• Claude Code CLI: The Tools, Tricks, and MCP Servers That Actually Matter

  A practical guide to Claude Code's most useful features — CLAUDE.md memory, Plan Mode, MCP servers, custom slash commands, hooks, and Git worktrees for parallel sessions — with real commands and config examples.

  11 min read   ·   March 23, 2026

  2026   ·   Claude Code   developer productivity   MCP   CLI   AI tools  

• Claude AI: The Features You're Probably Not Using (But Should Be)

  A practical tour of Claude's most powerful and overlooked capabilities — Artifacts, Projects, MCP, prompt engineering patterns, and the tricks that actually change how you work.

  9 min read   ·   March 23, 2026

  2026   ·   Claude AI   AI tools   prompt engineering   productivity   MCP  

• Speculative Decoding: Making LLMs 2–3x Faster Without Breaking Anything

  Technical diagram comparing standard autoregressive decoding (one token per forward pass) vs. speculative decoding (draft + parallel verify for 2-3x speedup)

  18 min read   ·   March 8, 2026

  2026   ·   LLM   inference   optimization   speculative-decoding   performance   vLLM   SGLang  

• GLM-5 vs Qwen 3.5 vs MiniMax M2.5: The Open-Weight LLM Showdown (2026 Edition)

  A deep dive into benchmark scores, real-world performance, and architectural strategies for deploying GLM-5, Qwen 3.5, and MiniMax M2.5 open-source LLMs in production.

  12 min read   ·   March 6, 2026

  2026   ·   AI   LLM   benchmarks   open-source   agents  

• China’s LLM Face-off: GLM-5 vs Qwen3.5 vs MiniMax-M2 — Who Wins in Code, Reasoning, and AI Agents?

  Conceptual diagram of a hybrid AI architecture using GLM-5 for reasoning and planning, Qwen 3.5 for multimodal UI and tools, and MiniMax M2.5 for high-speed code execution agents.

  20 min read   ·   March 6, 2026

  2026   ·   AI   LLM   GLM-5   Qwen3.5   MiniMax-M2   open-source   coding   agents   benchmarks