<?xml version="1.0" encoding="utf-8"?><rss version="2.0"><channel><title>MJ Zou</title><link>https://mjzou.com/</link><description>Working through AI ideas in public.</description><item><title>The General Tso's Chicken Theory of AI Adoption</title><link>https://mjzou.com/writing/general-tsos-chicken-theory/</link><guid>https://mjzou.com/writing/general-tsos-chicken-theory/</guid><pubDate>Tue, 17 Feb 2026 00:00:00 GMT</pubDate><description>Chinese food has been in America for 150 years. The real stuff is right there on the menu. And yet the default order is still a dish that doesn't exist in China. AI adoption has the same problem.</description></item><item><title>The Last Human Complaint About Bureaucracy</title><link>https://mjzou.com/writing/the-last-human-complaint-about-bureaucracy/</link><guid>https://mjzou.com/writing/the-last-human-complaint-about-bureaucracy/</guid><pubDate>Sun, 01 Feb 2026 00:00:00 GMT</pubDate><description>Bureaucracy persists because it’s organizational memory. AI will digest that memory—and mutate the iron cage into something faster and harder to appeal.</description></item><item><title>Anthropic Just Ate My Startup</title><link>https://mjzou.com/writing/anthropic-ate-my-startup/</link><guid>https://mjzou.com/writing/anthropic-ate-my-startup/</guid><pubDate>Mon, 19 Jan 2026 00:00:00 GMT</pubDate><description>I built an iOS app to bridge Apple Health data to AI. Two weeks later, Anthropic shipped the same feature natively. Here's what I learned about building in the AI era.</description></item><item><title>I'm Quitting Claude (But Not What You Think)</title><link>https://mjzou.com/writing/quitting-claude/</link><guid>https://mjzou.com/writing/quitting-claude/</guid><pubDate>Sat, 10 Jan 2026 00:00:00 GMT</pubDate><description>After losing six months of Claude context to an account reset, I rebuilt my AI workflow on local files. Here's why context ownership matters more than convenience.</description></item><item><title>Why I Had to Rewrite My MCP Server from Python to TypeScript</title><link>https://mjzou.com/writing/mcp-python-to-typescript/</link><guid>https://mjzou.com/writing/mcp-python-to-typescript/</guid><pubDate>Sat, 27 Dec 2025 00:00:00 GMT</pubDate><description>I tried deploying a Python MCP server to Cloudflare Workers. It didn't work. Here's what I learned about remote MCP deployment.</description></item><item><title>What Technical Stack Would I Choose If I'm Starting My Own GP in This AI Era?</title><link>https://mjzou.com/writing/ai-era-gp-tech-stack/</link><guid>https://mjzou.com/writing/ai-era-gp-tech-stack/</guid><pubDate>Tue, 23 Dec 2025 00:00:00 GMT</pubDate><description>After helping three friends build their GP tech stacks from scratch, here's how I'd approach it today—with AI agents, open APIs, and conversation at the center.</description></item><item><title>The Manus Paradox: What the Fastest-Growing AI Company Reveals About Agentic AI</title><link>https://mjzou.com/writing/manus-paradox/</link><guid>https://mjzou.com/writing/manus-paradox/</guid><pubDate>Fri, 19 Dec 2025 00:00:00 GMT</pubDate><description>Manus went from zero to $100M ARR in 8 months—then got cloned in 3 hours. What this paradox reveals about where value actually lies in autonomous AI agents.</description></item><item><title>I Vibe Coded an AI App — But There's No AI Inside</title><link>https://mjzou.com/writing/vibe-coding-ios-app/</link><guid>https://mjzou.com/writing/vibe-coding-ios-app/</guid><pubDate>Sun, 14 Dec 2025 00:00:00 GMT</pubDate><description>I spent two Saturdays building an iOS app with 98% AI-generated code. It exports running data for AI analysis and generates art prompts — but has zero AI features itself. Here's why that's the point.</description></item><item><title>From Two Subscriptions to One Weekend Build: The AI Architecture Pattern Everyone's Missing</title><link>https://mjzou.com/writing/ai-architecture-pattern/</link><guid>https://mjzou.com/writing/ai-architecture-pattern/</guid><pubDate>Mon, 17 Nov 2025 00:00:00 GMT</pubDate><description>I spent a weekend building what I was paying $200/year for. Here's what I learned about AI architecture—MCP, Skills, Subagents, and Prompts—and why context window is sacred.</description></item><item><title>Why Anthropic Wins with 5% of the Users</title><link>https://mjzou.com/writing/anthropic-wins-with-5-percent/</link><guid>https://mjzou.com/writing/anthropic-wins-with-5-percent/</guid><pubDate>Fri, 17 Oct 2025 00:00:00 GMT</pubDate><description>Claude has only 5% of ChatGPT's users but generates 40% of OpenAI's revenue. The secret? Anthropic's product team doesn't try to outrun OpenAI—they out-see them by finding combinations everyone else missed.</description></item><item><title>Writing is Thinking</title><link>https://mjzou.com/writing/writing-is-thinking/</link><guid>https://mjzou.com/writing/writing-is-thinking/</guid><pubDate>Mon, 28 Jul 2025 00:00:00 GMT</pubDate><description>Why writing remains an irreplaceable tool for human metacognition, even as AI writes more of our text.</description></item><item><title>I Asked AI to Help Me Read My Great-Grandfather's Homework. It Failed Spectacularly.</title><link>https://mjzou.com/writing/ai-meets-great-grandfathers-homework/</link><guid>https://mjzou.com/writing/ai-meets-great-grandfathers-homework/</guid><pubDate>Wed, 23 Jul 2025 00:00:00 GMT</pubDate><description>A 100-year-old Cantonese children's book exposed surprising blind spots in our most advanced AI systems. Sometimes the best stress tests come from the most unexpected places.</description></item><item><title>Context Engineering Lessons from Building Manus</title><link>https://mjzou.com/writing/context-engineering-lessons-from-manus/</link><guid>https://mjzou.com/writing/context-engineering-lessons-from-manus/</guid><pubDate>Sun, 20 Jul 2025 00:00:00 GMT</pubDate><description>Key insights from Manus's experience building AI agents - why context engineering has become the critical skill for 2025</description></item><item><title>Blanked Out During Claude's Demo. How Does It Just Work So Well?</title><link>https://mjzou.com/writing/blanked-out-claude-demo/</link><guid>https://mjzou.com/writing/blanked-out-claude-demo/</guid><pubDate>Fri, 18 Jul 2025 00:00:00 GMT</pubDate><description>From a jarring product demo to reverse engineering insights: why Anthropic's 'one model, many tools' philosophy creates seemingly effortless AI experiences.</description></item><item><title>The AI Honeymoon Period: What Context Engineering Taught Me About Hype vs Reality</title><link>https://mjzou.com/writing/ai-honeymoon-period/</link><guid>https://mjzou.com/writing/ai-honeymoon-period/</guid><pubDate>Mon, 14 Jul 2025 00:00:00 GMT</pubDate><description>A Twitter debate about AI's transformative potential reveals the gap between honeymoon-period optimism and measured reality. Why agile methodology and product managers might be the real winners in this AI transition.</description></item><item><title>The AI Productivity Paradox: Why Curiosity Beats Role</title><link>https://mjzou.com/writing/ai-productivity-paradox/</link><guid>https://mjzou.com/writing/ai-productivity-paradox/</guid><pubDate>Sat, 12 Jul 2025 00:00:00 GMT</pubDate><description>Why some people see massive AI productivity gains while others don't—and what I discovered about my own motivation when I tried to understand the difference.</description></item><item><title>The Best Interview Question for AI Product Managers</title><link>https://mjzou.com/writing/the-best-interview-question-for-ai-product-managers/</link><guid>https://mjzou.com/writing/the-best-interview-question-for-ai-product-managers/</guid><pubDate>Thu, 10 Jul 2025 00:00:00 GMT</pubDate><description>Why asking 'Show me your best prompt' reveals everything about an AI product manager's competency - and how to evaluate their answers.</description></item><item><title>Why Can't LLMs Say 'I Don't Know'?</title><link>https://mjzou.com/writing/why-cant-llms-say-i-dont-know/</link><guid>https://mjzou.com/writing/why-cant-llms-say-i-dont-know/</guid><pubDate>Wed, 09 Jul 2025 00:00:00 GMT</pubDate><description>A deep dive into why AI models struggle with uncertainty, what Anthropic discovered about hallucination mechanisms, and practical strategies for reducing AI hallucinations in both personal and enterprise contexts.</description></item><item><title>Three Months Later: From Prompt Engineering to Context Engineering (And Why I've Been Doing It With Humans All Along)</title><link>https://mjzou.com/writing/three-months-later/</link><guid>https://mjzou.com/writing/three-months-later/</guid><pubDate>Tue, 08 Jul 2025 00:00:00 GMT</pubDate><description>A three-month reflection on AI developments, the shift from prompt to context engineering, and how Product Management taught me to debug human communication.</description></item><item><title>Starting This Blog: Musings on AI, Learning, and Conscious Effort</title><link>https://mjzou.com/writing/my-first-post/</link><guid>https://mjzou.com/writing/my-first-post/</guid><pubDate>Sat, 05 Apr 2025 00:00:00 GMT</pubDate><description>My reasons for starting a blog: navigating the age of AI, fostering independent thought, improving communication, and documenting my learning journey.</description></item></channel></rss>