Google has released AX, an open framework for running and coordinating multiple AI agents. Here's what it does and why teams building agentic systems should take a look.
Startup PrismML is building a small-footprint language model designed to run locally on consumer hardware — a direct challenge to the assumption that useful AI requires massive cloud infrastructure.
A new WhatsApp Business MCP server lets developers hand off account setup, message template creation, and API testing to AI coding agents like Claude, Cursor, or ChatGPT.
Inception Labs released Mercury 2.5, a diffusion language model for code that delivers competitive benchmark results at significantly lower inference latency than comparable autoregressive models.
A new OpenAI model listing, GPT-6 Astra, has surfaced on OpenRouter and is drawing attention from developers. Here's how to evaluate it before wiring it into production.
A developer has rendered the entire city of San Francisco as a navigable 3D environment in the browser. It's a striking demonstration of what's now possible with public geodata and real-time rendering.
A lightweight open-source prompt collection forces Claude to drop the buzzwords, bullet-point overload, and hollow affirmations — producing tighter, more direct output for technical use cases.
Cerebras has announced the CS-4, its fourth-generation wafer-scale engine designed to dramatically accelerate large model training and inference. Here's what it means for teams running serious AI workloads.
Cloudflare has released Kitesurf, a cloud-hosted browser engineered for AI agents rather than human users. It consumes significantly less compute than Chromium for typical automation workloads, lowering the cost and complexity of browser-based agent development.
Tencent's Hunyuan3D team unveiled WorldClaw, a system for building expansive, explorable 3D environments through an agentic pipeline rather than manual scene assembly.
Hark is previewing a browser-use agent it says outperforms competitors on both speed and cost. Here's what builders should know about where this fits in the growing stack of web automation tools.
Cloudflare details how it serves open-weight models like Kimi and GLM more efficiently across its network. Here's why it matters for anyone building on serverless inference.
