Part III benchmarks Opus 5 on SlopCodeBench in the software-factory debate

teropa · x · 2026-07-28

Part III of the Why Software Factories Fail series

This follow-up continues the earlier two installments and focuses on benchmarking Opus 5 on SlopCodeBench. The author frames it as a continuation of the broader argument that software-factory style automation still runs into hard limits, even as benchmarks improve.

Original post →

More from coding & agent

coding & agent channel →