Microsoft successfully attracts Inflection AI co-founders Mustafa Suleyman and Karen Simonyan, along with colleagues, to lead new Consumer AI division. Inflection AI to focus on AI Studio business. Microsoft continues to showcase AI advancements.

Tencent unveils Tencent Hybrid in latest report, a breakthrough in AI. Hybrid model scaled to trillions, powering text-to-image and video. Tested in 400+ scenarios, deployed in core products. Collaboration with partners expands AI solutions to 20+ industries, promising commercial success. AI advancements set to drive future achievements.

Zhihu's Discovery Conference showcased new products like "Sea Salt Plan 6.0" and "DEEP TRUST Marketing System." Founder Zhou Yuan introduced "Discover·AI Search," emphasizing community trust and professional content. The introduction of the "Navigator" role aims to improve content quality and community management. Salted Stories launched the "Short Story 3A Plan" to evolve short stories through AI technology and multimedia dissemination.

In a recent update, Humane founders unveiled details about the highly anticipated AI wearable device, Ai Pin. Shipping starts in March with advanced features like hands-free communication and visual recognition. Concerns linger over usability, necessity, and price compared to existing technology. Developing CosmOS as an open-source AI system could hold more value than the hardware itself.

360 Company unveiled its latest achievement - Security Big Model 3.0, praised as first domestic large-scale security model with significant AI practical application results. It showcases unique architecture and ranks top in security industry data ranking, leveraging 360 Brain core and 20 years of security big data to build competitive advantage. It offers comprehensive security protection services for enterprises with powerful security intelligence system.

In a lecture at Oxford, Geoffrey Hinton stated that AI models are approaching human brain levels and will surpass them. Large language models have trillions of weights, enabling them to know thousands of times more information than humans. Elon Musk and Ray Kurzweil predict AI to surpass human intelligence by 2029. Hinton warns of AI risks, comparing them to atomic bombs.

Software development is revolutionized by Devin AI and Microsoft AutoDev, with Devin excelling in independent coding and AutoDev focusing on complex project management. The future holds collaborative potential between the two, reshaping software development with increased productivity and innovation.

GitHub introduces Code Scanning Auto-Fix, aiding developers in swiftly fixing most warnings with minimal editing. This AI tool, supported by CodeQL and GitHub Copilot, covers a wide range of warning types in popular languages like JavaScript and Java. By streamlining the vulnerability fixing process, GitHub aims to enhance application security and optimize developer experience, ultimately reducing "application security debt." Support for additional languages such as C# and Go is on the horizon.

OpenAI CEO Sam Altman critiques GPT-4, hopes GPT-5 will exceed expectations. New ChatGPT model expected next year. Altman emphasizes need for continual improvement in AI technology. buzz around GPT-5 release date and features. Sora, OpenAI's text-to-video generator, generating interest with realistic videos.

At NVIDIA GTC, NVIDIA introduced DGX SuperPod with GB200 superchip, supporting trillions of AI model parameters. Each system has 36 CPUs and 72 GPUs connected via NVLink, with 11.5 exaflops AI power. Powered by Quantum-X800 InfiniBand, BlueField-3 DPU, and SHARP technology. GB200 will be part of DGX Cloud on AWS, Google Cloud, and Oracle Cloud, contributing to the Ceiba supercomputer project with 20,000 GPUs for over 400 exaflops AI power.

On March 18th, NVIDIA's 2024 GTC conference began in San Jose, California with CEO Jensen Huang emphasizing its developer focus over entertainment.

OpenAI launched the GPT Store for third-party developers to create conversational AI apps but faced challenges like lack of users, analytics support, and restricting payment. Despite focusing on various AI projects, GPT Store's slow start could hinder OpenAI's long-term growth. Developers struggle with user engagement and tracking, turning to forums for promotion.

Alibaba's Tongyi Tingwu receives important update, introducing new features like audio-video assistant "Xiao Wu", one-click AI rewriting, and mind map generation. Users benefit from enhanced functionality and improved user experience.

Stability AI's SV3D improves 3D tech with better quality and viewpoint consistency. SV3D_u generates trajectory videos from a single image, while SV3D_p supports single image and trajectory viewpoints. The use of video diffusion enhances output generalization and viewpoint consistency. SV3D has vast applications in VR, AR, and product visualization. Commercial use available through Stability AI membership.

A group of Apple computer scientists developed a new LLM model that interprets images and data, advancing AI capabilities. The MM1 series integrates text and image data for more accurate interpretations, with each model containing up to 30 billion parameters.

Artificial General Intelligence (AGI) represents a major advancement in AI, surpassing human abilities. CEO Jensen Huang discussed challenges, emphasizing the importance of accurate information to combat AI illusions. AGI may be achievable within 5 years, depending on specific tests.

The IFOW study in the UK found negative correlation between well-being and exposure to new workplace technologies like robots and AI. Some technologies improve quality of life, but advanced tech leads to decreased happiness. AI used to boost productivity, with IBM planning to replace 8000 jobs by 2023, starting with HR.

Engineers at the University of California are developing a device the size of one square inch to help people speak without vocal cords. The wearable tool captures muscle movements and converts them into sound signals using machine learning. Testing showed 94.68% accuracy, with plans to expand vocabulary and assist those with speech impairments.

The annual GPU Technology Conference (GTC) by NVIDIA is back offline with the highly anticipated B100 Blackwell AI chip, set to revolutionize AI computing with faster training speeds and advanced packaging technology. Stay tuned for more updates on the B200 GPU.

Moonshot AI announces breakthrough in Kimi Intelligent Assistant, processing up to 2 million words in ultra-long, lossless context. Beta testing phase initiated for users to experience advanced language processing capabilities firsthand.