Staff Chinese tech giant Alibaba commits $53 billion to AI infrastructure development, declaring Artificial General Intelligence its primary objective. The historic investment, larger than the company’s entire past decade of AI spending, positions Alibaba to compete with U.S. tech leaders while signaling China’s renewed support for tech innovation.
New icons discovered in Claude’s iOS app suggest Anthropic is preparing to launch significant AI upgrades, including enhanced reasoning capabilities and web search support. CEO Dario Amodei’s recent hints at these features, combined with the app findings, point to an imminent release that could reshape the AI assistant landscape.
In a groundbreaking move that challenges AI industry norms, DeepSeek announces plans to open-source five core repositories of its R1 platform’s code and data. The R1 model, boasting 671 billion parameters while using only 37 billion per token, has already claimed top spots in coding and math benchmarks. This unprecedented level of transparency stands in stark contrast to competitors’ closed approaches and could reshape how AI development happens in the future.
Microsoft unveils Magma, a breakthrough AI model capable of understanding and acting in both digital and physical environments. This first-of-its-kind foundation model can navigate user interfaces, control robots, and execute complex tasks autonomously, marking a significant advancement in multimodal AI technology that could transform industries from manufacturing to healthcare.
Former OpenAI CTO Mira Murati emerges from stealth mode with Thinking Machines Lab, a startup focused on making AI more accessible and aligned with human needs. With a star-studded team of AI veterans and a commitment to open science, the company aims to reshape how humans collaborate with artificial intelligence.
Google today introduced “AI co-scientist,” an innovative AI system built on Gemini 2.0 that helps researchers generate hypotheses and develop research plans. Why it matters The tool could significantly speed up scientific discoveries by streamlining literature reviews and suggesting experimental approaches. How it works The AI co-scientist accepts natural language inputs from researchers, then: Between … Read more
Why it matters: Meta signals its ambitious open-source AI strategy with a dedicated developer conference, potentially challenging closed AI ecosystems as Zuckerberg aims to make Llama “the most advanced and widely used AI models” in 2025. The big picture Meta announces today its first-ever AI-focused developer conference, LlamaCon, scheduled for April 29, 2025. What to … Read more
Why it matters: Generative AI is fundamentally transforming how we find information online, potentially upending the advertising-based internet economy while creating new ways to access knowledge. The big picture MIT Technology Review announces today a roundtable discussion on “Generative AI Search and the Changing Internet,” featuring Editor-in-Chief Mat Honan and Executive Editor Niall Firth. Key … Read more
OpenAI CEO Sam Altman unveils plans for a unified AI platform with GPT-5, eliminating complex model selection while introducing a three-tier subscription system including a free tier. The announcement comes alongside the cancellation of the standalone o3 model as the company streamlines its product offerings.
Why it matters: The “AI, Science and Society” conference brings together global experts to shape AI’s future ahead of the larger AI Action Summit, addressing critical challenges at the intersection of technology and humanity. The big picture École Polytechnique announces today an international scientific conference titled “AI, Science and Society” taking place February 6-7 in … Read more