[gsdatta]

Candidly, we're disagreeing.

The framework is explicitly titled "Measuring developer productivity with the DX Core 4". Diffs/engineer, even if not at the individual level, is still meant as a proxy for individual throughput. The question they are trying to answer is "how do we make Ganesh X% more productive?"

My argument is that that is the wrong question to be trying to optimize for, regardless of how you measure it. It also does not capture the risk dimensions with AI adoption (and the reduction in humans-in-the-loop) which my paper tries to directly address.