Things I'm trying, testing, or have abandoned.
Running compute closer to users
Exploring offline-capable, sync-enabled apps
Using LLMs as a thinking partner