Building With Agents Today 🤖 — with Charlie Guo
Summary
Luca interviews Charlie Gruo, a developer experience engineer at OpenAI, about the evolution of AI coding tools and team dynamics. They discuss how Codex shifts focus from traditional IDEs to agent orchestration, comparing current AI integration to early mobile web adaptations. The conversation covers strategies for sharing AI skills within teams, the rise of generalist roles, and the importance of deterministic guardrails like linters to ensure code quality.
Topics discussed
Introduction: Codex philosophy and guest Charlie Gruo
The flat structure of the Codex team at OpenAI
Enforcing workflows and values in AI-assisted coding
From CLI tools to native apps and AI as the OS
Native AI patterns vs. adapting the web for agents
Context management, MCPs, and skill sharing challenges
Turning individual AI workflows into team assets
The rise of generalists and cross-functional support
Harness engineering: Defining boundaries for AI quality
Multi-agent roles and organizational structures
Using AI agents for Slack context and documentation
Design workflows with new image models and Codex
Model selection, pricing controversies, and cost efficiency
Listen ad-free on Castria