I want to say something before, today I set up an Alexa device, and soon I noticed an inconsistency with what they promised, they promised automations and a friend, but they only delivered automations, chatgpt for many and a friend, besides a listener, I mean … if openai develops a device like alexa, or others on the market, it would be the leader, and the preferred one for generations, due to the fact that the answers are processed differently and after the question
People use chatgpt on a website, but why not embed a chatgpt system on your smartphones and doth devices? it would be the apex of evolution, gpt chat could reduce the number of words for each answer on these devices, making them more practical, efficient and better, but they would need to develop automation, such as opening apps and making alarms…
but this is the simplest step, I have programmer knowledge, and I believe that the price of a chatgpt device would be extremely cheap for what it is capable of doing, a good and clear suggestion for you! think about developing a smart doth device, like alexa, with chatgpt, I would be the first to buy!!
Thank you for your attention, I am attentive after this opinion of mine told you! 
such a system would be easy to improve over time, in addition to the idea of an AI, it literally helps us on a daily basis, it is the best of ALL!
Been done.
It is underwhelming because people don’t “speak” in AI instructions, and the large language model AI takes considerable time to answer.
Also, the leader in what, losing $10 billion a year even when there is a direct connection to shopping and services?
sorry, I wanted to say that I would love to have a chatgpt available to me on a daily basis, without the complexity of a website or computer, there could be a way to make it closer, like in an app to answer by voice, or just the app would be a big breakthrough. even if it is paid, the agility in everyone’s day to day would be tremendous, I even talked about an alexa model with ai, that would be good. I was not specific in technical data, I’m sorry
Are you interested in two different options for interacting with ChatGPT using voice?
Option 1: Type your input, and ChatGPT responds using voice. This option is relatively straightforward to implement, and you can find numerous projects on GitHub that allow you to clone and run them. The simplest implementation involves using Chat API for the response and the browser’s Speech API for generating voice, and you only pay for the Chat API usage.
Option 2: Use voice as input, and ChatGPT will respond using voice as well. This option is more complex. To achieve this, you will need Whisper API for Speech to Text conversion, then use Chat API for generating the response and the browser’s Speech API for generating voice output. Response might not be real-time and there could be some delays. Like option 1, you can find projects on GitHub for this setup, but keep in mind that you will be billed for both Whisper and Chat API usage.
OpenAI does not offer a dedicated Text-to-Speech (TTS) API that is why these samples I gave uses the browser’s Speech API although some might use other TTS solutions.
You can easily run these projects in your computer. If you want to use your smartphone/tablet devices within your local home/office setup, assuming you have a router, just run the projects in a HTTPS server.
Thanks my brother, Have you seen the new chatgpt app? that’s what I’m doing, facilitations, such as an intelligence response and input by voice, like siri, implemented directly on the cell phone or an app that does that, where you say it… hey chat, and he answers, you ask go he answers, you ask him to mark something on your agenda he marks it would be the future ksks
I have not seen the new chatgpt app. As for this, like siri, implemented directly on the cell phone or an app that does that, where you say it… hey chat, and he answers, you ask go he answers, this is doable. You can use function calling to check whether the user said hey chat and it will start conversing if it is not already doing it.
There are apps built on top of chatGPT (or the underlying models) and speech recognition/transcription models, such as Whisper.
If the main concern of your request is “I’d like to say ‘hey chat’ to start chatting, rather than having to press an icon on my phone home screen,” then that sounds very much like a tertiary kind of concern compared to the main value proposition of chat interactions with software. It’s not like the device gatekeepers (Google, Apple) will give up their users rapt eyeballs (earballs?) for some random third party to take center stage.
From an apt package manager $ sudo apt install mycroft-core then add chatgpt as a fallback … put it on a raspberry pi shove it in whatever case you want and you have your device
It looks like rabbit r1 had now the potential to fill this slot!