An operational philosophy in MLOps that subjects AI agents to standard software engineering best practices, including version control, automated unit testing, CI/CD pipelines, and bug tracking.

Used by MLOps leads to remind developers to stop treating AI prompts like magic spells and start testing and versioning them like standard python code.

• Common Sense Rebranding: Framing basic software engineering hygiene (testing and version control) as a revolutionary new philosophy.
• Pragmatism Posture: Stating the obvious rule that code interacting with AI should be tested like any other software code.
• House-Cleaning Reality Check: Like washing a plastic cup the same way you wash a glass cup and calling it "treating every container like dishware."

Halt and Catch Fire (2010s) scene where engineers enforce strict software development hygiene, treating every hardware interface component like standard, testable software code.

Applying standard software testing and version control rules to AI prompt scripts.

“Treating every agent like software prevents unvetted prompt changes from breaking production apps.”

“We put all our chatbot prompts into Git code repositories and ran automated tests on them before pushing updates live.”