Technical
The Skill Soup Thesis, Crystallized After Eleven Months
For almost a year I have written about skill soup. Eleven months of evidence later, I can state the thesis without hedging.
The Thesis
In an era of AI coding tools, the highest-leverage individual is not the deepest specialist. It is the broadest integrator. The person who can wire together backend, frontend, AI, infra, and business context into a single shipping thing. Not because they are better at any one layer than a specialist. Because they are the only one who can move the work from intent to deployed in a single session.
Why It Works Now
AI collapses the per-layer depth tax. You no longer need to memorize Django internals, React internals, Terraform internals, and SQL tuning tricks. You need to know enough to direct, verify, and correct. The AI provides the depth. You provide the breadth and judgment.
Before AI, breadth was a penalty. You knew a little about everything and were slow at each thing. After AI, breadth is a multiplier. You knew a little about everything and the AI makes you fast at each thing.
The Evidence
Every client engagement this year that succeeded had one person on my side who could hold the whole picture. Every engagement that stalled had work fragmented across specialists who could not see each others context. The generalist multiplied. The specialist blocked.
The Counter-Thesis I Tested
I tried hiring specialists for a quarter. Better code per layer. Slower shipping overall. The integration cost ate the depth benefit. Data is in my notebook.
What Skill Soup Is Not
It is not dabbling. It is not shallow. It is not a generalist brag. Skill soup is a deliberate capability portfolio where each ingredient meets a minimum bar and the combination beats the sum. The soup metaphor matters: not a salad where ingredients stay separate.
The Minimum Bar Per Skill
- Backend: can build a production API with auth, persistence, and observability
- Frontend: can build a responsive app with forms, state, and routing
- Infra: can deploy serverless, monitor, and rollback
- AI: can direct agents, verify outputs, and catch silent drift
- Business: can write a proposal, scope a project, and deliver it
Below the bar: pay a specialist. Above the bar: do it yourself with AI.
What This Means for Hiring
Senior individual contributors are the new multipliers. Juniors without breadth will struggle to stay useful. Teams that bet on specialists will be slower than teams that bet on integrators. Read Gergely Orosz for longitudinal data on this shift.
Eleven months in, the thesis holds. The soup wins.
RELATED READING
The Consulting Shift I Am Making In Year Two
After a year of writing and building, my consulting practice is changing shape. Shorter engagements. Sharper outcomes.
ReadThe Frontend Shift: Shipping Less JavaScript In Year Two
A year ago I reached for Next.js for everything. This year I often reach for nothing.
ReadThe Serverless Lesson I Would Write On A Sticky Note
After a year of shipping serverless projects, one rule explains most of the wins and all of the losses.
Read