Hi folks.
Thought to do this myself, but im not familiar with plugins for several browsers, so maybe someone more experienced can get interested for the idea.
Basicly (in most simplified way) what if i want to play “guess who” game with GPT? I actually do that and “true or false” but this can be extended for other gaming also.
For now, GPT cant pass any data unless user does see it in chat. What if GPT could pass data what user cant see? “Guess who” comes honest “guess who” because for now GPT has to guess himself too what he thought in previous message.
If GPT could pass data inside hidden element, it could give some pretty good extensions to use it?
[message]
Im thinking of Finnish ex president who is highly awarded due hes peace work
[hidden-for-ai]Im thinking of Martti Ahtisaari]/hidden-for-ai]
[/message]
Now, GPT knows who he is thinking for. We only need to hide <hidden…></hidd…> part from chat so user cant see it.
Now extend this idea for simple space gaming etc.
I believe this could be cool extension. Im just too lazy to do it lol.
Its easy with OpenAI api but with browser things, im bit lost.
You could do all of this with the API and some code, would need to find a Developer to move the idea forwards, or you could ask ChatGPT to create the code for you…
Working with API via nodejs or python is so much different than creating a plugin for a browser at chatp.openai.com that i ratherly give the idea for some more proper developer.
But its just idea, lets see does it fly anywhere.
Just gonna leave this here cause Im new and to tired to keep looking for where it should go - It would be awesome if users were able to use their cameras to scan data into chatgpt for analysis help, such as reading data and/or solving work projects. I mean sure someone could plug in the data from a spread sheet and ask for help, but for someone to open there camera and get an active ai response to what they are working on. With ways to either solve, enhance or dismantle the data in front of them would be amazing.
This feature is coming, called multi modality and image handlining or image input, it was demonstrated at the GPT-4 launch but has not been a public feature, it is in R&D and will be a feature for people to use in the future.
1 Like