Author of Babashka Conf & Dutch Clojure Days talks sabotage: overthinking kills projects. Last weekend, host Marcin and I built a weekend woodworking shelf—simple, scoped, done. Success? Jamming on a project with a friend.

Contrast: spent 4 hours researching structural/semantic diff tools, overcomplicating to replace existing workflows—no prototype, time wasted. Key insight: success criteria must be specific and internalized. Example: wanting a nicer diffing workflow in Emacs—scope: build it for myself, 4 hours max. Else, spend hours on background research but no usable prototype. Conservation of scope creep: building Finda-style fuzzy path search, LLM suggested Nucleo library—good start, but added unneeded anchor functionality, threw it out. Lesson: any increase in speed is offset by unnecessary features.

Structural diffing: line-by-line diffs ignore higher-level structure—problem to solve.