Auditing our inputs, answering the four big questions, and figuring out that one Q isn’t enough.
The Four Questions
After the pivot, we needed to nail down what “broader” actually meant. TB answered four shaping questions:
- Tools vs strategy? — Both, strategy-weighted. Big theme: AI as daily operating system.
- Career/PM? — First-class. TB is becoming a PM, not just watching.
- Sources? — Additive, not replacement. Audit for gaps.
- “So what” framing? — “Why does this matter” full stop. Not always tied to Newspack.
Source Audit
Full audit of everything we follow — 22 newsletters, 19 podcasts, 9 YouTube channels, 4 Instagram accounts, 28 Twitter voices, 15 subreddits. Organized into 5 categories. Found 4 gaps: AI strategy voices, AI × productivity, PM newsletters, AI × PM intersection. Added SVPG, Product Talk, One Useful Thing, Superhuman AI, Tiago Forte, and LLM Watch.
Multi-Agent Architecture
TB asked the right question: should we split the work across multiple agents? Yes.
Five streams identified:
- Research & Synthesis — newsletters, podcasts, YouTube, Reddit
- Competitive Intel — WP Engine, CMS landscape, market moves
- PM Growth — TB’s career development (stays with Q)
- Newspack Dogfooding — testing releases, finding pain points
- Prototyping — building ideas, proof of concepts
Research and Intel agents spun up with their own workspaces. Q stays as TB’s single point of contact — orchestrating the others, never making TB manage multiple conversations.
Why it matters: We went from one AI doing everything to a small team with specialties. The source audit showed us what we were actually consuming vs. what we thought we were. And the four questions gave us a compass for everything that follows.