With the advent of LLMs, the temptation to churn out a flood of unit tests for a false
veneer of productivity and protection is stronger than ever.
My colleague Matthias Doepmann recently fired a shot at AI-generated tests that don’t
validate the behavior of the subject under test (SUT) but instead create needless ceremony
around internal implementations. At best, these tests give a shallow illusion of confidence
in the system’s correctness while breaking at the smallest change. At worst, they remain
green even when the SUT’s behavior changes.

Planet Golang is a community-driven aggregator that compiles the latest and most relevant content about the Go programming language from blogs, forums, and websites across the globe. It provides developers with a centralized platform to discover tutorials, technical articles, news, and updates, making it easier to stay connected with the Go ecosystem. 

Planet Golang

State-based testing validates system behavior by checking outcomes rather than implementation details. Mock-based interaction testing creates brittle tests that break on harmless refactors while missing real bugs. Handwritten fakes that implement domain rules provide better test maintainability and reliability than generated mocks. LLMs often produce poor interaction-checking tests that add maintenance overhead without meaningful validation.

Test state, not interactions