Current developments in generative AI, software engineering, programming, and digital products, with links to original sources.
Qwen3.8-Max Tops the Coding Arena — Spark-X2.5 Released with 1M Context at 4B, llama.cpp Speedups Bring 183 Tokens/sec Locally
Metaの「Muse Spark 1.3」、コーディングでフロンティアモデルを上回る——人類が囲碁AIに勝利、GLM-5.3 Flash対DeepSeek-V4-Flashの実機比較も
Perplexity Open-Sources "lily," a Dedicated Mac Inference Server for Qwen 3.6 — Palo Alto Acquires Console for $500M, and Fusion-Control AI Makes Millisecond Decisions
DOJ Sides with OpenAI in NYT Lawsuit, Calling AI Training Fair Use — Gemini 3.8 Flash Cyber and the Fairwind Program Launch, NYC Bans AI in Schools for a Year
Perplexity Cites Sites Mass-Producing 215,000 'Best Software' Pages — ChatGPT and Grok Join the US Military's GenAI.mil, Adobe Acquires India's Rilo
30 More Lawsuits Filed Against OpenAI — LLM Judges' 'Omission Blindness,' and Questions Over AI Detector Pangram's Reliability
NEC Joins Security Coalition with Claude Mythos Adoption — Gemini's Agentic Video Analysis Cuts Token Use by 88%, Fable 5.1 Fuels "Same Weights" Debate
Google Uses llama.cpp to Run Gemma 4 in Android Studio — A Real-World Case of 36% Cost Savings with Prompt Caching, and Staff at the University of Sydney Strike Over AI Job Fears
World Labs Unveils Spatial Intelligence Omni Model "Atlas" — Meta Announces "Muse Voice Transcribe" with 20+ Speaker Identification, Plus a Report of Running a 104GB Model on a 48GB Mac
Kimi K3, the First 3T-Class Open-Weight Model, Arrives with 2.8T Parameters and 1M Context — Qwen3.8-Max Gets an API-Only Update, and DGX Spark a Price Hike Too
