Description
FLock.io provides a comprehensive, full-stack co-creation platform designed to streamline the entire AI journey, from data preparation to model deployment. The platform emphasizes data ownership, allowing for local training and hosting where data remains local, and fairly rewarding contributors for crowdsourced data. It enables the fine-tuning of foundation models like LLMs and Stable Diffusion, either individually or collaboratively with other FLock nodes, which helps to lower data requirements. The infrastructure is designed to be scalable, offering fast and compute-saving fine-tuning through technologies like LoRA. Furthermore, the platform is community-owned, with fine-tuning and RAG (Retrieval-Augmented Generation) coordinated on-chain, ensuring that contributors of data, feedback, and compute share in the rewards. FLock.io seeks to decentralise training and value alignment, ensuring that AI objectives match the public’s ethics and societal aims, and that decision-making falls to communities. It allows developers to provide models, data, or compute in a modular way, democratising AI agent training, fine-tuning and inferencing, while enabling equitable distribution of rewards and widespread governance through blockchain technology.