Koop Handshake (SN58) in Mexico

Dippy Speech, also known as Subnet 58 on the Bittensor network, is a decentralized AI project focused on creating advanced open-source speech models. Developed by the team behind the Dippy AI companion app, this project aims to build a system for immersive, lifelike, and empathetic voice interactions. While the team previously established Subnet 11 to focus on roleplay language models, Subnet 58 is specifically dedicated to the vocal aspect of AI companionship. The primary objective of Dippy Speech is to move beyond the limitations of traditional voice assistants and closed-source software. Most existing systems rely on static reference recordings which can limit creative flexibility. In contrast, Dippy Speech uses natural language prompting to manage speaker identity and emotional style. This approach allows the AI to produce more dynamic and personalized voices that are better suited for roleplay and genuine human connection. The project operates through the Bittensor ecosystem, where a global network of miners and validators collaborates to train and refine these speech models. This decentralized structure is designed to leverage collective intelligence and open-source innovation, ensuring that the resulting technology remains transparent and accessible. The development of the subnet is planned in multiple phases. The early stages involve launching a robust pipeline for roleplay-specific text-to-speech models and implementing evaluation metrics such as human-likeness scores and word error rates. Later phases focus on refining these models to produce highly expressive and human-like outputs. Ultimately, the team plans to integrate these advanced speech capabilities directly into the Dippy app, providing users with seamless, end-to-end audio experiences for their AI interactions. By prioritizing empathy and natural expression, Dippy Speech seeks to address the growing demand for digital companionship that feels authentic. The project emphasizes community-driven development and the use of synthetic data pipelines to scale its intelligence, positioning itself as a key contributor to the broader movement for open-source, decentralized artificial intelligence. Leer meer