</aside>

<aside> <img src="notion://custom_emoji/1033d2c0-f40d-4c2f-9583-313aa8d2b337/1a080ee7-e5ee-8009-889f-007a7a44cbfb" alt="notion://custom_emoji/1033d2c0-f40d-4c2f-9583-313aa8d2b337/1a080ee7-e5ee-8009-889f-007a7a44cbfb" width="40px" /> MIZU is the first Edge AI Data Network, enabling self-hosted AI models and local data agents to run on personal devices—laptops, phones, and edge servers. By creating a network of interconnected data agents, MIZU breaks down data silos, allowing agents to communicate and share data beyond restrictive APIs. This unlocks the true value of personalized data, enabling seamless AI-driven automation across platforms.

</aside>

Introducing the MIZU Pool

The AI, of the people, by the people, for the people

In today's rapidly evolving technological landscape, artificial intelligence (AI) has become a cornerstone of innovation, driving advancements across various sectors. However, the journey to harnessing AI's full potential is often hindered by significant barriers, notably the substantial computational resources and financial investments required. This challenge is particularly pronounced in the realm of Edge AI, which focuses on processing data locally on devices rather than relying solely on centralized cloud servers.

Recognizing these challenges, MIZU is thrilled to announce the launch of the "MIZU Pool," a groundbreaking platform designed to democratize AI by enabling individuals to effortlessly build and manage edge AI networks. This initiative empowers users to host AI models on a decentralized network of edge devices, making advanced AI accessible to all.

The Imperative for Democratizing AI

The development and deployment of AI models are resource-intensive endeavors, often necessitating high-performance hardware and substantial energy consumption. This reality has traditionally confined AI capabilities to well-funded organizations, leaving smaller entities and individuals at a disadvantage.

Concurrently, Edge AI is gaining momentum, with the market projected to grow from $27.01 billion in 2024 to $269.82 billion by 2032, reflecting a compound annual growth rate (CAGR) of 33.3%. This surge underscores a significant shift towards localized data processing, which offers benefits such as reduced latency, enhanced privacy, and decreased reliance on centralized infrastructure.

There's a burgeoning trend towards self-hosting, as organizations and individuals seek greater control over their data and computational processes. This movement raises an intriguing question: can we unite our collective computational power to serve not just individual needs but also the broader community?

Introducing the MIZU Pool: A Collaborative Edge AI Ecosystem

At its core, the MIZU Pool is a collection of edge devices collaboratively serving a single AI model. This innovative approach allows for the distribution of computational tasks across multiple devices, optimizing resource utilization and enhancing accessibility.

Key Features of the MIZU Pool:

Universal Accessibility: Anyone can create a MIZU Pool, emphasizing inclusivity and broad participation.
Ownership and Control: Pools can be designated as "ownable," granting creators the authority to determine who can contribute to and utilize the pool.
Monetization Framework: Users compensate pool owners with MIZU credits for accessing the AI models. Pool owners can set a distribution ratio (e.g., 20%), where a portion of earnings is retained, and the remainder is allocated to contributors based on their input.

Group 2085663328.png

MIZU Credits: Facilitating Seamless Transactions

To streamline interactions within the ecosystem, MIZU employs a credit-based system. Each pool establishes a pricing model, setting rates per million tokens for both input and output. MIZU meticulously tracks token usage, ensuring transparent and fair compensation between users and pool owners.

Early contributors from Phase 1 can use points to unlock free credits, giving them the opportunity to explore, engage with the system, and earn rewards.

Asynchronous Inference Framework: Ensuring Reliability in a Decentralized Network

Edge devices, by nature, can be unpredictable—subject to intermittent connectivity and varying performance levels. To mitigate these challenges and uphold service quality, MIZU has developed an asynchronous inference framework.

Core Components of the Framework:

Asynchronous Request Handling: Users submit requests and retrieve results in an asynchronous manner, with synchronous APIs available for those who prefer real-time interactions.
Stateless Edge Devices: Devices operate without retaining state information, processing tasks from MIZU and returning results efficiently.
Comprehensive Management: MIZU oversees all critical aspects, including pool management, time-to-live (TTL) settings, request failover mechanisms, and data storage solutions, providing a holistic and user-friendly experience.

Embarking on Your MIZU Journey

Whether you're looking to create a pool or utilize existing models, MIZU offers a straightforward pathway:

For Pool Creators:

Register with MIZU: Begin by creating your account on the MIZU platform.
Establish Your Pool: Select the AI model you wish to deploy, utilizing platforms like Ollama to streamline the process.
Integrate Your Devices: Add your devices to the pool by running the MIZU web application, thereby contributing to the collective computational effort.

For Model Users:

Register with MIZU: Create an account and obtain your unique API key.
Select a Pool: Choose the pool that aligns with your requirements.
Access the API: Retrieve the API host endpoint and integrate it with any OpenAI-compatible application to begin utilizing the model.

Real-World Applications: Harnessing the Power of Collaborative AI

The versatility of the MIZU Pool enables numerous compelling use cases:

Personal Hosting: Host AI models on your desktop and access them seamlessly from your mobile devices, ensuring consistent performance across platforms.
Private Group Pools: Collaborate with peers by creating a pool comprising multiple devices, tailored to serve the specific needs of your group.
Public Service Pools: Assemble a robust pool capable of handling public traffic, enabling applications such as PDF-to-markdown parsers powered by specialized language models.

In essence, the MIZU Pool represents a paradigm shift in how we approach AI deployment and accessibility. By leveraging collective edge resources, MIZU is making advanced AI models not only more accessible but also more manageable for individuals and organizations alike. This initiative embodies the principle of AI "of the people, by the people, for the people," heralding a new era of collaborative technological advancement.