The Personal Index
A growing capability needs a way to find its own judgment again—without turning a person’s life into a database.
01 / the retrieval problem
Activity is not yet a capability.
The personal index is the practice of making a Stack Person’s useful questions, sources, decisions, and unfinished work retrievable without pretending that every piece of context should be stored or shared.
AI can make it unusually easy to produce a wake of material: conversations, drafts, generated options, links, notes, recordings, and tasks. The pile can feel like progress because it contains traces of real work. But a pile has no point of return. When the next decision arrives, a person may remember that they solved something similar and still be unable to find the reason, the source, or the limit that made the earlier answer trustworthy.
App People learned the navigation supplied by a product: a project tab, a folder tree, a search box, a recent-items list. Stack People work across changing applications and accumulated methods. Their challenge is not merely to search more. It is to leave a small enough trail that a past piece of judgment can become present capability.
A personal index is not a record of everything a person has done. It is a deliberate set of return points that lets a growing stack recover what it learned, why it mattered, and what must be checked before it is used again.
This is a distinction about practice, not a promise that software will remember a whole life. A model’s recall can be useful, but it is not the same as a person being able to name where a claim came from or why a decision was deferred. The index keeps that responsibility visible.
02 / four return points
Index the work at the moment it becomes useful.
A strong index does not demand a second job of documentation. It asks for a few signals at the moment a piece of work changes status: when a question becomes reusable, when evidence changes an answer, when a boundary matters, or when a future person will need a clean way back in.
Name the question.
Capture the decision or recurring problem in ordinary language. Product names and clever prompt labels can be useful details, but they should not be the only handle on the work.
Keep the trace.
Save the source, assumption, permission, or human review that made the answer responsible. The trace can be brief; it needs to be enough to prevent confident reuse of an ungrounded result.
Mark the boundary.
Say what cannot travel: a client detail, a private conversation, a provisional claim, or a context owned by an employer or collaborator. An index is valuable partly because it makes restraint findable too.
Set the return.
Give the item a plausible future doorway: a project, a person, a decision type, or a date to reconsider it. Otherwise useful context becomes an archive that nobody can re-enter.
03 / a lightweight method
The index should reduce remembering, not create a new performance of remembering.
The wrong response to scattered context is comprehensive self-surveillance. It can turn every conversation into a filing obligation and every useful experiment into another form to complete. That practice exhausts the person and often captures material that should remain private, temporary, or intentionally forgotten.
A better rule is selective return. Save only the items that change a future decision: a source standard worth applying again, a hard-earned instruction, an unresolved question, a meaningful handoff, or a boundary that a new collaborator could otherwise cross by accident. The index earns its place when it shortens a future search for judgment, not when it proves that someone was busy.
AI can help sort, summarize, and surface those return points. It should not quietly become the author of their importance. A Stack Person still decides what deserves a name, what context is protected, and when an old answer has expired.
04 / a social question
A person’s method can be portable without making every context portable.
Organizations often react to personal stacks with another false choice. Either every worker keeps a private, inscrutable archive, or every useful trace is copied into an institution’s approved system. Neither arrangement recognizes the actual shape of responsible work.
Some material must live in shared systems because a team, client, student, or public obligation needs access and accountability. Other material is part of a person’s portable method: how they frame a recurring question, how they recognize a weak source, how they stage a careful review. The personal index can make this line clearer. It points to shared records when they are required while helping a person carry a practice that is not reducible to a company folder.
That matters for careers and education. The old application model rewarded people who could repeatedly learn the interface of the institution they entered. The app-to-stack transition will reward people who can accumulate a method, locate its history, and bring the useful parts into a new setting without carrying what they do not own.
It also gives managers and teachers a better question than whether someone has adopted the approved AI tool. Can the person explain how a result became reusable, what evidence it rests on, and where its context properly lives? Those are the beginnings of capability that other people can trust.