Coming to think of technical debt in ML systems leads to the additional overhead of ML-related issues on top of typical software engineering issues.