I was confused how ChatGPT’s DALL-E 3 image generating feature and its new vision feature worked together. Instinctively, you group those two capabilities together in your mind and expect to use them interchangeably. I also saw other savvy tech people asking about how this works. The answer is, they’re separate features. To generate an image with DALL-E 3 in your ChatGPT Plus account: * Start a new chat * Select the GPT-4 model * In the dropdown menu under the GPT-4 model, select DALL-E 3 * In your chat window, ask GPT-4 to generate any type or style of image * Mobile and desktop work the same To upload an image to be seen and interpreted by GPT-4 (requires ChatGPT Plus): * Start a new chat * Select the GPT-4 model * In the dropdown menu under the GPT-4 model, select “Default” * In your chat window, click the image icon to the left of “Send a message” * Upload your image and ask anything you want about it * This works essentially the same way in the mobile app as it does in the desktop browser How do you turn these two features on? * You don’t turn the vision feature on; the image icon simply appears in your chat window when you have “Default” selected in the GPT-4 model dropdown * You do turn DALL-E 3 on in your account’s “Settings & Beta”, and then you’ll see the “DALL-E 3” option in the GPT-4 model dropdown menu These two features are flat out incredible. Enjoy! (I wrote this post on Linkedin and it seems to help people so cross-posting it here.) --- This post was migrated from connect.com and was originally published at an earlier date.
Thanks for sharing! I didn’t even know about image generation within GPT 4 until now. --- This post was migrated from connect.com and was originally published at an earlier date.
The richness and clarity in DALL-E 3 is amazing. --- This post was migrated from connect.com and was originally published at an earlier date.