All writing / Private Worlds
Essay 02 / 04 · Abundance & attentionWhen You Can Build Anything, What Deserves to Exist?
As making gets easier, deciding what to keep becomes a more personal discipline.
A music studio, a household dashboard, a personal knowledge graph, an agent taskboard: each begins with a reasonable thought. It would be useful if this existed.
Once you have agents helping with implementation, that thought acquires unusual force. An idea can become a project before you have fully decided what role it should play in your life. The distance between imagining and maintaining gets shorter.
That is exhilarating. It also changes the nature of the constraint.
When building is difficult, implementation provides a filter. Plenty of ideas disappear because there is insufficient time, money, or expertise. This filter is crude: it eliminates wonderful things alongside pointless ones. Removing it is a real gain. But whatever selection it performed still has to happen somewhere.
The decision moves closer to the person.
Every working system creates a small constituency for its own continuation. It has data worth preserving, dependencies worth updating, and a few obvious improvements. Because it already exists, the next hour spent on it feels easier to justify than the original hour did. A project acquires momentum that can be mistaken for purpose.
Agents can accelerate that momentum. Give one a backlog and it can help make progress. Ask another to improve the backlog and it can produce a more coherent plan. Soon the machinery for getting things done becomes impressive enough to obscure the unresolved question of what any of it is for.
The answer cannot be a demand that every project prove its economic value. A strange song, an interactive landscape, or an impractical experiment may be among the best uses of a computer. Play is a sufficient reason. Curiosity is a sufficient reason. Making something because you want to see it is a sufficient reason.
What matters is being able to recognize that reason—and notice when it has changed.
A personal system needs ways to finish, pause, archive, and forget. It needs an acceptable state between “actively improving” and “failed.” Otherwise, abundance becomes a collection of obligations disguised as possibilities.
The week stays
the same size.
Move the sliders to see how tending a growing collection of projects can fill the time that easier building was meant to free.
Your available project time
Of 10 hours available, 6 hours go to upkeep and 4 hours remain uncommitted.
4 hours remain for making, playing, or doing something else.
Illustrative arithmetic, not measured behavior. Weekly upkeep = active projects × hours per project. Uncommitted time = available hours minus upkeep, with a minimum of zero. This assumes equal upkeep and that archiving removes it entirely; real projects vary and archives may still need care.
Reading the comparison. At six projects, one hour of upkeep each, and ten available hours, four hours remain. Archive one project and five hours remain. At twelve projects, upkeep exceeds the same ten-hour budget by two hours.
This may be one of the central disciplines of working with capable AI: establishing the conditions under which further progress would be unwelcome.
The difficult instruction is often not “build this.” It is “this is enough.”
A person surrounded by tireless assistants will need to exercise that judgment deliberately. There will always be another improvement available. Some of the freedom gained from easier building should be spent leaving good ideas unfinished—and going to do something else.