With egui, ort and webgpu. using a 60 USD used Thinkpad to run AIs for computer vision and bioacoustics in the GPU (2 cores, integrated GPU)

My laptop kept freezing under memory pressure with no dedicated GPU VRAM, so I wrote a userspace daemon to fix it

Building a commercial OFX plugin for DaVinci Resolve in Rust (wgpu + vello): GPU interop war stories

A packed-bit GPU cellular fabric in Rust: Conway's Life at 6.7T cell-updates/sec, bit-exact vs CPU

memory ssa for numa/gpu

push2talk: a Tauri v2 + whisper-rs push-to-talk dictation app — raw evdev for hotkeys, ydotool for typing, GPU-accelerated whisper.cpp

[Show & Tell] RamShared — idle GPU memory as a backup cushion on Linux/WSL2 (when RAM is tight, borrow the GPU — give it back if the GPU needs it)

How to build a GPU

WebGL Without a GPU

Popping the GPU Bubble