Notes on trustworthy AI.
Essays on personal AI systems, delegation, and building assistants you can actually rely on — from 30 years of making technology serve people.
The Agent Blind Spot: Who's Watching Your AI Fleet?
Your agents make thousands of LLM calls a day — and nothing inspects them. Why I built ClawNex, a real-time SOC for AI agent fleets, and open-sourced it.
9 min readRead article →The Assistant That Lives on Your Computer: Why OpenClaw Changes Everything
Most AI assistants live in a browser tab. OpenClaw runs on your computer, connects to your messaging apps, and acts on your behalf—without being asked.
9 min readRead article →The Scorecard Problem: How to Know If Your AI System Is Actually Working
You built the AI system. It's running. But is it actually making you better—or just making you feel productive? Here's how to measure what matters.
10 min readRead article →The Agentic Orchestration Layer: Why Your AI Tools Need a Conductor
Most teams have 5-10 AI tools. None of them talk to each other. Here's why that's a problem—and how agentic orchestration fixes it.
12 min readRead article →Building a Personal AI Assistant That Actually Knows You
Generic AI is impressive. Personalized AI is transformative. Here's how to build an assistant that understands your context, preferences, and workflows.
10 min readRead article →Digital Second Brains: Why Memory Matters More Than Intelligence
Your AI doesn't need to be smarter. It needs to remember better. Here's how to build a system that captures, organizes, and retrieves your knowledge.
11 min readRead article →The Context Problem: Why AI Forgets and How to Fix It
Every AI conversation starts from scratch. That's not a feature—it's a bug. Here's how to build systems that maintain context across interactions.
9 min readRead article →From Prompt to Process: Automating Your AI Workflows
Prompting AI is useful. Automating AI workflows is transformative. Here's how to move from manual prompts to automated processes.
10 min readRead article →The Trust Layer: Building AI Systems You Can Rely On
AI makes mistakes. That's not the problem. The problem is not knowing when it's wrong. Here's how to build trust into AI systems.
11 min readRead article →Designing AI Interfaces That Don't Suck
Most AI interfaces are just chat boxes. That's lazy. Here's how to design interfaces that match how people actually work.
9 min readRead article →The ROI of Personal AI: When Does It Actually Pay Off?
Building a personal AI system costs time and money. Here's how to calculate if it's worth it—and when it's not.
10 min readRead article →