WaterPlum Targeted Developers via Fake Job Interviews
North Korea's WaterPlum infected 30,000 developer machines via fake interviews and VSCode repos. Here's what they stole and the 5 checks to run ...
Atria Dawn Preview: MIT-Licensed 744B Agentic MoE
Shanghai AI Lab quietly dropped Atria Dawn Preview — a 744B MoE model with MIT ...
AgentVerse-OS: Run Claude Code on Your Own Server (2026)
AgentVerse-OS is a new open-source personal cloud OS for running AI coding agents on your ...
Colibri: Run a 744B MoE Model Locally on 25GB RAM
The headline writes itself: run a 744-billion-parameter model on a laptop. The fine print is ...
Plugin4Shell: Zero-Click RCE in AI Coding Agents
Plugin4Shell is a zero-click RCE vulnerability in the plugin ecosystems of all four major AI ...
Sakana Fugu Max: Can a $6 Orchestrator Beat Frontier Models?
Sakana Fugu Max routes tasks across open-weight specialists at $6/M output — 60% below Sonnet 5. Here's the architecture, the benchmarks, and the ...
Grok 4.6 in Cursor: What the SpaceX GPUs Actually Do
Grok 4.6 is live in Cursor on all plans. Here is what SpaceX GPUs actually ...
.NET 11 RC1: C# 15 Union Types Are Production-Ready
.NET 11 RC1 ships with go-live support and C# 15 as default. Union types replace ...
GitHub Copilot Auto Model Tiers: Efficiency, Balance, or Intelligence?
GitHub Copilot auto model selection now has three tiers — Efficiency, Balance, and Intelligence. Here ...
Portless: Vercel Labs Kills Port Numbers for Developers and Agents
Vercel Labs shipped Portless — a CLI tool that replaces localhost port numbers with stable ...