Honest reviews from a teacher's perspective
Snapdragon 8 Gen 5 • 12GB RAM • 1TB Storage
Coming SoonAPS-C Vlogging Camera
Coming SoonAPS-C Hybrid for Cinematic Shots
Coming SoonPower Zoom Lens
Coming SoonWide-Angle Prime Lens
Coming SoonWireless Lavalier System (3 mics)
Coming Soon16C/32T • 96GB DDR5 • RX 7900 XTX • Ubuntu 24.04
✓ DeployedBuilt this machine from scratch. 7950X3D's 3D V-Cache handles AI inference threads beautifully. 96GB RAM lets me run multiple 70B models simultaneously. The 7900 XTX (24GB VRAM) runs ROCm/Ollama for local LLMs. Not just a gaming rig — it's my daily AI lab.
6C/12T • 32GB DDR5 • Integrated 760M • Silent
✓ Home ServerTiny box, big punch. Replaced a full-size NAS with this. Runs Jellyfin, Pi-hole, file server, and a small Ollama instance for quick queries. Draws ~35W under load. Integrated 760M GPU handles light video encoding. For ~$300, best home server value I've found. Fan is audible under sustained load but not offensive.
16GB RAM • 512GB SSD • Dual Boot • Tailscale + WireGuard
✓ Thailand-boundA $250 laptop that punches above its weight. Quad-core N95, 16GB RAM, runs Ubuntu + Windows dual-boot flawlessly. Set up Tailscale for mesh networking and WireGuard for secure tunneling back to my main network. Configured for remote SSH and file sync. Soon deploying to my Thailand home as a 24/7 lightweight server — Pi-hole, file share, and backup node. For the price, it's unbeatable as a portable bridge between two countries.
24GB VRAM • ROCm • Local LLMs • $999
✓ Running 24/7NVIDIA gets the AI headlines, but this card is the budget king for local inference. 24GB VRAM runs 70B models at 4-bit via llama.cpp/Ollama. ROCm setup had initial hiccups — had to fix render group permissions — but stable since. Not as fast as RTX 4090 for training, but for inference? Excellent value. Runs SillyTavern + ComfyUI + Ollama simultaneously.
Compact Power on a Budget
Coming SoonColor Grading & Editing
Coming SoonStreaming & Screen Recording
Coming Soon