AI News: ChatGPT Pulse, Gemini Robotics, Qwen3-Max, Stargate, OpenAI and Nvidia, and more!

ChatGPT has launched 'Pulse,' a new proactive feature that delivers personalized updates and research, while Alibaba, Kimi Moonshot, Google, and Meta have released advanced AI models for image editing, video generation, coding, and robotics, amidst significant investments in AI infrastructure like the Stargate project.

• ChatGPT Pulse offers proactive, personalized updates based on user history and connected apps for Pro users.

• New AI models from Alibaba, Kling, Google, and Meta advance text-to-video, image editing, code generation, and robotics.

• The Stargate project and Nvidia's $100 billion investment are driving massive expansion in AI data centers and compute capacity.

OpenAI introduced ChatGPT Pulse, a proactive AI feature that personalizes updates based on user history and connected apps, initially for mobile Pro users. Concurrently, major advancements in AI models include Alibaba's Qwen image editing and Qwen3Max for coding, Kling and Alibaba's Qwen 2.2 Animate for text-to-video, Google's Gemini Robotics ER 1.5 for embodied reasoning, and Meta's Code World Model for code generation through execution learning. Large-scale infrastructure projects like Stargate, involving OpenAI, Oracle, SoftBank, and Nvidia's $100 billion investment, are rapidly expanding AI compute capabilities.

ChatGPT Pulse Launch

• 00:00:00 ChatGPT has introduced 'Pulse,' a new proactive feature designed to personalize user experiences by thinking overnight, analyzing conversation history and memory, and generating a list of topics or articles users might find interesting. This mobile-only feature for Pro users leverages connected apps like calendars and emails to curate personalized updates, effectively shifting AI from reactive to proactive research based on user feedback and preferences. OpenAI's blog post describes Pulse as delivering personalized updates based on chat feedback and connected applications, allowing users to curate researched topics.

Advanced AI Models for Media and Robotics

• 00:02:33 Several new AI models have been released, showcasing advancements in text-to-image/video generation and robotics. Alibaba's Qwen 2.2 Animate allows for high-fidelity character animation by replicating expressions and movements from reference videos, seamlessly integrating characters into original scenes. Kling AI 2.5 Turbo Video Model also demonstrates excellent text-to-video capabilities with superior clarity, character consistency, and speed. Furthermore, Google's Gemini Robotics ER 1.5 is a state-of-the-art, high-level reasoning model specifically for robots, offering fast spatial reasoning and orchestrating advanced agentic behavior, while Meta Fair’s Code World Model learns code generation by executing and learning from code rather than relying solely on existing code, aiming for better world understanding.

Qwen3Max and AI Performance Benchmarks

• 00:07:09 Alibaba's Qwen3Max demonstrates significant advancements in coding and agentic skills, rivaling top models on benchmarks such as SweeBench, Tau2Bench, SuperGPQA, LiveCodeBench, and Amy2025. With 'MaxThinking' equipped with tool use and deployed in 'heavy mode,' Qwen3Max achieves near-perfect scores on key benchmarks, including a 100% score on certain Python-use scenarios. These phenomenal results indicate that Chinese AI models have considerably closed the performance gap with leading Western models like GPT-5 Pro and Grok 4, especially in areas like reasoning and code generation.

Massive AI Infrastructure Expansion

• 00:13:02 The Stargate project, a $500 billion initiative for building new AI data centers, is expanding with five new sites, putting it ahead of schedule to secure its 10-gigawatt commitment by late 2025. Key locations include Abilene and Shackleford County in Texas, Dona Ana County in New Mexico, and a forthcoming Midwest site, projected to create over 25,000 jobs. Additionally, Nvidia is investing $100 billion into OpenAI, establishing a landmark strategic partnership to deploy at least 10 gigawatts of Nvidia systems, highlighting a complex web of financial commitments and collaborations among major tech players in the AI ecosystem.