SC 015 — In the Lab: Experimenting with dbt’s State Management & Writing Online

Summary

Two threads: (1) experimenting with dbt’s state management for local development, noting the gap between dbt Cloud’s polished Slim CI workflows and the rougher local experience; (2) reflecting on a 12-week Write of Passage Runway experiment — publishing one article or three Twitter threads weekly.

Results: hit publishing targets consistently (five weeks of three threads, five weeks of newsletters). Twitter excelled for serendipitous connections with industry figures. Substack offered creative freedom. Lowering quality standards paradoxically enabled publication. Neither platform generated meaningful follower growth.

Key Arguments

• dbt’s state management is powerful in Cloud but underdeveloped locally — a real gap for practitioners
• Lowering quality standards is a publication enabler, not a compromise
• Twitter for connections, Substack for depth — different platforms serve different purposes
• Consistency matters more than any single post’s performance

Writing Style Notes

Honest about mixed results — no inflated metrics or false optimism. The “lowering quality standards paradoxically enabled publication” insight is characteristically self-aware. Experimenter’s mindset throughout.

Connections

• 01-projects/newsletter/index — meta-reflection on the newsletter practice itself
• 06-reference/2026-04-03-curiosity-consistency-newsletter-growth — consistency as growth strategy