Three Durable Capabilities

for human work in an AI-augmented world

CST395 develops three capabilities that together enable students to create value in an AI-augmented world. They aren’t separate skills — every sprint activity exercises multiple capabilities at once. By the end of the course, students can answer two questions with evidence.

Superagency

“What problems are now within my reach that I would not have attempted before?”

Human Value Proposition

“What would be worse about my solutions if I had simply handed the problems to AI?”

SDL

Self-Directed Learner

Developing a meta-learning architecture for rapid expertise

What It Means

Self-Directed Learners develop a meta-learning architecture—a systematic approach to acquiring expertise in unfamiliar domains rapidly. This capability requires more than desire to learn; it’s knowing what you need to learn versus what you can strategically ignore, and using AI as learning infrastructure to build the required just-in-time depth.

Example in Action

When a student starts work on building patient communication tools, they need to understand medical terminology, HIPAA privacy regulations, clinical workflows, patient psychology under stress, and how different age groups interact with technology—in days or weeks, not semesters.

How we think about developing this in learners

We treat self-directed learning as a muscle that strengthens through deliberate, low-stakes practice. Students reflect by hand before consulting AI — slowing the loop down so the question “what do I actually need to learn here?” gets asked. Each sprint expands the territory: from learning about themselves, to learning about a partner’s domain, to learning from indirect sources when no expert is available. By the end, students are using AI as learning infrastructure — not as an answer machine — to build just-in-time depth.

IS

Integrative Solver

Operating at intersections where human value concentrates

What It Means

Integrative Solvers apply self-directed learning capacity across domains, developing T-shaped (or even m-shaped) expertise—deep knowledge in one technical area combined with broad boundary-crossing competencies. In the AI age, the horizontal bar becomes critical: they operate at intersections between systems of knowledge where AI struggles but human value concentrates.

Example in Action

When analyzing customer complaints, an Integrative Solver doesn’t just build the requested sentiment analysis dashboard. They spend time with the complaint handlers, discovering that 50 patients generate 60% of complaints—but nobody realized this because complaints were filed by date, not patient. They bridge from technical capability to human workflow, delivering simple frequency analysis that creates more value than a sophisticated tool.

How we think about developing this in learners

We enable students to operate across boundaries from the start. Structured peer conversations (16+ across 9+ unique partners) give them practice asking questions of people, not just systems. An explicit Assumption Audit makes the distinction between observation and interpretation a named cognitive move. In Sprint 2, students enter someone else’s domain; in Sprint 3, they work without easy access to experts. Across all four sprints, the Human Value Statement asks the same question: “what did you specifically contribute that AI couldn’t?”

AB

Adaptive Builder

Executing through cycles of building, testing, and adapting

What It Means

Adaptive Builders execute under uncertainty—when full information isn’t available, when the right answer can’t be reasoned out in advance. They synthesize self-directed learning and integrative solving through cycles of building, testing, learning from failures, and adapting. They exercise restraint in complexity, discovering through iteration that simple solutions executed well often deliver more value than sophisticated ones.

Example in Action

An Adaptive Builder starts with basic frequency analysis rather than ambitious sentiment analysis, testing whether it solves the core problem. Through iteration, they learn to meet user needs without over-engineering. The cycle teaches them to distinguish symptoms from root causes and to iterate toward value rather than complexity.

How we think about developing this in learners

We scaffold iteration explicitly, then withdraw the scaffold. A structured cycle (UMPIRE) gives students a vocabulary for their build process. Across Build v1 → Build Review → Revised Plan → Build v2, students test, hear feedback, and revise — under uncertainty, not as polish. Build Logs document reasoning at the time of decision; demonstrations require working evidence, not plans. The pedagogical bet: students learn to iterate well when they can’t reason their way to the answer in advance.

How They Work Together

These capabilities aren’t separate—they’re integrated:

Foundations

Techniques: how thinking is expressed Reading · Writing · Talking · Listening

Thinking isn’t invisible — it shows up in how we read, write, talk, and listen. This course exercises all four.

📖 Reading

Extracting meaning from structured information. Recognizing patterns, following arguments, identifying what matters.

Exercised through: Course content, documentation, AI outputs

✍️ Writing

Articulating your thinking clearly. Structuring arguments, explaining decisions, documenting reasoning.

Exercised through: Reflection notebooks, goal setting, demonstrations

🗣️ Talking

Explaining ideas out loud. Defending positions, articulating value, making the implicit explicit.

Exercised through: Peer conversations, check-ins, presentations

👂 Listening

Understanding others’ perspectives. Asking clarifying questions, receiving feedback, finding the real problem.

Exercised through: Peer conversations, stakeholder interviews, thought partnering

Habits: what the AI age demands Slow Down · Know Yourself · Take the Lead

AI makes speed easy. These habits make speed valuable.

⏸️

Slow Down

Resist the urge to immediately generate. Pause before prompting. Think before asking AI to think.

Built through: Handwritten reflections, wait time, deliberate process

🪞

Know Yourself

Understand your patterns, gaps, and growth edges. Self-awareness enables self-direction.

Built through: Reflection practice, peer feedback, evidence collection

🎯

Take the Lead

You direct the work. AI assists. Maintain agency over goals, quality standards, and decisions.

Built through: Sprint autonomy, self-set goals, portfolio defense

How techniques and habits combine into capabilities SDL · IS · AB

Techniques and habits combine into observable capabilities — the evidence of growth.

SDL

Self-Directed Learner

Develop meta-learning architecture for rapid expertise acquisition. Know how to learn what you don’t yet know.

Built from: Reading + Writing + Know Yourself + Take the Lead

IS

Integrative Solver

Operate at intersections where human value concentrates — connecting domains, stakeholders, and contexts.

Built from: Listening + Talking + Slow Down + Know Yourself

AB

Adaptive Builder

Iterate toward value through cycles of building, testing, and refining. Ship, learn, improve.

Built from: Writing + Talking + Take the Lead + Slow Down