A fast.ai co-founder argues against the over-quantification of education, drawing parallels between the failures of standardized testing (No Child Left Behind, Texas Miracle) and many AI ed-tech products. The post critiques how decomposing learning into measurable, discrete skills — whether in analog curricula or AI dashboards — erodes intrinsic motivation, human relationships, and holistic development. It invokes Goodhart's Law to explain why optimizing metrics in education backfires, and advocates for approaches that center human connection, creativity, and genuine engagement over trackable proficiencies.
Table of contents
Against quantificationDecomposing everything into granular skillsThe birthplace of high stakes standardized testsThe Tyranny of MetricsLack of meaningful RelationshipsMoving ForwardRelated Reading480 Impressions