Resolves YES if:
The web UI at chat.openai.com allows image inputs within some conversations, for any persons not affiliates of OpenAI or Microsoft.
Resolves 50% if:
The web UI at chat.openai.com is announced to support inline images in some conversations, or there is strong evidence that the UI supports it for some non-affiliates. However, I am unable to verify a non-affiliate's access. Then, the market will be held open for up to 6 additional months to confirm it is eventually released to non-affiliates and resolved 50%.
Resolves NO if any of:
No verified access in 2023, and no announcement or strong evidence of private testing.
No verified access in 2023, and regardless of announcements or evidence, no verified access through June 2024.
ChatGPT is permanently taken down during 2023
Image inputs are accessible in Playground but not chat.openai.com
ChatGPT is renamed to a different product, even if the UI looks similar.
Definitions:
An "affiliate of X" is an employee, owner, investor, contractor, or vendor signing an NDA of X. Customers bound by a Terms of Service are not considered affiliates.
"Strong evidence" means a press release, journalist article, interview with an affiliate.
"Verify" means either personally using it, observing someone I believe to be a non-affiliate using it, or seeing a journalist report access.
"ChatGPT": OpenAI is allowed to change their domain name, but the product must be substantially similar: A list of conversations each of which supports tree-branching, rerolling of responses, automatic summarization of the conversation. ChatGPT-branded mobile or desktop or other apps do not count, only the web interface accessible at its primary domain. If "ChatGPT" is renamed with a minor change (e.g. "GPTChat"), it's considered the same product; if it's renamed to something very different (e.g. "Open Chat Experiences"), then it's considered a different product even if the UI looks similar.
"Image Inputs" means either uploading of images from disk, uploading of images taken from camera, or hotlinking of images. If novel forms of image input are discovered(uploading Turtle programs?), I retain the option to count those too.
Related questions
π Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | αΉ517 | |
| 2 | αΉ152 | |
| 3 | αΉ122 | |
| 4 | αΉ82 | |
| 5 | αΉ81 |
@RahulShah Can you name or link a plugin that enables any model on ChatGPT to accept an image as input using its multimodal support? BeMyEyes, Bing, and certain b2b contracts for dedicated instances are the only cases of GPT-4 accepting image inputs that I know of currently.
@Mira Sorry I was thinking of PDF plugins that can parse images. Not directly images but it works
ChatGPT was announced to have several new capabilities, including uploading of images for use in an embedded Python terminal and plotting of images.
This doesn't yet count for resolution because ChatGPT itself is not receiving the image inputs using its GPT-4 multimodal support: It's writing code to manipulate the images.
However it's an obvious next step for OpenAI to implement.
@Mira "OpenAI is adding support for plug-ins to ChatGPT βΒ an upgrade that massively expands the chatbotβs capabilities and gives it access for the first time to live data from the web.
Up until now, ChatGPT has been limited by the fact it can only pull information from its training data, which ends in 2021. OpenAI says plug-ins will not only allow the bot to browse the web but also interact with specific websites, potentially turning the system into a wide-ranging interface for all sorts of services and sites."
From the verge article : https://www.theverge.com/2023/3/23/23653591/openai-chatgpt-plugins-launch-web-browsing-third-party