Unreliable Narrators: Calibrate What Readers Can Infer
Separate a narrator's account from reader inference with an invented scene, alternative readings, and precise revision checks for deliberate ambiguity.
Guides and maker's notes from the AZET team — what azet.io is, how the AI assistant and browser automation work, and the AZET product family.
Separate a narrator's account from reader inference with an invented scene, alternative readings, and precise revision checks for deliberate ambiguity.
Plan a wordless comic with visible causes, readable actions, and deliberate gaps through an original six-panel exercise and a focused review prompt.
Compare literal meaning, comic timing, and lost ambiguity when explaining a bilingual pun built from English and an explicitly invented language.
Give fictional idioms stable meanings and flexible uses through an invented workshop scene, contrasting examples, and a practical context review.
Build invented place names from a small set of language rules, then test compounds, exceptions, and map labels in an original worldbuilding exercise.
Build a photographic diptych around one intended relationship, test image order and crop, and distinguish visible echoes from an invented shared story.
Trace a hand-drawn gesture through numbered poses, separate a stray motion point from an intentional turn, and plan a small, checkable redraw.
Compare front, side, and back character drawings through shared landmarks, visible seams, and deliberate exceptions in a fictional turnaround exercise.
Review a tiny fictional sprite at native size using silhouette, gaps, and pixel clusters, with a worked revision and a focused critique prompt.
Build reader performance cards for fictional audiobook characters, using modest delivery contrasts, unchanged dialogue, and a practical return-to-voice check.
Write an original duet exchange where both speakers change the scene, using response roles, a worked lyric revision, and checks beyond equal line counts.
Mark thought groups in an original spoken monologue, compare alternate pause placements, and keep performance notes separate from changes to the words.
Draft short spoken identities for an imaginary radio station, compare name placement, and test clarity without inventing broadcasts or audience evidence.
Review confirmation dialogs by naming the affected object, stating the actual consequence, and checking that button labels match a fictional action brief.
Distinguish a new collection from an empty search result, draft the right next step, and review empty state copy against a fictional interface brief.
Draft form error messages around the field, the problem, and the next step, then test invented examples without making unsupported technical claims.
Review unfinished page copy with a placeholder inventory, distinguish deliberate examples from accidental leftovers, and track each replacement to its source.
Plan a filename change with a reviewable mapping, spot collisions before touching files, and preserve an undo path using a small illustrative example.
Describe pottery surfaces through located visual observations, separating light effects and personal impressions from unsupported material claims.
Keep printmaking edition notes consistent by separating sheet IDs, maker-assigned labels, observed differences, and corrections that remain traceable.