The rise of large generative models has led to the centralized control and monetization of personal data by AI corporations, posing serious risks to privacy, data security, and regulatory stability. In response, Rainfall introduces the first Self-Sovereign, Decentralized, Generative AI Platform, featuring the Kindred Bit—a user-controlled digital twin that models identity, preferences, and behaviors using dynamic, self-supervised learning. Kindred Bits collaborate securely via a trusted compute and storage infrastructure to create Crowd Sourced Intelligence (CSI) and Real-time Social Intelligence™ (RSI), while ensuring fair compensation for contributions.
Kindred Bit is a play on Kindred Spirit: By fusing "kindred" with "bit," the term "Kindred Bit" becomes a playful yet profound nod to the concept of a "kindred spirit" - someone with whom one shares a deep and natural affinity, including similarities in beliefs, attitudes, and feelings. In a digital context, a Kindred Bit is thus envisioned as a digital twin or avatar that embodies the essence of one's identity, mirroring not just superficial characteristics but the core attributes and values that define a person.
This etymological exploration reveals that a Kindred Bit is more than just a digital persona or avatar; it is a manifestation of one's essence in the digital realm. It is a personal, self-sovereign entity that carries the depth, complexity, and uniqueness of the individual it represents, acting in “the digital world with the same integrity and authenticity as its human counterpart would in the physical world. Through this lens, the Kindred Bit is a digital kindred spirit, navigating the vast expanse of the digital universe, ensuring that our interactions, decisions, and digital footprints are true reflections of our deepest selves.
The vast amount of data each person generates can reveal patterns and preferences, and even predict future behaviors. By actively auto constructing our Kindred Bits, we reclaim control over how we're represented in the digital realm. Constant interactions with the Kindred Bit make it truer, smarter and more powerful.
Moreover, as digital environments become increasingly sophisticated with the advent of virtual reality, augmented reality and AI-driven platforms, having these self tuned digital personas opens up a myriad of usage scenarios that equip every individual to participate in the wonders of these connected, digital worlds.
Kindred Bits are not just task oriented agents, but dependable representations that adhere to our values and principles, all without the need for direct supervision.
Kindred Bits are created and updated in near real time by the Rainfall Nets on the edge from the data streams shared by the user. These include physical and social media activities, online shopping behaviors, health data, financial transactions, direct shares from devices and conversations with the Kindred Oracle. These data points, indexed by the context in which they occur, serve as the inputs to the neural autoencoders. These encoders are optimized to learn the user personality from the variegated and seemingly unconnected inputs using specialized embeddings.
This comprehensive integration allows the Kindred Bit to learn the user's preferences, habits, states and moods in multiple contexts and situations (situated cognition). Crucially, the Kindred Bit is designed to reside on the user's personal devices, such as desktops, laptops, and smartphones, or in dematerialized space provided by a MPC Blind Compute Fabric, ensuring that users maintain complete control over their digital personas without the risk of external manipulation or loss of agency. Furthermore, the architecture of the Kindred Bit is built with privacy and security at its core, providing a secure, user-controlled approach to personal data management and digital representation. Through this innovative framework, the Kindred Bit empowers users to navigate the digital world with confidence, knowing their digital twin acts in their best interest with precision, privacy, and protection.
A Kindred Bit is designed to emulate the user in specific situations. The objective of the Kindred Bit is to react to a stimulus in a specific context with a response that mirrors the user it represents.
Examples of stimuli and responses could be as follows:
| Context | Stimulus | Response | |
|---|---|---|---|
| 1 | It’s 2 pm on a rainy October afternoon outside the office | Next meeting at 2:45 pm | all pumpkin spice latte, banana bread |
| 2 | It’s 6 pm on a rainy October afternoon outside the office | Meet Paula at Fog City Diner in 10 min | Get an Uber |
| 3 | It’s 6 pm on a balmy September afternoon outside the office | No more meetings | Walk home |
| 4 | Dinner at Laura’s day after at 7 pm | Gift | Dog Point, Sauvignon Blanc from Wine Trading Co. |
| 5 | 7 pm on Steven’s Creek Boulevard (unseen context) | Dinner | Vegan café, burgers and wine by the glass (abstracted preferences) |
Contexts can be historical, current or even prescient. The accuracy of responses decreases as the context shifts away from the present.