The Hitchhiker's Guide to the K-Shaped Economy
AI changed what we can build. It also changed what we need to know. Here are the six skills that matter now.
The cost of building just dropped by an order of magnitude. Everyone is talking about that. Almost nobody is talking about what it actually demands of us.
The job market is splitting into a K shape. The roles that rise are the ones that involve knowing what to build: product thinking, architectural judgment, problem framing, quality standards. The roles under the most pressure are the ones defined entirely by executing what someone else already specified. Not because the people in those roles lack talent, but because the economics of execution changed under them.
This is not a developer story. A product manager, a marketing director, a founder, an operations lead: the same shift is hitting everyone. Execution got cheap. Judgment got expensive. And the skills that bridge the gap are not the ones most of us were trained in.
This series is about those skills. Six of them. Not technical skills. Human skills. The ones that determine whether AI makes us more capable or just more productive at building the wrong thing.
The Six Skills
1. Specification: “What do you actually want?”
The ability to describe what you want clearly enough that something which cannot read your mind will build it correctly. AI builds exactly what you ask for. The problem is that what you ask for is often not what you meant. This skill turns “build something like this, you know what I mean” into a spec precise enough to execute on. Everything else in this list depends on it.
2. Judgment: “Is this actually good?”
AI has capability without taste. It can produce a hundred options in the time it used to take to produce one. The bottleneck is no longer generation. It is evaluation. Knowing whether the output is good, whether it solves the right problem, whether it will hold up in the real world: that is judgment. It is the skill that took you fifteen years to develop, and it just became the most valuable thing you own.
3. Decomposition: “How do I slice this?”
A large ambition handed to an AI agent whole will produce chaos. The same ambition, broken into pieces small enough to carry, produces results. The skill is not solving the problem. It is slicing it into tasks that are specific enough for a machine to execute independently, and sequencing them so the pieces fit back together. This is the difference between someone who has an idea and someone who can actually get it built.
4. Orchestration: “How do I run this team?”
AI agents are extraordinarily fast, moderately reliable, and have no sense of direction. Managing them is not a technical skill. It is a management skill. Knowing when to intervene, when to let them run, how to coordinate multiple agents working on different pieces of the same problem: this is the new management, and it applies whether you are an engineer, a founder, or a department head.
5. Intent: “What do I actually mean?”
“Prompting” used to be a single skill. It has since split into at least four distinct capabilities, and most people are only practicing one of them. Intent goes deeper than prompt engineering: it is the ability to express what you mean at multiple levels, from a single query to a system-level specification that governs how an agent behaves over days of autonomous work. The gap between people who understand this and people who do not is already enormous and widening.
6. Evaluation: “How do I know this is wrong?”
AI is confidently wrong some percentage of the time. Not randomly wrong. Confidently, plausibly, articulately wrong, in ways that are difficult to catch unless you know what to look for. The skill is not checking every line. It is knowing which lines to check, recognizing the patterns of confident error, and building verification habits that catch problems before they ship.
Why This Order
These are listed in order of dependency, not difficulty.
You cannot judge output until you have specified what good looks like. You cannot orchestrate agents until you can decompose the problem they are working on. You cannot express intent effectively until you understand what you are trying to achieve at multiple levels. And you cannot evaluate results until you have the judgment to know what right looks like.
Specification is the foundation. Everything else builds on it.
Who This Is For
If you write code, this series will change how you work with AI tools.
If you manage people who write code, it will change how you evaluate and support them.
If you do not write code at all, and you are a product manager, a designer, a founder, a director: this series might matter more to you than to anyone, because these skills are not programming skills. They are thinking skills. And the people who develop them first will have a compounding advantage that grows with every generation of AI tools.
The tools are improving every month. The human skills that make them useful are improving much more slowly. That is the gap. This series is about closing it.
First deep dive, on Specification, drops in two days. Follow along.
