我仅在2025年靠设计就赚了多次六位数。没有资金。没有团队。只有我、一台笔记本电脑,以及我经过多年试错建立的流程。
如果我今天,也就是2026年,带着现在知道的一切重新开始,我会做得非常不同。不是因为基础发生了变化,而是因为捷径变了,工具变了,而大多数现有的建议还没有跟上。
这就是我实际上会做的。
2026年的大多数设计建议已经过时了
常见的建议仍然是:学习 Figma,建立作品集,申请工作,参加设计活动进行社交。
那套操作指南进展缓慢。它假设设计价值来源于执行、让东西看起来整洁,以及孤立地调整像素。
那已经不是价值所在了。
到了 2026 年,执行变得快速且廉价。AI 可以做原型,AI 可以生成界面。仍然稀缺的是品味、判断力,以及交付真实产品的能力。这就是我从第一天起就会专注培养的能力。
增加产出量。这才是质量真正形成的方式。
我会以每天完成两个完整设计为目标。不是概念草图。不是情绪板。而是完整的界面。仪表盘、移动应用、网站。两个。每天。
我知道这听起来很多。确实如此。这正是关键所在。
培养审美最快的方法就是提高产出量。你无法仅靠思考就得到好的设计。你必须做出大量糟糕的作品,才能明白它们为什么糟糕。然后你开始做出好的作品。接着你开始理解它们为什么好。
控制你的自尊心。点击发布。然后继续下一个。
大多数设计师每周只做一件作品,却不明白为什么没有进步。你没有进步,是因为你没有创作足够多的作品。产量才是基础,质量随之而来。
设计真正的产品,而不是Dribbble。
我会在第一天就停止查看Dribbble。我会把它从我的书签中删除。
Dribbble 上充满了美丽但无法落地的设计。闪亮的渐变色。幻想风格的界面。那些作品在真实的工程讨论中根本无法存活。
相反,我会挑选三个我每天都会使用的真实产品。类似 Linear、Vercel 和 Stripe。我会研究它们。我会模仿它们。我会理解它们的设计语言、间距决策、排版选择以及组件行为。
然后我会构建我自己的版本。不是为了抄袭,而是为了理解。
真实的产品有真实的约束。它们是由必须面向数百万用户发布的人设计的。在一个真实产品中能学到的,比一百个 Dribbble 作品还要多。
打造值得付费的技能。从真正的产品开始。
尽早学习 TailwindCSS 和 Shadcn。
在 2026 年,这一点不是可选项。
如果你想设计真正能被构建出来的产品,你需要理解它们是如何被构建的。TailwindCSS 和 Shadcn 是现代产品开发中使用最广泛的两个库。不是将来。就是现在。
我不是在说要学会编写所有代码。我是说要熟悉这两个工具。理解 Tailwind 的间距是如何运作的。理解 Shadcn 组件的行为方式。当你带着这些知识去设计时,你做出的设计会更接近可直接投入生产使用。你的交付会更加清晰。你与开发者的协作会更顺畅。你用 AI 生成的代码也会真正与设计匹配并可用。
尽早开始。这会在以后为你节省数月时间。
不要害怕模仿其他设计师。
每个你欣赏的设计师都曾模仿过别人。他们喜欢的设计师也曾模仿过更早的人。整个设计史就是人们相互借鉴、不断改进,并最终称之为自己风格的过程。
太阳底下无新事。
去模仿你欣赏的设计师。不是为了把它当成自己的作品发布,而是为了理解他们是怎么思考的。为什么他们用了那样的间距?为什么他们选择了那种字体?为什么这个卡片布局看起来比另一个更通透?
模仿就是学习。不要太矫情。不断去做、去产出就好。
别他妈总想着和其他设计师搞人脉。先让自己成为值得别人来结识的人。
在我职业生涯早期,我曾以为人脉经营是必须的。去参加活动。在 LinkedIn 上建立联系。私信别人。
这大多是在浪费时间。
你真正想认识的人都很忙。他们不会去活动上结识随机的设计师。他们只会回应那些做出值得被注意作品的人。
正确的做法是把自己做到足够优秀,让别人来找你。发布你的作品。写下你知道的东西。分享你的创作过程。当你的作品足够出色时,对的人会找到你。合作会自然发生。你不需要去追着它跑。
别再一味想着去拓展人脉,先让自己成为值得被认识的人。
尽可能早地开始学习 vibe coding。
Claude Code、Codex、Cowork。选一个,然后开始学习如何使用它。
我不是在说你一定要成为开发者。我是说,要学会如何引导 AI 把你的设计变成真正能运行的产品。这是当下设计领域最大的变化之一。
当你能够在 Figma 中设计某个东西,然后使用 AI 在同一天发布一个可用版本时,你会成为一种完全不同类型的设计师。不是等待开发者的设计师,而是能发布作品的设计师。
当你学会这个时,会发生两件事。首先,你了解产品背后的技术是如何运作的。其次,你开始设计那些真正可交付的东西,而不仅仅是视觉上光鲜亮丽的东西。当你知道交接后的事情是怎样的,你的设计直觉会发生改变。
早点学会。它会快速积累。
要有耐心。给自己整整一年的时间。
不要期待在前三个月看到结果。
设计职业的发展并不是线性的。会有很长一段时间,你在努力工作却几乎看不到回报。大多数人都会在这里放弃。他们把这叫做“没效果”或“这不适合我”。其实都不是。这只是事情开始真正起作用之前的阶段。
如果你能坚持十二个月——每天产出,每天学习——你就会超过大多数和你同时开始的设计师。不是因为你更有天赋,而是因为你坚持了下来。
给自己整整一年的时间,不设任何期待。每天出现,做好工作。在这一年结束之前,不要对自己所处的位置做出任何评判。
完全忽略的东西
那些批评初级设计师的大型 Twitter 账号。有些拥有大量粉丝的设计师通过嘲讽新手来吸引关注。不要给他们你的注意力。他们与你的成长毫无关系。他们只是噪音。
觉得自己已经足够优秀的感觉。我已经做了多年的设计,但仍然会发现自己的作品在某些方面不如自己期望的好。停止学习的设计师,就是停止成长的设计师。保持低调,继续努力。
跳过设计系统。在 2026 年,没有使用设计系统工作会是一个真正的劣势。不仅因为它会让你的作品缺乏一致性,还因为当你把设计交给 AI 时,如果你没有提供系统,它会自己发明一个系统,而你可能不会喜欢它发明的结果。尽早开始使用设计系统,即使只是最基本的层面,比如颜色标记和字体比例。早早养成这个习惯。
简短版
每天产出两个设计。研究真实产品,而不是 Dribbble 上的作品。尽早学习 Tailwind 和 Shadcn。模仿你欣赏的设计师。停止四处建立人脉,开始让自己变得值得被认识。学会用 AI 交付产品。给自己一年的时间。
这就是我会遵循的路线图。不是因为它听起来很好,而是因为它确实真正有效。
显示英文原文 / Show English Original
I made multiple six figures from design in 2025 alone. No funding. No team. Just me, a laptop, and a process I built over years of getting things wrong first. If I had to start over today, in 2026, with everything I know now, I would do things very differently. Not because the fundamentals changed. But because the shortcuts have changed. The tools have changed. And most of the advice out there has not caught up. This is what I would actually do. Most design advice in 2026 is outdated The common advice is still: learn Figma, build a portfolio, apply for jobs, network at design events. That playbook is slow. It assumes design value comes from execution. From making things look clean. From pixel-pushing in isolation. That is not where the value is anymore. In 2026, execution is fast and cheap. AI can prototype. AI can generate UI. What is still rare is taste, judgment, and the ability to ship real things. That is what I would focus on building from day one.
Produce volume. This is how quality actually develops. I would aim to produce two complete designs every single day. Not concept sketches. Not mood boards. Full screens. Dashboard, mobile app, website. Two. Every day. I know that sounds like a lot. It is. That is the point. The fastest way to develop taste is through volume. You cannot think your way to good design. You have to make a lot of bad work to understand why it is bad. Then you start making good work. Then you start understanding why it is good. Hold your ego. Hit publish. Move on to the next one. Most designers produce one thing a week and wonder why they are not improving. You are not improving because you are not producing enough. Volume is the backbone. Quality follows. Design real products. Not Dribbble. I would stop looking at Dribbble on day one. I would delete it from my bookmarks.
Dribbble is full of beautiful, unshippable design. Glowing gradients. Fantasy interfaces. Work that would never survive a real engineering conversation. Instead I would pick three real products I use every day. Something like Linear, Vercel, and Stripe. I would study them. I would copy them. I would understand their design language, their spacing decisions, their typography choices, their component behavior. Then I would build my own version. Not to steal. To understand. Real products have real constraints. They were designed by people who had to ship to millions of users. There is more to learn in one real product than in a hundred Dribbble shots. Build skills that are worth paying for. Start with real products. Learn TailwindCSS and Shadcn as early as possible. This one is not optional in 2026. If you want to design products that actually get built, you need to understand how they get built. TailwindCSS and Shadcn are the two most widely used libraries in modern product development. Not eventually. Right now.
I am not saying learn to code everything. I am saying get familiar with these two tools. Understand how Tailwind spacing works. Understand how Shadcn components behave. When you design with this knowledge, you design things that are closer to production-ready. Your handoff becomes cleaner. Your relationship with developers becomes easier. Your AI-generated code actually works with your designs. Start early. It will save you months later. Do not be afraid to copy other designers. Every designer you admire copied someone. Their favorite designer copied someone before them. The whole history of design is people borrowing from each other, improving on it, and calling it their own style. There is nothing new under the sun. Copy the designers you admire. Not to publish it as your own work, but to understand how they think. Why did they use that spacing? Why did they pick that typeface? Why does this card layout feel more breathable than that one? Copying is studying. Do not be precious about it. Just keep producing. F*ck networking with other designers. Be someone worth networking to.
Earlier in my career I thought networking was a must. Go to events. Connect on LinkedIn. Slide into DMs. That is mostly a waste of time. The people you want to know are busy. They do not go to events to meet random designers. They respond to people who make work worth noticing. The move is to be so good that people come to you. Post your work. Write what you know. Share your process. When your work is excellent, the right people find you. The collaboration happens naturally. You do not have to chase it. Stop trying to network. Start trying to be worth knowing. Start learning vibe coding as early as possible. Claude Code. Codex. Cowork. Pick one and start learning how to use it. I am not saying become a developer. I am saying learn how to guide AI to turn your designs into working products. This is the single biggest shift in design right now.
When you can design something in Figma and then ship a working version the same day using AI, you become a completely different kind of designer. Not a designer waiting for a developer. A designer who ships. Two things happen when you learn this. First, you understand how the technology behind products actually works. Second, you start designing things that are actually shippable, not just visually polished. Your design instincts change when you know what happens on the other side of the handoff. Learn it early. It compounds fast. Be patient. Give yourself one full year. Do not expect results in the first three months. Design career growth is not linear. There is a long period where you are working hard and seeing very little. Most people quit here. They call it "not working" or "not for me." It is neither. It is just the part before things click. If you can stay consistent for twelve months, daily output, daily learning, you will outrank the majority of designers who started the same time you did. Not because you are more talented. Because you stayed. Give yourself a full year with zero expectations. Show up every day. Do the work. Let the year end before you make any judgments about where you are.
What to ignore completely The big Twitter accounts that criticize junior designers. Some designers with large followings build their audience by roasting beginners. Do not give them your attention. They have nothing to do with your growth. They are noise. The feeling that you are already good enough. I have been designing for years and I still find areas where my work is weaker than I want it to be. The designers who stop learning are the ones who stop growing. Keep your head down. Skipping design systems. In 2026, working without a design system is a real disadvantage. Not just because it makes your work inconsistent. Because when you hand your design off to AI, it will invent a system if you do not give it one. You will not like what it invents. Start using design systems at the most basic level you can, even if it is just color tokens and a type scale. Build the habit early. The short version Produce two designs a day. Study real products, not Dribbble. Learn Tailwind and Shadcn early. Copy the designers you admire. Stop networking and start being worth knowing. Learn to ship with AI. Give yourself one year. That is the roadmap I would follow. Not because it sounds good. Because it is what actually worked.