At the 2024 AI for Good Global Summit, OpenAI CEO Sam Altman predicted that China will have unique large language models and believes that AI may make humans more humble, prompting us to reevaluate our position in the universe.

The lip-sync video model PROTEUS allows pictures to sing! The AI model jointly launched by Stanford University and Apparate Labs realizes the function of generating realistic virtual characters from a single photo and singing and speaking in real-time.

What other hot topics in the AI industry at home and abroad are worth paying attention to over the past day? Let Raven take you to take a look.

/ 01 / Large Models

1) Developed by Stanford University! The lip-sync video model PROTEUS allows pictures to sing

The AI model PROTEUS jointly launched by Stanford University and Apparate Labs realizes the function of generating realistic virtual characters from a single photo and singing and speaking in real-time. The model features real-time generation of realistic characters, high frame rate video streams, multimodal interaction, and can be applied to personalized virtual assistants, virtual pets, customer service, and other fields.

2) VideoLLaMA 2: Upload videos to instantly recognize and interpret video content based on instructions

VideoLLaMA2 aims to advance spatiotemporal modeling and audio comprehension capabilities of large video language models. The project can help users better understand video content with fast and accurate recognition speed.

3) Midjourney version update, adding personalized fine-tuning functionality

According to official news from Midjourney, Midjourney has now released a new version, introducing a new personalized fine-tuning feature. Midjourney stated that the new feature will allow users to fine-tune the MJ model based on their aesthetic preferences. The model will fill in unspecified parts of the content based on user habits when users enter prompt words. These habits come from the image content generated by the user previously.

/ 02 / AI Applications

1) FontStudio: Easily create various textured and cool font effects

FontStudio is a new method that can help create beautiful font effects, making works more interesting and unique. It uses diffusion model technology to generate font effects on irregularly shaped canvases, introducing segmentation mask technology to maintain shape consistency.

/ 03 / Investment and Financing Intelligence

1) OpenAI CTO Reveals: The Latest Model Is Not Much Different from the Free Model

OpenAI Chief Technology Officer Mira Murati claimed at the Fortune Most Powerful Women Dinner held in San Francisco that "OpenAI's AI models are not much more advanced than publicly available models," which does not seem conducive to building investor confidence.

2) Market value of nearly 20 billion! Tencent harvests an AI super IPO, QuantumPharm lands on the Hong Kong Stock Exchange

QuantumPharm Inc. (2228.HK), an AI and robot-driven innovative technology company, officially listed on the main board of the Hong Kong Stock Exchange, becoming the first new stock listed company to comply with the 18C specialized technology rules. The closing price on the first day was HK$5.8, 9.85% higher than the listing price of HK$5.28. The price once surged by 24.6% intraday, with a full-day trading volume of HK$499 million and a latest market value of HK$19.76 billion.

/ 04 / AI Infrastructure

1) Altman Discusses AI Opportunities, Challenges, and Human Self-Reflection: China Will Have Unique Large Language Models

OpenAI CEO Sam Altman pointed out at the 2024 AI for Good Global Summit that AI has already played a positive role in enhancing productivity but has also brought issues such as network security. Altman promised to continuously improve GPT-4o, address language equity issues, and emphasized the importance of AI governance. He also predicted that China will have unique large language models and believes that AI may make humans more humble, prompting us to reevaluate our position in the universe.

2) It is rumored that MediaTek is designing Arm-based chips for Microsoft AI laptops

According to TrendForce, MediaTek is developing an ARM-based PC chip that will run Microsoft's Windows operating system.

Last month, Microsoft released a new generation of laptops equipped with ARM-based chips, providing sufficient computing power for running AI applications. Microsoft executives stated that this represents the future trend of consumer computing. It is said that MediaTek's latest ARM-based PC chip is specifically targeted at such laptops.