AI-assisted dictation can change what composing requires: a speaker forms a thought while transcription and surface correction happen elsewhere. This paper develops a conceptual argument about this redistribution of work, informed by retrospective accounts of the author's own sustained dictation practice and prior research on writing, augmentation, and cognitive offloading. The central question is which operations assistance removes and which people continue practicing. Batch correction offers one arrangement: defer transcript monitoring while retaining formulation, judgment, and consequential review. A practice repertoire and evidence map connect this arrangement to reported lower expression cost and a recalled change in following tangents during phone conversations. These accounts motivate, but do not establish, productivity gains or lasting cognitive effects. The framework evaluates efficiency, retained abilities, and freedom from continuous visual feedback in supported work and life beyond the screen. It proposes portable personal dictionaries and replacement rules so learned adaptation can travel across products. Author preprint. This manuscript is under review and has not been accepted for publication.
No takes yet. Share an insight, caveat, or question.
Anonymous (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: