From T-Shaped to Comb-Shaped

For two decades, the T-shaped professional was the gold standard. You went deep in one discipline — design, engineering, product, marketing — and built enough breadth across adjacent fields to collaborate well. Specialists with manners. The vertical bar of the T was your value. The horizontal bar made you bearable to work with.

That model is starting to break.

What changed

Generative AI has collapsed the cost of execution in almost every knowledge-work discipline. A single person with the right judgment can now produce a working prototype, a marketing plan, a system diagram, a legal first draft, and a financial model — all before lunch. The bottleneck has moved.

It is no longer about who can do the work. It is about who can decide what work is worth doing, frame it correctly, and connect it to everything else.

That decision-making lives at the intersection of multiple deep domains. And you cannot fake intersection with breadth alone.

The new shape

I have started calling it comb-shaped. Several deep teeth, connected by a strong horizontal bar of taste and judgment. The teeth I see mattering most right now:

• Design. Not visual polish — the discipline of shaping how a system meets a human need. AI can generate a hundred interfaces in a minute. Knowing which one is worth shipping is the hard part.
• Business. What the work has to do in the world. Pricing, distribution, unit economics, competitive position. Without this, AI-accelerated teams build beautiful products nobody pays for.
• Software Architecture. How the pieces fit together, what they imply about scale, and what they cost to maintain. Generative tools love producing plausible-looking code and plausible-looking diagrams. Software architecture is what tells you which ones survive contact with reality.

You do not need to be the best in the world at any one of these. You need to be genuinely competent in all of them — competent enough that when the AI gives you fifty options, you can throw away forty-eight without anxiety.

Why breadth alone fails now

The classic T-shape assumed the horizontal bar was just enough to ask good questions of specialists. That was fine when specialists were the expensive part of the system.

Now the specialists are cheap (or simulated). The expensive part is the integration — the judgment about how the pieces fit. Surface-level breadth cannot do that work, because the trade-offs live three layers deep in each discipline. You only see them if you have done the work.

A designer who has never owned a P&L will keep proposing things that cannot be funded. An engineer who has never run a usability test will keep building things nobody can use. A business leader who has never read a software architecture diagram will keep promising things that take ten times longer than expected. Each of them can ask AI to generate the missing piece — and each of them will accept the wrong answer, because they cannot tell good from bad in a domain they have never inhabited.

How to grow another tooth

If you are mid-career and good at one thing, the path forward is uncomfortable but not complicated:

1. Pick the adjacent tooth that matters most for your next decade, not the one that feels easiest. For most designers, that is business. For most engineers, it is design. For most business leaders, it is software architecture.
2. Ship something in that domain end to end, with stakes. Reading is not enough. You need the experience of being wrong in public.
3. Use AI ruthlessly to accelerate the parts you already know, and slowly on the parts you do not. The danger is letting AI mask the gap instead of closing it.

The horizontal bar of the comb — the connective taste that links the teeth — comes from doing this two or three times. There is no shortcut.

The honest part

This is harder than the T-shape. It takes longer. It is uncomfortable to be a beginner again at forty. And the people who refuse to grow another tooth will spend the next decade slowly watching their value evaporate, even as their tools get more powerful.

That is the trade. Generative AI did not make specialists obsolete. It made single-specialists obsolete. The professionals who matter now are the ones with two or three real depths and the judgment to weave them together.

Build the comb.