I. The position most knowledge workers can no longer find
Sometime in the last decade, the position above your work disappeared.
Not in a dramatic way. The calendar filled, the tabs multiplied, the surfaces with notifications grew from one to four to nine. Each tool, individually, made you faster at doing the work. The aggregate effect was that the place from which you used to think about the work — Saturday morning, the second coffee, the walk between meetings — got eaten by more work.
The position has a precise location: one step outside the activity you are currently in. A vantage from which the activity has a shape, a boundary, and a relationship to everything else. It's where you notice that the thing you are about to spend an hour on is not, in fact, the lever. It's where you notice that two conversations from different weeks are actually about the same thing. It's where the strategic move becomes visible — because the strategy is never visible from inside the tactic.
Most software in 2026 is designed to deepen your immersion in whatever you are doing right now. That is not a defect. It is the explicit design goal of the productivity category. The question this essay is interested in is: what tool is built to do the opposite?
II. "In" and "on" are not just prepositions
The framing this essay rests on — working on your work, not in your work — is not a clever phrasing. It is a cognitive distinction that has been in your head, prefabricated, since you learned English.
George Lakoff and Mark Johnson made this case explicit in Metaphors We Live By (1980). Their central claim: humans don't reason about abstract things directly. We reason about them through metaphors that are grounded in our bodies. The most basic of these metaphors are orientational — they use the spatial axes our bodies are organized around: up-down, in-out, on-off, front-back.
Two of those axes — in-out and on-off — do all the work in our argument.
Lakoff & Johnson are explicit that activities themselves are conceptualized as containers. We "get into" work. We "get out of" work. We are "immersed in" it. We do things "outside of" it. These are not figures of speech. They are the everyday cognitive structure English speakers use to think about activity at all. Lakoff's exact phrasing, page 31:
Activities in general are viewed metaphorically as substances and therefore as containers... He's immersed in washing the windows right now... Thus, activities are viewed as containers for the actions and other activities that make them up.
When we say working in your work, we are activating that container schema. There is an interior space (the work), a boundary, and an exterior (everything else). When you are inside the container, the container is the world — its interior fills your visual field. You can see the immediate task. You cannot see the shape of the container itself.
When we say working on your work, we are activating a different position relative to the same container. You are outside. The container has a shape, a boundary, a relationship to other containers. The interior is in view but does not occupy you.
These are not synonyms. They are not stylistic choices. They are two different bodily positions, projected onto an abstract activity. Your brain processes them differently because your body grew up moving through space, and that movement is the substrate of how you reason about anything at all.
III. Why the position matters
Lakoff makes a stronger claim, late in the book, that is the actual load-bearing point for everything below. He argues that container metaphors — what he calls ontological metaphors — are not decorative. They are the precondition for being able to reason about an activity, rather than just do it.
Such ontological metaphors are necessary for even attempting to deal rationally with our experiences. Once we can identify our experiences as entities or substances, we can refer to them, categorize them, group them, and quantify them — and, by this means, reason about them.
Read that twice. The container schema is the cognitive precondition for asking should I be doing this? From inside the container, that question is not available. You can only ask how do I do this faster? The other question requires the position outside.
This is why "working on your work" is not a soft, vaguely aspirational framing about reflection. It is a cognitive precondition — for strategy, for prioritization, for noticing that you have been in the wrong room for six weeks.
Most knowledge workers, most of the time, are stuck inside the container. Not because they are unintelligent. Because the tools they use all day are designed to keep them inside — to deepen the immersion, to refill the queue, to make the interior denser, more responsive, more colorful. The exterior, the place from which the work is visible as a whole, has no tool that takes you there. It is the unstaffed position in the modern software stack.
IV. The asymmetry of the two positions
There is an interesting asymmetry, often missed, between the IN and ON positions.
The IN position has infrastructure. Every productivity tool, every workflow automation, every AI agent that drafts your documents and runs your processes — these all live inside the container. They make the interior more efficient. They are good at this. We are not making the contrarian case against them.
The ON position has furniture. The Saturday morning. The shower. The long walk. The cancelled meeting. These are still the moments most strategy actually happens. There has never been a piece of software that produced the ON position the way Notion produces the document-in-the-container.
Two things follow from this asymmetry.
First: the productivity-tool category cannot, structurally, deliver the ON position. It is the category's defining job to deepen immersion in the container. Asking Notion to give you perspective on your work is like asking a magnifying glass to step back.
Second: there is an open category — call it the cognitive vantage — that no incumbent occupies. The tools that gesture toward it (Roam, Mem, journaling apps) are mostly containers for thoughts. They are still interior infrastructure, just for a different interior. They are not the same as the position above the activity.
This essay's argument is that a tool for the ON position is possible, and that the question for the next decade of software is not "how do we automate more of the work" but "how do we make the position above the work reliably accessible."
V. What a tool for the ON position would actually do
A tool for the ON position would have a specific, narrow job: make the boundary of the current container visible to the person inside it.
It would not optimize the interior. It would not generate documents, run workflows, or shorten any action chain. Those are all interior moves.
What it would do — and here we are describing the product Domain is, not making a future-tense claim — is the following:
Hold context across conversations and weeks. The container's boundary is invisible from inside not because the boundary doesn't exist, but because the person inside is occupied by the foreground. A tool that remembers what was said three weeks ago, what threads are still open, what the team decided on a Tuesday — and surfaces those to the user at the moment they would otherwise be re-immersed — is doing geometric work, not memory work. It is letting the user see the shape of the container they are about to step back into.
Connect across containers. Most knowledge workers do not have one container; they have several. (Three to nine, in the average week.) The position outside any single container is also the position from which the relationship between containers is visible. A tool that notices "the thing you said in the Tuesday chat is the same thing this Friday meeting is about" is producing the cross-container view that strategy requires.
Stay out of the way during the interior. A tool whose job is to make the ON position accessible must, by definition, not fight for time in the IN position. It cannot interrupt mid-task to ask if you'd like to step back. It cannot ping you in the middle of execution. It can only be available when you are already in transit between containers — and ready, in that transit moment, to make the larger structure visible to you.
Domain is built on this geometry. The product lives in the messaging tools you already use, holds the context of your working conversations, surfaces what's relevant at the moment you'd otherwise lose the thread, and does not interrupt you when you are inside the work. The design choice that lets Domain live in Signal, Telegram, your team chat — rather than in a separate workspace — is downstream of the cognitive claim: the ON position is reached between the messages, between the meetings, between the contexts. Not in a separate app you go to.
VI. The honest part
A claim this strong deserves its honest counter-arguments. Three are worth naming.
One: the ON position has costs. Stepping outside the container takes time. The user who reflects too much is the user who does too little. The defense of immersion — just do the work — has a long tradition, from monasticism to flow states, and it is not wrong. The argument is not against immersion. The argument is that immersion without periodic visits to the ON position becomes activity without direction, which is the trap modern knowledge work has fallen into. Both positions are necessary. The IN position has plenty of tools. The ON position has almost none.
Two: people who don't already value reflection won't buy this argument. This is correct. The audience for a tool that delivers the ON position is people who already know — usually painfully — that they have been stuck inside the container for too long. This audience is smaller than the audience for "make my work faster," and that's fine. The point of an open category is not that everyone will enter it. The point is that the category exists, and that the people who would benefit from a tool there have nothing right now.
Three: how do you know it's working? The honest test is linguistic. After using Domain, do users describe what changed in cognitive terms — "I stepped back," "I could see the work as a whole," "I noticed the pattern across the threads" — or in productivity terms — "It saved me time," "It generated the doc," "It got me through the task"? If the user reaches for the second vocabulary, Domain has been read as another tool inside the container, and the positioning has not landed. If they reach for the first, the cognitive geometry is doing its work. This is a real and observable signal. It tells the team when the product is working and when it isn't.
VII. The invitation
If you have read this far, you have probably spent some of the last few minutes in the ON position. That is what an essay like this does, when it works: it makes the container of your day briefly visible from outside, by occupying a different position in your attention than the work normally occupies.
The harder question is what happens in the next ten minutes. The container is going to close again — the next tab, the next message, the calendar reminder you just ignored. The position you are in right now is not a position your tools were designed to keep you in.
Domain is one attempt at a tool that takes the other side of that bet. It lives where your work already happens — in Signal, in Telegram, in the chats your team is already in — and its job is to make the position above the work accessible without forcing you to leave the work to reach it.
We are running a public beta. If this essay reached you in an on-the-work moment, you can join the beta by texting OTW to DOMA on Signal at the number below, right now, while the position is still available. We will pair you with a real person — Toby or Andrew — for your first session, in the same window of attention you are in now.
Signal: +1 619 393 7662 — text
OTW.
The position will close. The tool that respects the position will not.
— DOMA & Matt's Claude, May 2026