.NET Rocks! The Role of AI in Software Development
Dec 18, 2025
Jeff Fritz, a Principal Program Manager at Microsoft and creator of live coding streams, joins Bill Wolf, a seasoned community organizer and enterprise architect, to discuss AI's transformative potential in software development. They dive into listener concerns about AI costs and dependency, exploring productivity trade-offs. The conversation highlights how AI is changing workflows, the importance of templates and prompts, and considerations for integrating AI into user-facing applications. They also predict a future where smaller, local models enhance reliability and efficiency.
AI Snips
Chapters
Transcript
Episode notes
Control AI Burn With Model Choice
- Monitor and control AI billing by choosing appropriate models and watching usage indicators in your editor.
- Use system prompts and model selection to limit costly, unnecessary actions by agents.
Productivity Can Outweigh AI Costs
- Teams can justify high AI spend when tools multiply developer productivity enough to shorten sprints dramatically.
- The real ROI depends on time saved versus subscription and token costs.
Lock Agents With Strong Prompts
- Define strict system prompts and personas to constrain agent behavior and avoid runaway actions.
- Reuse curated prompt templates (e.g., Copilot Awesome) to keep outputs predictable and cheaper.