Search AI Products and News

At the BITA Conference held today in 2025, the Beijing Academy of Artificial Intelligence formally released the Wu World series of large models, comprehensively showcasing its latest research results and strategic layout in the direction of physical general artificial intelligence (AGI). The Wu World series aims to break through the boundary between virtual and real worlds and empower the physical world, covering four cutting-edge models: the multi-modal world model Emu3, brain science model Jianwei Brainμ, embodied intelligence brain RoboBrain 2.0, and microscopic life model OpenComplex2, constructing a framework from...

On April 15th, Hugging Face, the renowned open-source large language model platform, announced its acquisition of Pollen Robotics, marking its official entry into the physical robotics field. While specific transaction terms remain undisclosed, the acquisition will bring approximately 20 Pollen Robotics employees to Hugging Face. This represents the company's largest personnel acquisition to date, signifying its ambition in expanding its business areas. Hugging Face's co-founder...

Higgsfield, an innovative AI video generation company, recently unveiled Higgsfield Mix, a groundbreaking technology that completely overturns the physical limitations of traditional cameras. According to AIbase, this technology allows users to combine multiple motion controls in a single shot, creating dynamic effects impossible with real cameras. Higgsfield also introduced 10 new motion control modes specifically designed to enhance speed, tension, and cinematic impact, empowering film creation and numerous other applications.

From April 1st to 18th, approximately 100,000 ninth-grade students in Wuhan participated in their physical education entrance examination. This year's exam introduced a new AI intelligent judging system, marking a significant shift in physical examination assessment. Traditionally, physical education exams relied on manual timing and scoring. This year's exam utilized intelligent devices and cloud management to ensure fairness and efficiency. The exam incorporated several AI technologies, including facial recognition, electronic distance measurement, and wireless sensing, covering 14 items across three categories.

At the 2025 Consumer Electronics Show (CES), NVIDIA unveiled the new Cosmos platform, designed to accelerate the development of physical artificial intelligence (AI) systems, particularly for autonomous vehicles and robots. The Cosmos platform integrates a generative world foundation model (WFM), video taggers, safety mechanisms, and an accelerated data processing pipeline, enabling developers to create and optimize AI models with reduced reliance on real-world data.

Google recently announced the formation of a new team dedicated to developing artificial intelligence (AI) models capable of simulating the physical world. The team will be led by Tim Brooks, who previously co-led the video generation model Sora at OpenAI. Brooks stated on social media platform X that this new team will belong to Google's AI research lab, Google DeepMind. In his announcement, Brooks mentioned, 'DeepMind's plans are ambitious.'

After a successful entrepreneurial journey, former SmartThings CEO Alex Hawkinson embarks on a new startup adventure. His new company BrightAI has just completed a $15 million seed funding round exclusively led by Upfront Ventures, unveiling this tech company deeply focused on smart monitoring for infrastructure. BrightAI was founded in 2019 by Hawkinson along with Nathan Hanks and Doug.

Recently, researchers from ByteDance Research Institute and Tsinghua University jointly released a new study, pointing out that current AI video generation models, such as OpenAI's Sora, while capable of creating stunning visual effects, have significant flaws in understanding basic physical laws. This study has sparked widespread discussion on the capabilities of AI in simulating reality. The research team tested AI video generation models under three different scenarios: predictions under known patterns and predictions under unknown patterns.

San Francisco-based startup Physical Intelligence recently announced that it successfully raised $400 million, bringing its post-money valuation to $2.4 billion. This funding round attracted the attention of several prominent investors, including Amazon founder Jeff Bezos, OpenAI, Thrive Capital, and Lux Capital. According to a company spokesperson, Bond Capital also participated in this investment.

As the wave of technology sweeps across the globe, generative AI has undoubtedly become a focal point. However, Wu Yongming from Alibaba has presented a deeper insight: the true value of generative AI goes beyond creating dazzling applications; it lies in fundamentally changing the way we interact with the physical world. Wu keenly pointed out that people often fall into the thinking trap of 'overestimating short-term effects while underestimating long-term impacts' in the face of a new technological revolution. This phenomenon is particularly evident in the current fervor surrounding generative AI, with the public generally focusing on the immediate tangible effects that AI can bring.

Judge Kevin Newsom of the US 11th Circuit Court of Appeals innovatively used chatbots like ChatGPT to interpret the legal term 'ordinary meaning,' particularly concerning the dispute over whether a 'physical restraint' was involved in a gun robbery case. In this case, the judge agreed to uphold the severe sentence for the defendant DeLeon but noted that there may have been past misinterpretations of the definition of 'physical restraint' and emphasized that the term should be understood according to its ordinary, common meaning. Newsom asked GPT-4, Claude, and Gemini about the ordinary meaning of 'physical restraint.'

GPT-4, OpenAI's language model, passed the Japanese National Physical Therapist Exam without additional training, achieving an accuracy rate of 73.4%. It excelled in handling pure text questions with an accuracy of 80.5%, but performed poorly in solving practical problems and dealing with questions containing images and tables, with accuracy rates of 46.6% and 35.4%, respectively. The research reveals the limitations of AI in certain areas, particularly in understanding complex contexts and visual information. Despite being trained on English data, GPT-4 still performed effectively when processing Japanese input.

KAN2.0, an upgraded version of the neural network architecture, aims to deepen the integration of AI and classical physics. Researchers can customize the model to incorporate expertise for discovering key concepts in physical systems, such as the Lagrangian. The new framework enhances the model's expressiveness, practicality, and interpretability through three main features: MultKAN, kanpiler, and Tree Converter, addressing compatibility issues between AI and science. KAN2.0 excels in feature recognition, revealing module structures, and discovering symbolic formulas, particularly in fields that are difficult to represent with symbolic equations, such as chemistry.

World Labs, a startup founded by Stanford University AI professor Fei-Fei Li, has completed two rounds of funding in two months, reaching a valuation of over $1 billion and becoming a 'unicorn' company. The latest funding round was led by NEA, with a significant increase in valuation compared to the first round in April. World Labs is dedicated to developing AI models that can accurately estimate the three-dimensional physical properties of real-world objects and environments, aiming for detailed digital replication without the need for large-scale data collection, and is focused on addressing the scarcity of 3D data. The company's business model has not been validated by the market.