Breadth needs depth

The belief

Pure breadth is commoditized. A generalist with no specialty is now indistinguishable from a chat interface: wide, available, useful to a point, replicable by a prompt. The fix is the T-shape. Keep the breadth. Add a depth-axis.

The axis is a free variable - vertical (industry, domain) or horizontal (functional, specialization). The depth can go as humanward as EQ: if you know how to use AI you will be efficient, but if you know how to connect with humans you will be loved. What is not free is the choice to pick one. Breadth without a depth-axis is the failure mode, not breadth itself.

How to apply

1. Default to the T-shape when describing your value. Lead with the breadth - range of domains, pattern-matching speed across them. Then immediately name the depth-axis. An answer that stops at breadth is a resume. An answer that names the depth-axis is a pitch.
2. Pick a depth-axis before the market forces one on you. Domain-specific AI models are already fragmenting the generalist layer. A PM who chose a depth-axis before this fragmentation lands on the right side of it; one who waited is competing with specialized tooling.
3. Audit whether your depth is real or claimed. The test: can you write something - an analysis, a PRD section, a design critique - that holds up under review from specialists in that domain? Claimed depth that fails peer review is surface exposure. Real depth survives the review.
4. Treat EQ and human connection as a legitimate depth-axis. Not as a consolation prize for non-technical roles. Knowledge is becoming a utility. The depth that stays durable is being irreducibly human - the capacity to feel, to be affected, and to bounce back. An agent cannot replicate that substrate.
5. When advising others on career positioning, apply the same test. The generalist-without-step-up is a liability pattern. The prescription is not to abandon breadth - it is to name a niche and invest.

What this is not

• Not an argument that breadth is the failure mode. The belief is breadth-needs-depth, not breadth-is-wrong. The parent thesis - that cross-domain range is a genuine differentiator - is not contradicted here. It gains a floor: breadth without depth no longer clears the bar the way it did pre-2023.
• Not a requirement to specialize early and narrow fast. The axis is free. Pivots are permitted. The constraint is: choose a niche at some point and invest. The choice does not need to be permanent; it needs to be made.
• Not a claim that depth must be technical. EQ, communication craft, first-principles thinking, systems design judgment - all qualify. The axis is any dimension where you are building something that a chat interface cannot replicate from a single prompt.

Argues against

• "Being a well-rounded generalist is enough in an AI-augmented team - breadth covers what depth used to require."
• "Specializing too early limits career optionality; staying broad keeps all doors open."
• "The AI era makes human depth less relevant because the tools can compensate."

Where to go from here

If you want the parent thesis that post-ChatGPT breadth-needs-depth refines rather than replaces, go to breadth as differentiation. The parent stands; this belief adds the binding constraint.

If you want to understand how to build depth that transfers across the axis you choose, go to learn concepts not tools. The method is concept-first - tools change, concepts compound.

If you want the parent theme with the full evidence arc across the corpus, go to the breadth-as-differentiation theme.