Why: The Devin post (How I Built a Full Booking System as a Side Project) is the strongest signal yet. It's concrete (GymClass app), personal (martial arts academy), and has a clear value prop that differentiates from generic "vibe coding" content. This can become a series: "How to partner with AI" from the product-minded engineer's perspective.

Impact: High — establishes a unique voice at the intersection of product engineering + AI collaboration
Effort: Medium — 3-5 articles, each requiring 1-2 drafting sessions
Confidence: 9/10
Compounds: Yes — each article builds authority for the next, creates a searchable body of work

Sub-opportunities:

"Refactoring with AI: When Devin is better than Copilot" (deep-dive on the Vitest→Bun migration story)

"From Phone to Production: Async Engineering with AI" (the workflow pattern)

"The Product-Minded Engineer's Guide to AI Partners" (positioning piece)

Dependencies:

None active. Could expand with screenshots / real PR links from GymClass.

---

Why: This is Horacio's signature intellectual territory. He has:

State machines to visualize the work of agents (thesis)

State machines are great, but hard to establish (problem)

Notes about how the document machine should work (real implementation)

Devin post (real AI collaboration case study)

The missing piece is the synthesis: a talk or deep-dive article showing how state machines make agent behavior deterministic, observable, and composable. This is a conference talk waiting to happen.

Impact: High — positions Horacio as a thought leader at the intersection of two hot topics (state machines + AI agents)
Effort: High — requires writing a talk and/or article, but material is already scattered across short posts; needs synthesis
Confidence: 8/10
Compounds: Yes — this is the kind of original work that gets cited, shared, and invited

Concrete next action: Outline the talk. The three acts: (1) State machines fix broken UI logic → (2) Document machine in production → (3) Extending the same model to AI agent orchestration.

---

Why: The Devin post's buried lead: "I did all this while working full-time on another product." The meta-lesson — how to ship when you have zero large time blocks — is more valuable than the tool review. This could be a guide with the actual workflow: async prompting, video review from phone, multi-PR orchestration, delegation patterns.

Impact: Medium-High — solves a universal founder/developer pain point
Effort: Low — one solid article, already half-written in the Devin post
Confidence: 8/10
Compounds: No directly, but supports the AI partner series

---

Why: The Devin post describes a full-stack booking app (TypeScript, Bun, Hono, tRPC, Drizzle, Neon, Cloudflare) that's already in production at a real academy. If the codebase is clean enough, open-sourcing it as a "building a modern SaaS with AI" reference architecture could attract significant attention.

Impact: Medium — developer audience, but requires polishing code + writing docs
Effort: High — code cleanup, documentation, deployment guide
Confidence: 6/10 — depends on codebase quality and willingness to maintain
Compounds: Yes — an open-source project with real users is a permanent asset

---

Why: The content across short-posts + writing clusters naturally around a voice: frontend engineer who cares about product, building in public, experimenting with AI tools. A curated newsletter synthesizing these threads could build a direct audience.

Impact: Medium — slow build, compounding over time
Effort: Medium — weekly/biweekly curation + original content
Confidence: 5/10 — crowded space, requires sustained commitment
Compounds: Yes — newsletter audiences accumulate

---