EP 147 - AI Tools: CLEAR with Marcos Polanco
Wed Feb 04 2026
We’re continuing our AI Tools series with Marcos Polanco, engineering leader, founder, and ecosystem builder from the Bay Area, who joins Matt and Moshe to introduce CLEAR, his method for using AI to build real software, not just demos.
Drawing on decades in software development and his recent research into how AI is reshaping the way teams ship products, Marcos shares how CLEAR gives both technical and non‑technical builders a production‑oriented way to work with vibe coding tools.Instead of treating AI like a magical black box, Marcos frames it as an “idiot savant”: incredibly capable and eager, but with no judgment.
CLEAR wraps that raw power in structure, guardrails, and engineering discipline, so founders and PMs can go from prototype to production while keeping humans in control of the last, hardest 20%.
Join Matt, Moshe, and Marcos as they explore:
Marcos’s journey through engineering, founding, and AI research, and why he created CLEAR
Why AI tools like Bolt, Cursor, Claude, and Gemini are fabulous for prototypes but risky for production without a method
CLEAR in detail:
C – Context: onboarding AI like a new hire, using stories and behavior‑driven design (BDD) to articulate requirements
L – Layout: breaking work into focused, scoped pieces and choosing a tech stack so AI isn’t overwhelmed
E – Execute: applying test‑driven development (TDD), writing tests first, then having AI write code to pass them
A – Assess: using a second, independent LLM as a QA agent, plus a human‑run 5 Whys to fix root causes upstream
R – Run: shipping to users, gathering new data, and feeding it back into the next iteration of context
How CLEAR lowers cognitive load for both humans and AIs and reduces regressions and hallucinations
Why Markdown (with diagrams like Mermaid) is becoming Marcos’s standard format for shared human–AI documentation
How CLEAR changes the coordination layer of software development while keeping engineers central to quality and judgment
Practical advice for PMs and founders who want to move from “just vibes” to predictable, production‑grade AI development
And much more!
Want to go deeper on CLEAR or connect with Marcos?
CLEAR on GitHub: https://github.com/marcospolanco/ai-native-organizations/blob/main/CLEAR.md
CLEAR slides: https://docs.google.com/presentation/d/1mwwDtr7cCP5jLUyNVgGR5Aj-MBq8xsMlhSc0pvSQDks/edit?usp=sharing
LinkedIn: https://www.linkedin.com/in/marcospolanco
You can also connect with us and find more episodes:
Product for Product Podcast: http://linkedin.com/company/product-for-product-podcast
Matt Green: https://www.linkedin.com/in/mattgreenproduct/
Moshe Mikanovsky: http://www.linkedin.com/in/mikanovsky
Note: Any views mentioned in the podcast are the sole views of our hosts and guests, and do not represent the products mentioned in any way.
Please leave us a review and feedback ⭐️⭐️⭐️⭐️⭐️
More
We’re continuing our AI Tools series with Marcos Polanco, engineering leader, founder, and ecosystem builder from the Bay Area, who joins Matt and Moshe to introduce CLEAR, his method for using AI to build real software, not just demos. Drawing on decades in software development and his recent research into how AI is reshaping the way teams ship products, Marcos shares how CLEAR gives both technical and non‑technical builders a production‑oriented way to work with vibe coding tools.Instead of treating AI like a magical black box, Marcos frames it as an “idiot savant”: incredibly capable and eager, but with no judgment. CLEAR wraps that raw power in structure, guardrails, and engineering discipline, so founders and PMs can go from prototype to production while keeping humans in control of the last, hardest 20%. Join Matt, Moshe, and Marcos as they explore: Marcos’s journey through engineering, founding, and AI research, and why he created CLEAR Why AI tools like Bolt, Cursor, Claude, and Gemini are fabulous for prototypes but risky for production without a method CLEAR in detail: C – Context: onboarding AI like a new hire, using stories and behavior‑driven design (BDD) to articulate requirements L – Layout: breaking work into focused, scoped pieces and choosing a tech stack so AI isn’t overwhelmed E – Execute: applying test‑driven development (TDD), writing tests first, then having AI write code to pass them A – Assess: using a second, independent LLM as a QA agent, plus a human‑run 5 Whys to fix root causes upstream R – Run: shipping to users, gathering new data, and feeding it back into the next iteration of context How CLEAR lowers cognitive load for both humans and AIs and reduces regressions and hallucinations Why Markdown (with diagrams like Mermaid) is becoming Marcos’s standard format for shared human–AI documentation How CLEAR changes the coordination layer of software development while keeping engineers central to quality and judgment Practical advice for PMs and founders who want to move from “just vibes” to predictable, production‑grade AI development And much more! Want to go deeper on CLEAR or connect with Marcos? CLEAR on GitHub: https://github.com/marcospolanco/ai-native-organizations/blob/main/CLEAR.md CLEAR slides: https://docs.google.com/presentation/d/1mwwDtr7cCP5jLUyNVgGR5Aj-MBq8xsMlhSc0pvSQDks/edit?usp=sharing LinkedIn: https://www.linkedin.com/in/marcospolanco You can also connect with us and find more episodes: Product for Product Podcast: http://linkedin.com/company/product-for-product-podcast Matt Green: https://www.linkedin.com/in/mattgreenproduct/ Moshe Mikanovsky: http://www.linkedin.com/in/mikanovsky Note: Any views mentioned in the podcast are the sole views of our hosts and guests, and do not represent the products mentioned in any way. Please leave us a review and feedback ⭐️⭐️⭐️⭐️⭐️