Code as Data and the Credit Score

Measuring Code Quality in the Age of GenAI - Matt Van Itallie

Topic 2 of 11

Published Feb 10, 2025 · Clip 5:16–14:16 (9:00)

Matt explains Sema's credit score product for software engineering organizations, which analyzes many data sources into a 1-100 score. He gives the example of 250 high-risk security warnings, showing that interpretation depends on company size and stage quartiles. He also stresses that code is a craft, not a competition — data must facilitate a conversation rather than stand alone, and their reports are frequently used in high-stakes technical due diligence.