January 11, 2020

Design thinking as a model for learning

Seymour Papert, an AI pioneer and the creator of the Logo programming language, wrote that “Anything is easy if you can assimilate it to your collection of models. If you can’t, anything can be painfully difficult. What an individual can learn, and how he learns it depends on the models he has available."

For Papert, that model was differential gears. “Gears, serving as models, carried many abstract ideas into my head."

For me, that model is Design Thinking. Design Thinking asks for at least three things:

  1. Uncover a non-obvious insight for the context you are designing for
  2. Generate divergent ideas beyond what comes to your mind right away
  3. Make progress through enlightened trial and error

Here a few other things I have assimilated through the model of design thinking:


Jobs to be done is a more structured way to do “needfinding.” Jobs translate to a product or a feature more directly than needs do.

Design sprints

A design sprint is a strongly scaffolded design thinking project. With a clear outline for each day, and some tweaks like using drawing during brainstorming.

Meadow’s Leverage Points

Donella Meadows’ ‘Thinking in Systems’ lays down points to intervene in a system that can change the system’s behavior. Design thinking makes those leverage points actionable instead of evaluative. One of the highest points of leverage in a system is the mindset out of which the system emerges – understanding mindsets is one of the key objectives of practicing empathy. The other leverage points can serve as tools to ideate. Instead of starting with a generic brainstorming question, you can say “How might we design feedback loops to solve for x?”

First principles

First principles thinking became quite popular after Peter Thiel’s Zero to One came out, but where do you go and find first principles for human behavior? Non-obvious insight can become first principles for the subgroup you are creating products for.


Yes, I know, Design Thinking has become too buzzwordy these days and the very mention of the phrase seems to carry a miasma of bullshit, but look past that and it is an incredibly useful toolkit. The economic historian Carlota Perez says that most technologies follow a tight script. First, an ‘irruption’ phase, then ‘frenzy’, then ‘synergy’ and finally ‘maturity’. During frenzy companies often cut ethical corners to make a quick buck, and during synergy, the technology becomes more widespread. Models follow a similar script. Just because there are some bad actors in the frenzy phase, or unskilled operators in the synergy phase does not make the model wrong. (Perez’ framework found via Cenydd Bowles)