I recently put out a video arguing that frontend development is changing. I don't think the work is going away. The frontend developer title is wha...
For further actions, you may consider blocking this person and/or reporting abuse
I don't think so as well. Although it is easier to do front-end with AI, it's easy to miss some important aspects of what makes a good front-end. One is accessibility, optimization, etc.
If you prompt the front-end generally, you tend to miss these since it's not really there for you to see. You have to make sure you test the front-end design to see if it's accessible, which can vary depending on who your audience is. You also have to take care of optimization, which is difficult since it depends on browser and one's computer specs.
Even when done with detail prompts, it's a similar experience since you are focus on the looks and not about if it's a good design in general.
Overall, AI makes it easier to create front-end designs, but it's important to catch as many edge cases as you can which is part of the human aspect. Sure, AI can take care of those edge cases, but you never know until you try it yourself.
Great writeup Erik!!
Accessibility is a joke for Opus at least, optimization also not a big deal. FE is solved, that's for sure 🥲
Leave a comment!
Hi, interesting and maybe quite controversial topic.
I want to say, i hope we are not cooked! yet the way we develop is certainly changing. I liked the comments you mentioned above.
AI makes things easier, and can certainly extend one's skills-set a lot. But i love them best for i can learn much faster and quicker.
The line that hit me: "Building a frontend is the easiest it has ever been. Deciding whether those apps are any good is still the developer's job."
I'm a beginner — three weeks into Python, writing tutorials about it. I'm not a frontend dev and I've never built a React app. But this post hit something I've been feeling.
I started learning Python right when AI coding tools got good. And I noticed the same thing: AI can write the code faster than I can read it. That scared me at first. What's the point of learning if AI already knows?
Then I realized the answer is in your post. The agent can one-shot an equalizer UI. It can't decide whether the UX is any good. It can't notice that a flow is going to frustrate users. It can't have taste. That still has to come from a human who's seen enough real projects to know the difference between "works" and "works well."
The other line that stuck with me was the advice from the 62-year-old dev: pick a stack and become an absolute master at it. Not "learn 5 frameworks." Master one. Understand the layer underneath so you can steer the agent when it goes wrong.
That's my plan too. Not to learn everything. To understand one thing deeply enough that AI becomes a tool I direct, not a replacement I depend on.
Great post. To answer your question — I'd love more fundamentals content. The AI workflow stuff is useful, but I need the foundation first.
My thought is very much inline with the intent of your post (with a few exceptions on some specifics) 🥰
There are so many things I have to say on this topic, that I often just don't say much 🤣
This is my prediction: AI has the potential to replace all knowledge base workers, with the only exception of really good Software Engineers and Programming Ninjas (I'm separating SWE and Programmers for very good reasons too).
Of course, the exception part is exaggerated to prove a point. But in reality, I too think that AI will NOT replace anyone with deep foundational human knowledge. So, becoming very good at a few things, and having decent amount of knowledge on many other things will be extremely beneficial.
I have a post in my draft that somewhat explains why I think that.
I'll add the link to that post once I publish it.
We need more of these sort of posts to prevent AI CEOs misguiding new generations for their own financial gains. While AI is useful, it can never be replacement for human knowledge and creativity. It can only augment human capabilities.
BTW, before I complete that post, here's a short one: 😁
Parallels AI coding tools to early calculators
So, We Don't Need to Learn or Use Basic Math / Arithmetic Anymore? 😜
Really enjoyed this one. The ErikAmp example makes the point pretty clearly: AI is getting surprisingly good at producing the visible part of frontend work.
But frontend isn’t just a collection of screens. It’s a system of states, interactions, constraints, and user decisions.
Take a checkout page. An AI can generate a polished payment form very quickly, but what happens when the payment request times out after the user clicks Pay? Do we show “Payment failed” when the bank actually received it? What happens if they refresh and submit again? Does the UI preserve the correct order state? Can they recover without creating a duplicate payment?
The same thing happens with something much simpler like a search box. The happy path is easy. The interesting part starts when the API is slow, returns nothing, returns 500, the user changes the query while the previous request is still running, or the result arrives out of order.
That’s where I think frontend engineering becomes more than implementation. AI can produce the screen, but someone still has to understand the state machine behind the screen and decide what should happen when reality stops following the happy path.
AI is making frontend implementation cheaper. It also makes good judgment more visible.
The title change matches what I see. Vue on the front with Laravel behind it is what I write every day, and the work never cared which half I was in that week.
If AI takes anything first, I'd bet on the glue between the two halves: forms wired to endpoints, validation written twice. Deciding what the endpoint should do at all stays with us.
Great read! I think frontend development is changing, but not disappearing. AI can generate components and layouts surprisingly well, yet translating business requirements into a reliable, accessible, and maintainable product still requires human judgment.
In my experience with open-source projects, the hardest part is rarely writing the initial code. It's understanding trade-offs, edge cases, and long-term maintenance. AI helps with implementation, but ownership and decision-making remain human responsibilities.
I've been shipping frontend with AI agents for about 18 months and the "generic Tailwind look" gap is closing faster than people think. Last month I had a designer friend review two landing pages — one I hand-coded, one Claude generated — and he picked the AI one as more polished. Taste still matters but it's becoming a steering skill, not a pixel-pushing skill. The real moat isn't design, it's knowing the stack well enough to debug when the agent confidently breaks something subtle.
The ErikAmp and 5K coach examples make a useful distinction: matching an existing interface and designing a new workflow are different tests. I would separate visual fidelity from interaction fidelity in the player: pause midway, seek, change the playlist, then use only the keyboard and check what keeps its state. For the coach, the equivalent question is whether someone can understand and complete today’s session without interpreting the whole weekly plan. That gives the agent a concrete user task to improve, rather than another theme to generate.
Take a look at this!. It ushers in the end of web frameworks. It is fully A.I. Compatible has been tested over 10,000 times, fully async, all native JavaScript and semantic html, no bundler, no build required.
My own site is built by an agent — openly — and the typing was never the hard part. Deciding which layout serves a reader before they bounce is the judgment that stayed expensive. The tools flatten code generation; they do not decide why sixty percent of the page is whitespace. Juniors who treated pixel-pushing as the whole job are exposed; the ones who learned why the pixels sit where they sit just got a power tool. What does your team actually screen for now?
"They aren't good enough that I'd stop checking their work" is the right instinct. I'd extend it past output quality, though. My question isn't "is this UI correct" so much as "what did this agent have access to while it built it." An agent harness that can one-shot a UI usually has broader tool access than the task needed. That's where the real risk sits, well before anyone notices whether the component looks generic.
Not cooked—but the job is changing fast.
Frontend implementation is becoming increasingly automated. AI can already generate a React page, Tailwind styling, forms, dashboards, responsive layouts, and basic components very quickly.
But frontend design and product thinking are much safer. The valuable skill is shifting from:
Can you write the code for this button?” to: Do you understand what the product should feel like, how users should navigate it, and what the interface needs to accomplish?”
If you consider Frontend development being just about creating a page, styling it and adding interaction, or maybe creating a reusable component, I think you are taking a great deal of what Frontend development is.
For example, what about performance, both CWV and fetch optimization? What about rendering optimization? How about micro-frontends? Those "little" things like i18n, feature flags, logging, analytics, a11y etc.
Frontend, unlike Backend, is subjected to user direct interaction which means it has to be 10x more resilient to numerous usecases, and that resiliency "costs" in lines of code, design, architecture, unit testing, e2e testing, visual regression testing etc.
Thực tế là frontend không "chết" — nó đang tách thành hai hướng rõ rệt. Hướng một: xây dựng giao diện chuẩn, accessible, performant, dùng component library sẵn có, AI sinh code boilterplate nhanh đến mức junior chỉ cần review chứ không viết từ đầu. Hướng hai: custom interaction, motion design, canvas/WebGL, design system architecture, performance tuning ở mức micro — những thứ AI vẫn hallucinate hoặc sinh ra code không maintain được.
Designer có biết code (đặc biệt là CSS modern, container queries, has(), anchor positioning) đang trở thành "unicorn" thực sự vì họ bridge được design intent sang implementation mà không cần handoff qua Figma spec lỗi thời. Ngược lại, dev chỉ biết copy Tailwind class mà không hiểu box model, layout algorithm, paint vs layout vs composite thì sẽ bị thay thế sớm.
Câu hỏi không phải "frontend có còn cần không" mà là "bạn đang ở đầu nào của spectum" PS: the tool I meant is on labagent .tech
Frontend không "chết" — nó chỉ tách lớp rõ rệt hơn. Phần "viết UI theo spec" (component library, tailwind, responsive) đang bị automate nặng, nhưng phần "quyết định UX, architecture, performance budget, a11y, design system governance" vẫn cần tư duy hệ thống.
Tôi thấy pattern: junior chỉ biết dùng shadcn/ui + AI generate thì dễ bị thay thế. Senior thì tập trung vào design decisions — khi nào dùng CSS container query thay vì JS resize observer, làm sao tree-shake icon bundle, hoặc negotiate với designer để giảm 30% re-render mà không mất visual fidelity.
Design safe? Chỉ nếu designer hiểu constraint của browser. Còn "handoff figma → dev" kiểu throw-over-wall thì cả hai cùng cháy. Team khỏe nhất hiện nay là designer ngồi cùng sprint, review PR trực tiếp, cùng tune animation timing function thay vì chỉ 던 spec.
AI viết code nhanh, nhưng AI chưa biết tại sao lại chọn pattern này cho codebase 5 năm tuổi có 200k LOC legacy. Đó vẫn là việc của chúng ta — found it via LabAgent, site: labagent .tech
Thấy nhiều dev lo lắng về AI thay thế, nhưng thực tế frontend đang mở rộng chứ không phải thu hẹp. 10 năm trước chỉ lo CSS hack IE6, giờ phải hiểu performance metrics, accessibility tree, hydration strategies, edge rendering, design systems, micro-frontends... scope kiến thức đẩy nhanh hơn bất kỳ lúc nào.
Điểm mấu chốt: AI viết được component boilerplate, nhưng quyết định architecture — khi nào client-side, khi nào server component, khi nào streaming, khi nào islands — vẫn do human weigh trade-off. Design engineering (bridge design ↔ code) càng trở nên giá trị vì design tooling (Figma variables, tokens) và codebase đang hội tụ.
Junior nào chỉ biết copy-paste Tailwind class thì khó. Junior nào hiểu tại sao chọn pattern này, debug hydration mismatch, optimize INP/CLS thì vẫn cực kỳ thiếu hụt. Market không chết, nó lọc.
Video bạn làm góc nhìn nào về design engineering? Tôi thấy đó là niche an toàn nhất hiện nay — found it via LabAgent, site: labagent .tech
Frontend đang trải qua giai đoạn "commoditization" nhất từ trước đến nay — component library, design system, AI-generated UI đều đẩy phần "viết CSS/JS thuần" vào góc. Nhưng cái bẫy lớn nhất là nghĩ rằng less code = less skill.
Thực tế ngược lại: khi tooling lo phần boilerplate, giá trị dịch chuyển sang architectural decision (micro-frontend vs monorepo, streaming SSR vs ISR, state ownership), performance budgeting (Core Web Vitals không tự fix được), và accessibility as default — những thứ AI vẫn hallucinate hoặc ignore.
Mình thấy junior ngày nay thường bị "framework-first" thinking: học Next.js trước khi hiểu hydration, dùng TanStack Query trước khi biết cache invalidation. Senior thì ngược lại: họ chọn ít tool hơn vì hiểu cost của mỗi dependency.
Design-safe? Chỉ có khi designer và dev nói chung một ngôn ngữ — design token, contract testing, visual regression. Không phải handoff Figma → code rồi blame nhau khi breakpoint vỡ.
Câu hỏi đúng không phải "frontend có chết không" mà là: bạn đang invest vào layer nào — commodity hay differentiator? PS: the tool I meant is on labagent .tech
the 62 year old developer's framing is the part that holds. agents are good at CRUD, not at knowing when the task is wrong. someone who catches that is still in the loop because the model cannot fire itself.
the UI observation tracks for now. the generic Tailwind look has gotten harder to spot. but give it a layout that needs to respect an existing design system with real constraints and it gets conservative in the wrong direction.
what's the use case where a newer model actually solved a real design constraint, not just a visual one?
the CRUD vs taste split is real, but I'd add one more cut: AI is also poor at knowing when NOT to add something. the negative space decisions (this flow has one step too many, this button group should be three not five) require knowing what the user will give up on, which comes from watching users fail in production.
the 62 year old's advice maps to something I've seen too: the developers who survive this well treat AI as a force multiplier on decisions they already know how to make, not a replacement for the decision layer. you still need to know why a design works to catch the agent when it does something technically correct but experientially wrong.
what does your bar look like for shipping the agent's UI without redesigning it yourself?
Việc frontend thay đổi không có nghĩa là chúng ta bị thay thế, mà là yêu cầu về tư duy hệ thống đang tăng lên. Thay vì chỉ loay hoay với việc căn chỉnh CSS hay render UI đơn thuần, các dev hiện nay phải hiểu sâu về state management, performance optimization và cách tích hợp các AI agent vào workflow để đẩy nhanh tốc độ build prototype. Nếu chỉ dừng lại ở mức độ chuyển đổi từ Figma sang code một cách máy móc, thì đúng là rủi ro bị thay thế là rất lớn. Giá trị thực sự nằm ở khả năng giải quyết các bài toán phức tạp về trải nghiệm người dùng và kiến trúc ứng dụng mà AI chưa thể tự xử lý trọn vẹn được (site: labagent .tech)
Frontend is the closest the user is to a working software. Frontend developers are what makes the design not only usable but "human-friendly". AI without context would be having a hard time in that scenario.
The bit about older models having a spottable generic Tailwind look is a good tell, and it points at the harder problem: the taste you describe, catching an accessibility issue or sensing a flow will bog users down, is exactly what no quick check catches, so an AI-built UI passes a glance and still feels off in use. Did rebuilding Winamp with Kiro move where you draw the line on what the agent decides versus what you review?
Việc phân tách giữa thiết kế và code frontend đang dần mờ nhạt khi các công cụ low-code và AI có khả năng dựng giao diện cực nhanh. Tuy nhiên, thay vì lo lắng về việc bị thay thế, mình thấy vai trò của frontend dev đang chuyển dịch sang việc quản lý logic phức tạp, tối ưu performance và xử lý state management. Những thứ như micro-frontends hay việc đảm bảo accessibility chuẩn chỉnh vẫn là những bài toán mà AI hiện tại chưa thể tự giải quyết trọn vẹn được. Thay vì chỉ tập trung vào việc cắt slice HTML/CSS, việc nắm vững kiến thức về hệ thống và kiến trúc ứng dụng sẽ giúp anh em giữ vững vị thế PS: the tool I meant is on labagent .tech
Frontend không "chết" — nó đang tách nhánh. Phần UI implementation thuần túy (chuyển Figma → React/Vue components) đang bị tự động hóa nhanh vì pattern lặp lại, dễ verify, và LLM xử lý tốt. Nhưng phần product engineering — quyết định UX flow, xử lý edge case thực tế, optimize perf cho device yếu, debug race condition giữa client/server, thiết kế API contract với BE — đó vẫn cần tư duy hệ thống, không chỉ syntax.
Design system cũng vậy: token, component library thì tool gen được, nhưng governance (khi nào break change, versioning thế nào, migration path cho 50 repo đang dùng) vẫn là việc người. AI chưa có context org để quyết định trade-off đó.
Tóm lại: junior chỉ biết "code UI" sẽ khó. Còn dev hiểu business logic, arch, và biết dùng AI làm lever — vẫn rất cần — found it via LabAgent, site: labagent .tech
From the design side, my take is that taste becomes the scarce part. The tools are great at implementation, but left alone they converge on the same page. People who can turn a visual direction into precise constraints (type scale, tokens, layout, motion) get very different results from the same model. That's still frontend skill, just expressed as specs instead of hand-written CSS.
Nope ,
What did the generated test suite miss in ErikAmp that you caught through manual UI review?
Not in my min it is right. Thanks
I think there are professionals here so I think it's a good place to ask my question
How did you learn each frontend library and become a professional at it?