Summary
AI sharply lowers the cost of specialised execution, shifting the bottleneck from doing to deciding. The most durable profile isn't generalist or specialist — it's both: wide enough to apply AI across domains, deep enough to go beyond it where it breaks. Intelligence is becoming cheap; judgment is not.
The debate has been running for years. Is the future owned by the curious generalist who can move fluidly across fields, or by the deep specialist who knows one thing better than any machine?
AI doesn't resolve the debate—it reframes it entirely.
Why generalists are rising
AI dramatically lowers the cost of doing specialised tasks. A generalist today can write functional code with an AI copilot, run data analysis without a statistics background, and produce credible first drafts of legal, technical, or marketing content in minutes. These were once genuine moats. They are becoming commodities.
When execution gets cheaper, the bottleneck shifts. It moves from how to what—from knowing how to carry out a task to deciding which tasks are worth doing in the first place. That favours people who can see connections across domains, frame problems well, and make sound judgments under uncertainty. In other words, AI amplifies breadth.
Why specialists are still critical
That doesn't mean depth is obsolete. AI is powerful, but it isn't reliably original, fully trustworthy, or context-aware in edge cases. Somebody has to validate its outputs, correct its errors, and push past the frontier of what it can already do. In high-stakes domains—medicine, law, safety-critical engineering, frontier research—the cost of a wrong answer is too high to delegate to a model without expert oversight.
The more important distinction is between kinds of specialists. Specialists whose expertise is largely procedural, repetitive, and well-documented are genuinely at risk. If your value comes from executing a process that AI can replicate, then AI will compress that value. But specialists who are operating at the frontier—generating new knowledge, navigating genuine ambiguity, exercising judgment that can't be extracted from a training set—remain indispensable.
The demand for true depth isn't falling. The demand for the appearance of depth is.
The real winners: T-shaped and π-shaped people
The most durable profile combines both dimensions. Broad enough to move across domains and leverage AI tools effectively. Deep in one or two areas where the work goes beyond what AI can reliably do.
Product managers who can code and understand users. Investors who grasp technology, markets, and policy together. Engineers who also understand business and design constraints. These aren't unicorns—they're people who resisted the pressure to optimise narrowly.
The logic is straightforward: generalist breadth lets you use AI across domains; specialist depth lets you go beyond AI where it breaks down. The two advantages compound. Without breadth, you miss opportunities. Without depth, you have no floor.
A blunt summary
A pure generalist with no meaningful depth risks having their output replaced outright. A narrow, routine specialist risks having their value steadily compressed. A hybrid thinker—someone with genuine depth and wide curiosity—finds that AI multiplies their impact rather than threatening it.
What this means practically
The practical implication isn't to abandon either path. It's to be deliberate about where you're building depth and whether the domains you're curious about are genuinely connected to that core.
Keep building technical depth where it matters—not surface familiarity, but the kind of understanding that lets you catch what the model gets wrong. And actively expand into adjacent domains: policy, business, design, whatever sits closest to the problems you care most about. The goal isn't to know everything. It's to have enough range that you can frame problems well, and enough depth that you can actually solve them.
Above all: practise problem selection and decision-making, not just execution. In a world where execution is increasingly automated, the ability to decide what to build, what to fix, and what to leave alone is the skill that compounds most.
That's the real edge. Not generalist or specialist. Both, in the right proportions.