Pedagogical Efficiency in AI Training: From Reading Recovery to Retrieval Architectures

Recent empirical results from small language models demonstrate that carefully curated,textbook-quality training data can produce performance comparable to models trained onorders of magnitude more unfiltered data. This paper proposes that this efficiency gain is notaccidental but reflects well-established principles from educational pedagogy—specificallyMarie Clay’s Reading Recovery methodology. We argue that pedagogical structure, not rawdata scale, is the primary driver of efficient learning in both human literacy acquisition andartificial language models. Building on this insight, we propose a shift fromgenerative-dominant architectures toward retrieval-centric systems using CanonicalKnowledge Objects (CKOs), enabling transparent, low-resource, and trustworthy AIdeployment. This reframes AI efficiency as a curriculum design problem rather than ahardware or scale problem