Photo/Liu Xuemei (NBD)

NO.1 OpenAI Launches Vision Pro Version of ChatGPT

OpenAI recently announced on Platform X that ChatGPT has officially landed on Apple Vision Pro, and users can download the application from the visionOS App Store. It is reported that the application's display effect on visionOS is not much different from the iPad version, and it does not provide additional style effects for visionOS.

Commentary: This is another move by OpenAI to enhance its AI dialogue capabilities, and it is also the result of its cooperation with Apple in the AR field.

NO.2 Firefly AI Lands on Apple Vision Pro

Adobe announced on Friday that its AI text-to-image tool Firefly AI will land on the Apple Vision Pro headset as a native application, while its Lightroom photo editing software was already showcased during the headset's launch. The new Firefly experience is designed specifically for the visionOS system, allowing users to move and place images generated by the application onto real-world spaces such as walls and tables.

Commentary: This is a major breakthrough for Adobe in the field of AI image generation, and it also reflects its close cooperation with Apple in the AR field.

NO.3 27 EU Member States Representatives Unanimously Support AI Act

On February 2, representatives of the 27 EU member states voted unanimously to support the text of the AI Act, marking a major step forward for the EU in legislating and regulating AI. Thierry Breton, the EU's Internal Market Commissioner, said in a social media post that this shows that the 27 countries recognize that "negotiators have found the perfect balance between innovation and safety." The AI Act still needs to be approved by the European Parliament. If approved, the relevant rules will be implemented in stages, with some of the prohibitive rules taking effect 6 months after the law is passed, and some rules related to general AI models will apply from 2025.

Commentary: This is a major development in the EU's AI regulation, and it is also an important move for the EU to uphold its values and interests.

NO.4 Apple Opens Source Automatic Image Editing Tool MGIE

The Apple team recently open-sourced an image editing tool called MGIE, which uses multimodal large models to solve the problem of insufficient instruction guidance. Take a photo and enter text instructions to let the phone start automatic image editing. MGIE is composed of a multimodal large model and a diffusion model, which learns to obtain concise expression instructions and provides clear visual guidance. Through end-to-end training, the diffusion model is updated synchronously and uses the potential imagination of the target goal to perform image editing. In this way, MGIE can benefit from inherent visual reasoning and solve ambiguous human instructions, thereby achieving reasonable editing.

Commentary: This is an innovation by Apple in the field of AI image editing, and it is also a contribution to the open source community. The open source of MGIE will allow more developers and users to use and improve this tool.

Disclaimer: The content and data in this article are for reference only and do not constitute investment advice. Please verify before using.