, and most of them are unaware of doing it. 指责他人产出未经核实的内容，其本身就是一种未经核实的产出。换句话说……他们指责 Vibecoder 的行为，正是他们自己正在做的事：没有定义，没有测试，没有证伪。仅仅是一个断言，快速发布，从不核实。这种指责本身才是真正的“Vibecoded”内容。指责者自己就是那个 Vibecoder。他们指责别人犯下了他们自己正在犯的罪行，而他们中的大多数人甚至意识不到这一点。
The line in the sand
划下的界线
Their whole attack vector is built around the premise of defending a gate that doesn’t need defense, and a threat that doesn’t even exist. I can prove it. There are levels in this work. Level 1 is the typing. Syntax, semicolons, the years memorizing pointer arithmetic and which header file the function lives in. Level 2 is the verifying. The harness. The test suite. The reflex of rejecting the ninety attempts that almost work and shipping the one that does. Level 3 is the deciding. What to build at all. Which architecture survives contact with the real world. The three have never been the same thing. 他们整个攻击路径的预设前提是：守卫一道根本不需要守卫的门，面对一个根本不存在的威胁。我可以证明这一点。这项工作是有层级的。第一层是“打字”。语法、分号，花几年时间死记硬背指针运算以及函数存在哪个头文件里。第二层是“验证”。测试框架、测试套件。拒绝掉九十个“差不多能用”的尝试，只发布那一个真正可行的反射性动作。第三层是“决策”。决定到底要构建什么，什么样的架构能在现实世界的考验中存活。这三者从来就不是一回事。
The gate was never at Level 1. The gate was at Levels 2 and 3, where the work that holds together actually happens. AI lowered the cost of Level 1. It did not touch Levels 2 or 3. The gate is exactly where it always was. The Level 2 and 3 people can see this. They look at SoulPlayer and ask about the harness. They look at an AI music video and ask about the timing decisions or how consistency was enforced. They are at the gate. They know where the gate is. They recognize when someone else is also at the gate. 门从来不在第一层。门在第二层和第三层，那是真正让工作能够维持运作的地方。AI 降低了第一层的成本，但它没有触及第二层或第三层。门依然在它原来的位置。处于第二层和第三层的人能看清这一点。他们看 SoulPlayer 时会问测试框架的问题；他们看 AI 音乐视频时会问关于节奏决策或如何强制一致性的问题。他们守在门前。他们知道门在哪里。他们能识别出谁也同样守在门前。
The accusers cannot see this. They are not at the gate. They were at Level 1. Level 1 was their identity, their hours, their proof of belonging, their reason to feel at home in this profession. When AI made Level 1 cheap, it did not threaten the gate. It threatened them. Because they bet their self-worth on the layer that just got rented out. So they call the work vibecoded. They have to. The new artifact cannot be legitimate, because if it is legitimate, then Level 1 was never the gate, and they were never at the gate either. 指责者看不见这些。他们不在门前。他们一直停留在第一层。第一层是他们的身份、他们投入的时间、他们归属感的证明，也是他们在这个行业感到自在的理由。当 AI 让第一层变得廉价时，它威胁到的不是那道门，而是他们自己。因为他们将自我价值押注在了那个刚刚被“外包”出去的层级上。所以他们称这些工作为 Vibecoded。他们必须这么做。新的产物不能是合法的，因为如果它是合法的，那就意味着第一层从来就不是门，而他们也从来没有真正守在门前。
Exactly, and I feel sorry for everyone who is solely operating at Level 1 and with nothing left to contribute. If AI revealed they were never doing the deciding, that they were never doing the verifying, that they were the typist all along, then they are in deep trouble right now. And they are finding this out in public. By panicking at the wrong things. By calling other people’s work vibecoded because that is the only move left when the floor disappears under you. I do not feel angry at them. I feel sorry for them. 没错，我为那些只停留在第一层、再无其他贡献的人感到难过。如果 AI 揭示了他们从未参与决策、从未参与验证，而一直只是个打字员，那么他们现在正陷入巨大的麻烦中。他们正在公众面前发现这一点。通过对错误的事情感到恐慌，通过称别人的工作为 Vibecoded——因为当脚下的地板消失时，这是他们唯一能做的动作。我并不生他们的气，我只是为他们感到遗憾。
I will not use this accusation. I could. I have receipts. SoulPlayer has a ninety-test verification harness developed in Python that emits the C64 assembly and binary. Four bit-identical reference implementations have to agree before anything ships. I made sixteen AI music videos, weeks of work each, spent months creating our own custom inference toolchain in order to gain a maximum of control over the final frame. I have been the person you call when nothing else works, in rooms full of people who would never have called me on a good day. I know where the gate is. I have been at the gate since before this profession called it a gate. 我不会使用这种指责。我本可以的，我有证据。SoulPlayer 有一套用 Python 开发的、包含九十项测试的验证框架，用于生成 C64 汇编和二进制文件。在发布任何东西之前，必须有四个位完全一致的参考实现达成共识。我制作了十六个 AI 音乐视频，每个都耗时数周，并花了数月时间创建我们自己的自定义推理工具链，以便对最终画面获得最大程度的控制。我一直是那种在其他方法都失效时被叫去救场的人，即便是在那些平时根本不会找我的人的房间里。我知道门在哪里。早在这一行把那叫作“门”之前，我就已经守在那里了。
My demoscene background alone would allow me to turn around and call other people’s work vibecoded. I could punch down at every prompt-and-pray app on the internet and land a lot of hits. I will not. Because I know what the accusation costs the person on the receiving end. I have been the target of “you don’t really belong here” my whole life. Neurodivergent, Tics. Freelancer-not-by-choice. Demoscene-not-industry. Artist-not-CS. Disabled. The accusation is the same shape every time. It lets the accuser feel superior without doing any work. It makes the target spend their next week defending themselves instead of building the next thing. 仅凭我的 Demoscene（演示场景）背景，我就完全可以转过头去指责别人的工作是 Vibecoded。我可以对互联网上每一个“提示词祈祷型”应用进行降维打击，并且能命中很多目标。但我不会。因为我知道这种指责对被指责者意味着什么。我一生中一直是“你不属于这里”这类话的目标。神经多样性、抽动症、非自愿的自由职业者、Demoscene 而非工业界、艺术家而非计算机科学出身、残疾人。这种指责每次的形状都一样。它让指责者无需付出任何努力就能感到优越。它让被指责者不得不花下一周时间为自己辩护，而不是去构建下一个作品。
I am writing this essay right now instead of finishing two other projects. The accusation is already winning, even though it has no evidence, no definition, no falsification, and no Photoshops to support it. The accusation doesn’t have to be true. It just has to cost the target enough time and morale that the next person stops sharing their work. The accuser wants to harm and they succeed usually. People are afraid to say they used AI. Not because using AI is shameful. Because the accusation is shameful, and the accusation is cheap to make and hard to refute. The shame economy runs without any actual shameful behavior in the system. It runs. 我现在正在写这篇文章，而不是去完成另外两个项目。这种指责已经赢了，尽管它没有任何证据、定义、证伪，也没有任何 Photoshop 实例来支持。指责不需要是真的。它只需要让目标付出足够的时间和士气代价，从而让下一个人不敢再分享他们的作品。指责者想要伤害他人，而且他们通常会成功。人们害怕说自己使用了 AI。不是因为使用 AI 可耻，而是因为这种指责本身是可耻的，而且这种指责制造起来成本极低，却极难反驳。这种“羞耻经济”在系统中没有任何实际可耻行为的情况下依然在运转。它就这样运转着。