Sakana AI, founded by ex-Google experts, introduces Evolutionary Model Merge technique for creating advanced AI models. Already applied to language and vision models, leading to impressive results. Raised $30M in seed funding for Tokyo AI hub.

NVIDIA introduced LATTE3D, a model that quickly creates 3D shapes from text prompts. It accelerates design iterations, offering multiple options and quick optimization. Trained on animals and objects, it can be adapted for various industries like landscape design and robotics. Powered by NVIDIA A100 GPUs, it enhances robustness and detail rendering for a range of applications.

Alibaba's Tongyi Qianwen offers free long document processing up to 10 million words, benefiting experts in finance, law, research, medicine, and education. It outperforms global AI apps like ChatGPT, aiding professionals in efficient document analysis and decision-making.

Apple is in talks with Baidu to collaborate on generative AI models for Chinese iPhone users. Apple may integrate Google's "Gemini" model in iOS 18. Baidu's "Wenxin Yiyuan" model is widely recognized in China.

OpenAI to meet with film studios in LA to discuss potential partnerships for upcoming text-to-video generator Sora. Early access granted to A-list directors and actors. Sora expected to revolutionize film production process by simplifying text-to-video conversion. Market testing and content protection challenges remain.

The highly anticipated Rabbit R1 tech product will begin shipping on Easter Sunday, March 31st. Despite a minor delay, consumers are eagerly awaiting its unique design and powerful features. Rabbit has sold 100,000 units, reflecting the popularity and enthusiasm for innovative tech products.

Recent reports suggest that Anthropic may partner with Apple for AI, enhancing iOS performance. Speculation on deep integration of generative AI into iPhone to enrich iOS functionality and promote app ecosystem. iOS 18 to be highly customizable, promising significant software update with strengthened AI capabilities and improved user experience.

At the 2024 Global Developer Pioneer Conference, Leap Star debuted its Step-1V trillion-parameter multimodal model, outperforming GPT-4V on "Sinan". Founder Dr. Jiang Daxin unveiled the Step-2 MoE language model with advanced architecture and API interfaces for selected partners. Leap Star's successful training of trillion-parameter models highlights its technical prowess and ambition to rival OpenAI in AI development.

TacticAI, a "Football Tactical AI Assistant" co-developed with Liverpool FC, uses geometric deep learning to predict corner kick outcomes and offer tactical insights. Powered by Google DeepMind, it aims to optimize football strategies worldwide.

OpenAI's GPT Store hosts AI tools for various tasks but faces issues like copyright infringement and academic misconduct. Despite efforts to regulate, the platform is struggling with growing pains and legal challenges.

Baidu Intelligent Cloud introduced ERNIE Speed, ERNIE Lite, and ERNIE Tiny, smaller parameter models for specific scenarios. ModelBuilder offers ERNIE Character and ERNIE Functions for role-playing and tool invocation applications, simplifying model deployment for intelligent assistants.

Key members of Stable Diffusion team left Stability AI, causing another blow to the startup. Misunderstandings arise from integration with Stable Diffusion. CEO Emad Mostaque faces scrutiny over past experiences. Financial struggles and legal issues plague the company.

The 2024 GDC featured Tencent's GiiNEX game AI engine, blending generative AI and decision-making AI for comprehensive game lifecycle solutions. GiiNEX enhances content creation efficiency, offers creative support for developers and players, and showcases advancements in 3D city generation. Tencent aims to further advance game AI technology through GiiNEX, driving innovation in the gaming industry.

The official website of Taobao Xingchen, an AI model by Taobao Tmall Group, has been launched. Utilizing Transformer architecture and Megatron-ama framework, it enhances copywriting, knowledge Q&A, and decision-making in e-commerce. Taobao Xingchen offers creative generation, knowledge Q&A, decision-making, and coding capabilities. It is currently in internal testing but may open for external experience soon. This highlights Taobao Group's focus on AI e-commerce, aligning with industry trends.

Baidu Map's AI Guide exceeds 100 million users, offering human-like communication for convenient travel. It provides intelligent recommendations for attractions, ticket reservations, and personalized travel plans. AI Guide's multi-turn natural language interaction supports over 70% of regional dialects, enhancing the user experience. The continuous evolution and upgrades aim to be a top "AI-native map" globally.

Microsoft launches two new Surface devices, Surface Pro 10 for Business and Surface Laptop 6 for Business, with Intel's Core Ultra processors and NPU for AI features in Windows 11. Both offer Thunderbolt 4 ports and enhanced displays. Available for commercial customers on April 9.

Public Citizen advises against using AI tools for mushroom identification due to the high-risk nature of the activity. Local knowledge and skills are essential for accurately identifying mushrooms, as AI applications may provide unreliable information and misidentify toxic species. It's crucial to rely on expert foragers and not solely on technology.

OpenAI is set to release GPT-5 soon, building on the success of GPT-4 with enhanced capabilities. The new model is expected to interact with other AI programs, signaling a shift towards a collaborative AI ecosystem. Release date and availability details are yet to be confirmed.

YouTube unveils new disclosure tool in Creator Studio to differentiate realistic AI-generated content from fake. Labels will be rolled out next week to increase transparency and build trust with viewers. Enforcement actions may be taken against creators who consistently fail to disclose information.

Unity's new report reveals that 62% of game studios use AI for development, despite recent negative feedback. AI is used for prototyping, asset creation, and world building, improving operational efficiency by 71%. Some studios remain hesitant to adopt AI due to time constraints, lack of skills, and understanding. AI is expected to further revolutionize the gaming industry with tools like NVIDIA's AI upgrade and Microsoft's optimization.