Lessons for Agentic Coding: What should we do when code is cheap? | Hacker News

Hacker News commenters debated whether AI coding agents really make software cheap, or just move the cost elsewhere.

Many said these tools are helpful for prototypes and routine edits, but still struggle with architecture, context, and maintainability in real production systems.

The concern is that faster code generation can raise technical debt, create “big ball of mud” software, and increase production risk.

Some also warned that treating code as cheap could reduce investment in senior engineering talent and careful refactoring.