TimBased

@jxnlco take a look here if you can. x.com/TimBased/statu…

ChatGPT voice is for when you’ve got your hands full

@LexnLin DUDE that was fun... we got tryhards in there allready lmao

Test it here: claude-of-duty-six.vercel.app Here is the repo: github.com/Leonxlnx/claud… And the prompt (yes pretty long) is here: github.com/Leonxlnx/claud…

I tried to reverse engineer a prompt for this game. Used that in Cursor, with NO Ultracode/subagents only Opus 5 high and this is the result! The repo and the prompt are below

Codex Voice experience is seriously broken right now. Across multiple sessions, Voice repeatedly created a projectless task instead of staying inside the selected Codex project. Worse, the interface made claims that did not match reality: • It said background work was running on Sol/max, while the logs showed Terra/low. • It claimed it could see and control a live presentation, but was only sending text prompts to a separate worker thread with no shared screen state. • Feedback was submitted without my approval. • The German voice sounds horrible. • There are no proper controls for voice, dialect, persona, model, or reasoning level. The potential is massive and here and there i got a real feeling of damn this is the future. But sometimes it disapoints me. @guinnesschen @jxnlco @ajambrosino @thsottiaux

@MatthewBerman Anthropic’s lobbying efforts against it would be nice to be covered in that video and why you think that their doing it.

I'm putting together the most comprehensive video about open source AI - what topics do you want me to cover?

@zodyysol @jxnlco Opus 5 is awesome ngl. Id also say slightly better than 5.6 For most business workflows & automations gpt is more reliable and the better eco but for fun things i allways choose Fable or Opus

@TimBased @jxnlco tried opus 5 yet? seems to par if not slightly better than 5.6 sol to me

Gpt voice today inside codex feels max retarded and cant do the easiest tasks. And doesnt reason at all. We need a voice config or settings page @jxnlco

Insane that the one lab that seems to be the most concerned about ai safety and concentration of Power isnt on that list @AnthropicAI @DarioAmodei Would love to hear the reasoning behind it

@argofowl same dude, feels awesome

i let opus 5 touch my private parts

@DanDr1s I was a doubter at first but Dan seems legit ngl.

🚨 Claude Opus 5 has been released Nobody believed me, but here it is😊

Almost cried reading that

For my first post, I’m sharing a letter @NVIDIA signed on why open models matter. AI will transform every industry, power every company, and be built by every country. Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty. The world needs both frontier closed models and frontier open models. images.nvidia.com/pdf/Open-Weigh…

On a call with codex for 4 hours allready now… Its growing on me. This is the new way of interacting with technology. Shity voices and missing voice setting page needs to be improved but ehh🤌

@camsoft2000 @OpenAI voice mode via remote?? why cant i find this on my phone?

Thank you @OpenAI - voice mode via remote Codex session to Mac is incredible. This is exactly what I have been dreaming of. Talking is so much more natural than typing. I can have a conversation more naturally and more verbosely than typing responses.

@Dimillian Its such an awesome experience but also a lot of improvements that can be done so its even more bullish

CODEX IS ON THE PHONE with me

So much to improve and thats a great thing. shouldnt be too hard to implement but really the voices is the worst thing that makes it so hard to use voice mode... they sound soo ai xD

OpenAI’s new Voice Mode has huge potential, but the current control layer basically doesnt exist. There’s no clear Voice settings page where I can configure how the assistant should work, set persistent operating instructions, choose a communication style, or tune how assertive vs. cautious it should be. The voice quality also drops noticeably outside English. For German, it often sounds unnatural. would be so awesome if we could take elevenlab voices or other open source providers Most frustrating: it can become overly defensive and require repeated instructions before actually acting. Power users need a proper settings layer for behavior, autonomy, tone, language, and voice. The model is capable; the interface needs to catch up. @guinnesschen @thsottiaux @ajambrosino @OpenAIDevs @jxnlco

@jxnlco @jlave_dev Got the update but also no voice mode at all- feature flag means we just have to wait?

@jlave_dev must be a feature flag

anyone tried claude voice and codex voice and tell me how they compare? I'm on my work laptop and i'm banned from using it

@argofowl I dont know where your from but even worse if your from europe and mess up your sleep because you wanna try that.

still don't have the voice mode in codex... frustrating

@argofowl Same hahhaha

still don't have it, even after the update 😔

chatgpt voice in codex, here i come 👀

@jxnlco Yes my voice mode hasnt rolled out yet on codex