On April 5th, OpenAI made a number of updates and upgrades to ChatGPT on Wednesday, users can edit DALL.E pictures in ChatGPT through web pages, iOS systems, Android and other systems, directly integrating the SD diffusion model, editing pictures is equivalent to ControlNet's Inpainting function, providing users with the most convenient P picture service!
April 1st, April Fool's Day, OpenAI just announced that it allows users to use ChatGPT directly without registering for the service, greatly lowering the threshold for people to experience artificial intelligence, but due to the influx of a large number of users, OpenAI's server went down that day, and the API interface error rate increased, and many users who did not log in were temporarily unable to access ChatGPT, and then OpenAI fixed the error.
Last year, OpenAI put DALL· E 3 is integrated into ChatGPT, and after the update on April 1, ChatGPT has more new features, and OpenAI has launched an image editor (DALL· E editor), users can now use GPT-4 directly on the web or mobile (iOS/Android) to edit DALL· E images, now with two editing modes.
One is to select the area and then have a dialogue to make a P map, first use great force to generate a picture, select the area to be modified, and then give a Prompt prompt, you can see GPT-4 directly edit and generate the latest picture.
The other is to have a direct conversation, so that Prompt can be applied to the whole image, compared to PS software, OpenAI's DALL· The E 3 image editing function is more powerful, which can generate different styles of pictures, and then combine the above editing methods to adjust the P image.
At present, ChatGPT drawing has two entrances, one is to select GPT-4, or go to GPT Store to call DALL· E plug-in, in the field of Wensheng diagrams, the GPT-4 version of this update is mainly aimed at MidJourney and Photoshop beta versions, as well as the partial redrawing function of Stable Diffusion Inpainting, to achieve the goal of accurate P map redrawing.
After this update, GPT-4 also provides a mobile reading service, compared with the AI technology of Meta, Amazon and other large companies, ChatGPT's power lies in the integration of the integration of Wensheng diagram (DALL· E) Bunsen voice (Vioce engine) function a variety of technologies.
In the future, ChatGPT will most likely integrate Wensheng Video Sora to achieve more comprehensive and complex application scenarios, open up text-to-image, text-to-speech, text-to-image conversion and generation tasks, provide users with a richer and more diverse human-computer interaction experience, and promote the development of artificial intelligence application scenarios.