AI SDK is the TypeScript SDK for building AI applications, features, frameworks, and agents across any model provider. AI SDK 7 focuses on what it takes to run AI in production.

Explore how the GitHub Copilot agentic harness delivers strong results across multiple benchmarks and leading token efficiency.

Researchers introduce generative causal testing, which translates black box models into clear hypotheses and verifies them in the scanner, revealing what specific brain regions respond to in language.

A new OpenAI research paper shows how AI agents are transforming work, enabling longer, more complex tasks and expanding productivity across roles.

Zhipu AI's GLM-5.2 nearly matches Claude Opus 4.7 in a Snowflake benchmark with 103 coding tasks at one-fifth the cost per output token. But the Chinese model burns through nearly twice as many tokens per task. Still, that pricing gap is putting real pressure on Anthropic and OpenAI, and could rattle the valuations of Western AI labs.

Xiaomi's HarnessX autonomously rewrites AI agent harnesses mid-execution, delivering +14.5% avg performance gains — and +44% for smaller open-weight models.

Agent reliability is not a mysterious model property — it emerges from a control loop where correctness is continuously verified; open loops amplify drift.

Ultra Code is Claude Code

How Anthropic’s new tool, Claude Design, is creating overnight web-design clichés.

This new microchip architecture from IBM builds up, not out, to overcome the spatial limitations of scaling transistor density.

We’re on a journey to advance and democratize artificial intelligence through open source and open science.

Thariq Shihipar, engineering lead for the Claude Code team, recently published a blog post (Using Claude Code: The Unreasonable Effectiveness of HTML) arguing that HTML, with its richer visualizations, color, and interactivity, improves the productivity of human-agent communication in many settings, especially when compared to default Markdown outputs.

Cloud operations are entering a new era as AI-driven and autonomous agents become a larger part of modern software systems. As software becomes increasingly agentic, the challenge is no longer just managing greater scale and complexity. Operators must also contend with systems that evolve faster, act more autonomously and interact across an expanding network of...

In this article, you will learn why a large context window is not the same thing as agent memory, and how techniques like retrieval, compression, and summarization fit together in an agent’s cognitive stack.

OpenAI and Broadcom introduce Jalapeño, a custom AI chip built for LLM inference to improve performance, efficiency, and scale across AI systems.

Claude Tag gives enterprise teams a persistent, multiplayer AI presence in Slack — one that operates under its own identity.

Make Interfaces Feel Better An [Agent Skill]( based on the article [Details that make interfaces feel better]( This skill teaches AI coding assistants (Claude Code, Codex, etc.) the small design engineering details that compound into a great interface. What it covers - Text wrapping (`text-wrap: balance` / `pretty`) - Concentric border radius for nested elements -

A collaborative AI workspace, built on your company context. Build and orchestrate agents right alongside your team

I'm joining OpenAI next week!🥹 The job search turned out to be really challenging but also super rewarding, so I wrote a small blog to share what I learned along the way and hopefully make the process a little less mysterious for the next person.

One model to command them all

Introducing Clips - 100% free, open source, agent-native alternative to Loom Unlike Loom, agent's can fully understand Clips just from a URL. Every Clip comes with APIs and metadata for agents to explore their contents. Agents can "see and hear" anything in a Clip - not just transcripts, but everything visually in the video at any timestamp. Easily share bug reports, feedback, analyses, or anything else in a way that you can easily pass to agents to use to improve products, reports, or

Astro 7 is here! A new Rust compiler, a new Rust Markdown/MDX processor, Vite 8 and more. Get ready for 60%+ faster builds.

How AI coding models calibrate effort, token cost, and tool use on even the simplest task, and what that means for model selection and cost.

Explore the astonishing amount of music available to AI developers.

See how AMD Instinct GPUs deliver MLPerf Training 6.0 results across LLM workloads, multi-node FLUX.1 scale and partner validation.

Retrieval-augmented generation, or RAG, is a pattern where an application retrieves relevant source material and adds it to a model prompt so the model can answer from that context. A larger context w

Atlassian details the Forge billing platform built for usage-based pricing across its cloud ecosystem. It processes large-scale usage events with correct attribution, deduplication, and aggregation using a streaming pipeline, idempotent processing, and layered storage to enable accurate billing, near real-time visibility, and reliable reconciliation across distributed services.

Nvidia's Nader Khalil on backing OpenClaw, building agent blueprints, and why every enterprise will soon ship its own specialized AI agents.

Learn how GitHub built Qubot, our internal Copilot-powered analytics agent, to allow any GitHub employee to ask questions about our data in plain language.

Building small models for everyone

LlamaIndex is a simple, flexible framework for building knowledge assistants using LLMs connected to your enterprise data.

Most LLM applications need a clear workflow, not an autonomous agent. Here's how to build one in plain Python.

Cursor's Origin, GitLab's Project Switch and Zed's DeltaDB are racing to rebuild code hosting for AI agents as GitHub buckles under the load.

Arbor separates strategy from execution using isolated git worktrees, so engineering teams can finally trace which optimization actually moved the needle.

Lately, there's been more focus on creating specialized Small Language Models (SLMs) for high-throughput, real-time applications. But we seem to be at an impasse: we excel at fine-tuning these models,

Train Your Own Encoder-Free VLM in $100

Every enterprise company is seemingly trying to adopt the Model Context Protocol (MCP) to connect its AI agents to tools. But so

Sandboxes are usually designed around what we think agents will need. VMs are designed around what agents actually do: use computers.

An open specification for finding and verifying tools, skills, and agents across the web.Agents are ...

Stanford's DeLM lets AI agents coordinate without a central controller, cutting multi-agent inference costs 50% and beating SWE-bench baselines by 10.5%.

The Figma MCP server extends across our platform. From FigJam to Figma Slides, Figma Make, and the Figma agent, here are four ways we’re using it.

Just Shipped: Flue 1.0 Beta Flue is the TypeScript framework for building the next generation of agents, designed around an open agent harness with zero LLM lock-in. It’s like Astro, for agents. Flue 1.0 has been redesigned around three core primitives: 🔁 Workflows — structured automations designed for background work, where your code drives the agent from start to finish. 🧭 Agents (New!) — autonomous, stateful loops where the model drives itself to complete a given task. 📡 Channels

PlanetScale offers the world’s fastest and most scalable cloud hosting for Vitess and Postgres.

Kubernetes is a framework for feedback controllers: write down what you want, observe what exists, make the next change, and repeat.

Learn how to optimize your website for Google Search's generative AI features, including official best practices, technical SEO advice, and emerging AI agent guidance.

A detailed look at MCP that turned my scattered tool definitions into a stable, discoverable server

Create parallel Claude Code, Codex, and Cursor agents in isolated workspaces. See at a glance what they're working on, then review and merge their changes.

In this article, we explore what makes AI agent tools work well and the common design mistakes that cause failures. Learn how tool design affects an agent's ability to complete tasks accurately and consistently.

Z.ai launched GLM-5.2 on June 13, 2026, across every GLM Coding Plan tier. The headline is a usable 1-million-token context window plus High and Max effort levels. It drops into Claude Code, Cline, and OpenClaw through an Anthropic-compatible endpoint. No benchmarks shipped at launch, and MIT open weights are promised next week.

Google Open Knowledge Format standardizes how organizational knowledge can be shared between AI agents, tools, and teams.

Ponytail He says nothing. He writes one line. It works. <img

The directive would even bar Anthropic's own foreign employees from using Fable and Mythos. Anthropic called the government position "a misunderstanding".

UC Berkeley's PixelRAG renders pages as screenshots instead of parsing text, boosting RAG accuracy by up to 18.1% and cutting AI agent token costs 10x.

Use Claude on Apple platforms through the Foundation Models framework with the Claude for Foundation Models Swift package.

Mobile teams have been asking for a Core Web Vitals equivalent for years. The Core Mobile Vitals initiative is built using the same rigor, research, and user focus.