We tend to treat software as an asset. But every tool is also a liability. It demands to be checked, configured, and managed. You adopt it to save time, but often you just trade the work of doing for the work of watching.
I think about this whenever I see someone adopt a new project management tool. It saves them 30 minutes a day on coordination. But now they’re checking it between meetings, updating statuses, responding to comments, configuring permissions. If that takes more than 30 minutes, they’re underwater.
I’ve never seen anyone actually do this accounting. We measure the upside and pretend the downside doesn’t exist. Each tool adds a small tax, and the taxes compound.
Software companies don’t make money when you close their app. They make money when you open it. Their pricing, metrics, renewal rates - everything depends on you showing up. So they optimize for attention capture.
The cost is worse than it looks. When you check a dashboard, you don’t just lose the seconds it takes to look. You lose context, the thread of whatever you were actually working on. Cal Newport calls this attention residue. The dashboard check that takes 30 seconds might cost you 10 minutes of degraded focus.
What if software minimized your attention instead of capturing it?
Think about a washing machine. You don’t stand there watching the drum spin. You start it and leave. The machine is most valuable when you forget it exists.
Software could work the same way. But it requires a trust layer: the belief that problems will surface when they matter, that the system won’t fail silently. Most systems haven’t earned this. They require your attention as a reliability mechanism. You’re the monitoring system.
The difference is escalation. Good systems tell you when they need you and stay quiet when they don’t. Instead of demanding your attention, they earn your absence.
Companies built around engagement metrics can’t copy this. Imagine Slack deciding to minimize your time in Slack. They can’t. Their whole business depends on you showing up.
If engagement is the wrong metric, what’s the right one? I think it’s something like outcomes achieved per unit of attention spent. Most software ignores the attention cost entirely. I don’t know anyone who optimizes the ratio.
Attention is finite. There’s no Moore’s Law for human focus. But demand keeps growing, and nobody’s designing for the actual bottleneck. Either we get better at protecting our attention, or we drown in the demands for it.
The industrial age gave us physical leverage: machines did the lifting so you didn’t have to. Software should do the same for cognition. But we built tools that demand attention instead of saving it.
The real leverage is software that lets you leave.