AI
Breaking AI announcements, model launches, company news
The lead
Latest first
Browse AI by topic
Open-Source AI
Hugging Face, GitHub, Ollama, local models, agent frameworks
Explore 55 postsModel Launches
The latest in Model Launches — sourced and dated.
Explore 43 postsAI Agents & MCP
The latest in AI Agents & MCP — sourced and dated.
Explore 29 postsAI Use Cases
Real examples of what people are doing with AI
Explore 22 postsResearch
The latest in Research — sourced and dated.
Explore 10 postsPolicy & Safety
The latest in Policy & Safety — sourced and dated.
Explore 1 postAI Tools
Curated directory of AI tools by category
Explore 1 postLocal AI
The latest in Local AI — sourced and dated.
ExploreLatest in AI
How an Astrophysicist Uses Codex to Simulate Black Holes
Chi-kwan Chan uses OpenAI Codex to crack a decades-old plasma physics bottleneck, enabling the first black hole video for the Event Horizon…
US Bans Anthropic Fable 5 & Mythos 5 — Export Control Order
US export control forces Anthropic to suspend Fable 5 and Mythos 5 globally. Gov cites jailbreak; Anthropic calls it a misunderstanding. Venice…
Ollama’s MLX Mac update boosts LLM inference speed by 20%
Ollama's June 11 MLX update brings NVFP4 quantization, snapshot-based prefix caching for agents, and 20% faster output on Apple Silicon. Here's what…
DeepSeek V4-Pro’s MIT License Changes Everything
DeepSeek V4-Pro ships MIT-licensed with a 1M-token context and 93.5 LiveCodeBench — the open model that finally closes the legal gap with…
Run a Local LLM With Ollama in 5 Minutes
How to run a local LLM with Ollama in 2026: install, pick a model sized for your RAM, start chatting, hit the…
Power Apps MCP Learns From Its Mistakes
Microsoft's Power Apps MCP server turns user corrections into org-wide agent improvements — a structural fix for enterprise AI's stubborn last-mile problem.
Connect an MCP Server in Five Minutes
Step-by-step guide to connecting an MCP server to your AI assistant in 2026: prerequisites, config, first server, troubleshooting, and scoping it safely.
MiniMax M3 Cracks Open the 1M-Context Frontier
MiniMax M3 is an open-weight model with a 1M-token context window and 59% SWE-Bench Pro score — here's exactly what it means…
What Is Model Quantization? The 4-Bit Trick
Model quantization shrinks AI weights from 16-bit to 4-bit, cutting VRAM by 75%. Here's how it works, where quality breaks, and how…
What Is RAG? The Retrieval Problem Nobody Talks About
What is RAG? Retrieval-augmented generation makes AI search your own docs before answering — and the real failure mode is bad retrieval,…