The Three Tools
Full-day orientation and first build. Both teams brainstormed challenge areas, drafted Masterplan PRDs, and deployed v1 apps by end of day.
What Happened
Day 1 exceeded expectations. The morning orientation grounded everyone in the three-tool framework, and the live demo made the Chat → PRD → Build workflow concrete. By afternoon, both teams had narrowed their challenge areas and produced working first builds — rough but real.
Key Moments
- Students immediately connected domain expertise to app ideas
- Live demo of Chat → PRD → Lovable workflow landed well
- Both teams had deployed apps within two hours of starting
Student Highlights
“I didn't expect the AI to build something I could actually show a patient.”
— Danielle
Day Content (original)
The Three Tools
Day 1 introduces the three tools that combine to make this course work:
1. Your Brain
Domain Expertise. Four years of medical training. You’ve seen what works, what’s broken, what makes patients and students suffer needlessly. That’s not background knowledge—that’s your primary building material.
2. Design Thinking
User-Centered Design. A structured way to turn “this is frustrating” into “here’s what to build.” Brief and punchy overview—you’ll learn by doing all week.
3. Generative AI
LLMs as Pattern Engines. How they’re made (training data → patterns → generation). Prompt as steering wheel, context window as working memory, hallucination as confident pattern-completion without grounding. Not a CS lecture—just enough to demystify so you can use it with intention.
Afternoon: First Build
After lunch, teams draft their Masterplan PRDs in Claude Chat, then do a guided first build in Lovable. By end of day, every team has a deployed app—however rough.
- Step 1: Brainstorm with Claude — “What tools do med students need?”
- Step 2: Draft a Masterplan PRD in Claude Chat (30–60 min)
- Step 3: Guided Lovable first-build — paste PRD, deploy, share URL
Learning Targets (as planned)
Knowledge
- How LLMs work — training data, patterns, generation
- Prompt as steering wheel, context window as working memory
- What a PRD is and why it’s the spine of every build
- The daily workflow: Chat → PRD → Build → Fresh app
Skills
- Brainstorming with Claude Chat
- Drafting a Masterplan PRD
- First Lovable deploy from a PRD
Mindset
- Process over polish
- Domain expertise is the scarce resource
- AI is a pattern engine you’re steering, not a magic box
Schedule (as planned)
| Time | Block | What Happens |
|---|---|---|
| 9:00–10:00 AM | Solo Warm-Up | Complete Day 0 exercises if not done. Baseline pulse survey. Profile page check. Time for students who haven’t seen the pre-work email. |
| 10:00–10:45 AM | Who’s in the Room | Introductions, course energy-set, student profile pages |
| 10:45–11:00 AM | Break | |
| 11:00 AM–12:00 PM | The Three Tools & Live Demo | 1) Your Brain (domain expertise). 2) Design Thinking — turning “this is frustrating” into “here’s what to build.” 3) Generative AI — how LLMs work, prompt as steering, context window. Week 1 workflow overview. Matt live demo: problem → Claude → Masterplan PRD → Lovable → v1 app. |
| 12:00–1:00 PM | Lunch | |
| 1:00–3:00 PM | First Build | Group brainstorm: “What tools do med students need?” Narrow to two challenge areas for Team A and Team B. Teams draft Masterplan PRDs in Claude Chat. Guided Lovable first-build: account setup, paste PRD, deploy. |
| 3:00 PM | Deliverable | Post team Masterplan PRD draft + first Lovable deploy in Slack |
| 3:30–4:30 PM | Group Debrief | Reactions to first build. What did the AI get right? What was missing? Introduce Week 1 rhythm. |
Tomorrow we add the User Journey layer to PRDs.