AI is reshaping job skills — developers and non-developers need to learn what machines can't replace yet

A discussion on Hacker News (6 comments, 4 points) surfaced the question: which skills remain valuable when AI can now write code, draft emails, and automate routine tasks. The thread asks what's worth learning today to avoid obsolescence in an AI-first job market.

The core tension: technical skills like coding are automatable (AI can generate code), but judgment, cross-disciplinary thinking, and the ability to direct AI systems are harder to replace. Non-developers face a parallel shift — spreadsheet skills matter less when AI can analyze data, but asking the right questions and interpreting results matter more.

For you: if your job is pure execution (writing boilerplate code, formatting documents, basic analysis), AI tools are already cheaper and faster. Your payoff comes from becoming the person who tells AI what to build, catches its mistakes, and solves problems AI can't frame on its own — roles that require context, accountability, and judgment.