karminski/one-small-step

★ 6,933PythonAudience · generalComplexity · 1/5LicenseSetup · easy

Mindmap

mindmap
  root((One Small Step))
    What it does
      Short tech explainers
      Five minutes per article
      Chinese language
    Topic areas
      AI and LLMs
      Mathematics basics
      Hardware concepts
      System internals
    AI topics covered
      Transformer architecture
      Model quantization
      RAG
      LoRA fine-tuning
    Format
      Markdown articles
      Organized by date
      Updated 3x per week

mindmap root((One Small Step)) What it does Short tech explainers Five minutes per article Chinese language Topic areas AI and LLMs Mathematics basics Hardware concepts System internals AI topics covered Transformer architecture Model quantization RAG LoRA fine-tuning Format Markdown articles Organized by date Updated 3x per week

Click or tap to explore — scroll the page freely

Things people build with this

USE CASE 1

Read a five-minute article to understand a specific AI concept like RAG, LoRA fine-tuning, or model quantization without needing a math background.

USE CASE 2

Browse the hardware section to understand how NVMe SSDs, CPU caches, or PCIe memory work in plain terms.

USE CASE 3

Use the articles as reference material when learning about large language models or vector databases before starting an AI project.

Tech stack

PythonMarkdown

Getting it running

Difficulty · easy Time to first run · 5min

No software to install, just browse the Markdown articles directly on GitHub.

Open source under the MIT license, free to read, share, and use for any purpose.

In plain English

One Small Step is a Chinese-language technology education project that publishes short explainer articles on computing concepts. Each article is written to be readable in about five minutes. The project aims to make advanced technical ideas accessible to a general audience rather than assuming deep engineering background. The articles are organized into several topic areas. The largest section covers AI and large language models, with pieces explaining concepts like the Transformer architecture, model quantization, speculative decoding, fine-tuning with LoRA, retrieval-augmented generation (RAG), vector databases, AI hallucination, Flash Attention, and more. There are also shorter sections covering mathematics concepts (such as matrix rank and overfitting), system-level topics (like how Windows Task Manager reports memory), and hardware topics (including NVMe SSD design, CPU cache levels, and PCIe memory expansion). The repository is updated at least three times per week according to the README. Articles are stored as Markdown files organized by date and topic in subdirectories. Readers can browse the article list directly on GitHub. The project is written entirely in Chinese (Simplified) and is maintained by a single contributor who goes by karminski. Community contributions and corrections are welcome through issues or pull requests. The project is open source under the MIT license. The README does not describe any runnable software. The Python language tag in the repository likely reflects a small number of code examples or utility scripts rather than a standalone application.

Copy-paste prompts

Prompt 1

I'm reading the one-small-step article on RAG (retrieval-augmented generation). Explain the concept in even simpler terms and give me an analogy I can use to explain it to a non-technical friend.

Prompt 2

I want to understand model quantization before I start working with local AI models. Summarize what quantization does, why it matters, and what tradeoff it makes, in under 200 words.

Prompt 3

I read the one-small-step article on Flash Attention but want to understand why it matters for running AI models. Explain it using a simple analogy involving reading a book.

Prompt 4

I'm starting to learn about vector databases after reading the one-small-step intro. What should I learn next, and what small project could I build to practice the concept?

Open on GitHub → Explain another repo

← karminski on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.