Writing about calmer ways to run a day, what an agent should and should not do on your behalf, and the decisions behind how Drift works.
What belongs in a daily briefing, and what does not (Product, 2026-08-05). A daily briefing is a summary you did not ask for, which sets a high bar. What earns a place in one, what does not, and why the reason matters as much.
How to plan a day from your calendar, email and tasks at once (Guides, 2026-07-28). A daily planning routine that starts from the hours you actually have left, not from a list. Five minutes, three inputs, and one thing you commit to first.
A 10-minute inbox triage routine you can actually keep (Guides, 2026-07-15). Inbox zero fails because it is a filing system. This is a triage routine built on one decision per message, in ten minutes, that survives a bad week.
Why an agent should ask before it sends, and where the line belongs (Engineering, 2026-07-02). Confirmation is friction, and friction does not demo well. The case for an approval boundary an agent cannot argue its way past, and where to draw it.
AI assistant vs AI agent: the difference that shows up in your day (Explainers, 2026-06-24). An AI assistant answers the question you typed. An AI agent goes and gets what it needs, then does the work. Where the line falls, and why it matters.
What "connect your Google account" actually grants (Guides, 2026-06-08). The consent screen scrolls past in four seconds. Here is how to read what an AI tool is actually asking for, and the questions worth asking before you agree.
Task initiation is the barrier, not task tracking (Explainers, 2026-05-19). Most productivity apps are built for people who struggle to remember work. The harder problem is starting it. What changes when you design for initiation.
Why every productivity system works for eleven days (Explainers, 2026-05-06). Every system works at first. The ones that collapse share a property: they need to be maintained by hand, and the maintenance is charged at the worst time.