Define behavioral inclusion criteria, exclude bias sources like pros, and use a recruitment matrix for diversity to ensure external validity and avoid misrecruits costing time, incentives, and bad decisions.
Solo GM runs full product by writing only the roadmap; Claude generates PRDs, tickets with context/data/AC/tech notes from GitHub README in minutes, fed by user feedback/usage data.
When your company's default AI tool underperforms, don't complain—run a simple one-week test on a recurring job comparing it to a specialist tool. Measure time saved and quality to reframe your ask as evidence, not preference.
Tie AI token spend to specific business outcomes using 'AI x Outcome = Strategy'—without a clear outcome sentence, it's just wasteful token burning subsidized by VCs today.
In early hypergrowth, expand proven leaders' scope to fix specific problems serially. In late stage, hire new leaders to tackle skeptic demands everywhere at once—key for AI-era scaling.
Early hypergrowth solves specific problems serially by expanding proven leaders' scopes. Late hypergrowth demands parallel solutions for skeptics, requiring new specialized leaders instead of scope creep.
