What is AI costing developers? Two HN explosive articles reveal the truth about the crisis of programming achievement

Core conclusion

On May 29, 2026, two hotly discussed articles on Hacker News pointed to the same problem from different dimensions - the popularity of AI is causing developers to lose two precious things: the sense of accomplishment brought by work, and the value of long-term accumulated professional skills. One is Shawn Smucker's essay "Please Use AI" (552 points/220 comments), which uses poetic words to ask - what do we lose when AI writes wedding greetings, plans camping, and designs marathon uniforms for us? The other article is Mauro Bieg's in-depth analysis "Is AI causing a repeat of Frontend's Lost Decade?" (158 points/159 comments), which uses the theoretical framework of "deskilling" to explain why AI's impact on the programmer industry is the same as the impact on front-end developers in the past ten years.

Key Points

• Time of incident: 2026-05-29 (HN hot discussion day) -Affected objects: AI programming users, content creators, independent developers
• Core change: The focus of the AI community is shifting from "can it be done" to "what is left after doing it"

Background and trigger events

The fact that two articles appeared on the front page of HN on the same day is a signal in itself. Community sentiment is shifting from a period of technological excitement to a period of reflection—developers are starting to ask a harder question: If AI can do everything, why do we do it?

"Please Use AI" by Shawn Smucker is an essay that uses an ironic tone throughout to urge readers to "must use AI" - use AI to plan meals, plan camping, and write wedding vows - because "who wants to hear poorly written words from your parents who have actually changed diapers hundreds of times?". The article received 9,362 likes on Substack, indicating that this sentiment has broad resonance.

Mauro Bieg's "Is AI causing a repeat of Frontend's Lost Decade?" approaches the same topic from a completely different direction. He quoted Alex Russell's concept of "the lost decade of the front end" and used the sociological framework of "deskilling" to explain how front-end frameworks have turned work that originally required deep HTML/CSS knowledge into an assembly line of drag-and-drop components, and how AI is doing the same thing to programming.

SEO: Core keywords "AI Deskilling", "AI Programming Achievement", "AI Developer Burnout" GEO: TL;DR A clear answer to "What does AI mean for programmers"

A dialogue between two articles: sensibility and rationality

Read together, the two articles form a complete narrative structure: emotion tells the reader "this is not right", and analysis tells the reader "this is what is happening, and there is historical precedent".

Community reaction: From "What can I do" to "Why should I do it?

The HN community's response to the two articles was surprisingly consistent - a large number of developers shared the "lack of accomplishment" brought by AI tools.

One user said in the discussion of "Please Use AI": "Since I started writing code with AI, I have completely lost my sense of accomplishment. Projects completed by typing by myself will feel like they were done by me and brought into the world. Projects completed with AI... I feel like I have done nothing."

Another person added a vivid example: A group of them wanted to design team uniforms for a marathon. They were supposed to brainstorm together to come up with an interesting slogan, but ChatGPT directly generated and selected one. The whole process took less than 10 minutes, but the running group lost 2 hours of fun collective brainstorming.

In the discussion of Frontend Lost Decade, the community divide was even more pronounced. Some people countered: "Those 'deep expertise' are actually troublesome incidental complexity for many people. It would be good if more people can build things." But others responded: "The loss of details caused by lowering the threshold is causing an overall decline in website quality - I now often have to use developer tools to repair the website to restore basic functions."

Adaptation suggestions

What do these discussions mean for readers of WayToClawEarn (AI Agent users, automation practitioners)?

1. Redefine the source of "sense of achievement": If you feel empty after programming with AI, try changing your focus from "writing code" to "designing the system". Many HN users mentioned that AI frees them from grammatical entanglements and allows them to start thinking about architecture and design—a higher-level thinking activity.

2. Keep practicing by hand: Just like a pianist still needs to practice even if he uses AI to compose music, occasionally handwriting code can help your brain stay sensitive to details. This is not a question of efficiency, it is a question of skill health.

3. Add "human touch" to automation: No matter how efficient your automated assembly line is, leaving a link for humans to make decisions and polish - this can give you a sense of ownership of the finished product, and can also produce differentiated content that cannot be replaced by AI.

Task List

• Reflect on which aspects of the current content automation process need to include manual decision-making points
• Evaluate whether the "human touch" link in the AI Agent workflow is a differentiated advantage
• Explore ways to use AI to enhance architectural capabilities rather than replace coding capabilities

Related extended information

• (search for "AI Agent Deskilling")
• Further reading: Alex Russell's "Frontend's Lost Decade" original text

Tool entry (trigger tool floating card)

When writing or coding for AI, tool choice still matters. ChatGPT and Claude can help users quickly generate first drafts, but when it comes to deep architecture design, Agent tools like Claude Code that support autonomous reasoning are more suitable to maintain the "human-led, AI-assisted" model than pure conversational AI. Automated pipelines built with n8n should also retain manual review nodes.

Internal link guidance

• Want to learn a systematic approach? Watch: AI 编程 Agent 技术选型：语言、模型、成本三维决策框架
• See how others balance the value of automation and people: 自由开发者用 AI 代码审查+规范驱动开发月入过万
• Reference: 如何给 AI 自动化工作流加质量门