If i run kimi 2.6 in Workers AI on a complex task, I run into timeouts.
If i run it async (env.AI.run(model, input, { queueRequest: true })) I get: 8007: This model does not support request queueing
Is this the intended behavior @michellechen?
I never thought I am going to say this but I am going to start using Claude now that @thsottiaux and team made codex unusable with the 5 hour limit changes
Saying not knowing how to code gives you an advantage in building software (thanks to AI) is like saying not knowing anything about filmmaking gives you an advantage in making films (thanks to having a smartphone apps to edit stuff)
Ignore this stuff and keep learning building
Other day I was having issues installing open replay on GCE when using GCS and external Postgres. After hours, I almost gave up. Then I installed @OpenAICodexCli and asked it to do it. Ran for an hour, configuring, testing, installing, uninstalling till it got it perfect. FTW!!!
my new codex cli reality after 5.4
Error running remote compact task: stream disconnected before completion: error sending request for url (chatgpt.com/backend-api/code…)
@thsottiaux
.@OpenAIDevs so if i do not exit plan mode in Codex CLI when i am first prompted, i am stuck there for ever? codex cannot exit plan mode on my instructions and there is no slash command to exit? wtf???? @thsottiaux !!!
Understanding the agent stack might be the most valuable technical skill you can build right now.
Pi is the library that powers OpenClaw. We dive in layer by layer, progressively adding features until we have a custom agent & nothing about them feels like magic anymore.
The best thing about @OpenAIDevs incorrectly flagging my account and downgrading to 5.2 is that I got to discover kimi 2.5. Works great in openclaw and it’s free with the nvidia dev account. Sad to see @openai fumble at the 1 yard line
The @OpenAICodexCli is so much smarter than the codex app even though both are using the same 5.3 model. I am about to give up on the app. Also pro tip, having two cli point to the same repo, one as the orchestrator and another as the grunt is amazing.
Today was the hate phase of my love hate relationship with @OpenAIDevs codex. It introduced a subtle bug and then kept blaming cloudflare configuration and phantom excuses. It took me 8 hours, and finally Gemini, to figure it out.