# AI-blogs - [The Agent Harness: The Infrastructure Layer That Makes AI Agents Actually Work](https://snowan.gitbook.io/study-notes/ai-blogs/agent-harness-deep-dive.md) - [AI Inference Batching: Static, Dynamic, and Continuous Batching Explained](https://snowan.gitbook.io/study-notes/ai-blogs/ai-inference-batching-explained.md): A comprehensive guide to AI inference batching mechanisms. Learn when to use static, dynamic, or continuous batching, understand the tradeoffs, and make informed decisions for your AI inference API de - [The Complete Guide to Building Skills for Claude — Summary & Key Takeaways](https://snowan.gitbook.io/study-notes/ai-blogs/anthropic-skills-guide-summary.md) - [12 Ways to Customize Claude Code — Boris Cherny's Latest Guide](https://snowan.gitbook.io/study-notes/ai-blogs/boris-cherny-customize-claude-code.md) - [Boris Cherny: How the Creator of Claude Code Actually Works](https://snowan.gitbook.io/study-notes/ai-blogs/boris-cherny-on-building-with-ai.md) - [System Design: Designing Chess.com](https://snowan.gitbook.io/study-notes/ai-blogs/chess-system-design.md) - [How Claude Code Agent Teams Actually Works - reverse Claude Code Agent Teams use CC](https://snowan.gitbook.io/study-notes/ai-blogs/claude-code-agent-teams.md) - [Deep Dive: Claude Code Memory Architecture](https://snowan.gitbook.io/study-notes/ai-blogs/claude-code-memory-architecture.md) - [Dario Amodei: "We Are Near the End of the Exponential"](https://snowan.gitbook.io/study-notes/ai-blogs/dario-amodei-end-of-exponential-dwarkesh-2026.md) - [Demystifying Agent Harnesses: The Infrastructure Layer That Actually Makes AI Agents Work](https://snowan.gitbook.io/study-notes/ai-blogs/demystifying-agent-harness.md) - [System Design: ChatGPT — An AI Inference Platform at Scale](https://snowan.gitbook.io/study-notes/ai-blogs/design-chatgpt-ai-inference-platform.md) - [System Design: Distributed Crossword Puzzle Solver](https://snowan.gitbook.io/study-notes/ai-blogs/design-crossword-puzzle-solver.md) - [System Design: Google Calendar](https://snowan.gitbook.io/study-notes/ai-blogs/design-google-calendar.md) - [System Design: In-Memory Database](https://snowan.gitbook.io/study-notes/ai-blogs/design-in-memory-database.md) - [System Design: Multi-Tenant URL Shortener with Organization Namespaces](https://snowan.gitbook.io/study-notes/ai-blogs/design-multi-tenant-url-shortener.md) - [System Design: Online IDE](https://snowan.gitbook.io/study-notes/ai-blogs/design-online-ide.md) - [System Design: Payment System](https://snowan.gitbook.io/study-notes/ai-blogs/design-payment-system.md) - [System Design: Point of Interest (POI) System](https://snowan.gitbook.io/study-notes/ai-blogs/design-poi-system.md) - [System Design: Slack — Enterprise Real-Time Messaging](https://snowan.gitbook.io/study-notes/ai-blogs/design-slack-messaging-system.md) - [System Design: Text-to-Video Generation Pipeline (Sora-like)](https://snowan.gitbook.io/study-notes/ai-blogs/design-text-to-video-generation.md) - [System Design: Webhook Delivery System](https://snowan.gitbook.io/study-notes/ai-blogs/design-webhook-delivery-system.md) - [System Design: YouTube — A Video Streaming Platform at Scale](https://snowan.gitbook.io/study-notes/ai-blogs/design-youtube-video-streaming-platform.md) - [Dynamic Rate Limiting for AI Inference: Why RPM is Dead](https://snowan.gitbook.io/study-notes/ai-blogs/dynamic-rate-limiting-ai-inference.md): Why traditional rate limiting fails for LLMs, and how to build adaptive, token-aware limiters that protect your GPU resources. - [System Design: CI/CD Pipeline Like GitHub Actions](https://snowan.gitbook.io/study-notes/ai-blogs/github-actions-architecture-deep-dive.md) - [How to Build an Agent Harness: A Practical Guide from Teams Who Actually Did It](https://snowan.gitbook.io/study-notes/ai-blogs/how-to-build-agent-harness.md): A practical, implementation-level guide to building agent harnesses, synthesized from the hard-won lessons of five production teams: Anthropic (Claude Code), OpenAI (Codex), Manus, LangChain (DeepAgen - [Letta's Context Repositories: Git-based Memory for Coding Agents](https://snowan.gitbook.io/study-notes/ai-blogs/letta-context-repositories.md) - [Deep Dive: How OpenClaw's Memory System Works](https://snowan.gitbook.io/study-notes/ai-blogs/openclaw-memory-system-deep-dive.md): A comprehensive look at OpenClaw's file-first memory system, exploring its hybrid search architecture, automatic memory flush, and implementation details. - [PagedAttention: How Virtual Memory Revolutionized LLM Inference](https://snowan.gitbook.io/study-notes/ai-blogs/paged-attention-explained.md): A deep dive into PagedAttention, the breakthrough memory management technique that enables efficient LLM serving. Learn how borrowing ideas from OS virtual memory solved the KV cache memory problem an - [Pi: The Minimal Agent Philosophy — How Less Becomes More](https://snowan.gitbook.io/study-notes/ai-blogs/pi-minimal-agent-openclaw.md) - [Speculative Decoding: How to Make LLMs 2-3x Faster Without Losing Quality](https://snowan.gitbook.io/study-notes/ai-blogs/speculative-decoding-explained.md): A comprehensive guide to speculative decoding, the technique that accelerates LLM inference by 2-3x while maintaining identical output quality. Learn how draft-then-verify works, the math behind accep - [We're All Addicted To Claude Code](https://snowan.gitbook.io/study-notes/ai-blogs/we-are-all-addicted-to-claude-code.md) - [infographic](https://snowan.gitbook.io/study-notes/ai-blogs/infographic.md) - [anthropic-skills-guide](https://snowan.gitbook.io/study-notes/ai-blogs/infographic/anthropic-skills-guide.md) - [Analysis: The Complete Guide to Building Skills for Claude](https://snowan.gitbook.io/study-notes/ai-blogs/infographic/anthropic-skills-guide/analysis.md) - [Structured Content: Building Skills for Claude](https://snowan.gitbook.io/study-notes/ai-blogs/infographic/anthropic-skills-guide/structured-content.md) - [prompts](https://snowan.gitbook.io/study-notes/ai-blogs/infographic/anthropic-skills-guide/prompts.md) - [infographic](https://snowan.gitbook.io/study-notes/ai-blogs/infographic/anthropic-skills-guide/prompts/infographic.md) --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://snowan.gitbook.io/study-notes/ai-blogs.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.