Build vs buy — when to write it yourself vs use something existing

Almost every non-trivial feature has an existing library or service that claims to solve it — authentication, payments, date parsing, PDF generation. "Build vs buy" is the decision of whether to write that piece yourself or bring in something that already exists (a library, a paid service, an open-source tool) — and it's a real trade-off, not a question with a universally correct answer.

Pulling in a dependency isn't free even when it's open-source and costs nothing directly:

Writing it yourself has its own real, easy-to-underestimate costs:

Ask whether the thing in question is core to what makes your specific product valuable, or a solved problem every product like yours also needs. Payment processing, authentication, and PDF rendering are rarely what makes a product distinctive — they're usually "buy" (or use a well-maintained library). The specific business logic that actually differentiates your product from a competitor's is usually the "build" — that's exactly the part where writing custom code, tailored precisely to your case, pays off.

Neither extreme survives contact with a real project: refusing every dependency means reinventing dozens of already-solved problems badly; adding a dependency for absolutely everything means inheriting maintenance burden and security surface for things that didn't need it. The actual skill is asking the "is this core to what we do" question deliberately, case by case, instead of defaulting to either extreme out of habit.