Triage your backlog with LNO - product-minded engineering with Sam Julien

Kent C. Dodds introduces the episode and welcomes his guest, Sam Julien, a seasoned product professional with a journey from self-taught developer to director of product. They recount their first in-person meeting at a conference in San Francisco ten years ago, noting the changes and progress in their careers and the industry since then. Sam shares his diverse experience across coding, marketing, and product, emphasizing his current role at CopilotKit, which focuses on generative UI, and his authorship of an O'Reilly book, Agentic GraphRAG. The conversation sets the stage for exploring the evolving role of software engineers, especially as AI tools automate many coding tasks.

Sam discusses how AI advancements have leveled the technical skill playing field, making 'how' something is built less of a competitive moat. Instead, the critical differentiation now hinges on good judgment—knowing what to build, when, and why. Kent supports this view, emphasizing that being product-minded is increasingly vital, especially since the traditional scarcity of technical skills is diminishing. They explore why engineers should understand the product context deeply enough to challenge and refine specs, leading to better outcomes and more collaborative engineering and product processes.

A key theme is the value of engineers who can push back on given specifications, not out of defiance but to improve product outcomes. Sam explains that such engineers use telemetry, user feedback, and product understanding to recommend better options. Kent shares his own experience with AI agents, highlighting that the bottleneck has shifted from technical capability to decision-making and judgment. Both agree that a collaborative approach, where engineers understand user problems deeply, leads to more effective and impactful solutions—especially with the assistance of AI.

The conversation dives into frameworks like LNO (leverage, neutral, overhead), RICE, Jobs to Be Done, and North Star metrics as tools for engineers to assess their backlog. Sam emphasizes the practicality of LNO for triage: identifying work with high impact relative to effort, whether necessary, or candidates for automation. Kent echoes the importance of measuring value, impact, and effort, especially in AI-enhanced development, to avoid just shipping features without understanding their long-term benefit. They advocate for applying these frameworks regularly, even for small weekly reviews, to make better prioritization decisions.

The episode underscores the importance of defining problems, not just solutions, when working with AI agents. Sam advocates for providing context and explicit intent before engaging an agent, which improves the quality of outputs and reduces repetitive iterations. Kent observes how recent AI improvements allow for more transparency and debugging—asking the agent why it made certain decisions or what assumptions it's based on. They discuss the significance of clear communication, shared vocabulary, and intention setting to harness AI effectively while ensuring the engineering aligns with product goals.

Sam highlights the need for engineers to be adaptable in fast-moving AI markets, where product demand, customer needs, and capabilities evolve rapidly. Continuous discovery—regularly engaging with users, analyzing telemetry, and assessing impact—is vital to avoid building features that quickly become obsolete. Kent agrees, noting that this ongoing iterative approach ensures products deliver real value. They recommend focusing on impact, not just activity, and using frameworks like Teresa Torres' to maintain agility and relevance in AI-driven product development.

The episode concludes with practical advice: engineers should spend a brief time each week triaging their backlog using frameworks like LNO, asking why they are doing each task, and understanding the effort versus impact. Kent advocates for this habit to develop better judgment and product awareness, regardless of role. Sam encourages exploring resources like Teresa Torres' work for continuous discovery insights and checking out CopilotKit’s open-source projects that embody the discussed philosophies. The overarching message is to integrate product thinking into daily engineering practices to improve long-term impact.