Five Dated Predictions on Meaning, Machines, and Human Thinking (2027-2032) — with Public Annual Review (v1.0)

Abstract. Five falsifiable predictions with explicit deadlines and failure conditions: meaning-versioning in agent ecosystems by 2028; a standards-body specification for versioned concept identifiers by 2030; person-specific fidelity testing as an evaluation category by 2030; organizations reporting a human-capability measure alongside AI productivity by 2029; a leading world model exposing a human-readable semantic interface by 2032. Publicly scored every September; failed predictions remain on record.