Ship the feature before the user asks.

Anticipatory personal surfaces

Learn behavioral patterns well enough to predict the next want, then run an agent loop that builds the UI, aggregates the data, and implements the capability before that moment arrives.

Two stages

First, prediction: model the next action with enough confidence to act early.

Second, construction: an agent loop that materializes the surface — so when the user reaches for action A, the feature for A is already live.

Constraints

Anticipation without ambient authority: grants must bound what an agent can pre-build.

Wrong predictions should be cheap to discard, not sticky obligations.