Software Taste vs. Slop in the Age of AI – TWSoftwareDev26
The article contrasts "software taste" (craftsmanship, maintainability) with "slop" (thoughtless output) in AI-assisted coding. It argues that AI often produces slop unless guided by experienced engineers, and that human taste remains essential for building quality systems.
Background
The author, Tod Herlein, is a veteran software developer and blogger. This post is a response to a wave of developer culture debates around "taste" (knowing what good code looks like and when to apply principles) versus "slop" (careless, low-quality output). The core tension: AI coding assistants (like GitHub Copilot or Claude-generated code) make it trivially easy to produce large volumes of code that *looks* right but lacks structural wisdom. Herlein argues that real software taste — the hard-won skill of knowing when to add abstraction, when to keep things simple, when a design pattern fits — becomes *more* important as AI lowers the cost of generating code. Without taste, AI just accelerates the production of technical debt.