Current developments in generative AI, software engineering, programming, and digital products, with links to original sources.
Qwen3.8-27B Reportedly Beats Opus at Malware Analysis, Speculative Decoding Delivers 3x Speedups on Apple Silicon — AI News, August 15
Anthropic's Q2 Revenue Tops $11.5 Billion After 14x Growth, Claude Watermark Detection API on the Way, US Tells Allies to 'Pick a Side' — AI News for August 15
How Claude's Text Watermarking Works Is Now Technically Clear, macOS Screen Sharing Vulnerability Actively Exploited, and the Local LLM Scene Is Buzzing Over Qwen3.8-27B — AI News for August 15
DeepSeek Open-Sources Agent Harness "dsh", V4-Pro Hits 17,000 tok/s on B300×8, and Apple Core AI's MoE Gets a Metal Speedup — AI News for August 15
OpenAI and Anthropic Enter a Price War, Qwen3.8-27B Evolves by Leaps Without Changing Its Architecture, and Google to Allow AI Watermark Removal — AI News for August 15
SpaceX Completes $60B Cursor Acquisition, Four Major Open Models from China in One Month, French Top Court Rules Under-15 Social Media Ban Unconstitutional — AI News for August 15
Claude Code Runs Anthropic's Internal Software Maintenance at a 46% Merge Rate, llama.cpp Adds Sandboxed Execution, and a Philosophy Journal "Deliberately" Publishes an AI-Written Article — AI News for August 14
Liquid AI's 3B On-Device VLM, Qwen3.8-27B Model Card Appears, Bill to Ban 'Marrying' Chatbots — AI News for August 14
ChatGPT Remembers Your PC Activity, GPT-5.6 Sol Gets an Ultrafast Mode, GLM-5.3 Weights Released — AI News for August 14
Z.ai Announces GLM-5.3, Autonomous AI Cyberattack Hits Taiwan — AI News for August 14
