OpenAI Codex Review

Summary

The interface I want and the performance I'll have to wait for... Table of contents Codex: How it works Codex is currently a chat-first experience. You gain access by being invited or by paying for the Pro ($200/per month) subscription. Once you've got access, you start by enabling multi-factor authentication, which is required to use Codex. Next, you authorize the Codex GitHub app for each organization you want it to work with. Codex then clones your repositories into its own sandboxes so it can run commands and create branches on your behalf. If you maintain dozens of public and private repositories, this setup is fantastic because you can jump between projects and queue up tasks for each of them without leaving the interface. If you only keep a single repo or two, the overhead may feel heavier than just asking an LLM for help or working in an AI-powered editor like Cursor. Things I like about Codex Consciousness and desire are multi-threaded Codex feels like it was designed for me. This GitHub connection allows you to specify which repository and which branch your current instructions are for, because the primary chat interface is contemplated as a place for you to rapid-fire a day's worth of tasks into the interface to spin up multiple tasks in parallel. I took a swing through the Codex best practices guide, which encourages you to spin up as many tasks as you need. The current rate limits support you doing this. This is one of the things I like most about Codex and that I'm most excited for as the platform improves, because this gels with the way I work. By the time I start work, I tend to have a laundry list of items I want to complete, so initiating a ton of them in parallel via natural language feels like a reasonable interface. I think this will eventually support my dream untethered workflow As I wrote about in Walking and talking with AI in the woods, ideally I'd like to start my morning in an office, launch a bunch of tasks, get some planning out of the ...