Decision
hardcalls.co becomes the second property in the in-house content engine (after JOLT), as an internal property: Michael’s brand, no client billing. Goal: own the founder decision-quality query space in both classic search and AI citations (GEO), with a blog + one YouTube embed per article, feeding the newsletter and Stack doors.
Shape
- Client folder: ~/lore-content/clients/hard-calls/ (strategy, context, rule card, voice card, publishing, topic queue, exemplars from the site v2 canon).
- Two-layer architecture: site v2 canon pages are the hubs (never rewritten for SEO); blog articles are the spokes (long-tail, situational, comparison queries).
- Three launch clusters: A the ten named calls; B method vocabulary (tripwires, one-way doors, enough, etc.); C the documented record (Airbnb 2020, Zuckerberg/Yang, CUDA, Shopify 2022, Spanx).
- Editorial gate: Ben in lore-content-review, same flow as JOLT (notified 2026-08-08).
- Cadence proposal pending Michael’s approval: 3/week (2 Cluster A, 1 B/C), Tue/Wed/Thu.
- Blog scaffold: engineering PR in RyeValley/hard-calls AFTER the v2 site stack (#7/#8/#9/#10) merges. Route /blog. GSC verification at scaffold time.
- Weekly research cron covers hard-calls via research-playbook.md (added 2026-08-08; research-evidence-only until Michael approves the initial topic list).
- Voice: Michael’s print register; rule card stricter than JOLT (seesaw ban, unlocks-> earns, canon-vocabulary policy). Phase 2 video (Michael on camera) is Michael’s call.
Why
The v2 site is a definitional play; GEO is the distribution arm of the same thesis. The canon (fixed vocabulary, receipts-flagged cases, honest caveats) is exactly the content shape LLMs cite. Nobody owns founder decision queries today.