Notes d’épisode
What changes when AI makes code easier to produce—but human attention, judgment, and accountability remain limited? Shiv talks with Allan Macgregor, the author of the pragmatic CTO, about pragmatic technology leadership, the evolving craft of software engineering, and why solving the right problem matters more than writing more code.
Allan traces his journey from aspiring architect in Mexico to a software career in Canada spanning e-commerce, travel, and HR and payroll technology. Now CTO at LiORA, he discusses work on environmental intelligence, combining sensors and groundwater models to better understand soil and water contamination. Along the way, he shares experiments with AI assistants, approaches to reviewing AI-generated code, and career advice for both newcomers and experienced engineers.
In This Episode
• Curiosity over specialization: How working across different businesses shaped Allan’s ability to learn domains quickly and explain technical ideas through analogies.
• What makes a CTO pragmatic: Balancing ambition with limited resources, resisting shiny-object syndrome, and being comfortable saying “not right now.”
• Technology with real-world impact: Allan’s account of using frequent sensor readings and groundwater simulations to inform environmental remediation.
• Build versus buy in the AI era: Why the ability to create a software clone does not settle questions of maintenance, security, support, or whether the work deserves the team’s attention.
• Agents beyond the hype: Lessons from Madonna, Allan’s AI executive assistant, and an experimental agentic modeler for groundwater modeling.
• From engineer to orchestrator: Working with agents through clear outcomes, boundaries, and escalation rules rather than constant micromanagement.
• The code-review bottleneck: Why faster code generation creates new verification challenges, and Allan’s experiment with prioritizing review by risk and context.
• Value-driven engineering: Looking at problems solved and stakeholder outcomes rather than lines of code, and using quick prototypes to test ideas earlier.
• Career paths in a changing profession: Communication, systems thinking, and fundamentals for newcomers; forward-deployed engineering as a possible direction for experienced leaders.
• Writing as a grounding practice: How blogging, handwriting, woodworking, and metalworking help Allan make space for reflection.
Key Takeaways
• Stay curious about why things work—not just whether they work. Allan emphasizes understanding the underlying problem and building a mental model, rather than stopping at a working solution.
• Protect human attention. Cheap code generation does not make every project worth pursuing. Focus on work that advances the organization’s purpose.
• Delegate outcomes, with boundaries. Allan’s assistant experiment relies on defining what the agent can handle independently and when it should ask for help.
• Review where the consequences are greatest. His review experiment…
Site web: listen.hubhopper.com

