The question
What is productivity, and what is learning, in the age of AI?
Output is cheap now.Aim is not.
Machines can draft, summarize and plan in seconds, so how much you produce no longer says much. What they can't do is decide what matters to you, or learn on your behalf. That part stays yours: choosing what matters, understanding it well enough to judge, and seeing it through. Nocked makes tools for that part.
Early access to Nocked, a keyboard-first app for getting things done. Email only, kept private.
Read the argumentI. The noise
Everything got faster. Your attention didn't.
There has never been more to read, more to answer, more to try. Every app wants a moment, every feed refreshes itself, and now every tool can generate more: more drafts, more summaries, more suggestions, more plans. The supply of things you could do is, for practical purposes, unlimited.
Your day is not. You still get one stretch of focus at a time, and a few good hours on a good day. When producing becomes nearly free, effort stops being the scarce thing. Three things take its place.
- Attention
- Everything competes for it, and you have no more of it than you did before.
- Judgment
- Someone still has to decide what is worth doing, and whether an answer is right.
- Skill
- Knowing about something is not the same as being able to do it.
That is the trap of this moment. It is easy to feel productive while drowning: tabs open, tasks generated, answers collected, nothing finished. And it is easy to feel informed while learning nothing that stays.
II. Productivity
Productivity is not output. It is aim, then release.
An archer does not shoot faster to hit more. They choose the target, nock the arrow, draw, and let go once. We think work in the age of AI asks for the same discipline.
Get it out of your head.
Anything you are trying to remember is something you are not thinking about. Capture every commitment the moment it appears, in a place you trust, and your mind is free to work.
In the appPress Ctrl Space in any app, type, and it lands in your Inbox. Brain Dump turns a page of pasted notes into proposed to-dos.
Decide on purpose.
A list is not a plan. Your next action should be a choice you make with your context, time and energy in view, not whatever shouted loudest. Machines can help you see the options. Choosing is yours.
In the appClarify Inbox proposes where each item belongs. Ask Nocked suggests what to do next from your context, time and energy. You accept or skip.
Finish what you start.
A drafted plan, a half-read article and a generated outline all feel like progress. Done is the only state that counts. Fewer things, carried all the way through.
In the appToday shows what you chose for today and anything that is due, with Tonight below it. Everything else waits in Next, Scheduled and Someday.
III. Learning
An answer is not understanding.
When any question can be answered in seconds, it is tempting to stop learning and start retrieving. But an answer you did not work for rarely stays, and it does not teach you to tell a good answer from a confident wrong one. As machines grow more fluent, that judgment matters more, not less.
Learning the right way still looks the way it always has. You try, you get it wrong, you try again. You explain it in your own words. You come back to it after you have half forgotten it. And above all, you use it.
The question is no longer what you know. It is what you can now do that you could not do before.
So we treat learning and doing as one loop. What you read becomes a next action. What you try becomes a note for your weekly review. What you learn shows up in your work, where it belongs. AI is welcome in that loop as a patient sparring partner: it can quiz you, question your plan, draft a first version. It should never do the understanding for you.
In the appAnything you read can become a next action in a few keystrokes, and the guided Weekly Review is where you look back.
- Learn: take in one idea.
- Act: put it to work this week.
- Review: see what held, and adjust.
The answer, in short
Productivity is aim. Learning is practice. Focus is a choice.
IV. How we build
AI proposes. You decide.
Everything we make follows from the argument above. These are the rules we hold ourselves to.
Proposals, not actions.
AI can suggest, sort, draft and plan. In the app, nothing it produces takes effect until you accept it, and anything can be undone.
Your work stays yours.
Local-first by default. Your tasks live on your computer, not in our database. No account required, and no tracking.
Bring your own intelligence, or none.
Use the model you trust, hosted or running on your own machine. Or switch AI off entirely. The tools still work.
Attention is the interface.
Keyboard first and quiet by design. No badges, no streaks, no feeds. A tool for focus should not compete for it.
Open where it counts.
Our first app is MIT licensed, and its source will be published. We intend to open more of what we build, so you can read how your tools work, change them and keep them.
V. The first tool
Nocked, the app.
A keyboard-first app for getting things done, built on the principles above. It is the argument, made usable.
- Early preview 0.1.0
- macOS, Windows, Linux
- MIT licensed
Capture from anywhere.
Press Ctrl Space in any app and type. Dates written in plain English or Swedish land where they belong.
Lists that mean something.
Inbox, Today with Tonight, Scheduled, Next, Someday and Done. Start and due dates are kept apart, so when you plan to begin never pretends to be when it is late.
Projects that hold together.
Sections and checklists for work with many steps. Jump to anything by typing its name.
A weekly review that walks you through.
A guided pass over everything open, once a week, so nothing quietly goes stale.
AI that proposes.
Clarify your inbox. Turn a brain dump into to-dos. Plan a project. Draft the email a task needs; nothing is sent. Ask what to do next, with your time and energy in view. Every result is a proposal you accept or skip from the keyboard.
Private by default.
Your tasks live in one file on your computer. Use Claude with your own key, a local model through Ollama or any OpenAI-compatible server, or no AI at all; the app shows you exactly what it sent. A local MCP server, once you turn it on, lets agents like Claude Code and Claude Desktop read your lists and, if you allow it, add to your Inbox.
This is an early preview. Installers are not yet signed, and things will change. That is why we are starting with a waitlist.
Join the waitlistVI. What Nocked is
A small company with one long question.
Nocked is a product company, founded in Sweden. The app is our first answer to the question at the top of this page, not our last. We expect to build more tools, some paid and some free, and to open-source the parts others can build on.
We also want to help people learn well, not only work well: to share what we find about focus, practice and understanding, in plain language, without selling shortcuts. If a method holds up, we will write it down. If it does not, we will say so.
Everything we make should leave you more capable than before, not more dependent.
- Products
- Tools worth paying for, built to last and to stay out of your way.
- Open source
- Code you can read, run, change and keep.
- Learning
- Plain writing on focus, practice and understanding, made to be used.
Early access
Take aim with us.
Nocked is in early preview. Leave your email and we will write when early access opens.
One email when early access opens, and nothing else. We keep only your address and the date you joined, never share it, and never track you. Reply to that email to have your address deleted.