Stories by nathan-osman

How we made AWS Trainium 17x faster (for conv1d)

Show HN: Tiny Diffusion – A character-level text diffusion model from scratch

The Manuscripts of Edsger W. Dijkstra

Streaming DiLoCo: Towards a Distributed Free Lunch

Fast-DLLM: Training-Free Acceleration of Diffusion LLM

BERT is just a single text diffusion step

Rustree: Visualize Rust ASTs in the Browser

Diffusion is not necessarily Spectral Autoregression

The End of Computing Science (2000) – Edsger W. Dijkstra [pdf]

Planner: Generating Diversified Paragraph via Latent Language Diffusion Model

Running GPT-2 in WebGL: Rediscovering the Lost Art of GPU Shader Programming

Show HN: A Poor Man's Apple Intelligence

Intro to Autograd Engines: Karpathy's Micrograd Implemented in Go

Show HN: GPT-2 implemented using graphics shaders

Show HN: Visual AI research exploration to replace chatbots with concept trails

Train Custom Image Classifiers in Seconds with as Few as 100 Images per Class [video]

Fixing iOS code-block font-size rendering quirk

1 2