Resolve yes?

@esusatyo Why resolve yes? A couple people (like VERY VERY few from what we heard) got access to it, and nothing since then

@ShadowyZephyr GPT-4 can do this already

@esusatyo GPT-4 is not chatGPT. ChatGPT is a chatbot based on the GPT-3.5 or GPT-4 models. GPT-4 has image input capabilities, which are not included in ChatGPT.

@esusatyo No, this would resolve "No" as is, but it will resolve at midnight today and things could change in that time. The questions is about ChatGPT, not OpenAI's API.

MAYBE code interpreter supports this but unless I get access to that by midnight and it does, then this resolves No. The question description clarifies that I should be able to access it (even if I have to pay).

P.S. GPT4 API doesnt even support images atm, see: https://platform.openai.com/docs/models/gpt-4

OpenAI API

An API for accessing new AI models developed by OpenAI

@RahulShah Correct, it does not. There are a bunch of clickbait channels trying to claim you can get access, but we would all know from reputable sources if you could.

Plugins are already here. GPT4 is capable of handling image input, just the chat UI needs the update.

@Abhijeet I recently bought ChatGPT Plus; however I don't have access to Plugins -- they are still on a waitlist. The question resolves as if I -- a ChatGPT Plus user -- can use it, not if GPT4 theoretically supports multimodal inputs in some non-publicly released code. So I would not be so sure.

@RahulShah does you getting access to plugins resolve this positively? Or anyone getting access to plugins does? Because in the 2nd case, that's already a yes. And half the time if you prompt it enough it even takes external links as input without plugins as well.

@Abhijeet I do not trust anything that I cannot see for myself so I would need to get access for this to resolve yes.

does you getting access to plugins resolve this positively?

Yes, that is correct

@Abhijeet

And half the time if you prompt it enough it even takes external links as input without plugins as well.

Actually it cannot really "process" the link as it does not have internet capabilities. It is just going off of the characters in the URL (and text in the metadata popup linked if that exists) and bullshitting you from that. It is tricking you, in other words. And I do not believe the image inputs are accepted unless I can see so for myself 😛

@RahulShah did you at least apply for the access or are you just mana farming at this point ? 😛

@Abhijeet I have applied for both plugins and currently have access to GPT-4 via the API (which I also applied for last week).

Why I made this market:

I do research (and closely follow technical details of LLMs) and having access to a multimodal LLM would allow for large improvements so I am actually betting no just to hedge against myself: if it were not to come out I would be disappointed that my research project cannot extend to images but I'll be content since I get mana. If it does come out, I lose mana, but will be able to do a cool project so 😃

@Abhijeet I now have access to plugins: which allow image uploads?

@RahulShah Pretty sure none of them do

Seems someone else is already able to upload images? Might be limited access through a waitlist or something https://www.reddit.com/r/ChatGPT/comments/124vpgg/i_can_now_upload_pics_to_gpt4_taking_requests/

@btcprox Seems to be a plugin that only a random sample of users were given access to per comments on https://www.reddit.com/r/ChatGPT/comments/124vpgg/comment/je11x5q/?utm_source=share&utm_medium=web2x&context=3

If I get access to the image plugin then this will resolve as Yes, if I get some other non-plugin feature this resolves as Yes, if everyone gets access to the image feature this also resolves as Yes.

I cannot tell with certainty,
But here's my pithy poetry:

Uploading images, such a fuss,
Too much data, too much fuss,
But ChatGPT's got the skill,
To add this feature, oh what a thrill!