Introduced Lists, a new concept that let badgeholders curate projects within their area of expertise and share proposed allocations with reasoning. This turned 500+ projects from an individual evaluation burden into a collaborative filtering problem, where voters could build on each other's domain knowledge instead of reviewing everything alone.

Replaced percentage-based allocations with absolute OP values so voters could feel the weight of their decisions. "Give this project 0.3%" is abstract. "Give this project $120,000" is a commitment you can actually evaluate and defend. This single shift changed how voters reasoned about their ballots, moving from distributing a pool to making individual funding decisions.

Improved filtering, sorting, and project profiles so voters could evaluate nuanced differences between 500+ projects without reducing impact to a single score. The goal was to make large-scale evaluation feasible without flattening the complexity of what each project actually contributed.

Added review states, staged commitment, and confirmation checkpoints at submission so voters felt confident their ballot reflected their actual intent before actions couldn't be undone. Slowing people down at the right moments counterintuitively increased trust in the process.