If DSPy is So Great, Why Isn't Anyone Using It?

DSPy adoption is low despite solving real problems in AI engineering. Through a step-by-step walkthrough of a company name extraction system, the post shows how teams inevitably reinvent DSPy's core patterns — typed I/O, composable modules, prompt management, retries, RAG, and evals — but worse and after much pain. DSPy's abstractions (Signatures, Modules, Optimizers) are just software engineering fundamentals applied to LLM systems. The post argues you should either adopt DSPy or deliberately build with its patterns from day one, because the alternative is drowning in ad hoc complexity six months later.