The AI industry today faces a significant centralization problem. Large corporations dominate the field, controlling vast data and computational resources. This centralization limits innovation and creates barriers for smaller entities and individuals who wish to contribute to AI development. As a result, many talented individuals and small teams find it challenging to showcase their work and make meaningful contributions to the industry.

Sentient AGI addresses this centralization problem by creating a decentralized, transparent, and open AI landscape. It provides a platform where anyone can participate in AI development, regardless of their resources or background. By leveraging blockchain technology and the innovative Open, Monetizable, and Loyal (OML) model, Sentient AGI democratizes access to AI resources and opportunities.

What is Sentient AGI?

Source: Sentient AGI twitter

Sentient AGI is an advanced artificial intelligence system designed to operate decentralized. It leverages blockchain technology to create an open, monetizable, and loyal (OML) ecosystem. This system allows for the development, deployment, and management of AI models in a way accessible to a broader range of participants, including individual developers and small teams.

Key Features of Sentient AGI

• Decentralization: Unlike traditional AI systems controlled by large corporations, Sentient AGI operates on a decentralized network. This ensures that no single entity has control over the entire system.
• Open Access: Sentient AGI provides open access to its resources, allowing anyone to contribute to and benefit from the AI ecosystem.
• Monetization: Contributors can monetize their AI models and data, earning rewards for their contributions.
• Loyalty: The system incentivizes long-term participation and loyalty through various reward mechanisms.
• Transparency: All transactions and contributions are recorded on the blockchain, ensuring transparency and trust.

Difference Between Current AI and Sentient AGI

Current AI systems are typically centralized, with large corporations controlling the data, resources, and development processes. This centralization limits access and opportunities for smaller players. On the other hand, Sentient AGI decentralizes these processes, making AI development more inclusive and democratic. Utilizing blockchain technology ensures that the power and control over AI development are distributed among a wide range of participants. This decentralization not only democratizes access to AI resources but also fosters a more collaborative environment where innovation can thrive.

Sentient Foundation: The Team Behind Sentient AGI

The Sentient Foundation is a non-profit organization that supports the development and growth of open-source AI technologies. It aims to create a decentralized, transparent, and open AI landscape. The foundation is committed to establishing a new Open AGI economy where AI builders are the principal actors and significant stakeholders. By providing the necessary infrastructure and resources, the Sentient Foundation ensures that the AI revolution benefits all of humanity. Promoting the development of Open Monetizable Loyal (OML) Models, the foundation counters the rent-seeking behavior of centralized AI companies. It fosters a collaborative ecosystem where diverse contributions are valued and rewarded.

Sentient Foundation Steering Committee

• Pramod Viswanath, Co-Founder: Professor at Princeton University, specializing in the first principle understanding and design of blockchains and inventing communication algorithms via deep learning. He is involved in teaching courses on decentralized finance and blockchain principles at Princeton.
• Himanshu Tyagi, Co-Founder: Co-Founder of Witness Chain, which provides AVS services and acts as a watchdog for operational rollups, focusing on proof of location and bandwidth for decentralized infrastructure (DePin). He has academic backgrounds from the Indian Institute of Technology, Delhi, and the University of Maryland (UMD).
• Sandeep Nailwal, Core Contributor: A key figure in the Sentient Foundation, contributing to its strategic direction and development.
• Sensys: A venture studio that creates advanced products and applications for Sentient AGI, supporting its technological and commercial advancements.

How Sentient AGI Works (OML Model)

Sentient AGI is building a platform that allows AI builders to collaborate and monetize their innovations. The core of this platform is the Open, Monetizable, and Loyal (OML) model. This model ensures that AI models are open for collaboration, can be monetized by their creators, and remain loyal to the community that built them.

Open, Monetizable, and Loyal (OML) Model

1. Open: Sentient AGI promotes openness by making AI models accessible to everyone. This openness encourages collaboration and innovation, allowing developers worldwide to contribute to and improve AI models. By being open, Sentient AGI ensures that the best ideas and solutions can emerge from a diverse pool of contributors.

2. Monetizable: One of the key features of the OML model is the ability for AI builders to monetize their contributions. This is achieved through a blockchain-based incentive mechanism that rewards developers for their work. By providing a clear path to monetization, Sentient AGI motivates builders to create high-quality AI models and continuously improve them.

3. Loyal: Loyalty in the OML model means that AI models remain aligned with the community that built them. This alignment is enforced by the underlying blockchain protocol, which ensures that the benefits of the AI models are shared with the contributors. Loyalty mechanisms help maintain the integrity and trust within the community, fostering long-term collaboration and commitment.

Concept of OML

• Economic Alignment: The underlying blockchain protocol and incentive mechanisms provide the necessary economic alignment for the evolution of Open AGI. This alignment ensures that all participants, from individual developers to large organizations, have a stake in the success of the AI models. Sentient AGI creates a sustainable and scalable ecosystem for AI development by aligning economic incentives.
• Mass Collaboration: Sentient AGI’s platform is built to enable mass collaboration. The tools and systems developed by Sentient AGI facilitate communication and cooperation among AI builders. This collaborative environment accelerates innovation and allows for the rapid development and deployment of AI models.

Participants in Sentient Protocol

In the Sentient Protocol, participants are categorized into two main groups: users and builders.

Users

Users interact with AI artifacts by paying fees. They can use these artifacts in two ways:

• Purchasing and Downloading: Users can purchase and download AI artifacts, gaining access to the model’s weights and metadata. This allows them to customize the artifact for their specific needs.
• Querying: Users can send queries to AI artifacts without downloading them. In this case, they pay a fee for each query and receive responses, interacting with the artifact as a black-box service.

Builders

Builders are the creators and contributors to AI artifacts. They play a crucial role in the development and enhancement of these artifacts. Builders can:

• Submit New Models: Builders can upload new AI models to the Sentient Protocol, inviting open contributions from the community.
• Contribute to Existing Models: Builders can download existing AI artifacts, modify them, and submit new versions. This collaborative approach allows for continuous improvement and innovation.
• Compose New Artifacts: Builders can combine existing AI artifacts to create new ones, expanding the functionality and capabilities of the Sentient Protocol.

How Participants Interact with the Protocol

Participants interact with the Sentient Protocol through a decentralized platform facilitating collaboration and monetization. The interaction process includes:

• Uploading and Downloading: Builders upload their AI artifacts to the protocol, making them available for other participants. Users can download these artifacts for customization or interact with them through queries.
• Contributions and Modifications: Builders can contribute to existing artifacts by downloading, modifying, and re-uploading them. This process ensures that AI models are continuously improved and updated.
• Ownership and Rights: AI artifacts can have multiple owners with varying ownership percentages. This shared ownership model encourages collaboration and ensures that contributors are fairly rewarded for their efforts.

Incentives for Participation

The Sentient Protocol provides several incentives to encourage participation from both users and builders:

• Monetary Rewards: Builders earn fees from users who purchase or query their AI artifacts. This monetization model provides a direct financial incentive for creating and improving AI models.
• Ownership Shares: Contributors to AI artifacts receive ownership shares, which entitle them to a portion of the revenue generated by the artifact. This shared ownership model ensures that all contributors benefit from the success of the AI models.
• Recognition and Reputation: Active participants in the Sentient Protocol can build their reputation within the community. High-quality contributions are recognized and rewarded, enhancing the visibility and credibility of the contributors.
• Access to Resources: Participants gain access to a wide range of AI models and tools, enabling them to leverage the collective knowledge and innovations of the community.

Sentient Protocol Architecture

Storage Layer

The Storage Layer is crucial for securely storing AI model data, ensuring reliable public access while maintaining exclusive writing access for specific upper-layer modules. It includes:

• Versioning: This module tracks new versions of AI models using a versioning tree. Each node in the tree is represented by a unique ID, typically a cryptographic hash (e.g., SHA256). This ID acts as a public commitment to the model, allowing anyone to verify the version they possess and track the evolution of subsequent versions.
• Ownership: This module records the ownership percentages of all contributors for each model version. It ensures that the rights of primary contributors are preserved and respected in all descendant versions of their AI models.

Distribution Layer

The Distribution Layer manages disseminating AI models in a protected format, ensuring traceability and proper attribution to the original builders. It links new models to their parent versions and distributes them in an OML format. Key modules include:

• Converter Nodes: These nodes transform models into a traceable format. Various solutions, such as model fingerprinting, are implemented to ensure models can be tracked.
• Evaluator Module: This module evaluates the value of the model. Builders determine acceptable evaluation methods for assessing future versions, allowing flexibility based on the model’s evolving nature and context.
• Ownership Module: This module allocates ownership percentages for new model versions. The allocation is based on the ownership of the parent version and the relative value of the new version compared to its predecessor.
• Challenge Module: This module ensures integrity by allowing owners of a parent model to challenge the assignment of ownership rights if a builder attempts to select an incorrect parent version or bypass lineage tracking.
• Distributor Module: This module handles the distribution of the formatted model to users, ensuring the process adheres to the path and guarantees chosen by the builders.

Access Layer

The Access Layer is responsible for tracking model usage while ensuring censorship-free user access. It authorizes users to access and use downloaded models by providing a signature. The Access Layer guarantees:

• Tracking Model Usage: Monitors how models are used to ensure proper utilization and compliance.
• Preventing Unauthorized Access: Ensures that only authorized users can access the models, protecting the intellectual property of the builders.
• Ensuring Monetization and Loyalty: Verifies that users have paid the required fees and are loyal before granting access, ensuring that builders are fairly compensated.

This layer consists of a network of nodes, known as “access nodes,” which handle user requests, verify their validity, grant access, and report usage to the Incentive Layer. The owners of a project can decide on the number of access nodes and who runs them. Each access node maintains a secret key (sk), allowing it to grant access to users. Owners can configure the access nodes in two ways:

• Single Node: One node manages access, centralizing trust in that node.
• Multiple Nodes: A network of nodes grants access through aggregated or threshold signatures, assuming a majority of nodes are honest. This setup ensures that access can be guaranteed even if some nodes act maliciously, provided a certain portion of nodes remain honest. It also ensures that users who do not pay fees and whose usage is not tracked will not get access.

Incentive Layer

The Incentive Layer ensures transparent usage tracking, fair governance, and trust-free ownership, ensuring proper incentivization and control for builders. It includes:

• Usage Tracking and Payment: This module receives query data from the Access Layer, including the user’s query and the access nodes’ signatures. It verifies whether the required fee has been paid for the query and finalizes the payment process. Once the payment is validated, it distributes the fee among the model owners, as well as other participating nodes, such as converter nodes and access nodes. The ratio for fee distribution is determined by the model owners. The module extracts the model version from the query and identifies the owners that need to be rewarded from the public storage ownership module.
• Ownership: The ownership of each model version is tracked and reported to the public storage, ensuring transparency for all participants. The versioning tree stored in this system illustrates the lineage of versions, helping determine the owners of new versions. Ownership can be gained by contributing to create a new version or through ownership transfer. Owners can transfer partial or full ownership of a specific version to others. Ownership allows owners to earn fees from the usage of that version and participate in its governance.
• Governance: A voting mechanism enables model owners to make collective decisions about the future of their model. Voting power is assigned in proportion to ownership percentage. Different versions of a model may have varying ownership distributions, so each version can have distinct decisions reflecting the preferences of its respective owners.

Sentient Roadmap

Sentient AGI has a clear roadmap outlining its key milestones and future plans. This roadmap is divided into short-term, midterm, and long-term goals, each designed to advance the development and adoption of Sentient AGI.

Short-term Goals

• OML White Paper: Publishing a comprehensive white paper detailing the Open, Monetizable, and Loyal (OML) model, its benefits, and its implementation.
• Early Platform and Curated Builder Demos: Launching an early version of the platform with curated demos from selected AI builders to showcase the potential of Sentient AGI.
• Open AGI Hackathons: Organizing hackathons to engage the community, encourage innovation, and identify new talent and ideas for the Sentient AGI ecosystem.

Midterm Goals

• Community Contributions to AGI Development on Platform Beta: Enabling community contributions to the development of Artificial General Intelligence (AGI) on the beta version of the platform.
• Sentient Protocol Launch: Officially launching the Sentient Protocol, providing a decentralized infrastructure for AI development and collaboration.
• The Sentient Reasoning Layer: Developing and integrating the Sentient Reasoning Layer, which enhances the cognitive capabilities of AI models.
• Early Foundational Model Experiments: Conducting experiments with early foundational models to test and refine the Sentient AGI framework.

Long-term Goals

New Foundational Models Built by Community Contributions: Encouraging the community to build new foundational AI models, leveraging the collaborative and decentralized nature of the Sentient Protocol.

• Incentives and Monetization for a New AI Builder Economy: Establishing robust incentives and monetization mechanisms to support a thriving AI builder economy.
• Community-Built Open AGI: Realizing the vision of a community-built Open AGI, where diverse contributions from around the world drive the development of advanced AI systems.

Sentient Funding Journey

Sentient AGI successfully closed an $85 million seed round, co-led by Founders Fund, the renowned venture capital firm led by Peter Thiel, alongside Pantera Capital and Framework Ventures. This round included contributions from a diverse group of investors, such as Founders Fund, Pantera Capital, Framework Ventures, Ethereal, Robot Ventures, etc

The funds raised in this round will support the hiring of engineers and the development of Sentient’s open AI platform. This substantial investment reflects the investor community’s confidence in Sentient AGI’s potential to revolutionize the AI industry by fostering a decentralized, transparent, and open AI landscape.

Conclusion

Sentient AGI represents a transformative approach to artificial intelligence, addressing the centralization problem and creating opportunities for individuals and small teams to contribute meaningfully to the AI industry. By leveraging the Open, Monetizable, and Loyal (OML) model, Sentient AGI fosters a decentralized, transparent, and collaborative ecosystem. The Sentient Foundation, with its dedicated team and strategic partnerships, is driving the development of open-source AI technologies. As Sentient AGI continues to evolve, it aims to democratize AI development, making it more inclusive and accessible.