Every teacher knows that Marcus needs a different text than Priya. That awareness was never the problem. The problem was always one person, thirty students, and a planning period that ran out before the last item on the list. This episode reframes what AI actually solves in a classroom.

This episode draws in part on research in:

• Zone of proximal development and instructional calibration (Vygotsky, 1978)
• The 2 sigma problem: group instruction versus one-on-one tutoring effectiveness (Bloom, 1984)
• Cognitive load theory and the cost of instructional mismatch (Sweller, van Merrienboer & Paas, 1998)
• Matthew effects in reading and the compounding of early skill gaps (Stanovich, 1986)
• Teacher workload, planning time, and instructional capacity — TALIS 2018 (OECD, 2019)
• Intelligent tutoring systems and personalized learning outcomes (VanLehn, 2011)