Reference Fees and Content Buybacks

Vector Databases and AI Media Models

AI models, particularly those used in semantic search, recommendation systems, and generative AI, increasingly depend on vector databases for efficient storage, retrieval, and analysis of high-dimensional data. As a result, these databases have become an essential part of modern AI infrastructure.

• Vector databases are specialized systems designed to store, manage, and search data represented as vectors—numerical arrays that capture the essence of information.

• Vectors act as unique "fingerprints" for various types of data, such as text, images, or user interactions.

• Their strength lies in efficiently identifying similar data points by measuring their proximity in vector space, making them indispensable for numerous AI-driven applications.

Using this model for our Worldstate allows us to organize our information in a way that is optimized for the core user of this data set: generative media models. By storing this information securely we force media creation engines to pay in order to reference this community created content. An example of this in the modern market is relationships between social media companies and massive LLM providers.

Reference Fees

To align incentives between content creators, token holders, and those who utilize the content in outputs, DreamNet implements Reference Fees. Whenever a verified piece of content from Worldstate (character, place, etc.) is queried and used by an external application or AI output, a small fee in $<universe> is charged to that application. This is analogous to paying a licensing fee or royalty for using an IP, except it’s automated on-chain and feeds directly back into the community economy.

The mechanics of Reference Fees are straightforward and powerful:

• The external creator (say a game developer pulling in a key agent, or a streaming platform featuring a community-created place) pays a fee in $<universe> for accessing those Worldstate entries via DreamNet.

• A portion of that fee is used to buy back the specific Agent or Place Token that was referenced. This is done automatically by the protocol on the open market. Essentially, whenever content is used, the market cap of the underlying idea/character is effected by its usage, reflecting its real-world value.

• The remainder of the fee is retained by the system to maintain and improve the platform. This ensures we can continue scaling DreamNet (covering AI compute costs, database hosting, etc.) as demand grows.

This Reference Fee model means popular content yields direct financial rewards to its investors and creators without a manual royalty system. For example, if a fan-created character becomes the star of a hit AI-generated web series (pulling data from Vector Worldstate each episode), every episode will generate buybacks of that character’s token. The character’s creator (who likely holds some of their token and the NFT, earning trade fees) and any investors in that token will see their asset appreciate.

It’s an automated feedback loop: more usage → more interaction with tokens → greater engagement → incentivizes more great content.

By structuring it this way, we ensure the platform sustains itself while maximizing value accrual to creators and content holders where even indirect monetization (like an AI agent borrowing your character in a story) still pays the original community creators.