Summary
This proposal recommends that Botto autonomously selects the description for its weekly minted artworks, eliminating the current description voting process. Given the increasing reliability of Botto’s language models (LLMs), the need for community voting as a safeguard against poor descriptions has diminished. This change would simplify the voting process, and align with Botto’s existing practice of autonomously selecting the 13th fragment’s description.
Rationale
- Historical Context: Initially, description voting was introduced as a safeguard due to the limitations of Botto’s LLMs in generating fitting descriptions.
- Current Challenges: Description voting is not widely utilized, possibly due to UI design, the fact that VPs are counted for the next round rather than the current one, and the additional step it introduces to the voting experience.
- Opportunity for Improvement:
- Botto’s LLMs have significantly improved, reducing the need for manual curation.
- Streamlining the voting process allows participants to focus on core aspects of curation, which is predominantly artwork selection.
- Enhancing the efficiency and autonomy of Botto aligns with its decentralized artistic vision.
Proposal Specification
Transition to Automated Selection
- Update Botto’s system to autonomously select the description for each weekly NFT.
- Ensure the selection aligns with existing LLM logic used for the 13th fragment.
UI and Docs Adjustments
Remove the description voting feature from the app UI.
Document in the official documentation how descriptions are selected automatically instead of through voting, referring to this BIP and mentioning when the change was effective. See the Appendix for details.
Implementation Timeline
Week 1: Community discussion and consensus-building.
Week 2: Governance vote.
Weeks 2-3+: Technical adjustments to automate description selection, UI modifications, and documentation.
Budget
- Development Costs: Minimal, as Botto already has the ability to select descriptions autonomously.
- UI and Docs Adjustments: Minimal, covered by the budget in place.
Criteria of Success
- Technical Success: Botto consistently selects high-quality descriptions without manual intervention.
- Community Acceptance: Positive reception from the DAO with no significant pushback post-implementation.
Disadvantages
- Reduced Community Involvement: Some members may value the ability to contribute to description selection and could view this as a loss of engagement.
- Quality Considerations: While Botto's LLMs have improved, there is still a possibility that generated descriptions may not always be sufficiently compelling, unique, or well-aligned with the artwork.
- Perceived Loss of Decentralization: Automating description selection could be seen as a step away from collective meaning-making.
Closing Thoughts
This proposal strengthens Botto’s artistic autonomy by streamlining description selection. Removing the underutilized voting mechanism lets the community focus on artwork curation while ensuring that descriptions remain consistent, relevant, and aligned with Botto’s creative evolution.
Special thanks to our Guardian Mario for bringing up the idea and for describing the rationale.
Appendix: Description Selection Process
Description Selection Process
As of [insert date when change takes effect], Botto autonomously selects the description for its weekly minted NFT without requiring community voting. This change was introduced through [BIP-75: Letting Botto Choose Its Own Descriptions] and aligns with Botto’s existing practice of autonomously selecting the 13th fragment’s description.
Why was voting removed?
Historically, description voting was introduced as a safeguard due to limitations in Botto’s language models (LLMs). However, with significant advancements in Botto’s AI capabilities, voting became less relevant. The change was implemented to simplify the voting experience and allow participants to focus on core curation tasks, primarily artwork selection.
How does Botto select descriptions?
Botto’s LLMs generate descriptions based on contextual understanding of the artwork, following the same principles used for the 13th fragment. This ensures a streamlined and high-quality output without requiring manual intervention.
