Addy Osmani recently made a point that stuck with me: narrow specialization has become a risk. AI commoditized the purely technical parts of software engineering. The engineers who thrive now are T-shaped — deep expertise in one area, broad competence across many others.
I’ve been thinking about this a lot, because it describes exactly what happened to me.
The shape of a career
The T-shaped model is simple. The vertical bar is your depth — that one corner of the stack where you feel invincible. The horizontal bar is your breadth — every adjacent area where you’re fluent enough to ship work without tapping someone else on the shoulder.
Most of us have a strong vertical. It’s the natural outcome of doing the same kind of work for years. You find your thing, you get really good at it, and it feels great. But as the industry explodes with new complexities — cloud infra, pipelines, containerization, design systems, data engineering — most of us do the same thing: we stay in our lane. Not because the rest isn’t important, but because at the end of a long day, you gravitate toward what you know and love. Let someone else handle the rest, right?
It was a comfortable way to work. Until it wasn’t.
The gap you can feel
Eventually, you hit the invisible wall. You design a solid architecture, you deeply understand the user’s problem, but you can’t actually ship the thing without depending on someone else. You know exactly what needs to be done, but you lack the operational muscle to do it yourself.
Maybe it’s deploying your own side project. Maybe it’s designing a decent interface for the tool you built. Maybe it’s automating a tedious workflow, or building a CLI that would save your team hours every week. You have the idea. You have the context. But you’re stuck at the edge of what you know.
And the thing is — it’s not just a knowledge problem. Even if you’re willing to learn, there are only so many hours in a day. You already have a job. You already have deadlines in your area of expertise. When would you find the time to get fluent in infrastructure, or design, or product thinking? The gap isn’t just what you don’t know. It’s what you’ll never have time to learn.
So, we accepted the gap.
AI changed the equation
When I first tried AI coding tools, I was just looking for a faster horse — a way to write my usual code with less boilerplate. But the real magic wasn’t speed. It was will.
Suddenly, those dark corners of the stack I had been avoiding for a decade weren’t so scary anymore. Setting up deployments? Building a blog from scratch? Editing photos for a project? Creating CLI tools? Automating the boring stuff that I’d been doing manually for years? I wasn’t flying blind anymore. AI became the ultimate pair programmer for all the things outside my comfort zone.
None of this replaced my core expertise. My vertical bar is still where my deepest knowledge lives. But AI dramatically expanded the horizontal bar. It filled in gaps I had been carrying for years — not by making me an expert overnight, but by giving me enough fluency to actually build things I never would have attempted before.
The superpower no one talks about
The conversation around AI in software engineering tends to focus on productivity. “Ship faster.” “Write less boilerplate.” “Do more with fewer people.” All true, but missing the point.
The real superpower is range.
Not just writing code faster in your comfort zone, but being able to step outside of it entirely. Starting that side project you always postponed because the backend felt intimidating. Designing a UI that actually looks good, even though you’re a backend engineer. Deploying your own infrastructure instead of waiting for someone to do it for you. Building tools, automating workflows, exploring creative work — things that were always technically possible but practically out of reach.
To be clear: being T-shaped doesn’t mean doing the job of five people. There’s a valid critique that “full-stack leadership” is often just corporate cover for understaffing. That’s not what this is about. It’s not about replacing specialists — it’s about having enough fluency to not be blocked by the absence of one.
AI didn’t make me a specialist in everything. It made me dangerous enough in everything that my deep expertise actually compounds. That’s the difference.
The new career math
The old equation was simple: go deep, stay deep, become the expert. And that still matters — without a strong vertical, there’s nothing to anchor the T.
But the horizontal bar is no longer optional. AI commoditized narrow technical execution. What it can’t replace is the engineer who sees the whole picture and can actually move across it — not just in theory, but in practice.
The future isn’t engineers versus AI. It’s T-shaped engineers with AI versus everyone else.
If you’ve been avoiding the parts of this craft that don’t excite you, now is the time to reconsider. Not because you need to master them. But because AI made it possible to be fluent in them — and that fluency changes everything.