Hypothesis
Name the learning problem precisely : usually one the default course shape can’t solve.
How the lab works
Start with the problem. Build just enough to test a better answer.
Most learning ideas die as slide decks. Here they get built. Every experiment starts as a sharp question about how people actually learn or decide : then becomes a working prototype we test in the open, keep what holds, and turn into a reusable pattern.
→
Prototype
Build a functioning version : often vibe-coded with AI : fast enough to learn from.
→
Test in the open
Run it as a learner, a designer, and a builder. Keep what holds up under real use.
→
Pattern
What survives becomes a reusable method : feeding the Synthlet, the products, and COSMOS.
What this place is for
Explore what happens when a learning problem is investigated, prototyped, and tested instead of treated as another request to fulfill.
The Innovation Lab attracts curiosity : it's proof of how we think, not the main event. The Lab and Synthlet are where that curiosity turns into ongoing support, community, and real, structured work.
The experiments
Working builds, with their status in the open.
6
Showing all experiments
Research threads
The questions every experiment keeps circling.
These aren’t shipped products : they’re the open inquiries that shape what we prototype next. Each one already has published thinking and tools behind it.
Judgment
Decisions under pressure
Can practice build the judgment that information alone never does? → Signal Room
MemoryDesigning for forgetting
What does retrieval, reinforcement, and transfer look like by design? → Library
AIAI & design judgment
Where does AI accelerate the build : and where must judgment stay human? → Perspectives
FeedbackFeedback that coaches
Can feedback help people reason and recover : not just reveal right or wrong? → Perspectives