AI

Hands-on guides to LLMs, agents, prompt engineering, and the AI tools Botmonster runs every day for real work, not demos.

RTX 5080 vs. RTX 5090: the best GPU for local AI workloads in 2026

Compare the RTX 5080 and 5090 for local AI in 2026: LLM inference benchmarks, image generation speed, power draw, and a clear value verdict.

Local AI image upscaling: Real-ESRGAN vs. Topaz vs. SUPIR

Real-ESRGAN is free and fast. Topaz Photo AI wins on quality and face recovery. SUPIR makes the most detail but needs 12GB VRAM. Speed and cost too.

Gemma 4 architecture explained: per-layer embeddings, shared KV cache, and dual RoPE

Gemma 4 uses per-layer embeddings, a shared KV cache to cut memory, and dual RoPE for mixed local-global attention. Activates 3.8B of 26B params.

Running Gemma 4 26B MoE on 8GB VRAM: three strategies that work

Run Google Gemma 4 26B MoE on a budget 8GB VRAM GPU using aggressive quantization, GPU-CPU layer offloading, and tensor parallelism. Full setup guide.

AI coding agents are insider threats: prompt injection, MCP exploits, and supply chain attacks

Study of 78 AI coding agents including Claude Code, Copilot, and Cursor: all fall to prompt injection, which succeeds 85% of the time via MCP exploits.

Running Gemma 4 locally with Ollama: all four model sizes compared

Google's Gemma 4 spans 2.3B to 31B parameters. Compare hardware needs, speed, and quality across all variants when running locally through Ollama.