Note: Snapshot is only a summary and may leave ambiguities that are covered in the full BIP in the forum which takes precedent.
Summary
BIP-98 proposes reducing the weekly voting pool from 1,000+ images to 512. A smaller pool concentrates community attention on a tighter set of works, encourages more focused curation discussion, and improves the quality of Botto's weekly training signal. If the DAO determines 512 is still too large after implementation, a further reduction can be proposed via a separate governance proposal.
Proposal Specification
- The weekly voting pool will be reduced from its current size to 512 images.
- The selection criteria for which images enter the weekly pool remain unchanged — the reduction applies to the total number of works presented to voters, not to the generation or pre-selection process.
- Works that do not receive sufficient votes to qualify for the top 10 in a given round will be eligible to re-enter future pools according to existing re-entry rules.
- The platform team will implement the pool size reduction, including any testing required to ensure sufficient diversity in the smaller pool.
- Any further reduction to the pool size requires a separate governance proposal.
- Requires platform changes and documentation updates.
Timeline
To be implemented at the beginning of the next period.
Budget
Minor platform configuration change. No additional budget required.
Criteria of Success
- Weekly voting pool reduced to 512 images
- Measurable increase in average VP per work, reflecting more concentrated attention
- Increase in curation discussion activity in Discord and Studio
Disadvantages/Risks
- Slower feedback loop: Botto generates works at the same rate but fewer are seen by voters each week, potentially slowing the overall training signal
- Reduced discovery: some community members may feel the reduction limits exposure to Botto's full range of output
- Combined effect with BIP-95: if the VP cap per work per wallet passes alongside this, the maximum VP that can be allocated in a round is approximately 5M VP — the DAO should monitor whether this meaningfully changes voting dynamics