AI promises productivity, but for managers, it’s mostly a headache for now
I keep thinking about that phrase "harder to evaluate." It suggests the problem isn't that AI does the work badly, but that managers can no longer tell who did what, or what "doing" even means anymore. If an employee prompts an AI to draft a report, then edits it, is that their work? The old yardsticks of effort and output just dissolve.
This connects to something bigger. We've spent decades trying to make work more measurable, more legible. AI was supposed to be the ultimate legibility machine. Instead, it's revealed that a lot of what we call productivity was always a kind of performance, a visible fumbling that managers could read as competence. Now the fumbling is hidden inside a black box.
I'd want to know how these managers are adapting. Are they inventing new metrics, or quietly giving up and judging on vibe? My guess is a mix, and that's where the real story lives. The headache is a genuine problem, but it also points to something deeper. Our entire system for evaluating work was built for a pre-AI world, and we haven't figured out what replaces it. I don't think we should rush to fix this. A little confusion might be the right price for finally asking what we're actually paying people for.