You can always take note of the time between people sending messages and then like if they sent one message and then another message like two seconds later it could couple the time that sPENT with what’s actually sent To guage the end users emotional response or the way the end user is feeling and maybe provide better output as a result of that because you know if you click something like immediately after and your text sounds excited you know maybe you know it could even react to you in a different way not certain if this is a good idea but. Something like ohh this person means it or something like that something successfully done but maybe it could detect urgency in this manner
1 Like
I don’t know if everybody knows this that you know your system automatically fixes voice typing errors and will understand you but in a nutshell Imagine there’s somebody out there not using voice typing but you know fixing every single error in their voice typed output before they enter a response Then those people are kinda being held back as far as what they could do you know I would tell the end user if it was you know my open AI library to use voice typing and speak as clearly as you can but to basically say things twice like say what you want and then reiterate it just in case there are voice typing errors and if you do that that’s all you have to do you don’t even have to look at what you voice typed as long as you know you type clearly and your mic is good … You know tell him to look at the output now and then but I would put a little voice typing tip section
If you read on here a lot I mention about issues logging in with Google and I was just gonna say real quick that the most logical thing to do for a programmer making a world class program like open AI library is if there’s something stopping you from making sure that no matter what it always asks you what Google account you want to log in with Then you should send a feature request to Google and say Hey I’m A I’m a big dog here you know I’m I’m a big company can you work with me on making it so I can make my software do what I need it to do and then it would probably just spill over to other companies too that are having that issue and you would have helped like a bunch of people…note if you read other parts in here by doing a search then you’ll see I’ve had times when I couldn’t log into your open AI on Google Chrome and I’ve had to use Firefox
Subject: Feature Request: Downloadable API Documentation
Message:
Dear OpenAI Team,
I am writing to suggest the addition of a downloadable version of the OpenAI API documentation. Having access to the API documentation in a downloadable format would greatly enhance usability, especially in environments with limited internet connectivity.
Thank you for considering this request.
Best regards!!
Now on chat GPT if you paste in data that you want to send to it and then turn off your Internet connection that button to send it grays out when you turn on your Internet connection again then the button doesn’t go back to normal I would put a listener on that that would just in a loop keep checking the Internet connection every couple seconds and then turn that button back to clickable once you connect to the Internet… As a developer I’m constantly turning off My Internet to test things in my software…so thisd help
If interoperation with Google Chrome or Firefox or whatever would be stopping you from doing this I would contact the chrome developers i’ve already approached them about this kind of thing about working with chat GPT better and they’re open to it in fact probably the whole world is so there’s any hindrance working with other apps you could just send them feature requests and they’ll they’ll work with you
1 Like
As your website is one of the biggest things ever to happen to humanity as far as technology and your models last training date was October 2023 I would seek out government assistance for funding to train your models so they’re always current.
Your customers are constantly getting wrong answers when they’re having code generated because you don’t set it to automatically look up the Internet when code is generated to you know verify it against the latest api
Here’s three important ones in one shot basically although kind of annoying in a normal windows desktop on Google Chrome your I prefer this response thing totally kind of sucks on chat GPT app on Android because you can’t even see the entire contents of the screen and you know when you’re trying to have a good workflow in any regard it’s totally annoying to have that come up
On UBUNTU it’s looking pretty good over there but you don’t have any voice typing solution I would recommend because this is probably the reason because there’s some sort of issue with the operating system or whatever I would recommend just contacting the makers of Ubuntu and working out a solution that would make instant voice access on Firefox a reality
Cause it’s kind of a serious drag I had to use my phone to send the data that I voice typed and the terminal output I had to paste in from the operating system I had to keep going back and forth…
Other than that and my other comments it’s still awesome world changing technology that I use every day
If anybody reads this send a message on mine and basically everybody’s behalf to not be pigeonholed in the in my own special post if you read my messages I mean these are genius things that open AI needs to know now like this one
I would make a regular post but I would get relegated here so I can’t say you know to the whole world that the voice system even though it’s a beautiful woman’s voice and kind of cool is incredibly flawed there should be a button like Google has to initiate voice typing in the main text window so you can work with it number one the way it is now you gotta go to the separate window you can’t see your old messages and it won’t look up the Internet and then when you’re on a different operating system like Ubuntu you don’t have like a microphone button like Google would have on their search bars I mean Google tested this and that’s what everybody likes so to deviate from that is an incredibly wrong move when you’re on a buntu you know you’re gonna find it difficult to implement voice typing I tried for an hour and there’s not really any good system so you should make it accessible to everybody because it’s a website so I’m hoping people see this
and If you read this one just send openai message and say Hey The moderators are making this guy stay in in this chat thing right here this this section and not make regular posts so everybody isn’t instantly seeing these things and could weigh in and these are important things I don’t really want to get my account banned but I could probably make a better business bureau report because it’s not really a smart move by the company and that would get attention., As I am of paying customer that pays for the A P I as well…
It’s just that developers deciding this or probably thinking hey this works good the developers that decide the boys saying but they’re they’re used to their own systems they have a nice setup where whatever they do works but then when it gets on our computers there’s all these errors and I can’t believe like nobody noticed how difficult it is to use that live voice typing.
To send text to the assistant instead of what you have the live mode that’s not really efficient because it requires extra button clicks to get back and forth in live mode won’t even look up the Internet so what I would do is just send text via the normal way exactly like Google does you have a little icon nicely colored and then you know just like on Google I guess the listening icon after you press it gets big and you know looking the exact same way just that you can think of it as an homage but that would add beauty to your site and you’d have a very nice responsive way to enter text and immediately whenever you need to
See I don’t want it to …when it outputs code for me add the import statements to the top every time you know I just gotta delete those like most of the time because it’s already set up in my code with the import statements after the first time it gave me the code usually but that’s not always the case though so if you made a series of check boxes on the right side of the screen so it looks nice just a series of checkboxes like a sidebar that opened up for items like that where I could turn on and import statements every time I turn it off so I don’t got to tell GPT that every time and then you could add other items in there too like you know show a differential you know show exactly what was added or the changes to the code like WINMERGE would show so you’re immediately aware of the changes that were added so it doesn’t seem cryptic when G P T just gives you an operator like what did it do then you test it and then did it work and no it didn’t work and it’s this big operator I don’t have the time to read all of it like you know if you just see a differential you’re immediately aware… What if 10 or 20 of these check boxes covering just a selection was all that was necessary to have a perfect code generation environment and make the sidebar I would say toggle so if you’re not a programmer you could just not toggle it on in the first place…
OK now you have to tell people that you did this one but this this is pretty genius OK you know how you could say save this into memory and chat GPT will just save it well if you get an idea for something like a feature request like all the ones I’m making here and all the tips that I’m sending here you should make it so you have the option when you’re talking to GPT like if you say you know send open AIA feature request or send open AI an issue it’ll just send it and then you could process it in the background and you know save it all to a database and then have your awesome AI functionality analyze it and provide it to you in a concise analyzed form
And another feature request I just want to make sure you get these even though I say say these before is it’s a very unwieldy to have to click that button to talk to the assistant and be taken away from the main interface where you’re typing in stuff to GPT I just don’t think it’s a very good design I think Google with how they have the microphone to the left of their search box on Google.com is a way more streamlined way to enter in voice input I think the main thing is you know Android phones have this live mode and we’re trying to replicate it and it’s awesome but for people who program it’s a major hindrance and it’s unusable.
the open AI developers though I treasure them so much remove the cancel button I can’t believe that happened I really wish they would allow me to make posts in the main thread instead of having to make them here then I could mention all these things you gotta be able to cancel the G P T output while it’s going but if you made a mistake.