There's a class of engineer that doesn't have a name yet. I work with them, I would like to hire more of them, and every time I describe what they actually do, I run out of vocabulary halfway through. So I'm going to give them one. I'll call them Product Innovation Engineers, and I think they're about to be the most valuable people in tech.
A traditional engineer takes a spec and turns it into software. A Product Innovation Engineer asks a different question: what would actually help this person, and am I the one who's going to find out? They are not waiting on a PM to validate the problem. They are not handing the thing off to marketing once it's built. The surface area of their job has expanded — dramatically — and the humans they spend most of their time thinking about are not their teammates. They are the users on the other end of whatever they're shipping.
Until very recently, this role was impossible. Building software well was a full-time job. So was writing the launch copy. So was making the landing page that sold it. One person could attempt all of it, but only by doing most of it badly.
That math has changed. With Claude Code on one side and Anthropic's Claude Design on the other — a tool that turns a prompt into a prototype, slide, or one-pager that respects your design system — the raw materials for an entire product launch are within reach of a single person who knows what they want to achieve. The bottleneck has moved. It is no longer execution. It is taste, judgment, and the willingness to own the whole arc.
This is what I want engineers to hear, and what I want founders to internalize: the comforting line about AI — that the safe jobs are the deeply human ones — applies to engineers more than most people realize. It just applies differently than they think. The most valuable engineer in 2026 is not the one who writes the most maintainable code; AI is steadily closing that gap. The most valuable engineer is the one with a thesis about a user, the conviction to build the thing, and the range to make people care about it once it exists.
What I find most interesting about this role is that it is defined less by the tools than by a mindset. Curiosity about the user. Ownership of the outcome. Taste in what "good" means. A refusal to hand the work off the moment it gets uncomfortable. The tools are table stakes. The mindset is the moat.
The way I put it to engineers who feel destabilized by all this: you're still an engineer. You're just no longer engineering software. You're engineering human value. The verb didn't change. The object did.
I want to be careful not to confuse this with "full-stack" in some bloated, do-everything sense. It is closer to the opposite. AI lets you stop pretending the org chart matters as much as it used to. If you have a clear view of what would help a user, you can now build it, position it, show it, and ship it — without convening a committee. The job is more concentrated, not more diluted.
I think that's what we should call the next generation of engineers. Not full-stack. Not 10x. Product Innovation Engineers — the ones whose value finally has nothing to do with the maintainability of the code they leave behind.