local

@thsottiaux codex app on linux!

With GPT-5.4 out. What should Codex ship or improve next?

@pashmerepat surprised as well

i am using codex way too much

@banteg yeah i have a similar suspicion

codex has completely lost its mind after i expanded agents.md with detailed best practices. here you can see the two versions of a fix it did, just horrible. im doing an experiment where i completely delete agents.md. codex has a very special relationship with that file and treats it too seriously. will report if it helps.

@ajambrosino linux people bootstrapped aur.archlinux.org/packages/codex…

fine I’ll be more specific. we need to land 3 more PRs and then we go to the group of 10k

@arrakis_ai @thsottiaux

Codex is incredible. But there’s one thing that’s starting to feel like friction. The model choice. For small, almost trivial tasks, using xhigh feels like overkill. It’s powerful, but it’s also expensive and unnecessary for quick edits or lightweight work. At the same time, every time I start something new, I catch myself pausing for a second, wondering which model is actually the right fit. That hesitation adds up. It’s subtle, but it’s starting to feel like the new bottleneck. Not capability. Not speed. Routing. If Codex had a true Auto mode, something like a route.md behind the scenes that understands the intent of the task and automatically sends it to the right model, it would remove that mental overhead entirely. No second guessing. No resource waste. No constant micro decisions. We shouldn’t be managing models. We should just be building. If OpenAI solves routing well, that might be the real productivity breakthrough.

@bskdany works as worded

@dkundel why no linux acknowledgement 🥀🥀 x.com/0xlocal/status…

For those of you waiting for the Codex app on Windows: it's coming! Sign up to get notified: openai.com/form/codex-app/

@ajambrosino @kanavtwt what about linux? we can it run fine i mean but official support would be nice x.com/0xlocal/status…

@kanavtwt x.com/ajambrosino/st…

Codex is macOS exclusive??? Windows users really have no respect anymore

@JamesZmSun Is this ever coming to Linux?

Another week of shipping at Atlas! Excited to share these with you: - Save prompts that you use frequently and access them anywhere with @ and in the bookmarks bar - Rename your tabs - Device tab in Dev Tools so you can see how websites respond on different devices - Smarter transition between agent mode/chat so you are not stuck in the wrong mode - Bug fixes across autofill, extensions, side chat suggestions, and picture-in-picture Just hit the update button on the top right to get the latest goodies. chatgpt.com/atlas

it works lol, just asked codex to run it codex app on arch, everything seems to be working as normal, haven't seen any bugs yet

@embirico linux?

No specific date to share, but I will say that we built it in Electron, sharing code the with VSCode extension, specifically so we can get to Windows quickly. (Initially we were even hoping to get out at the same time.)

@ah20im does no one uses linux at openai

The team really cooked on this one!

@pvncher i always use "implement"

Just experienced one of the quirks of codex - "Build the spec" turned into spending 20 minutes fleshing out my spec markdown vs implementing the code...

@adamhfry linux lil bro

Atlas weekly release is here, last one before the holidays: - Multi-profile: Multiple browser profiles ('personal', 'school', 'work') for the same ChatGPT login; multiple ChatGPT logins coming soon - Selectively import things like extensions or passwords, even after onboarding - Hover to see vertical tabs when they are closed - More DevTools fixes (no devices tab yet!) - Lots of bug fixes and polish (especially for empty responses, tabs, or high memory use) We're working on tab groups, and they're feeling great, but will be out shortly after the holiday. Just hit "update" in the top right. chatgpt.com/atlas

@GregorySchier @itsnoahd Is there a way to prettify json on the request side? I have seen it on the response body, was curious if it's on request body as well.

@itsnoahd Would be happy to have your feedback. And yes, GRPC has support for manually specifying proto files, and also server reflection

I've done the impossible. I can retire now

hot air rising in my chill house

@mert @haste888

man I'm starting to think NFTs aren't coming back

@stonkrates where is this

Dying of laughter as I set up this Hype/uFART LP to farm HyperEVM Fartcoin + Hype, what more could you want?

@mert @haste888 coded

I've been quite surprised that NFTs still have not made a comeback seems like a relatively less-rigged and more fun game that serve roughly the same niche

@0xFrenxbtdotxrp @inversebrah owbserwve

Stone cold.