I tried AgnesAI after watching the demo at E27 Echelon. Here is what made the difference between images that looked AI-generated and ones that held together.
Notes
Essays, field notes, and thinking in progress. Older notes drift — I leave them up anyway.
AI-assisted coding creates documentation debt faster than any team can pay it down manually. A structured workflow changes the economics of catching up.
Dynamic workflows in Claude Code let you describe an outcome and hand off the process design. Here is what that actually looks like in practice.
AI is making building more accessible. The interesting question is no longer which app everyone should use. It's what to build around the way you actually work.
A personal website should make the next note easier to publish.
A small home for notes on AI, tech, building, and useful personal tools.
The Markdown vs HTML debate misses the point. They serve different jobs in an AI-assisted workflow. One is for agents, one is for humans.
Skills are process, and process is local. What to look for before adopting a skill someone else built.
Real agent adoption starts with the unglamorous parts. The repeatable workflow wrappers that sit inside the normal software cycle.
Most reviews of Claude Design focus on the outputs. The practitioner signal is the opposite. The interesting part is what you feed into it.
The easiest way to overcomplicate a custom skill is to start from the word 'skill' instead of the work you already do.
The language around agentic skills creates complexity where none exists. Strip it back and you find something much simpler. And more useful.
Defining human value as 'what AI can't do yet' is defensive and shrinking. A stronger definition is needed.
The baseline for hiring has changed. If an agent handles the first pass, the human hire has to bring something above execution.
AI agents are not replacing headcount. They are changing the first question a founder asks. From 'who owns this?' to 'what workflow are we trying to run?'
AI saves time. The real signal is what a person does with it. Finish faster and wait, or reinvest it into something better.
Remote teams see only the output, not the day. AI makes it easier than ever to hide weak thinking behind a polished artifact.