What compounds in thinking

"Compound thinking" does not mean saving everything. Saving everything creates clutter. The question is more selective: which parts of the work make future work easier?

Those are the parts that compound.

The rest should be allowed to disappear.

Decisions compound

A decision is valuable because it removes a branch from the future. Once you know which direction is chosen, the next draft, plan, or question can use that constraint.

But decisions only compound when the reason is preserved. If you keep the answer without the reasoning, the decision becomes brittle. A future version of the project may not know when to keep it or when to revisit it.

The reusable unit is not only "we chose X." It is "we chose X because Y was true, and we would revisit it if Z changes." That small amount of context prevents future work from treating old decisions as mysterious rules.

Definitions compound

Good definitions reduce ambiguity. They let a team, a student, or a model use the same term consistently.

Definitions are especially useful when they can be applied independently. A clear definition can travel into a draft, a study task, a comparison, or a review session.

That is why definitions deserve their own Points. They are not just text inside a larger answer. They are reusable context.

Definitions also reveal disagreement early. If two people use the same term differently, the project can resolve that before the difference leaks into drafts, designs, or decisions.

Rubrics compound

A rubric turns judgment into a repeatable operation. Once you define the criteria for evaluating a product idea, research source, or writing draft, you can apply that rubric again.

The rubric can also improve over time. Each use reveals whether the criteria are too vague, too strict, or missing an important dimension.

Rubrics are especially useful with AI because they let you ask for a judgment without reinventing the standards every time. The model can evaluate against the same criteria, and you can revise the criteria when the evaluation feels wrong.

Questions compound

Unanswered questions may look like unfinished work, but they are often the most valuable part of a project.

A good question gives direction. It tells you what evidence to find, what decision is blocked, and what context should be connected next.

When a question is answered, the answer should not erase the question. Keeping both shows the path from uncertainty to clarity.

That path matters. It helps a future teammate understand why the answer exists and which evidence changed the state of the work.

Context compounds when it stays reusable

The useful unit is not always a whole document or a whole chat. It is often a smaller piece: one decision, definition, rubric, question, example, or source note.

BaseHalf is built around that unit. Put the reusable piece in a Point, revise it into a clearer Block, and connect it with References when it should inform another part of the work.

The goal is not to remember more. The goal is to make the right pieces available again.

What does not compound

Some material should be temporary:

• drafts that were clearly rejected
• summaries that repeat the source without adding structure
• clever phrases that do not guide future work
• broad risks with no test or mitigation
• questions that nobody intends to answer

Let those pieces leave the workspace. Compounding depends on selection.