Count the apps you touched before lunch today. Not the ones you used deeply. The ones you touched.
A task manager to see the day. A calendar to see the meetings. A notes app for the thing you wrote down on Tuesday. Slack or email for the requests other people call urgent. A habit tracker waiting to be fed. An AI chat where you actually figured out the hard problem, in a conversation you will never find again.
Each app earns its place in isolation. Together they form a stack, and the stack charges rent.
The tax has three parts
The switching itself. Decades of attention research agree on the basics: context switching is not free, and recovering focus after an interruption takes real time, often measured in minutes, not seconds. App switches are small interruptions you administer to yourself, dozens of times a day.
Knowing where things live. Was the renovation budget in Notes, Notion, or that chat thread? Routing every new thought to the right container is a decision, and retrieving it later is a search across containers. The more tools, the more of your memory is spent on a map of your tools.
Maintaining the federation. Every app has its own inbox, its own backlog, its own notion of “today.” Keeping them roughly in sync is invisible work. This is how you end up with a task in the task manager, the same task as a calendar block, and a note about it in a third place, all slightly different.
The productivity stack, in other words, recreates the problem it was bought to solve. You wanted less to hold in your head. Now you hold the contents and the topology.
The wrong fix: the everything app
The industry’s answer is consolidation: one workspace app with tasks, notes, docs, databases, calendars, and now an AI assistant in a sidebar. The pitch is fewer apps; the reality is the stack reassembled indoors. The modules still have seams, the seams still need maintenance, and now the tool itself has a learning curve measured in weekends. An everything app does not eliminate the tax. It collects it under one roof.
The better fix: fewer jobs, not more features
The tax is not proportional to how many apps you own. It is proportional to how many of them claim a piece of your attention for the same job. The fix is to give each job exactly one home, and to pick homes that do not ask for tending.
For tasks, that means one place that answers one question: what am I doing now? Not a dashboard. Not a workspace. A list, ideally one that resets itself every morning so its maintenance cost rounds to zero. That is the entire job description Ember was built against: two views, Now and Next, and nothing else competing for the visit.
The AI seam is the newest tax
The newest line item deserves its own mention. If you use an AI assistant seriously, a familiar loop has appeared in your day: you review your tasks in one app, then switch to the AI, re-explain a task from memory, paste the results somewhere, and go update the task list.
That seam is exactly backwards. The intent already existed in your task list; you should not have to re-perform it in a chat window. Ember’s approach is to let your agent come to the list: tag a task for delegation, and the AI you already use picks it up over MCP, asks its questions, and reports back. One job stays in one place, and the switching tax on it drops to a single tap.
An honest accounting
You do not need zero apps. You need each one to hold a whole job, do it without maintenance, and stay out of the jobs that are not its business. A calendar for time. A notes app for reference. One list for what you are doing now, that cleans itself.
When you audit your stack, do not ask which app has the most features. Ask which apps you maintain, and what they would look like after a week of neglect. The ones that punish absence are charging you the tax.
We built Ember to be the one that does not.