The return of the generalist
For a decade we were told to specialize. AI just changed the math.
- Note
- June 2026
- 4 min read
Deep, narrow specialization was the safe career bet for years, but AI now does many narrow tasks cheaply. The durable advantage is shifting to the generalist who can hold strategy, craft, and judgment at once and decide what is worth doing across domains.
The advice I heard my entire career was the same: specialize. Go deep, own a niche, become the person who does one thing better than anyone else. I never fully bought it, and the ground has just shifted hard enough that I no longer have to argue the point. The math behind that advice has changed.
Specialization is what automates first
Here is the uncomfortable part. A narrow, well-defined, repeatable task is precisely what a machine learns to do fastest. The deeper and more bounded the specialty, the more cleanly it can be captured. The very thing that made specialization feel safe, its narrowness, is what makes it the first to be automated. The moat was an illusion the whole time.
When the narrow tasks get automated, the connective judgment is what is left.
What does not automate is the connective work: linking strategy to form, knowing which problem is worth solving, carrying taste across domains so the parts add up to something coherent. That is the generalist position, and for years it was treated as the consolation prize for people who could not commit. It is about to be the most defensible place to stand.
This is not a case for being shallow everywhere. Depth still matters; you cannot connect things you do not understand. It is a case for being the person who sees the whole and decides, with enough depth to be credible and enough range to be useful when the narrow work gets done by something else. Hold strategy and craft in the same hands. That pairing is the job that is coming back.
Asked & answered
The balance is shifting toward generalists. Narrow, well-defined tasks are the first thing AI automates, so the durable advantage moves to people who can connect strategy, craft, and judgment across domains and decide what is worth doing.
Not worthless, but less protected. Deep skill still matters, but the safest position is pairing it with the generalist ability to see the whole problem and choose where the depth should go.