Just-in-Time Ontologies: What Cheap Intelligence Means for Data — Hrishi Olickel, Southbridge AI
Cheap AI inference makes data transformation cheap, enabling just-in-time ontologies over static universal schemas.
“I believe we have overcome the last hurdle in the transformation—general intelligence or judgment—and made it available on demand along the entire Pareto frontier.”
Hrishi Olickel argues that cheap AI inference eliminates the core assumption behind modern data infrastructure—that transformation is expensive—enabling 'just-in-time' ontologies instead of costly universal schemas. He introduces 'Pareto serving,' where models dynamically tune pipelines for cost, speed, and accuracy simultaneously. A concrete example shows entity resolution costs dropping from $10,000 to $800 by integrating a new model, illustrating the practical impact of this architectural shift.