Descriptions:
Greg Isenberg brings back power user Riley Brown for a comprehensive walkthrough of OpenAI’s Codex platform, exploring whether it can replace or complement existing developer stacks like Claude Code and Cursor. The episode covers Codex fundamentals—how it differs from ChatGPT, how to organize work into projects and threads, and the distinction between plugins (official OpenAI integrations requiring approval) and user-created skills stored as markdown instruction files.
Riley demonstrates several concrete workflows: a YouTube Researcher skill that pulls and summarizes transcripts from any channel, Notion connectivity for knowledge management, Remotion for programmatic video creation, and early browser-use automation that Riley argues is approaching human speed. The conversation also touches on GPT-5.5, which had just launched at time of recording, and how the Codex interface layers that model into an agentic desktop environment.
Isenberg starts as a declared skeptic—he has never downloaded Codex—which gives the episode an honest compare-and-contrast dynamic rather than pure advocacy. The hosts are candid about interface rough edges (the skills vs. plugins naming inconsistency) while making a strong case that consolidating coding, research, and automation work into one platform has real productivity advantages. Useful for anyone deciding whether to migrate from Cursor or Claude Code, or looking for practical on-ramp strategies for building personal automations with a ChatGPT subscription.
📺 Source: Greg Isenberg · Published April 27, 2026
🏷️ Format: Tutorial Demo
