Every month, AI reshuffles the productivity deck and — more quietly — redefines job descriptions.
AI is eating execution across every department: it codes, designs, specs, writes, compiles, reports, debugs — and the list goes on. I see it in every scale-up transformation I'm in.
So where does that leave us? Copy-pasting: specs from Claude to Notion, design from Replit to Figma, code from Claude Code to GitHub (okay, Claude handles that one on its own), and so on. I know it sounds reductive — but it's pretty close to the truth. AI executes, the human copy-pastes. Welcome to 2026!
But as an engineer, if I'm not the one coding, who am I? As a designer, if I'm not the one designing? As a PM, if I'm not the one writing the spec? You get the picture.
AI is here to stay. And the real challenge — supposedly technological — actually turns out to be deeply human. How do we recalibrate what an organization expects from its teams when AI is doing a growing share of the work the job description is supposed to cover?
Beyond AI as a skill — which, for real mastery, can still command a 20 to 40% salary premium depending on the sector — and short of falling into micromanagement, clarifying expectations from day one — who does what between humans and AI, and how we measure it — could be what separates the companies that get a 10x in productivity from the ones that get a 10x in frustration.
Clarifying who does what between humans and AI — and how we measure it — could be what separates the companies that get a 10x in productivity from the ones that get a 10x in frustration.