Developer's practice: AI coding shifts the bottleneck from implementation to design

4310sy · x · 2026-09-05

A Japanese developer shares his current workflow with AI coding agents like Claude Code: now that agents run autonomously and in parallel, he barely writes code himself—the most time-consuming part of development has moved from implementation to pre-implementation design.

Key practices:

His takeaway: the more autonomous the agent, the more design matters—engineer value is shifting toward articulating requirements and design.

Original post →

More from coding & agent

coding & agent channel →