William Leandro

probably human

[William Leandro @c00rvus]

@ajambrosino where is the reset button?

[William Leandro @c00rvus]

The only problem I have with the Codex app is that when opening a large thread, the thread title changes to "new task" and the chat area goes completely blank until the data loads. This lasts about 15 seconds. @thsottiaux @ajambrosino

Windows, btw.

[William Leandro @c00rvus]

@razdriz Run this in PowerShell and open the codex in the ChatGPT app. It's still buggy, but some things work:
& "$env:LOCALAPPDATA\OpenAI\Codex\bin\codex.exe" app-server --enable remote_control --listen ws://127.0.0.1:4503

[William Leandro @c00rvus]

I just asked Codex to connect the Codex Mobile app on Windows.

[William Leandro @c00rvus]

@lucas_montano Uma idéia legal para o perssua. Algo que os androids(recentes) tem com o gemini. Tem algo na tela em inglês, eu aperto um atalho e tudo na tela(um uma area que eu escolher) é traduzido para o idioma que eu escolhi. Os textos traduzidos ficam na posição do texto original.

[William Leandro @c00rvus]

@ajambrosino @thsottiaux @Matthew_AMS @OpenAIDevs same problem here

[William Leandro @c00rvus]

It's impressive how AI tools like Codex have been helping me create things. I've already built several apps for Android, iOS, Windows, macOS, and Linux, even without deep or professional coding experience. +

[William Leandro @c00rvus]

now we can export a session and send it to a developer, who can import it and inspect exactly where the problem is. It worked really well, and today we use it in our daily workflow as part of our QA and debugging process.

[William Leandro @c00rvus]

@ajambrosino

[William Leandro @c00rvus]

@thsottiaux The thing is If token speed goes up, rate limits need to go up too. Faster output means faster token burn, and since tasks finish sooner, we spend less time waiting and more time actively prompting. In practice, that just makes the cap feel like it runs out even faster.

[William Leandro @c00rvus]

@ajambrosino

[William Leandro @c00rvus]

@emanueledpt @Sayyidalijufri

[William Leandro @c00rvus]

I decided to create an app to control Codex through Android, just like @sigkitten and @emanueledpt, just for fun.
But maintaining a stable connection without any issues while Codex is running has been a pain. Still, I’m optimistic.

[William Leandro @c00rvus]

@SIGKITTEN It’s like a kind of cat Tamagotchi inside the Litter app?

[William Leandro @c00rvus]

@joaosenzi trabalho em uma empresa de IA focada em diagnósticos. Faça o teste de puder: https://t.co/zCAjWCCxoD 🙂

[William Leandro @c00rvus]

@NikTek

[William Leandro @c00rvus]

@ajambrosino well...
https://t.co/ktpNWSkOFo

[William Leandro @c00rvus]

hear me out. Make “Tibo” one of the possible names for subagents in Codex, with a 0.1% chance of that name appearing. And whenever it does, the rate limit gets reset. @thsottiaux