embirico's profile picture. Codex @OpenAI

Alexander Embiricos

@embirico

Codex @OpenAI

Pinned

Codex is finally shaped like a teammate: 1. Start pairing with it on your computer (CLI; IDE) 2. Begin delegating longer tasks (cloud) 3. Plug it in to team events so it works proactively (code review) Baby steps to the future 🚀

We’re releasing new Codex features to make it a more effective coding collaborator: - A new IDE extension - Easily move tasks between the cloud and your local environment - Code reviews in GitHub - Revamped Codex CLI Powered by GPT-5 and available through your ChatGPT plan.



Alexander Embiricos reposted

Bloody hell, I'll say this GPT 5.2 Codex Extra High is a methodical beast It's updating the OpenCode OpenAI Codex OAuth plugin Literally not leaving any stone unturned This is the first model that feels like it's building for itself ie leaving the door open for future work


You can also add skills to projects! Codex will pick up .codex folders nested in a git repo. See docs: developers.openai.com/codex/skills

Please add an option to manage skills and prompts on the project level (local) and not only in global config



Yes, Codex rate limits just got reset. Enjoy! Why: We're rolling a new way of counting usage to fix bugs where users previously saw inconsistent amounts of usage left. As part of this, usage is counts are resetting to zero.

embirico's tweet image. Yes, Codex rate limits just got reset. Enjoy!

Why: We're rolling a new way of counting usage to fix bugs where users previously saw inconsistent amounts of usage left. As part of this, usage is counts are resetting to zero.

Excited to bring skills to Codex! They unlock a ton of use cases, speed, & reliability for agents. Skills in Codex follow the agentskills.io standard (H/T @AnthropicAI for the great work), but we've extended them to support more user control with explicit invocation.

🆕 Codex now officially supports skills Skills are reusable bundles of instructions, scripts, and resources that help Codex complete specific tasks. You can call a skill directly with $.skill-name, or let Codex choose the right one based on your prompt.



Hiring on all fronts. Research, eng, sales, product and more. Do apply on the job site and reach out here!

Codex is getting extremely good and will rapidly improve. If you want to help make it 100x better next year, the team is hiring. Crazy adventure guaranteed; success probable.



Getting to API is priority for us. But it's going to take a little longer than the prior Codex models. Due to the higher cybersecurity capabilities, we're taking this very carefully and iteratively.

I hope it won't take too long to release it to the API.



Alexander Embiricos reposted

We're piloting trusted access for cyber capabilities in upcoming and more permissive models. It's invite-only for now but if you're interested, please let us know: forms.gle/RyJNzv5wEj2gZW…

GPT-5.2-Codex is more cyber-capable than GPT-5.1-Codex-Max, and we expect future models to continue on this trajectory. This helps strengthen cybersecurity at scale by giving defenders more powerful tools, but also raises new dual-use risks that require careful deployment.

OpenAIDevs's tweet image. GPT-5.2-Codex is more cyber-capable than GPT-5.1-Codex-Max, and we expect future models to continue on this trajectory.

This helps strengthen cybersecurity at scale by giving defenders more powerful tools, but also raises new dual-use risks that require careful deployment.


🤔

BREAKING 🚨: OpenAI is preparing to upgrade Codex, likely with GPT-5.2-Codex-Max. One last leap this year? Caribou 👀

testingcatalog's tweet image. BREAKING 🚨: OpenAI is preparing to upgrade Codex, likely with GPT-5.2-Codex-Max. One last leap this year? 

Caribou 👀
testingcatalog's tweet image. BREAKING 🚨: OpenAI is preparing to upgrade Codex, likely with GPT-5.2-Codex-Max. One last leap this year? 

Caribou 👀


Belated update: We're fully available in Europe now 🇪🇺

Codex in Linear is out to Plus users again (as well as Pro, Business, Enterprise and Edu). Still not available to Plus & Pro in Europe though. Working on that.



FYI We're seeing elevated error rates for Codex. Working on a fix. Updates here: status.openai.com/incidents/01KC…


Still just an opinion of mine, but here's a major learning over the past 1.5 years as we worked on various flavors of agentic assistants on desktop, with multimodal computer use, and now with Codex

OpenAI Codex Product Lead @embirico: "We are getting to the point where if you want to build any AI agent, maybe you should be building a coding agent. If we're gonna build a super assistant, it has to be able to do things. For models to do stuff, they're much more effective…



How we used Codex to build Sora for Android in 28 days Check out the latest @openai eng blog post by @hum_patrick & @RJMarsan. Lots of concrete advice from how they achieved insane speed to build a production app, from scratch, to #1 in the Play Store. openai.com/index/shipping…


Loading...

Something went wrong.


Something went wrong.