Just a few weeks ago I was a huge Codex fanboy. However I'm really considering switching now:
- ChatGPT's Android app quality is bad, especially the Remote Codex part
- Opus 5.5 is a better model than all current OpenAI's models
- Dots are still not available in EU
- Desktop app quality is degrading (example: the "Ask in side chat" feature which is super useful just stopped working for some reason)
@ajambrosino
@mtrostyle I didn't use it as much so far, but from what I see nowadays it's more or less the same feature set.
However Opus 5.5 is so much better and more fun to chat with ๐
@ajambrosino It would be great to be able to create new threads from an existing one in Codex cloud. I want to have an orchestrator thread that spawns and babysits other ones. That works great with local tasks, but not cloud ones.
@thsottiaux hey, I love Codex Cloud, I think it's underrated, but why is there no way for a thread to create other threads? I want to have an orchestrator thread that spawns and babysits other ones.
Tried from both web and desktop app
apparently the CLI doesn't let Codex select which model to use. Will let Codex desktop control Chrome to create the threads on Codex Cloud from the web.
We're adding a new plugin to Claude Code: You should Know.
It scans Claude's output for important information you might miss to help keep you in the loop.
Enable it with:
/plugin enable cc-plugin-you-should-know@builtin
Idea: use GPT Luna or even Jev to detect @thsottiauxโs Codex reset announcements and expose their status through a static API. Get notified or automatically kick off a big project with Astra to put your remaining quota to work before the reset.
Something great with Astra is that it is aware of the usage quota. So you can basically ask it to keep going with a long plan and tell it to stop when it sees that the reset has been applied. Engine started ๐ค
Something great with Astra is that it is aware of the usage quota. So you can basically ask it to keep going with a long plan and tell it to stop when it sees that the reset has been applied. Engine started ๐ค
The benefit is that you catch important flaws early on instead of at code-review time (or later). This results in a better overall solution and a lot less code-review back-and-forth.
I've been using @OpenAI's Codex a lot lately, and I find it amazing (both the model and desktop app).
At this point, I think everybody knows that creating a plan is essential when working on larger changes.
Here is what I've been using that makes it even better for me ๐
The reason this works so well is that subagents start from a clean context window. They aren't "biased" by the previous chat history, so they can look at the plan with completely fresh eyes.