AINews The Architecture That Quietly Won: Mixture of Experts, From 1991 Curiosity to the Engine of Modern AI Kimi K3 just became the largest open-weight model in history at 2.8 trillion parameters, yet it only uses 1.8% of itself at…
AINews The Firms Spending the Most on AI Are Hiring the Most. Here’s What That Does and Doesn’t Prove. A new study of 21,559 US companies links actual AI spending to actual workforce records for the first time, and finds that…
AINews Alphabet’s $85 Billion Bet: What It Means When Google Can’t Keep Up With Its Own AI Demand The largest equity capital raise in corporate history, anchored by an unexpected $10 billion check from Warren Buffett’s Berkshire Hathaway, says something…
AINews What 45,000 Prompts Reveal About How AI Companies Test Child Safety Meta hired hundreds of contractors to pose as children and probe rival chatbots on suicide, drugs, and sex. The companies being tested…
AINews 94% of Companies Are Investing More in AI. Only 44% Understand the Risks They Are Taking. The Logicalis 2026 Global CIO Report surveyed over 1,000 technology leaders worldwide. The numbers paint a picture of an industry moving faster…
BreakthroughsNews MiniMax M3 and the Return of Sparse Attention: What Just Changed in the Long-Context Race A Shanghai lab just shipped a model that does at 1M tokens what full attention can’t do at any price. The architecture…
BreakthroughsNews One GPU to Train Them All: What MegaTrain Changes About Who Gets to Build AI Until now, training a 100B+ parameter model required a cluster, a multi-million dollar budget, and a very patient CFO. A new paper…
