Week 7 · Tuesday, March 4
The Narrowing Move
The hardest skill in problem solving — and the one that separates class projects from real work.
Week 7 · Tuesday, March 4
The hardest skill in problem solving — and the one that separates class projects from real work.
The Destination
"What problems are now within my reach that I would not have attempted before?"
AI expands what's possible for you. But only if you can identify worthwhile problems, break them down, and maintain direction through complexity.
"What did I specifically contribute that AI could not have done alone?"
Superagency without human value is just delegation. You need to articulate the judgment, context, and integration that only you provide.
These aren't philosophical questions. By the end of the semester, you answer them with evidence from things you actually built.
The Path
Sprints 3 and 4 are where the real-world, portfolio-worthy work happens. But they only work if you've built the muscles in Sprints 1 and 2. You can't solve for a stranger if you haven't learned to solve for someone you know.
Where We Actually Are
I know some of this has felt like busy work. Reflections, audits, learning plans — they can feel like hoops to jump through when you want to just build something.
I also know the submission rates dropped hard this past week. Some of you are behind on Domain Learning and Demo Design. That's data I can see.
Here's what I want you to understand: every one of those "exercises" is building a specific muscle. Discovery taught you to listen for what people don't say directly. The Assumption Audit taught you to separate evidence from story. Domain Learning teaches you to let science redirect your instincts.
These are not academic exercises. They're the exact capabilities that make you valuable in a world where AI can write code, generate designs, and produce content. What AI can't do is sit across from a person, hear what they're actually struggling with, and narrow it to something buildable. That's a human skill. And it only develops if you do the reps.
The pivot we need to make ourselves useful in the age of AI requires fundamental capability building. That takes time and reps. Sprints 3 and 4 are where it pays off — but only if the foundation is there.
Today's Session
I've put together a structured activity that walks you through the hardest move in problem solving — going from "I understand my partner's situation" to "I'm building THIS specific thing, for THIS specific reason, and here's how I'll know if it worked."
Two parts:
Interactive examples built from real problems in this class. You pick the one closest to your partner's problem and step through the narrowing process. 10 minutes.
Open the Dojo. AI guides your thinking in 4 rounds. Your partner checks your thinking at 3 pause points. You walk out with a locked MVP statement.
If you're behind on Domain Learning — that's OK for today. The Dojo will weave research into the conversation so you learn as you narrow. But you'll need to go deeper before you build. This catches you up, it doesn't replace the work.
What This Leads To
This isn't just "turn in a prototype." The Build Log is where you show the thinking. Every decision you made — what form the solution takes, why you narrowed to this component, what research informed your design, what your partner said that changed your direction — that's the substance.
Today's session — the narrowing, the partner conversations, the domain learning the AI pushes you on — is directly building the material for this assignment. What happens in the next hour is your build log's backbone.
Due: Monday, March 9. Your partner should be using what you built before Demo Day.
Sprint 2 · Where You Are
Part 1 · 10 minutes
I've built three examples from real problems in this class. Pick the scenario closest to your partner's situation. Step through it — you'll answer questions about the thinking before you see what happens next.
This isn't a reading exercise. Each step asks you to identify the thinking move before it reveals the answer. Pay attention to where the student's first instinct was wrong and what changed it.
Three scenarios: Task Delay, Overcommitment, Job Search. Pick one.
Part 2 · Your Turn
Same structure you just saw — but with your partner's actual problem. The Dojo will guide you through 4 rounds. Your partner will check your thinking 3 times.
Go to CST395: S2 - Practice Dojo - Learn and Solve in the Symbiotic Thinking Dojo tool.
Open the Dojo. The AI will ask you about your partner's problem, challenge your first instinct, help you find the sub-problems, push you on what research says, and help you define exactly what to build.
At three points, it will tell you to stop and go talk to your partner. It won't continue until you report back what they said.
These pauses aren't optional. The AI pushes your logic. Your partner is the only person who knows if your logic matches their reality.
AI pushes the thinking. Partner provides the truth. Neither works alone.
Activity · 40 minutes
Describe your partner's problem. AI challenges your first instinct — are you projecting?
Close the tab. Ask partner: "Walk me through the last time this happened — minute by minute."
Report what partner said. AI helps decompose into sub-problems + pushes you on what research says.
Close the tab. Name the sub-problems. Ask: "Which one, if it disappeared, would change things most?"
Define what to build. AI pushes on form — match your partner's life, not your CS instincts.
Close the tab. Describe what you'd build. Ask: "Would you use this? How would we know it worked?"
Lock your MVP statement. Specific component, what you're building, how they use it, measurable outcome.
While You're Working
Ask: "What did you do instead of the thing you were supposed to do? What happened right before that?" People remember avoidance behavior more clearly than the trigger.
The "last Tuesday at 3pm" test: Can you name the specific time and place your partner would use what you're building? If not, it's still too broad.
Ask partner: "If this worked perfectly, what would be different about next Thursday compared to last Thursday?" Their answer is your outcome.
They're being polite. Try: "I'm going to describe two options — tell me which you'd actually open on a Tuesday night when you're tired." Forced choice reveals real preference.
Before You Leave
If you finished the Dojo session — read your MVP statement to your partner one final time. If they nod, you're locked.
If you're not fully through the rounds, write where you are and what you still need to resolve. You can finish the Dojo conversation tonight.
1. Your MVP statement — specific component, what you're building, how your partner uses it, measurable outcome. If it takes more than 3 sentences, it's still too broad.
2. What form does v1 take? — Not "an app." The specific thing: a Google Sheet, a daily text protocol, a physical card, a decision flowchart.
3. When does your partner start using it? — Agree right now, while you're sitting next to them.
This Week
S2: Productive Reflection #5
What happened during the narrowing? What did your partner say that surprised you? Where did the AI push you that your own thinking wouldn't have gone?
S2: Prototype v1 + Build Log 20 pts
Build the MVP you locked today. Your partner should be using it before Demo Day. Log every decision — what you chose, what you rejected, why. The log IS your demo material.
If you're behind on Domain Learning: Do it this week as part of your build process. When you're deciding what form your solution should take, look up what research says about why your partner's problem is hard. That research will directly improve what you build — and it goes in your build log.