zeroshotmind/open-blog

The open content layer for zeroshotmind.com — a technical blog focused on ML systems, inference, training internals, interpretability, and the math behind modern AI.

Posts in this repo are published directly to zeroshotmind.com. Anyone can contribute: open a pull request, and if it fits the site's depth and quality bar, it goes live.

What gets published here

zeroshotmind.com covers topics like:

• LLM inference systems (batching, KV cache, memory/compute trade-offs)
• Training internals (optimizers, normalization, attention variants)
• Mechanistic interpretability
• Reinforcement learning for language models
• Diffusion models and generative systems
• GPU kernels and systems-level ML
• Mathematical foundations of deep learning

Posts should be technical and substantive. Cheatsheets, deep dives, visual explainers, and interactive posts are all welcome.

How to contribute

1. Fork this repo
2. Add your post to posts/YYYY/ (see CONTRIBUTING.md for the full spec)
3. Open a pull request

See CONTRIBUTING.md for frontmatter requirements, component rules, and style guidance.

Repo structure

posts/
  2024/          # flat .mdx files (no custom components)
  2025/
    slug.mdx                   # flat post
    slug-with-component/       # post + its React component
      index.mdx
      MyComponent.tsx
  2026/
POST_TEMPLATE.mdx   # starter template
CONTRIBUTING.md