SE Radio 711: Scott Hanselman on AI-Assisted Development Tools

2026-03-11 · Source: Software Engineering Radio - the podcast for professional software developers · Field: Technology & Digital — Software Development & Engineering, Artificial Intelligence & Machine Learning · Depth: Intermediate, extended

Summary

Scott Hanselman, VP of Developer Community at Microsoft, discusses AI-assisted development tools, framing them as an evolution from syntax highlighting and autocomplete. He highlights the "ambiguity loop" inherent in LLM programming, contrasting it with deterministic traditional coding. Hanselman emphasizes the need for developers to express clear intent and specificity when prompting models like GitHub Copilot CLI, Claude, or Gemini. He shares experiences, including modernizing the 20-year-old Windows Live Writer to .NET 10, stressing the importance of foundational knowledge and defining success for AI agents. The discussion also covers verifying generated code, sandboxing agents, managing context windows (e.g., 200,000 tokens), and the cost-effectiveness of these tools, noting that a "Ralph loop" can reproduce 384 bugs for minimal expense.

Key takeaway

For AI Engineers or Software Development Managers integrating AI tools, recognize that AI-assisted development demands high specificity and robust verification. Treat AI-generated code like any other external contribution, requiring thorough testing and review. Your deep understanding of fundamentals and clear intent are critical to steering these tools effectively, preventing architectural debt, and ensuring the quality and reliability of production systems. Do not "vibe code" into production; prioritize verifiable outcomes.

Key insights

Specificity and foundational knowledge are crucial for effectively steering AI-assisted development tools and managing inherent ambiguity.

Principles

• LLMs excel at resolving programming "toil" due to their ambiguity handling.
• The programmer's role is to express precise intent and define success for AI agents.
• Correctness in AI-generated code requires rigorous verification and testing.

Method

Employ "ambiguity loops" or "Ralph loops" with clear, verifiable stop conditions and comprehensive test harnesses to ensure AI-generated code meets quality standards.

In practice

• Provide AI agents with detailed architectural context and documentation.
• Implement extensive testing (e.g., 70% code coverage, 300 tests) for AI-generated code.
• Utilize source control (Git) to manage and revert any problematic AI-generated changes.

Topics

• AI-Assisted Development
• Agentic Loops
• Code Generation
• Software Testing
• Developer Tools
• LLM Interaction
• Software Modernization

Best for: Software Engineer, AI Engineer, Director of AI/ML

Related on AIssential

• AI tools for hearing difficulties — helpful or harmful for language learning? — AI assistive technology presents a dilemma for language learners with hearing difficulties, balancing aid with skill development.
• How Enterprise Marketing Teams Are Actually Using AI Agents in 2026 — Enterprise marketing is shifting from human-assisted AI tools to autonomous AI agents that redesign entire workflows.
• [AINews] Microsoft Build: MAI-Thinking-1 and MAI Family models — Microsoft's MAI models showcase a transparent, full-stack AI strategy emphasizing clean data and local agent execution.
• Are we getting better at explaining things because of our interaction with LLMs? — Crafting clear LLM prompts may enhance human explanatory and communication abilities.
• OpenAI expands Codex with role-specific plugins to build a general-purpose app for non-developers — OpenAI is transforming Codex into a general-purpose, role-specific AI work app, expanding beyond coding for non-developers.
• How to Use Claude Managed Agents? — Anthropic's Claude Managed Agents simplifies AI agent deployment by providing a hosted, stateful, and secure execution environment.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Software Engineering Radio - the podcast for professional software developers.