Sprint 2 · Due · 10 points
Your partner's problem lives in a domain you haven't studied.
What does science say about why this problem is actually hard?
If your partner struggles with procrastination and your solution is "a better planner" — you haven't learned the domain. You've just applied your assumptions with a tool on top.
Most of you are building solutions for problems that have been studied for decades — procrastination, overwhelm, overeating, exercise consistency, motivation. Researchers have identified why these problems are hard, what interventions fail, and what non-obvious factors most people miss.
If you skip the domain research and just build from your own experience, you'll build something that solves the version of the problem you imagine — not the version your partner actually has.
Start here: Read the Domain Research Reference → — it covers the science behind the most common problem domains in this class, with sources and non-obvious insights. Find the section that matches your partner's problem. Then go deeper.
Read the Domain Research Reference section for your partner's problem. Then find at least 2 additional reliable sources (academic papers, research summaries, expert-written articles — not Reddit, not AI-generated summaries). You can use AI to help you find and understand sources, but you need to actually read them.
What you're looking for: something that surprises you. Something that contradicts your assumption about why this problem exists or what will fix it. If nothing surprises you, you haven't gone deep enough.
The research is general. Your partner is specific. The domain learning plan isn't a book report — it's about translating what science says into what it means for this person and this problem.
Go back to your discovery conversation and your assumption audit. What did your partner actually say? How does the research reframe that?
State explicitly: what were you going to build before this research, and what are you going to build (or build differently) now? If the answer is "nothing changes" — either you already knew the science (cite it) or the research didn't go deep enough.
You're encouraged to use AI (Claude, ChatGPT, the Dojo, etc.) to help you find sources, understand academic papers, and push your thinking deeper. A good prompt: "I'm building a solution for someone who [partner's problem]. I've read that [what you found]. What am I missing? What do experts say about why simple solutions to this problem usually fail?"
But the document you submit must be written by you, in your own words. The AI helps you think — it doesn't write the deliverable.
"My partner procrastinates. I learned that procrastination is bad for academic performance. I'm going to build a reminder app that sends notifications before assignments are due."
No sources. No surprising insight. The "learning" is something everyone already knows. The solution doesn't connect to any research finding.
"My partner procrastinates on assignments where they don't know how to start. I assumed this was a time management issue, but research by Sirois & Pychyl (2013) shows procrastination is primarily an emotion regulation problem — people avoid tasks that trigger negative feelings, not tasks they don't have time for. My partner said starting 'feels uncomfortable' — that's the emotional trigger, not poor scheduling.
This changes my approach. Instead of a planner (which addresses scheduling), I'm exploring something that reduces the emotional friction of starting — like breaking the first step into something so small it doesn't trigger the avoidance response. The research on implementation intentions (Gollwitzer, 1999) supports this for one-time actions but warns it doesn't work well for repeated behaviors, so I need to think about what sustains it beyond the first use."
Specific sources. A finding that contradicts the initial assumption. The solution changes because of what was learned. Acknowledges limitations.
A written document with the three sections described above. 1–2 pages is the target — quality over length.