Cloudflare Unweight: Lossless LLM Compression on H100
Cloudflare's Unweight compresses LLM BF16 weights 15–22% with zero quality loss. Open-source H100 GPU kernels, 3 GB VRAM savings on Llama-3.1-8B. Here's how ...
Meta Muse Glimmer 30B: Local AI Agent on One GPU
Meta released Muse Glimmer on August 10 — a 30B open-weight model built for local ...
X Open-Sources Its For You Algorithm: Developer Guide
X released the full source code for its For You feed ranking algorithm on GitHub ...
DeepSeek V4 Pro 0813 Goes GA: What Developers Need to Know
DeepSeek V4 Pro 0813 is now GA with 1M context, enhanced agent benchmarks, and API ...
GeoServer Zero-Day: No Patch, Active Exploits, Act Now
An unpatched SQL injection zero-day in GeoServer is being actively probed. No CVE, no patch. ...
Stagehand v4: Browser Agent SDK Now 2x Faster
Stagehand v4 ships August 10 with an extension-based architecture that is 2x faster than Playwright. Python and Go SDKs now at full feature ...
DeepSeek Harness v0.1: Open-Source Agent, Plugin by Plugin
DeepSeek Harness v0.1 is live in developer preview — MIT-licensed, model-agnostic, and built on a ...
Cloudflare OS Is Open Source: What Developers Get
Cloudflare open-sourced its AI agent workspace on August 5, 2026. Here is how the Gatekeepers ...
uBlock Origin Gives Up on Facebook Ads — Use This Instead
uBlock Origin's filter team quit chasing Facebook ads this week. Filter lists can't win this ...
Meta Muse Glimmer: The 30B Open Agent Model That Runs on Your Own GPU
Meta just open-sourced Muse Glimmer — a 30-billion-parameter model built specifically for autonomous agent workflows, ...