See data in everything
Design the invisible
See a world as data — document, analyze, find insights
Design things worth keeping
Translate a place into something someone else can feel
Articulate your design principles
Design the collaboration, not just the product
Honest iteration — early work is material, not precious
Edit ruthlessly — what you remove matters as much as what you keep
Know who you are as a designer — and show it
Week 7
Chengwei Liu — How to Be a Smart Contrarian (HBR 2021, forthcoming book). Most people reject unfamiliar ideas by reflex. Smart contrarians evaluate them on merit instead.
Class guest Monday: Chengwei Liu joins us in person.
Supporting frame: Sutton & Rao, The Friction Project — smart contrarianism in action is friction that makes the wrong things harder.
Pixar’s Braintrust — institutionalized smart contrarianism.
Gita Robot & Off Their Plate (Altringer Eagle) — designing for someone else’s life means being a smart contrarian about your own assumptions.
Design Intelligences (Altringer Eagle) — ownership inflation as the failure of contrarianism. MakerCoach.
“Designing Useful Friction”
Design AI’s role as your smart contrarian. Return to your Week 3 rabbit hole and design for the person you met there. AI argues back — productive disagreement, not compliance. “World where this is normal” visualization.
Process log + visualization + brief due Monday by 12pm.
Chengwei Liu
How to Be a Smart Contrarian (HBR 2021; book forthcoming)
Class guest, Monday. Professor of strategy and behavioral science; studies how individuals and organizations escape the default of dismissing unfamiliar ideas.
Sutton & Rao
The Friction Project — supporting frame
Read the HBR article before Monday. The book extends the argument.
Most people reject unfamiliar ideas by reflex. There are good reasons to be suspicious of new ideas — many are unrealistic — but over time the defensive posture hardens into habit and limits what you can see. Naive contrarianism is rejecting consensus to feel different. Smart contrarianism is evaluating ideas on their merits, including the unfamiliar ones, especially when the crowd has already dismissed them.
(1) Dismissing assumptions that challenge your frame. (2) Stereotyping the people who hold unfamiliar positions. (3) Discounting anything unfamiliar before evaluating it. Each one closes a door that was worth opening. The smart contrarian asks: “What if this were true? What would have to be true for this to be the right move?”
If you ask AI “help me design X,” you get consensus — a fluent, average answer. If you ask it to argue against your direction, or to represent your subject’s perspective against your assumptions, you get something that can actually move your thinking. Design the role. Treat AI as a smart contrarian, not a search engine: deliberately introduce productive disagreement into the loop.
Smart contrarianism is friction designed to make the wrong things harder — harder to default to your own assumptions, harder to accept the first idea, harder to skip past unfamiliar perspectives. Bad friction makes the right things harder. Good friction makes you slow down at exactly the moments you would otherwise rush through.
Before Monday: Read the HBR article. Bring one moment from your Week 3 research where you (or AI) might have been too quick to dismiss something.
A group of trusted peers who give brutally honest feedback about the project, not the person. The director doesn’t have to follow the feedback — the power stays with the maker. Liu connection: the Braintrust is institutionalized smart contrarianism — a structure that makes it safe to challenge consensus and survive it. Takeaway: Decoupling criticism from authority lets people hear hard truths without defensiveness.
A cargo-carrying sidewalk robot at Piaggio Fast Forward required Material + Artificial + Sensory + Service Systems all at once — and designing for someone else’s daily life, not your own. Liu connection: the hard part was being a smart contrarian about your own assumptions about how people move through their day. Takeaway: Real design problems don’t stay in one system.
Pandemic relief designed for people in crisis, under extreme time and resource constraints. Liu connection: under pressure there’s no time to be wrong about who the design is for; smart contrarianism becomes a survival skill, not a luxury. Takeaway: Designing for someone else means getting out of your own head about what they need.
148 creative professionals, 6 continents: smart teams have predictable failure modes. 31 biases mapped. Ownership inflation is an absorbing state — once a team enters it, they almost never leave on their own. Liu connection: ownership inflation is the failure of contrarianism — you stop being able to evaluate the idea you own. MakerCoach helps teams design the contrarian role back in. Takeaway: Design the collaboration, don’t just have a conversation.
Three weeks ago you went deep into a rabbit hole. Since then you’ve designed a 30-year object, co-designed a sensory experience, and (will have) curated your taste and articulated your design principles. You have new eyes.
This week you have a chance to return to that world from your interviewee’s perspective — but this time you’re not just researching. You’re designing for someone in it: the person you interviewed, or the persona you built from your research. And you’re not doing it alone. AI joins your team — but only if you design its role.
Note: if you have a strong alternate idea, propose it to Dr Eagle. Make these assignments work to help you advance your professional goals.
A cargo-carrying sidewalk robot required collaborating across robotics, industrial design, and urban planning, and designing for someone else’s daily life, not your own. Material + Artificial + Sensory + Service Systems all at once.
The hardest part was not the engineering (okay, it was at times!). But long-term, it was understanding how very different people actually move through their day and designing something that fits into a life you don’t live.
There is not one answer. It could help healthy people lug groceries or yoga mats. It could help people with accessibility needs. How does a team decide?
Chef Tracy Chang · 2020 pandemic relief
Service Design Under Pressure
Pandemic relief designed for people in crisis, under extreme time and resource constraints. When you can’t iterate slowly, you have to listen faster. Every assumption about what people need gets tested immediately by reality.
Smart teams have predictable failure modes. Ownership inflation is an absorbing state: once a team enters it, they almost never leave on their own. The structure has to be designed to interrupt the pattern.
Starting in 2010, my study of 148 creative professionals across 6 continents (designers, architects, strategists, and engineers at firms like IDEO, Frog, and Arup) asked a simple question: what makes creative collaboration work, and what makes it fail? Fifteen years in, two findings keep coming back.
First: the conditions for creative intelligence are designable. Woolley’s collective intelligence work showed that group IQ is real. Some teams are genuinely smarter than others, and individual talent barely predicts which ones. Structure does: turn-taking, social sensitivity, diversity of perspective.
Second: smart teams have predictable failure modes. The same biases and cascades show up across firms, countries, decades. Same patterns, same outcomes. People inside the team can’t see them happening.
The Design Intelligences project mapped 31 distinct biases in creative team decision-making. Some are familiar (confirmation bias, sunk cost). Others are specific to creative work: ownership inflation (once you feel ownership of an idea, you stop being able to evaluate it honestly), aesthetic capture (a beautiful execution masks a weak concept), expertise anchoring (the most experienced person’s first suggestion becomes the frame everyone else works within).
Bias density (how many biases are active simultaneously) predicts decision quality better than any individual bias alone. Ownership inflation is an absorbing state: once a team enters it, they almost never leave on their own.
Catmull built the Braintrust around exactly this. Pixar’s rule that “the director doesn’t have to follow the feedback” exists because once you feel you own the idea, you can’t hear honest feedback about it. The Braintrust isn’t politeness. It’s structure doing the work the team can’t.
This research led directly to building a tool (MakerCoach) that helps teams deliberately design their collaboration. Not just who does what, but what role each team member plays in the creative process. Who is the critic? Who advocates for the user? Who checks assumptions?
This week, AI joins your team. The question is not “can AI help?” (of course it can). The question is: what role should AI play? If you just ask AI to “help me design X,” you get compliance (the AI equivalent of a team with no Braintrust). If you deliberately design AI’s role (critic, advocate, systems checker), you get productive friction.
These patterns don’t disappear when AI joins the team. They show up faster, because AI is fluent and confident by default. The collaboration itself is what you design. The conversation is the result.
Open your Week 3 interview notes. Open any AI chat. Try one of these contrarian prompts:
Design Challenge 7
Design AI’s Role
Cast AI as your smart contrarian, not your assistant
Design for Person or Persona
Return to your Week 3 rabbit hole (or approved alternate)
Version 1 Package
Simple proposal, persona, set of weekly milestones, and journey map showing a world where your idea is now normal
Version 1 Milestone completion due next Monday
1. Cast AI’s role as your smart contrarian. Before you start, decide: is AI your critic? Your subject’s advocate? Your assumption-tester? Design the collaboration, don’t just have a conversation. 2. Design something for the person you interviewed (or the persona you built) from your Week 3 rabbit hole research. Your choice, based on your relationship with the person you spoke with. 3. Use AI as a disagreeable partner: feed it what you learned, push it to disagree, stress-test whether your design serves THEM or YOUR ASSUMPTIONS. 4. Create a simple proposal, persona, set of weekly milestones, and journey map showing a world where your idea is now normal. Idea must engage at least two systems. 5. Submit: brief on the person (in their words), process log (3+ friction moments with AI said / you did / why, first vs. final direction, one sentence on what AI saw that you didn’t), visualization, principle transfer statement, plus your 3 weekly gallery slides.
Liu’s point: most people dismiss unfamiliar ideas by reflex. Your job is to design a process that interrupts that reflex. Friction that makes it harder to default to your own assumptions and easier to see your subject’s actual needs. If your final design is the same as your first idea, the contrarian wasn’t doing its job. Either AI was too compliant, or you were too quick to dismiss what it said. The process log documents the moments where you actually engaged with something unfamiliar instead of waving it off.
Before Monday: read Chengwei Liu, How to Be a Smart Contrarian (HBR 2021). Re-read your Week 3 brief, rabbit hole notes, and interview with fresh eyes. Monday guest: Chengwei Liu, author of How to Be a Smart Contrarian (HBR 2021; and forthcoming book). Be prepared to discuss a moment where you (or AI) might have been too quick to dismiss something. Wednesday studio (90 min): design + AI studio to get to milestone version 1, clear milestones, making plan, persona definition, user journey maps (old world) and (new world where your idea is now normal), document friction moments as you go. Working question: “Is my design for THIS PERSON or for my idea of them? Is my plan feasible by the end of class presentations? What one thing should I prioritize making each week and why?”
WIP Doc — Week 7 section. See Challenge slide for full description.
Final project plan: describe your final project idea, your teammates (if working on a team), and clear goals for V1 (your work this week — the seed), V2 (Week 8), and V3 (Week 9). Final presentation: Week 10.
Concept tags — mark which apply to your work this week:
Week 7: Smart Contrarianism (Liu) · Useful Friction · AI as Disagreeable Partner · Designing the Role · Ownership Inflation · Productive Disagreement
Week 6: Curation as Self-Knowledge · System Clustering · Body of Work
Week 5: Multisensory Design · Frame Innovation · Emotional Design · Obsessive Sensory Attention · Place as Story · Making Invisible Perceptible
Week 4: Constraints Increase Variability · Emotionally Durable Design · Craftsmanship · Constraint as Identity · Durability as Desirability · Repair as Understanding
Week 3: Disappearing Interviewer · Curiosity as Structure · Interviewing Users · Invisible Systems Made Visible · Dual Desires · Going Deep
Week 2: Structure Mapping · Far-Field Analogies · Technology Brokering · Experience Economy · Affordances · Anticipation as Design · Controlled Reveal · Sensory Arc · One Resonant Idea
Week 1: Service · Energy · Material · Interactive/Sensory · Natural · Artificial · Longevity
Weekly Gallery Slides — your name section (4 slides):
Slide 1: Brief on the person or persona you designed for (in their words, not yours) + the design direction you chose. What did your Week 3 research reveal?
Slide 2: Your process log. 3 friction moments (AI said → you did → why). First direction vs. final direction. One sentence on what AI saw that you didn’t.
Slide 3: Principle transfer statement. What concept(s) from this week’s theory or case studies you engaged with, and how you applied it in your design.
Slide 4: Final project plan — your idea, your teammates (if any), and clear goals for V1 (this week, the seed), V2 (Week 8), V3 (Week 9). Final presentation Week 10.
Week 7
Chengwei Liu — How to Be a Smart Contrarian (HBR 2021, forthcoming book). Most people reject unfamiliar ideas by reflex. Smart contrarians evaluate them on merit instead.
Class guest Monday: Chengwei Liu joins us in person.
Supporting frame: Sutton & Rao, The Friction Project — smart contrarianism in action is friction that makes the wrong things harder.
Pixar’s Braintrust — institutionalized smart contrarianism.
Gita Robot & Off Their Plate (Altringer Eagle) — designing for someone else’s life means being a smart contrarian about your own assumptions.
Design Intelligences (Altringer Eagle) — ownership inflation as the failure of contrarianism. MakerCoach.
“Designing Useful Friction”
Design AI’s role as your smart contrarian. Return to your Week 3 rabbit hole and design for the person you met there. AI argues back — productive disagreement, not compliance. “World where this is normal” visualization.
Process log + visualization + brief due Monday by 12pm.