This site is not a news feed—it is a topic-organized knowledge base. Content is split into four tracks with different difficulty levels and audiences. Choose your path first, then start reading.

Four Content Tracks

TrackWhat It IsBest For
FundamentalsCore AI concepts, 8 articles from basics to productionEveryone—start here
Frontier TechMCP, Skills, Harness, and other industry standardsAfter fundamentals; some articles are engineering-heavy
ModelsText, vision, audio, embeddings, roboticsUnderstanding the model landscape
Case StudiesHow Cursor, Claude Code, Shopify, and others build real productsAnyone who wants to see “how others did it”

Beginner Path (zero background / non-technical)

1. This article (you are here)
2. What Is AI? Three Metaphors
3. How LLMs "Talk": Tokens, Context, and Probability
4. Prompt Basics
5. MCP Intro (frontier track, but written accessibly)

After this path you will:

  • Know what AI can and cannot do
  • Use prompts to get better answers
  • Understand why MCP is everywhere

Skip for now: Fine-tuning vs RAG, Harness deep dives, Eval systems—come back when you have a real use case.

Engineer Path (programming experience, want to ship AI)

Fundamentals 01–08 (focus on 04 RAG, 05 Agent, 08 Production Checklist)

Full frontier track: MCP Deep Dive → Skills → Harness → Context Engineering → Eval

Case studies: Cursor ecosystem → Claude Code → OpenAI Responses API

Architect / Tech Lead Path

Fundamentals: read only 06 (architecture choices) and 08 (production checklist)

Frontier: Harness + Context Engineering + Eval (these three set your team's engineering ceiling)

Read all case studies; compare against your own product to find gaps

About the Lv.1–5 Difficulty Labels

  • Lv.1–2: No programming or math required
  • Lv.3: Basic engineering concepts (APIs, databases) helpful
  • Lv.4–5: For engineers building AI products

About the Beginner / Professional Labels

  • Beginner: Deliberately avoids formulas and code; uses metaphors and flowcharts
  • Professional: Assumes you have run a RAG or Agent demo; focuses on design and pitfalls
  • All: Readable by both; professionals may find some articles shallow

How to Read Efficiently

  1. Do not read cover to cover—jump by path
  2. Cross-check case studies with official docs—articles link to primary sources
  3. Each article ends with “Next up”—follow order within a track when possible
  4. Questions? Email [email protected]

Pick a path and get started.