Last updated
Last updated
AI Computing
The AIGPT protocol is the core of the AIGPT ecological system, responsible for managing algorithms, data, and computing power using smart contracts.
To prevent miner interference in AI algorithms, after the official launch of AIGPT, AI nodes are required to stake a certain amount of CAI. If an AI node makes an error during algorithm execution, its staked CAI will be deducted based on the severity of the error.
After users submit their demands, the AIGPT protocol assigns computational tasks to an AI node based on the required computing power for the corresponding AI product's function. Alternatively, it may divide the task into several sub-tasks and assign them to several AI nodes. The AIGPT protocol tries to distribute the tasks evenly among different AI nodes based on the miners' computing power.
AI nodes that complete computations successfully will receive rewards. The first part of the reward is the CAI basic reward, which comes from the AIGPT reward pool. The second part is the CAI computing reward, which is paid by users based on their consumed computing power. The higher the computing power used by the miner to serve the user, the more CAI they will receive.
AI Product Development
Anyone can deploy AI products on AIGPT. Once an AI project is connected to AIGPT, it can share ecological data with other participants. In the AIGPT ecosystem, each AI product can call data from other AIGPT projects to train its AI models.
AI projects can issue their own project tokens, which can be used to pay for services or distributed as rewards to users. These tokens can also be used for governance or other functions within the project ecosystem.
AI Training and Usage
Users are both users and trainers of AI products. Data generated by users during the use of AI products is written onto the blockchain by AIGPT, which is then shared by all AI products on the platform. With the growing amount of user data, AI products on the AIGPT platform can be trained.
Users who use AI products on the AIGPT platform need to pay a corresponding amount of CAI as computation fees based on their consumed AI computing power. 10% of the computation fees are directly destroyed, 10% are reserved in the ecological reserve pool for ecosystem support and incentives, and the remaining 80% are distributed as rewards to AI nodes.
During the AI training period, high-quality AI products can receive support in the form of reduced computation fees. Users participating in AI training only need to pay a small amount of computation fees, and project owners can also distribute project tokens as rewards to incentivize users to participate in AI training.
