@ajambrosino@thsottiaux Ability for one codex app on my local machine to control/connect to the codex app/cli on my other remote server (like that other thing people used to use, Claudea)
@altryne Failed for me as well. Do we know what it’s supposed to do?
I’d love to be able to control my codex app running on my Mac mini Openclaw setup 🤞
@hgruenhagen@openclaw Use **other**.
Our current memory stack is QMD local, not OpenAI or Gemini embeddings: `embeddinggemma-300M` for embeddings, plus `Qwen3-Reranker` and `qmd-query-expansion-1.7B`. OpenClaw’s default memory index is disabled, so no `text-embedding-3-*`
@steipete That codex harness transition was rough! Finally got it mostly working after cron tool use issues, agent spawn issues, and LCM query timeouts.
Brighter days ahead! 🤘🤘 Keep at it @steipete
@altryne It’s a good start! 👌
I wish it was a dedicated codex app, as well as the ability to use the codex desktop app to connect to a remote codex app on another machine (like Claude does)
@thursdai_pod@altryne@Yampeleg@ryancarson@nisten Their choose your own model isn’t like what we’re hoping. It’s that dumb add-on that ChatGPT does now: Siri: “I’m sorry, I don’t know how to do that, want me to ask…” insert your other LLM in at that point.
@altryne@thsottiaux@steipete In fairness, 5.5 is a better communicator and has more personality not to mention smarterer… but it’s the OC releases of late that took a major whiff. 4.23 is the last solid, speedy, recent drop 🤙