
v0's Design Systems 2.0 Stops Making Teams Re-explain Their Components Every Chat
v0 Design Systems 2.0 lets teams teach v0 their real component library once, then generate on-brand UIs from any chat
20 stories

v0 Design Systems 2.0 lets teams teach v0 their real component library once, then generate on-brand UIs from any chat

World Labs brings its Marble Pro world model to SIGGRAPH 2026 via a free two-day hackathon, putting spatial AI tools directly in builders' hands

Goodfire erased a language model's German ability by tuning a single number — and left French and Spanish almost completely intact

Ai2's token-level analysis of OLMo 3 vs. OLMo Hybrid reveals exactly where recurrent layers beat attention — and where they don't

Gemini 3.5 Flash now has computer use built in as a native tool, letting agents see and control browsers, mobile, and desktop interfaces alongside Search and Maps in a single model call.

Microsoft's MAI-Image-2.5 lands at #2 for image editing and #2 for text-to-image on Artificial Analysis's Arena, signaling a credible first-party challenge to OpenAI and Google in the image generation market.
General Intuition raises $320M Series A at a $2.3B valuation to train large action models on billions of ground-truth-labeled gameplay clips from Medal's 17M monthly users — and deploy them in robots, drones, and beyond.

Gemini's new Study notebooks add adaptive quizzes, personalized lesson plans, and NotebookLM sync — free for all users globally

Manus launches a native Semrush MCP connector, letting users pull live SEO data, competitor analysis, and automated reports through plain-language prompts
Hugging Face spotlights its Local Apps ecosystem, where 45K+ GGUF models can be run privately on your own hardware with one command.

KTransformers now runs DeepSeek V4 Flash at 1M context on 2 consumer GPUs by splitting MoE experts between GPU VRAM and system RAM

Liquid AI's 230M-parameter open-weight model runs at 213 tok/s on a phone CPU, beats 1B-parameter rivals at tool use and data extraction

Liquid AI's 230M-parameter LFM2.5 runs at 213 tok/s on a Galaxy S25 Ultra and beats models twice its size on tool use and data extraction

Vercel's AI SDK 7 lands with durable agents, unified reasoning control, tool approvals, MCP Apps, and a new harness abstraction for Claude Code and Codex

Apify's open-source agent-skills repo gives Claude Code, Cursor, and Gemini CLI instant access to 30,000+ web scraping tools with zero glue code

mcp-unity gives AI coding assistants like Cursor and Claude Code direct, live control over the Unity Editor — no more copy-pasting code and clicking manually

TensorRT 11.0 drops legacy APIs, ships multi-GPU inference as GA, and makes strongly typed networks the new default

Qwen releases the first open language world model that can simulate 7 agent environments and directly improve downstream agent performance across benchmarks it was never trained on.

nanobot packs a full personal AI agent into ~4,000 lines of Python, with 20+ LLM providers, 14 chat channels, and a novel two-stage memory system

A 9B GGUF build of Ornith-1.0 is live on Hugging Face, offering an open agentic coding model that runs on a single GPU.