Nesa's Utility Suite

Nesa accessorizes its system with a hub for external AI tools that help facilitate model querying, training, upload, and evolution (Figure 11.2). As the point of final validation for any model inference query, Nesa sits at the very bottom of the stack and is therefore able to plug an assortment of third-party AI services on top of it.

These services are serially executed and consigned to a dedicated adapter within the system. The Utility Suite includes provisioned computational power (TPUs, GPUs), almost all major publicly available open-source models, the major API-based LLMs, an assortment of Information Oracles, DAO tooling, and Relayers.

Nesa will continue to grow its collection of AI-Web3 partner services and products that can be invoked during training, upload, orchestration, and inference before Nesa executes final query validation.

Nesa Utility Suite Architecture. Nesa connects a hub of utility services that help facilitate model querying, training, upload, and evolution for AI model developers in the Nesa ecosystem. A utility service is initialized, run, and then securely hands off data to Nesa’s Execution Layer to assist in a component of the preparation, orchestra- tion, or computation of the request. Inference is then conducted by Nesa Core and the results are sent to validator nodes for consensus before rollup to settlement.

Some examples of Nesa's future Utility Suite include:

1. Efficient sharing of computational resources on-chain, such as Render, Akash.

2. Decentralized AI tooling and infrastructure services, such as Olas, Bittensor.

3. Information oracles for off-chain knowledge relayance, such as Chainlink, Band.

4. DAO tooling management systems, such as Gnosis, Aragon