One problem I have with some of these measurements of value is that there's little talk of alternative tools.<p>I see a value add from AI, but I also believe that value could have already been there - and potentially <i>much</i> more reliably and less expensive, if the right people cared to implement it before this point.<p>My own observation is that a lot of engineering orgs have been neglectful in good practices and/or tooling, and AI being the hot new thing is filling in where those things were missing. Kind of like measuring the productivity of a starving laborer vs a laborer on cocaine and never thinking to measure what happens if you give him 3 meals a day.<p>I would like to see more of these posts discussing exactly what the value they are seeing from AI is, and evaluating if AI is even the correct tool for that value-add.