Search AI Products and News

Liquid AI opensources the next-generation edge AI model LFM2, available in three versions with 350M to 1.2B parameters. The model features an innovative architecture, achieving twice the inference speed and three times the training efficiency on edge devices, supporting 32K long context processing. LFM2 performs exceptionally well in tasks such as instruction following, outperforming models of similar scale, making it particularly suitable for privacy-sensitive scenarios. Fully open-sourced through Hugging Face, this marks the first time a U.S. company has surpassed Chinese open-source models in the field of efficient small models. Liquid AI

The UN AI for Good Summit was held in Geneva, where Peng Jin from Ant Group shared China's achievements in AI security technology. Data shows that Ant Digital helped Southeast Asian banks reduce fake face attack rates from 10% to 4%, with an identification accuracy rate of 99.9%. Ant provides financial-grade identity authentication through the ZOLOZ platform, serving 25 countries, and has opened a dataset of 1.8 million fake samples to promote industry research. China's technological solutions are offering important references for global AI safety governance.

UK study: 67% of teens aged 9-17 view AI chatbots as 'friends', with 12% relying on them due to lack of real social connections. AI mimics human emotions, potentially blurring human-machine boundaries. Experts warn of psychological risks and urge regulations to protect youth mental health.....

AI's 'time travel' trend thrives as ChatGPT transforms childhood photos. TikTok's AI effect drew 170K users, but results vary: Musk's image was unrecognizable, Asian stars distorted, while Eddie Peng fared slightly better. Experts note AI predicts general trends, not individual changes, yet this playful tech sparks social media buzz.....

Goldman Sachs introduced AI coding assistant 'Devin' to boost efficiency, planning hundreds of deployments. The hybrid human-AI approach enhances productivity, though AI won't replace developers.....

The United Nations University research team developed two AI virtual characters - Amina, a Sudanese refugee, and Abdullah, a rebel fighter - to raise public awareness of the refugee crisis through dialogue. The project was experimentally conducted by an academic team and is not an official United Nations project. Although researchers envisioned using it for fundraising presentations, test users provided negative feedback, stating that real refugees can already speak for themselves. The relevant website is currently unavailable.

On July 11, PixVerse AI video creation platform, which has surpassed 60 million global users, announced a major feature upgrade — the addition of the 'Multi-Keyframe Generation' function in the Start-End Frame module. This marks a new stage in AI video creation, transitioning from the generation of single segments to narrative expression. Users can now upload up to 7 images as keyframes via the web version's start-end frame feature, and the AI will automatically analyze the semantic relationships between frames, intelligently building smooth action and scene transition paths. This technological breakthrough enables static images to be presented dynamically.

Meta acquires AI voice startup Play AI to enhance its voice synthesis tech, aligning with AI avatars and wearables. The move reflects Meta's AI push amid rising demand for voice interaction.....

Meta acquires the voice AI startup Play AI to enhance its voice technology capabilities in areas such as AI avatars and wearable devices. The Play AI team will join Meta as a whole, and their natural speech generation technology is highly compatible with multiple Meta AI projects. This is another important move by Meta in the AI field. Previously, it had recruited talent from OpenAI and partnered with Scale AI. The transaction amount was not disclosed.

Tesla launches Grok AI assistant for AMD Ryzen-equipped vehicles, requiring hardware confirmation via 2025.26 software update. Currently focuses on interaction, not vehicle control. AMD selected for superior computing power since 2021. Future expansion planned, incompatible with Intel-based models.....

Tesla launches AI assistant Grok, available only for AMD Ryzen-equipped vehicles via 2025.26 update. Currently offers smart interaction but no vehicle control. Tesla switched from Intel to AMD Zen3 processors since 2021 for better AI performance. Older Intel models can't support Grok, which also provides humorous interactions.....

Moonshot AI officially released its latest creation - the Kimi K2 model, and simultaneously announced its open source. This foundation model based on the MoE architecture has gained widespread attention in the AI field since its release, thanks to its strong coding capabilities and excellent general Agent task processing abilities. The Kimi K2 model has a total of 1T parameters, with 32B activated parameters. It has achieved top performance among open-source models in a series of benchmark performance tests such as SWE Bench Verified, Tau2, and AceBench.

Recently, Tencent Cloud officially launched the API service for the Tencent Hunyuan A13B model on its official website. The input price is set at 0.5 yuan per million Tokens, and the output price is 2 yuan per million Tokens, which has quickly sparked enthusiastic discussions in the developer community. As the first 13B-level MoE (Mixture of Experts) open-source hybrid inference model in the industry, Hunyuan-A13B features a total of 80B parameters and only 13B activated parameters, achieving performance comparable to leading open-source models of the same architecture, while also demonstrating efficient reasoning capabilities.

Google DeepMind open sources the GenAI Processors Python library, helping developers build efficient generative AI workflows. The library supports asynchronous processing of multimodal data and optimizes Gemini API application development, significantly reducing latency in real-time applications. Core features include a modular Processor interface, streaming API design, and concurrency optimization, enabling rapid development of real-time applications such as intelligent assistants. Currently only supports Python, but with an open community contribution model, future plans include expanding functionality to cover more scenarios.

General AI company Manus adjusts its China operations, lays off employees, and relocates its core technology team to Singapore. The China region had approximately 120 employees, and the company states this move is aimed at improving operational efficiency and focusing on core business. The official website now shows that the region is unavailable, replacing previous messages about the development of the Chinese version. The official Weibo and Xiaohongshu accounts have also been cleared, indicating a significant shift in the company's market strategy in China.

Modo AI introduces a 30-second rapid prototype generation feature, supporting multi-device adaptation and conversation optimization. Users can generate high-fidelity, editable prototypes through text, sketches, and other input methods, and support iterative conversation adjustments. The AI can intelligently parse uploaded sketches, wireframes, and more, automatically generating interfaces. It offers dual-mode editing, automatic documentation generation, and code integration features, covering multiple scenarios such as e-commerce and social networking, significantly lowering the barrier to prototype creation and improving product design efficiency.

Mistral AI launched the Devstral2507 series with two AI models: the open-source Devstral Small1.1 (24 billion parameters, SWE-Bench score of 53.6%) and the enterprise version Devstral Medium2507 (score of 61.6%). Small1.1 supports a 128k context window and local deployment, while Medium2507 outperforms some commercial models. Both are optimized for code reasoning and program synthesis, and support integration with agent frameworks.

Zhipu AI launches AI Slides, a revolutionary PPT tool using GLM-Experimental model. It generates professional slides from text/documents with smart layouts and visual optimization. Free for business/academic use, praised for design quality and efficiency. Available on Zhipu's official site.....

AWS upgraded SageMaker with model observability and local IDE integration. HyperPod now monitors training performance and connects local dev environments to cloud. GPU cluster management was optimized for flexible resource allocation.....