Gyo is a designer who builds apps, digital products and writes about how systems work.
Landing wheels beneath a large aircraft.
Ranking is followed by selection, visibility checks, conversation deduplication and feed assembly. A high-scoring candidate still has several decisions ahead of it.
X’s prediction model turns reader history and candidate features into action estimates and time-valued outputs. Here is how those estimates reach the ranking formula.
A close portrait in a white turtleneck.
Before filtering and ranking, X has to find candidates for a reader. Following, interaction history, retrieval services and cached posts offer different routes in.
A black outfit beside a sunlit window.
A black dress in a narrow doorway.
Pink geometric walls in hard sunlight.
X’s candidate filters check identity, history, access and budgets before this ranking stage. Their order explains more than a list of weights can.
A seated figure against pale blue.
An old television in a colorful room.
Copy links, negative feedback, same-author discounts and age limits, read from the code with the switches and assumptions left in view.
A rider leaning over a white horse.
A helmeted figure in a red flower field.
A cherry beneath red lips.
A raised hand above an empty plaza.
A black rider silhouette beneath flying birds.
A black top in a sunlit garden.
A red ship among floating ice.
A white robe beside colored mats.
A seated figure with a reflective helmet.
An aerial view of repeating suburban streets.
Dark animals on a snow-covered slope.
A pale aircraft and car in the dark.
