Kimi K3 Launches—Moonshot’s 2.8T Open Model Challenges Labs
Moonshot AI released Kimi K3 on July 14, 2026: a 2.8T open model with a 1M-token context that challenges the proprietary labs and ships weights by…
Archive
Category: Artificial Intelligence
Coverage of AI models, machine learning research, generative AI, and artificial intelligence industry developments
Thinking Machines Ships Inkling—a 975B Open-Weight AI Model
Thinking Machines Lab shipped its first frontier model. Inkling is a Mixture-of-Experts transformer with 975B total parameters and 41B active per token, Thinking Machines…
Google’s Gemini 3.5 Pro Delayed—Coding Fell Short, Bloomberg Says
Alphabet shares sank roughly 4% on Thursday after a report that its most powerful AI model, Gemini 3.5 Pro, is “months behind schedule,” CNBC…
Anthropic Launches Claude for Teachers—Free Premium AI for US Educators
Claude for Teachers gives verified US K-12 educators free premium Claude, teaching skills, and FERPA-safe handling of student data in class.
Bonsai 27B Runs on a Phone—the First 27B-Class Model Small Enough for an iPhone
PrismML compressed a 27B vision-language model to ~1.125 bits per weight, fitting it on an iPhone and running chat, vision, and tools locally.
Spotify Launches ‘Talk to Spotify’—a ChatGPT-Like Assistant for Your Music
Spotify's new conversational assistant lets Premium users discover, control, and learn about music, podcasts, and audiobooks by voice or text.
ChatGPT Now Shows Live World Cup Odds—Via Kalshi
OpenAI has begun displaying Kalshi's World Cup prediction-market odds inside ChatGPT search, its first reported tie-up with a betting-style forecast platform.
The Brain Decides Earlier Than We Thought—and AI Could Learn
University of Illinois researchers found the brain begins deciding in its earliest sensory regions, a clue for building more efficient artificial intelligence.
60% Say AI in Brand Messaging Is a Turnoff, Not a Feature — WordPress VIP Research Reveals the Trust Gap
Dek: New Future of the Web 2026 survey finds consumers are hitting “bot fatigue” in 40 minutes, 86% still verify sources after AI summaries,…
What is ProgramBench: The Uncheatable Benchmark Where Every AI Model Scores 0%
Meta's new benchmark hands AI agents a compiled program and its documentation—no source code, no internet—and asks them to recreate it from scratch. Nine of the…