Skip to content
How I AIHow I AI

The power user’s guide to Codex | Alexander Embiricos (product lead)

Alexander Embiricos, the product lead for Codex at OpenAI, shares practical workflows for getting the most out of this AI coding agent. In this episode, he demonstrates how both non-technical users and experienced engineers can leverage Codex to accelerate development, from making simple code changes to building production-ready applications. Alex walks through real examples of using Codex in VS Code and terminal environments, implementing parallel workflows with Git worktrees, and creating detailed implementation plans for complex projects. He also reveals how OpenAI uses Codex internally, including how they built the Sora Android app in just 28 days, and offers insights on automated code review and the future of AI-assisted development. *What you’ll learn:* 1. How to set up and use Codex in VS Code and terminal environments for both simple and complex coding tasks 2. A practical workflow for running multiple Codex instances in parallel using Git worktrees to avoid conflicts 3. How to create detailed implementation plans using the Plans.md technique for complex engineering projects 4. Why context is critical when prompting Codex—and how to provide the right information for better results 5. How OpenAI uses automated code review to accelerate development while maintaining high quality standards 6. The key differences between vibe coding for prototypes versus building production-ready applications with AI 7. How the new GPT-5.2 model improves Codex’s capabilities with faster reasoning and better problem-solving *Brought to you by:* Brex—The intelligent finance platform built for founders: https://brex.com/howiai Graphite—Your AI code review platform: https://graphitedev.link/howiai *Detailed workflow walkthroughs from this episode:* • 3 Advanced Codex Workflows for Faster, Smarter Development with OpenAI’s Alex Embiricos: https://www.chatprd.ai/how-i-ai/advanced-codex-workflows-with-openai-alex-embiricos • How to Use OpenAI Codex to Understand and Modify a New Codebase: https://www.chatprd.ai/how-i-ai/workflows/how-to-use-openai-codex-to-understand-and-modify-a-new-codebase • How to Architect Complex Software Projects with OpenAI’s Plans.md Technique: https://www.chatprd.ai/how-i-ai/workflows/how-to-architect-complex-software-projects-with-openai-s-plans-md-technique • How to Manage Parallel Development with AI using Git Worktrees and Codex: https://www.chatprd.ai/how-i-ai/workflows/how-to-manage-parallel-development-with-ai-using-git-worktrees-and-codex *In this episode, we cover:* (00:00) Introduction to Alex and Codex (02:06) Getting started with Codex (04:54) Using Codex for parallel tasks (07:34) Understanding Git worktrees (09:51) Terminal shortcuts and command-line efficiency (12:16) How OpenAI built the Sora Android app with Codex (15:37) Using PLANS.md for problem solving (20:23) Using Codex for prototyping (22:22) Deciding between what needs a plan and what doesn’t (26:42) How to multiply the impact of Codex (28:08) Implementing automated code review with GitHub (30:01) Codex adoption at OpenAI (32:08) Challenges and innovations in AI integration (36:38) Recap and the Codex harness (43:49) Atlas and personalized AI interactions (49:09) Conclusion and final thoughts *Tools referenced:* • Codex: https://openai.com/blog/openai-codex • VS Code: https://code.visualstudio.com/ • Cursor: https://cursor.com/ • Git: https://git-scm.com/ • GitHub: https://github.com/ • Atlas: https://openai.com/atlas • ChatGPT: https://chat.openai.com/ • Slack: https://slack.com/ • Linear: https://linear.app/ *Other references:* • Sora app: https://openai.com/blog/sora • GPT-5.2 model: https://openai.com/index/introducing-gpt-5-2/ • SWE-bench: https://openai.com/index/introducing-swe-bench-verified/ *Where to find Alexander Embiricos:* LinkedIn: https://www.linkedin.com/in/embirico X: https://x.com/embirico *Where to find Claire Vo:* ChatPRD: https://www.chatprd.ai/ Website: https://clairevo.com/ LinkedIn: https://www.linkedin.com/in/clairevo/ X: https://x.com/clairevo _Production and marketing by https://penname.co/._ _For inquiries about sponsoring the podcast, email jordan@penname.co._

Alexander EmbiricosguestClaire Vohost
Jan 11, 202653mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

Power user playbook for Codex: planning, worktrees, and reviews

  1. Codex is presented as a thorough coding agent that shines on complex tasks, used both for answering questions about unfamiliar codebases and for implementing changes via natural-language prompts.
  2. The episode demos a zero-to-one setup in VS Code and the terminal, then shows how to run multiple tasks in parallel safely using Git worktrees.
  3. For larger efforts, Embiricos advocates structured planning (via a shared Plans.md template) and iterating on the plan in the same chat before executing implementation.
  4. They highlight “harness” differentiation (product + workflow around the model), especially automated GitHub code review that catches issues with high confidence and can even fix them on request, plus broader integrations (web/cloud, Slack, Linear) and personalization via Atlas/Sidechat.

IDEAS WORTH REMEMBERING

5 ideas

Treat Codex as a teammate for both questions and code changes.

Embiricos emphasizes that a high-frequency use case is simply asking questions—how to run a repo, what a feature flag does, or whether something shipped—reducing friction and avoiding unnecessary pings to engineers.

Parallelize work, but manage conflicts intentionally.

Small independent tasks can run in parallel chats; when changes might collide, switch to serial execution or isolate work using Git worktrees so each change remains reviewable and non-conflicting.

Use Git worktrees as the “safe parallelism” primitive.

Worktrees let one repo track multiple working directories (e.g., “French” and “German” variants) so Codex can execute concurrent edits without stepping on the same files/branch state.

For hard problems, planning beats one-shot prompting.

The Sora Android app story underscores that agents don’t remove architecture work; teams moved fast by defining the shape of the system, then using detailed plans as the contract Codex executes against.

Adopt a reusable planning spec (Plans.md) to standardize outcomes.

Copying a “how to plan” template into Plans.md makes plans self-contained and milestone-driven; Codex’s thoroughness becomes an advantage when it’s guided into consistent structure.

WORDS WORTH SAVING

5 quotes

People love how thorough and diligent Codex is. It's not the fastest tool out there, but it is the most thorough and best at hard, complex tasks.

— Alexander Embiricos

We used Codex to build the Sora app for Android in twenty-eight days, and it immediately became the number one app in the App Store.

— Alexander Embiricos

With coding agents, it doesn't get easier, but you just move way faster.

— Alexander Embiricos

If you have a chat where Codex is producing these plans and you want to change something, it's actually really nice for the model if you just use the same chat to ask for changes to the plan.

— Alexander Embiricos

The bottlenecks are kind of like thinking about what code to write and then making sure that code is good, reviewing it and landing it.

— Alexander Embiricos

Installing and starting Codex (VS Code, CLI, web)Using Codex to understand/run existing reposParallel task execution and separation of concernsGit worktrees for concurrent, non-conflicting changesPlans.md planning workflow for complex buildsPrototyping vs production engineering (“vibe” vs “architect”)Automated code review in GitHub and review loopsHarness differentiation: UX, latency, compaction, model updatesAtlas personalization and Sidechat page-aware assistancePrompting tactics: context, ambiguity, and when to restart chats

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

Add to Chrome