OpenAI Codex Masterclass
summary
Claim: Codex becomes a full software engineering system, not just a coding helper, using Codex app and CLI, plugins, automations and MCP servers. Demos show Imagen and Playwright generating game sprites and running interactive tests, a Google Drive plugin updating 57 event rows in two minutes, and subagents parallelize tasks (20 reviewers spawned, collated results). Guardian approvals, hooks, sandbox modes and 3 million weekly active users are cited as production‑level features.
AI-written summary · about
chapters (19)
- 0:00Start
- 0:14Introduction to the workshop and speakers
- 1:48What is Codex? (Overview of software engineering agent capabilities)
- 2:18Foundation models (GPT-5.3, Spark, GPT-5.4)
- 4:11The evolution of models and performance improvements (Websockets, Fast Mode)
- 7:04Codex app features, projects, and work trees
- 8:37Automations overview
- 12:28Plugins: Skills, apps, and MCP servers
- 15:13Game and Web development plugins (Playwright and Image Gen)
- 16:26Demo: Game studio plugin and generating assets
- 17:23Demo: Google Drive plugin for codebase data
- 18:52Demo: Setting up automations for Slack and Gmail
- 27:14Code Review features and integration with GitHub
- 32:39Subagents: Parallelizing tasks and custom personas
- 36:18Demo: Using subagents to review persona files
- 44:52Creating custom subagents
- 49:29Bleeding edge features: Guardian approvals, hooks, and personality settings
- 56:11Codex security and Cloud Code plugin
- 57:10Q&A session
related talks
47mHow OpenAI Uses Codex to Change How We Build
48mHow to Verify AI-Generated Code with Codex
36mThe evolution of Software engineering
13mShip Faster with Codex
24mCoding for the Future Panel
30mBuild Apps in Minutes 🤯 No Coding Required | Codex Skills Explained
01h 28mAI Coding Agents and how to code them
35mBeyond the Hype and Hyperbole