Chinese AI firm Moonshot AI releases Kimi K2 Thinking, an open-source reasoning model that outperforms GPT-5 and Claude Sonnet 4.5, achieving record 44.9% on Humanity's Last Exam benchmark and marking a potential turning point as Chinese models now rival leading US systems.
Chinese AI lab Moonshot releases Kimi K2 Thinking, a free open-source reasoning model that reportedly outperforms GPT-5 and Claude Sonnet 4.5 on major benchmarks while costing just $4.6 million to train, potentially disrupting the billion-dollar AI market and intensifying US-China competition.
Google Research unveils Nested Learning, a revolutionary AI paradigm that solves catastrophic forgetting by treating models as interconnected, multi-level systems where components update at different frequencies, enabling machines to learn new tasks without losing previous knowledge.
Physicists successfully eliminate imaginary numbers from quantum mechanics equations while preserving all theoretical predictions, overturning 2021 studies that claimed complex numbers were essential to the field, though the elegant mathematical beauty of imaginary numbers continues to puzzle scientists about their natural fit with quantum reality.
MIT researchers unveil a revolutionary coding framework that transforms AI-generated software by organizing code into modular 'concepts' and explicit 'synchronizations,' dramatically improving safety and reliability while enabling AI assistants to propose new features without dangerous hidden side effects.
Nvidia CEO Jensen Huang warns China is winning the AI race despite U.S. chip bans, citing that 50% of global AI researchers and most leading open source models come from China while Beijing subsidizes energy costs for local alternatives.
NVIDIA releases open-source NeMo Automodel library that delivers breakthrough 280 TFLOPs/sec GPU performance for AI training, enabling developers to train massive mixture-of-experts models in PyTorch with simplified infrastructure while achieving near-linear scaling across thousands of GPUs.
Microsoft launches new MAI Superintelligence Team led by AI chief Mustafa Suleyman to develop advanced artificial intelligence applications for healthcare diagnostics, education, and renewable energy, marking the tech giant's push to reduce OpenAI dependence and compete with Meta's similar research investments.
Oxford researchers discover nearly half of 445 AI benchmarks lack scientific rigor and oversell performance, with many failing to define what they measure or conduct proper statistical analysis, prompting calls for stricter evaluation standards.
Google dramatically expands NotebookLM AI assistant with massive 1 million token context window, 6x longer conversational memory, 50% better response quality, and customizable personas that can act as specialized roles like PhD advisor or marketing strategist.