AI News

NVIDIA CEO Jensen Huang met with MiniMax founder Yan Junjie in Beijing, praising China's AI innovation. MiniMax, founded just two years ago, has made breakthroughs including the open-source M1 model, Hailuo02 video tool, and $300M funding at $4B valuation. The meeting highlights potential for US-China tech collaboration.....

Meta reorganized its AI structure to establish a Superintelligent Lab, integrating 3400 employees, led by Alexandr Wang as Chief AI Officer. The new structure is divided into four teams: AGI Basic Research, AI Product Development (including Meta AI Assistant), the Basic AI Lab led by Yann LeCun, and a group focused on Llama5 development. Meta is offering high salaries to attract talent from companies like OpenAI and Apple, but this has raised doubts within the original team about the influx of high-paid outsiders. Recently, two AI leads from Apple have joined.

Li Auto received the first domestic batch of dual safety certifications for automotive generative AI at the 2025 China Automotive Forum, becoming the first automaker to pass the national standards GB/T 45654 and GB 45438-2025. The certification was jointly issued by the CCIA Automotive Cybersecurity Working Committee and the AI-Generated Content Identification Service Platform, covering the fields of content security and identification. This achievement marks Li Auto's leading position in the industry regarding the safety of in-vehicle AIGC technology, setting a benchmark for the safe development of intelligent vehicles, while enhancing consumer confidence.

OpenAI releases macOS ChatGPT Plus app with 120-minute audio recording, generating structured notes, action items, and code. Privacy-focused: audio deleted post-processing. Currently macOS-only, potential future expansion. Boosts meeting efficiency but faces multi-speaker recognition limits.....

Lightricks releases open-source LTX-Video13B, a 13B-parameter video generation model with multi-scale rendering, achieving 30x faster speeds. It runs on consumer GPUs, supports 1216×704 real-time generation, and offers text/image/video-to-video modes. The model enhances coherence and detail, enabling keyframe control and style transfer. Free for SMEs, it includes training tools and optimized versions to democratize AI video creation.....

Perplexity partners with Bharti Airtel in India, offering 360M users free Pro service for a year. Downloads surged 600%, MAUs up 640%. Collaborating with Paytm, it aims to lead India's AI market despite monetization challenges.....

Tencent's Yuanbao App integrates QQ Music, enabling voice search and song recommendations in v2.30, enhancing AI's multimodal capabilities for entertainment.....

Apple's MLX framework now supports CUDA, marking a strategic shift to embrace NVIDIA's ecosystem. This enables developers to train models on NVIDIA GPUs and deploy on Apple devices, enhancing efficiency and bridging platform gaps.....

Tesla has released its new Dojo 2 chip, with performance 10 times that of the first generation, and computing power close to NVIDIA B200. The chip is manufactured by TSMC and uses advanced packaging technology, solving power consumption issues. Dojo 2 will assist Tesla's FSD autonomous driving system training, processing 16 billion video frames per day, achieving technological self-sufficiency. Musk revealed that next year a more powerful Dojo 3 will be launched, and he joked that Dojo 2 can run 'Crysis' at one billion frames per second. This breakthrough will reduce Tesla's reliance on NVIDIA and may be offered externally.

Claude Code faces backlash as paid users encounter unexplained usage limits, sparking transparency concerns. Subscribers report frequent restrictions without clear explanations, disrupting projects. Anthropic's vague responses and repeated service failures intensify trust issues.....

Mistral's Le Chat AI chatbot upgraded with deep research mode, enhanced security, multilingual support, and new project/image tools to rival OpenAI.....

On July 18, Xiaodu officially launched its first MCP Server that supports interaction with the physical world, completing the MCP upgrade of terminal devices and core IoT capabilities, bringing a new transformation to AI application development and leading the industry towards a new era of 'all things intelligent interconnection'. The Xiaodu Open Platform has innovatively introduced two core services. One is the Xiaodu Speaker MCP Server, allowing developers to control devices, deliver voice announcements, and perform remote monitoring through text commands; the other is the Xiaodu IoT MCP Server, which opens up full-home device control.

Swedish AI platform Lovable raised $200M in Series A, reaching $1.8B valuation. With 2.3M users, it enables natural language coding, generating $75M annually. Backed by Accel, it's popular among non-tech users, creating 10M+ projects.....

Mistral AI launches Le Chat, challenging ChatGPT with features like deep research mode, voice interaction, and advanced image editing.....

New malware LameHug uses Alibaba Qwen2.5 large model to attack Windows systems, spreads through email attachments, and can dynamically generate data stealing instructions. The software collects system information and steals sensitive files, with multiple variants already discovered. Experts warn that this is the first publicly known AI-based malware and recommend users to remain vigilant and update their protection measures.

Moonshot AI's open-source model Kimi K2 excels in programming, matching Qwen3-235B-A22B and nearing o3-mini-high & Claude-3.7-Sonnet. With 1T MoE, 128k context, 65.8% accuracy on SWE-bench, and $0.14/M tokens, it's ideal for coding. Supports web generation, workflows, vLLM/Hugging Face deployment, MIT-licensed.....

Yang Jianchao, the head of ByteDance's Visual Large Model team, announced a temporary leave due to family reasons, with Zhou Chang, former technical leader of Alibaba's Tongyi Qianwen, taking over. This personnel change occurred during a period of adjustment in ByteDance's AI department, sparking concerns about the stability of the technical roadmap. Yang Jianchao's work information remains in the internal system, and Zhou Chang will lead the global Seed team to continue research on visual multimodal generation. The company emphasized its continued investment in basic research and hopes that the new leader will bring innovative energy. This change highlights the importance of balancing work and health in the fast-paced tech industry.

NVIDIA's Canary-Qwen-2.5B sets a 5.63% WER record on Hugging Face OpenASR. This CC-BY licensed model combines FastConformer encoder with Qwen3-1.7B LLM decoder for efficient speech-to-text and NLP. Supports multi-GPU deployment for cloud/edge applications.....

Suno AI V4.5+ introduces vocal/instrumental swap, playlist-inspired composition, 8-minute tracks, and enhanced audio for paid users.....