January 27, 2026
Code vs Chatbot: Fight Night
Ask HN: How to avoid skill atrophy in LLM-assisted programming era?
Are chatbots making coders rusty? The crowd fires back
TLDR: A Hacker News thread asked how to avoid losing coding chops as AI chatbots help write code. The crowd split between “ditch the bots,” worries about sloppy work and interview prep, and “this is just progress”—solution: keep building your own projects and review everything
On Hacker News, a simple question lit up the comments: are AI chatbots (aka “large language models,” tools that can write code) making programmers rusty? The crowd split fast. One camp shouted: stop outsourcing your brain. User bitbasher flexed, “I don’t even use an lsp,” meaning not even editor helpers. Reviewer icedchai ranted that heavy AI users “blindly open PRs full of slop,” dumping messy work on teammates. Another anxiety: interviews. Dysoco said coding puzzle prep has slipped; raw_anon shot back: were you practicing that at work anyway?
Others brought popcorn and perspective. Lowbloodsugar mocked the panic with history memes—spinning jenny, automobiles, typewriters—basically: new tools always freak people out, then become normal. Kylehotchkiss offered the wholesome fix: build a personal site and passion projects, hand‑code, and don’t open Claude for every little thing. Amadeuswoo claimed the only skills atrophying are the ones you weren’t really using.
The original poster confessed the real temptation: ask AI to design options first (fast, but brain fade), or think it through yourself then ask AI for feedback (slow, but sharper). Notepad0x90’s middle ground: let AI assist, but keep reviewing, second‑guessing, and deciding. OGEnthusiast’s parting shot: ignore the grifters and the hype
Key Points
- •An Ask HN post asks how to prevent skill atrophy when using LLMs to assist in programming.
- •The post questions whether technical skill will continue to matter in AI-assisted coding.
- •It focuses on balancing LLM productivity benefits with maintaining critical thinking and code understanding.
- •No concrete solutions are provided in the post; it solicits strategies from the community.
- •The thread is hosted on Hacker News as a request for practical advice on integrating LLMs without losing skills.