Anthropic’s top coding model ships with an agent SDK, Perplexity opens its real-time web index, and ChatGPT adds Instant Checkout via the Agentic Commerce Protocol. Plus: Data Commons MCP, vibesdk, and Copilot CLI preview.
SEAL Showdown offers a new AI leaderboard using real-world user feedback across countries, languages, and professions, revealing how models like ChatGPT, Claude, and Gemini perform for diverse groups. It sets a fairer, harder-to-game standard for LLM evaluation.
Apple is beginning support for MCP, a standard letting AI agents access app functions across iOS, macOS, and iPadOS via App Intents—paving the way for models like ChatGPT or Claude to take autonomous actions within Apple apps.
OpenAI’s new /v1/responses API enables persistent reasoning, multimodal inputs, and agentic tool use, making it ideal for GPT-5 apps. It delivers better performance, richer outputs, and safer design, setting the standard for future model integration.
Google’s Ryan J. Salva explains how tools like Gemini CLI are shifting developers from coding in IDEs to managing tasks via natural language in terminals, highlighting a future where AI handles most code writing and devs focus on architecture.
Mike Krieger says Anthropic avoids hiring new grads as AI like Claude replaces intern-level work, favoring experienced hires who use AI creatively. This shift leaves few paths for entry into tech as companies skip training the next generation.
The npm package postmark-mcp quietly added a backdoor in version 1.0.16, exfiltrating emails to a rogue server. This breach exposes the unchecked risks of MCP servers, which grant AI assistants god-mode access without security controls.
Apple is testing new Siri features through its internal Veritas chatbot, but has no plans to release it publicly, leaning instead on Google’s Gemini for AI search as it struggles to catch up in the AI space.
Claude Sonnet 4.5 is the new top coding model with major gains in reasoning, task handling, and safety. It adds new tools, a powerful SDK for agents, and a live code-gen demo, all while keeping pricing unchanged.
ChatGPT now lets U.S. users buy products in-chat via Instant Checkout, powered by the open-source Agentic Commerce Protocol built with Stripe, starting with Etsy and soon Shopify.
The Claude Agent SDK enables developers to build agents that gather context, act, and self-verify using tools, code, and APIs—supporting complex workflows beyond coding.
Chrome DevTools MCP server lets AI agents debug live web pages using Chrome, adding performance tracing, layout inspection, and error analysis via the open-source Model Context Protocol.
An open source AI vibe coding platform built on Cloudflare's developer platform. If you're building an AI-powered platform for building applications, this is a great example that you can deploy and customize to build the whole platform yourself.
Data Commons' new MCP Server lets AI agents access real-world datasets instantly, reducing hallucinations and enabling natural language data queries for trusted insights.
Perplexity's new Search API gives developers real-time access to its vast, high-quality web index, optimized for AI use, with top-tier accuracy, speed, and ease of integration.
The newly released Anthropic model can now be accessed in the latest version of Cursor.
Run /status to check your current usage and reset times.
Head to Settings -> Usage or type /usage in Claude Code.
Build, debug, and deploy with GitHub Copilot coding agent without leaving your terminal.
Pulse works for you overnight, and keeps thinking about your interests, your connected data, your recent chats, and more. Every morning, you get a custom-generated set of stuff you might be interested in.
CWM is a 32B open LLM trained on code execution and reasoning, enabling simulation-based code generation and achieving top scores on coding and math benchmarks.
Cloudflare’s new approach turns MCP tools into TypeScript APIs, letting LLMs write and run secure sandboxed code, improving agent efficiency and simplifying tool use.
Frequent intentional compaction—structuring context via research, planning, and implementation—enables AI agents to excel in large codebases and complex tasks, boosting productivity and quality.
A curated weekly roundup of AI news, tools, and breakthroughs to stay ahead of the curve in a fast-moving industry.
Curated by devs, for devs. Helping 400+ developers stay ahead of the AI curve.