Thoughtworks Technology Podcast How developers can get the most from new AI coding workflows
153 snips
Nov 13, 2025 In this engaging discussion, Brandon Cook, a Principal Technologist at ThoughtWorks, delves into how generative AI is revolutionizing software development. He explores the nuances of coding agents and highlights the importance of sensible defaults and TDD practices in guiding these tools. Brandon emphasizes the balance between adopting new AI technologies and maintaining team efficiency. He introduces the concept of an autonomy slider to manage trust in these agents, and stresses the necessity of solid engineering foundations before leveraging AI to enhance workflows.
AI Snips
Chapters
Transcript
Episode notes
Give Agents Clear Guardrails
- Provide agents explicit project context, standards, and expertise via configuration or agent README files.
- Embed your team's guardrails so agents follow your engineering practices automatically.
Tool Hopping Led To A Better Prototype
- Ken describes using an AI tool to prototype and getting closer to code than in years.
- He switched tools after a colleague suggested a better option that quickly felt superior.
Treat Agent Tools Like Architectural Choices
- Agent tooling evolves rapidly; teams must choose when to adopt new mechanisms for context and memory.
- Apply familiar engineering judgment to pick patterns that fit your team's risk and workflow.