[romankolpak]

The scope and complexity of the change is crucial here. There is a sweet spot where spec-driven agent offers great value: the change is small and simple enough that you can describe it reliably with a spec (i.e. you have complete understanding how to do it), while implementing it yourself is a lot more work than writing the said spec plus reviewing and correcting the agent.

For me currently this sweet spot is TINY. It's so small that my usage of Claude Code has dropped to almost none. It's simply more practical to let myself have the agency and drive the development, while letting AI jump in and assist when needed.