AI News Feed
Page 5 of 19
China proposes export controls on 23 AI technologies, Meta tightens scraping terms, Nvidia launches RTX Spark, and Google Research shows agents running wet-lab experiments.
In this digest
Tencent open-sources Hy-Memory 2.0, beating Claude Opus 5 on memory benchmarks. Plus: HybridAgent from Stanford and Meta, and faster llama.cpp CPU inference.
In this digest
ChatGPT can now see your open Chrome tabs, IBM ships modular Granite Switch 4.1, and Anthropic gives METR employee-level access for safety evals.
In this digest
Meta ships an 816B open-weight multimodal model optimized for Mac Studios, plus llama.cpp v0.7.0, NVIDIA Nemotron 3.5 Super VL, and FLUX.3 Image Klein.
In this digest
StepFun's 600B MoE flagship leads a run that includes Microsoft's local Copilot model, Fish Audio's Drama 3 TTS, and Jais 3's Arabic LLM family.
In this digest
Eight states sue OpenAI over ChatGPT child safety, Flock's CEO contradicts its own site, and open-source fine-tunes claim sub-30B frontier.
In this digest
YouTube labels AI-generated music, Anthropic warns of self-replicating AI, NVIDIA drops an open 74B MoE model, and Baker lab designs novel proteases.
In this digest
NVIDIA open-sources the full Run:ai GPU orchestration platform under Apache 2.0, while a new C runtime trains LoRA adapters directly on quantized GGUF weights.
In this digest
An OpenAI engineer was arrested for planting a hidden recording device, Google launched Gemini 4 Argon, and a judge ordered OpenAI to hand over hiring records
In this digest
Google's Gemini 4 Argon leads a packed run with vibe-coded Play Store apps, Pluto Security's $42M raise, and a new on-device MoE model.
In this digest
Put practical data boundaries and usage policies in place before AI makes existing oversharing easier to find.
Explore Secure AI Adoption