The competency map
Preserve, deepen, learn, deprioritize: four decisions about where your learning time goes.
4.1Four decisions
Where does your next learning hour go?
The proposed curriculum sorts every competency into one of four moves. They describe marginal investment: a skill can remain necessary and still deserve less manual practice.
01
Preserve
Keep the minimum mastery needed to understand the system and intervene.
02
Deepen
Invest more where your judgment steers many delegated decisions.
03
Learn
Acquire new practices: context, tools, agents and evaluation.
04
Deprioritize
Spend less time on repetitive execution and easily replaced memorization.
4.2The competency map
Sort the competencies
Place each competency where the report puts its priority move. Some deserve two moves; here you pick the main one.
The competency map
1 / 10 · Score 0Understanding the user and the domain
4.3The center of gravity moves
From doing the task to making it solvable
The contrast describes ways of organizing work — it doesn’t claim every engineer worked the same way a year ago.
Flip the switch — or tap a single row.
Key takeaways
- 1
Four moves — preserve, deepen, learn, deprioritize — describe where the next hour of study goes.
- 2
Deepen the competencies whose judgment guides many delegated decisions: domain, architecture, debugging, operations.
- 3
Learn the new practices: specification, context engineering, delegation, evaluation.
- 4
The center of gravity moves from doing the task to making it reliably solvable.
Check your understanding
Question 1 of 2
What does “deprioritize” mean in the competency map?
Read the full section of the original reportThe complete, unabridged text — with every source and nuance.
4.1. Four decisions about one's own development
The proposed curriculum organizes competencies into four moves:
- Preserve: maintain the minimum mastery needed to understand the system and intervene.
- Deepen: invest more in capabilities whose judgment guides many delegated decisions.
- Learn: acquire new practices around context, tools, agents, and evaluation.
- Deprioritize: reduce time spent on repetitive execution and easily replaceable memorization.
These categories indicate marginal investment. A skill can remain necessary and, at the same time, deserve less practice in manual execution.
4.2. Main matrix
| Competency | Priority move | Present-day expression of excellence |
|---|---|---|
| Understanding of the user and the domain | Deepen | Identify the real need and formulate criteria that represent value. |
| Code reading and semantics | Preserve and deepen | Explain behavior, recognize inconsistencies, and investigate failures. |
| Architecture and interface design | Deepen | Define boundaries that enable evolution and independent execution. |
| Algorithms, data, and systems | Preserve with selective depth | Judge scale, consistency, cost, and the consequences of failures. |
| Debugging and investigation | Deepen | Produce distinguishable hypotheses and connect symptoms to causes. |
| Work specification | Learn more explicit practices | Turn goals into contracts, constraints, and evidence of completion. |
| Context engineering | Learn | Make relevant, reliable, and current information available at the right moment. |
| Delegation and orchestration | Learn | Split work, calibrate autonomy, and integrate results. |
| Testing and agent evaluation | Deepen and learn | Design evidence that detects relevant errors and measures consistency. |
| Security of tools and environments | Deepen and learn | Control authority, data, execution, and the reach of consequences. |
| CI/CD, observability, and operations | Deepen | Close the loop between change, validation, deployment, and real effect. |
| Communication and collective learning | Deepen | Make decisions and improvements usable by other people and agents. |
| Memorization and mechanical production | Deprioritize | Automate and look things up, keeping enough understanding to evaluate. |
This matrix is a proposal of the report, not a statistical classification of the sources. Its intent is to guide decisions about study and practice.
4.3. What changes relative to the previous center of gravity
The contrast below describes ways of organizing work. It does not claim that all engineers worked the same way a year ago.
| Previous center of gravity | Center of gravity suited to the manifesto's premise |
|---|---|
| Solving a task mainly through one's own execution | Making the task solvable by a guided, verifiable execution system |
| Knowing enough detail to produce quickly | Knowing enough principles and context to direct, discriminate, and correct |
| Reviewing everything with the same procedure | Distributing verification according to properties, risk, and evidence |
| Improving one's own speed | Improving the capacity of the flow and the team |
| Solving each difficulty again | Reducing recurrence through context, contracts, tools, and tests |
| Studying to expand one's implementation repertoire | Studying to expand domain mastery, delegation power, and the reach of decisions |