Talk through work with ChatGPT Voice while it sees your screen, works across connected apps, and keeps tasks moving in the …
Talk through work with ChatGPT Voice while it sees your screen, works across connected apps, and keeps tasks moving in the …
Insane
(2026 – 2030) Phase 1: Let AI do the work for everyone in the workplace.
(2030 – 2035) Phase 2: Once the AI has learned everything, let it replace all humans.
(2035+) Phase 3: Everyone is unemployed, has no money, and Universal Basic Income (UBI) is created. Then use this system to control everyone.
I didn't say anything, but it felt like I was in the Matrix. Maybe I was just dreaming…
open ai is cooking hard
So basically codex in voice mode
love this! but worried about privacy & security. what information goes through the server vs stays local? data encrypted?
Thank you for implementing this feature! I requested it almost a month ago.
I can’t figure out if this is just not available on iPad and mobile or if this is a feature that it’s not here yet or just on desktop or what?
This looks genuinely impressive, exactly the kind of fluid, context-aware AI workflow a lot of us have been waiting for.
But it's frustrating that for the second year in a row, real screen-context support is still stuck on macOS. Windows users make up most of the desktop market, and we still don't have anything like Screen Context or Appshots.
If you're not on a MacBook, a big part of this "voice-first work" vision just isn't usable, and this gap is really killing the vibes
You guys should collaborate with Framework to make your own laptops.
it's a great idea, and executed really poorly. While it is FAIR at maintaining context, it is AWFUL at thinking deeply even on High. I cannot use it, sadly, due to its inability to wait around for 25 minutes for a long running project to finish or alert me that something in a subagent or child thread will require a long time and that it will alert me when the time comes. It's not responsive enough for me to use it across my workday, reducing the utility of my $200 plan's unlimited time limit on the model. It's really lazy, and lies as if it's running on 5.6-Luna-Light instead of 5.6-Sol-High which is what I would actually want it working on.
indeed
Memory save very buggy and very hard to remove with somehow glitch. I hope can fix it
what model/effort do you recommend for this workflow?
Works like this only for desktop app?
Why can’t I use the voices in GPTs can you guys fix that please it’s very jarring for a male or female assistant to be defaulted into shimmer not cool guys.
F’ing finally. Hope there’s an 11labs integration
When will this work on Windows computers? I don't want to switch to Apple products just to use this service.
Life changing feature especially with the Codex Micro ❤
CallGPT
Desktop fixes
1. Fix add to project for new gbt app. Recent chats are no longer have the function to add to a project
2. Add function that a previous / mention chat can enter any thread and not just a “ new chat “
3. Fix sharing projects. When sharing or receiving invite it states “ you don’t have access “ when you try to join
This gentleman needs speech coaching urgently. Great presentation; terrible diction
便利そうだ
These guys are cooking hard.
Amazing
chat GPT interrupts me in the middle of my question, so I can't even finish my question. Useless
Super annoying voice that pretends to be busy reading while talking
How about putting the circle to the right bottom of the screen instead of just in the middle?
Could It plan with Sol and implement with Luna for example ?
When is this coming to Windows?
The voice is irritating – bring back the old Ember
you keep teasing us w these stickers but i dont see them for sale!!!
It remains ineffective and dangerous. It excludes users only because of their personal identity and assumes suicide when this is not the case. Despite being told clearly that this is not the case, he continues to insist. Just for saying that I am emo he would label me as suicidal and put anti-suicide contacts on me. This is unacceptable. An AI should never do something like this just because of how a user identifies themselves. Stop ignoring this, because when the law falls on you, it will be your own fault for allowing discrimination
Thanks, it will help us, like small YT channels 👍.
*ATTENTION*
Voice models need dedicated hardware to reach their full potential. I would gladly pay $200–$300 for a small desktop device whose sole purpose is to run an always-on, low-latency voice model.
It would work somewhat like Alexa, with an optional wake word, but use natural, full-duplex conversation similar to modern live voice modes. You could interrupt it, speak over it, clarify yourself, and carry on an actual conversation rather than issuing rigid voice commands.
The real breakthrough would be treating it as an alternative interface for your computer, alongside the mouse and keyboard. It would not simply control the computer through voice commands. It would control the computer through conversation.
You could ask it to open programs, locate files, explain what is on the screen, help complete tasks, operate software, or guide you through complex workflows without constantly switching back to typing and clicking.
A dedicated conversational controller like this could fundamentally change how people interact with computers.
Multi billion dollar company any they can’t hire an actor who can read a script naturally?
Chatgpt bullied me😢
👁️
Arbor, wonderful voice!
Well can’t wait for OpenAI’s bankruptcy!
カメラ共有追加してください
Once chat gpt is gotten to where they want it where it can be used to invent things theycwill lo k it away and use it only for themselves this wont be free forever only while they meed is to train it.
guessing what ChatGPT would have in 2029:
sora access in plus, atlas version(chatGPT can interact with my desktop's programs,games & websites, 3d generation, game generation, Live feed(allows ChatGPT what you see & hear in live time), analyzation(ChatGPT can watch videos & listen to music)
Can I Request? Please Listen, But Me And My Friend Enjoys The Group Chat Feature, But I Accidentally Deleted It, Now I Tried To Create A Group Chat Again, It Isn't There Anymore, Please Add The Group Chat Feature Again. 😢😢😢😢😢
Hi OpenAI Team! 📻
I have a feature idea that I think would be really fun and make ChatGPT even more interactive.
Imagine a built-in AI Radio feature where users could create and listen to their own personalised radio stations inside ChatGPT. You could choose topics like technology, gaming, movies, sports, history, science, music, or even your own custom interests, and ChatGPT would generate a live-style radio experience with AI hosts discussing those topics.
Some ideas for the feature include:
🎙️ AI radio hosts with different personalities and voices.
📰 Daily news and trending topics explained in an easy-to-understand way.
🎮 Gaming news and reviews.
🎬 Movie and TV discussions.
📚 Educational segments for students.
🌎 World news and interesting facts.
🎵 Optional background music between segments (where appropriate).
🎤 The ability to ask ChatGPT questions and have the radio host answer them as part of the show.
A notification feature could let users know when a new personalised radio episode or daily briefing is ready to listen to.
There could also be a “My Radio” library where users save favourite stations, customise hosts, and organise different shows.
I believe a feature like this could help people learn, stay informed, and enjoy entertaining conversations in a completely new way. It could make ChatGPT feel like a personal radio station that’s tailored to every user.
Thank you for continuing to improve ChatGPT and for listening to community ideas. I hope this inspires a future feature! 🚀
— Jacob Curran
Pls i want list to update
Ok so how can I use that ???
What settings do you have your voice set too? The cadence is interesting.