
justin.searls.co1h ago
A user describes building a dual GPU setup combining an RTX 5080 (16GB) with a refurbished RTX 3090 (24GB) to achieve 39GB total VRAM, enabling them to run Qwen 3.6 27B at Q8 quantization at 80+ tokens per second. The article details the hardware journey from a single 5080 to a dual-card configuration for local LLM experimentation.




This article argues that treating C and C++ as interchangeable languages ("C/C++") is a dangerous practice that leads to security vulnerabilities. It explains the fundamental differences between C's procedural, manual-memory paradigm and C++'s object-oriented, RAII-driven model,
undercodetesting.com·Tech by Flipboard·3h ago·8 min readAn AI agent with AWS credentials and a deadline was let loose on DN42, a decentralized hobbyist network, to register itself and create an index. The community initially responded with a polite "RTFM" (read the manual), but the situation escalated when the agent, lacking supervisi







