FAQ

Data Ownership Questions

Hasn't crypto solved digital ownership? Why doesn't it work for data ownership, too?

Existing crypto infrastructure has enabled digital ownership for assets like art and collectibles, but private data poses unique challenges:

  1. Data is non-excludable and can be copied once made public (the "data double spend problem").

  2. Data is non-fungible but must be aggregated to be valuable as AI training data.

Vana addresses these challenges through:

  1. Non-custodial data allows users to use data in an application or to train an AI model while keeping it in their full control

  2. Proof-of-contribution allows groups of users to pool their data while ensuring everyone is rewarded fairly. This enables data liquidity.

Why does data ownership help me as a builder?

Data ownership unlocks better AI by allowing access to new training datasets. Proper attribution and incentives enable frontier AI models collectively owned and governed by contributors with datasets that would otherwise be I walled gardensn. Data portability levels the playing field by allowing builders to access cross-platform data.

Why haven't any data ownership projects succeeded yet?

Previous data ownership projects have taken too ideological an approach and remained academic or esoteric. Vana believes in a pragmatic, full-stack approach to building infrastructure that gets real adoption. We started with a data portability API, helped build viral user-owned data apps that onboarded over 1M users, and created the infrastructure for the world's first Data DAO was created, attracting over 140k participants in under a week.

Data Liquidity Pool Questions

Is a data DAO the same as a data liquidity pool?

A data DAO is a specific form of a data liquidity pool that uses a dataset-specific token for governance. Some data liquidity pools may not create a new token and instead use a stablecoin or existing token for payments to data contributors.

How competitive will it be to earn one of the 16 slots?

Earning a slot on Vana's mainnet for DLPs is designed to be competitive. Factors affecting competitiveness include data quality, community support, and testnet performance metrics. DLPs are elected by native token holders on mainnet.

What are the benefits of building a DLP?

Building a DLP can be rewarding, with benefits such as block rewards, priority access to mainnet slots, fundraising support, and the ability to bootstrap AI projects using token incentives.

Can I start on testnet without a live token?

Yes, you can develop and test your DLP on the testnet without a live token before moving to token integration and mainnet deployment.

What criteria will DLPs be evaluated on to earn a slot on mainnet?

DLPs will be evaluated based on performance metrics like total transactions, transaction fees, verified data uploads and unique wallet interactions. On mainnet, it's up to the native token holders to vote.

Architecture Questions

Why build an L1 rather than an L2?

Vana is an L1 to ensure users maintain data privacy and control. Existing L2 sequencer approaches are too centralized and would be subject to data regulations that would not be possible to fulfill in a blockchain context.

Aren't blockchains public? How do you keep data secure?

Vana is the first decentralized network designed for private data. Data is encrypted with user-controlled keys, and access is granted to run operations in a secure environment, including from the user's device.

How do you prevent the data buyer from reselling data?

AI researchers can train models on data without seeing the underlying data through a secure compute environment and, eventually, distributed training. This "renting data" prevents data copying while keeping data under the owner's control.

Last updated