2023 witnessed an explosion of artificial intelligence (AI) models, largely driven by the widespread success and adoption of applications like OpenAI’s ChatGPT. Within five days of its launch, ChatGPT attracted over a million users, setting the record as the world’s fastest-adopted online tool. The impact was swift and profound, representing a watershed moment in the everyday use of AI tools in the years to come.
That said, the existing AI ecosystem is far from perfect, facing challenges such as the absence of robust Service Level Agreements (SLAs), centralized APIs, high compute costs, and an oligopolistic market structure.
Enter Ritual — a strategic response weaving the principles of cryptography, game theory, and mechanism design into its solution. Drawing inspiration from the blockchain and cryptocurrency space, Ritual emerges as an open-source, decentralized AI execution layer. Its goal? To become an AI coprocessor and enable developers to build fully transparent decentralized finance (DeFi) applications, self-improving blockchains, autonomous agents, generated content, and more.
With the integration of AI into the mainstream, some key questions arise: how can AI supercharge the growth of the crypto industry? And, how is Ritual leading this revolution? From the basics of AI to a closer look at Ritual and its product, Infernet, here’s everything you need to know about Ritual.
What is AI?
Amid all the talk about AI, what exactly does the term cover?
AI refers to the development of computer systems or software that can perform tasks that typically require human intelligence. These tasks may include problem-solving, learning from experience, understanding natural language, recognizing patterns, and making decisions. In essence, AI aims to create machines that can mimic or simulate certain aspects of human cognitive functions.
AI systems can be categorized based on their functionalities:
Machine Learning (ML): This is a subset of AI where systems learn from data and improve their performance over time without being explicitly programmed. Algorithms enable the system to identify patterns, make predictions, and adapt to changing circumstances.
Natural Language Processing (NLP): NLP focuses on enabling machines to understand, interpret, and generate human language in a way that’s both meaningful and contextually relevant. This technology is used in applications such as chatbots, language translation, and sentiment analysis.
Computer Vision: Computer vision involves teaching machines to interpret and make decisions based on visual data, such as images or videos. This is used in facial recognition, object detection, and autonomous vehicles.
What is Ritual and how does it work?
Credits: Ritual.net
Ritual is designed as an open and modular infrastructure for hosting and executing AI models. Its mission is to democratize access to AI, putting the power of AI into the hands of all.
Imagine generating transactions effortlessly, interacting with contracts in natural language, or having AI take the reins in managing risks for lending protocols, all in real-time as the market unfolds. With Ritual, the horizon of new use cases becomes boundless. But how does Ritual achieve this?
Ritual brings together a distributed network of nodes, each incentivized to provide computational resources and access to AI models globally. Meanwhile, the platform empowers AI model creators to host and monetize their models within Ritual’s model marketplace. This synergy allows users to access a variety of AI models — be it an LLM or an ML model — through a common API.
Moreover, the platform incorporates cryptographic proof, adding an extra layer of assurance to make sure of computational privacy and integrity. Developers seeking to incorporate AI into their decentralized apps can employ Ritual’s SDK for a quick and seamless experience.
Ritual Superchain: understanding Ritual’s architecture
Serving as the underlying infrastructure of Ritual, the Ritual Superchain consists of multiple layers and components:
Modular Stateful Precompiles (SPCs): SPCs are specialized smart contracts that can access state and efficiently handle complex AI functionalities, ranging from knowledge distillation to fine-tuning and inference. These SPCs are designed to be seamlessly integrated into other types of virtual machines (VMs) from the community.
General Message Passing (GMP) layer: Allows for interoperability between existing blockchains and the Ritual Superchain, enabling Ritual to serve as an AI coprocessor for all blockchains.
AI VM: Houses SPCs and base-layer infrastructure for optimized execution of AI models.
Node set: Consists of different classes of nodes with varying functionalities and resource requirements, including full nodes, validator nodes, proof nodes, model caching nodes, and privacy nodes.
Models and model storage: In Ritual, models are like special AI tools for various tasks. Think of these models as apps on your phone that you can use whenever you need them. They can be kept on the phone (node) or downloaded from a special place where everyone can get them (model registry). Ritual’s models are used by SPCs and stored in a permissionless, censorship-resistant storage layer.
Portals: Allows for the evaluation of models on source chains before interacting with the Ritual Superchain.
Guardians: Through Guardians, nodes serve as gatekeepers, making sure only relevant and acceptable models are integrated and used.
Shared sequencer: Enables all the layers in Ritual to operate in a seamless and organized manner.
Data Availability (DA) layer: In Ritual, users get control over how their information is stored and accessed.
Permanent storage: Required for proof, privacy, and full nodes to reconstruct computation deterministically.
Routers: Directs Web2 AI tasks to Ritual nodes based on user preferences for quality of service, privacy, computational integrity, and cost. Users can customize the settings or rely on auto-selected configurations.
Initial security through Eigenlayer: Ritual will tap into Ethereum’s Layer 1 security through Eigenlayer, which acts as a bridge. As the Ritual ecosystem evolves, it aspires to develop its own security measures.
Ritual’s Infernet: bringing decentralized AI to on-chain applications
Infernet takes the spotlight as Ritual's inaugural product, leveraging the Ritual Superchain to bring decentralized AI to the world of on-chain applications. This marks the beginning of Ritual's journey, aiming to create a space where smart contracts can seamlessly access AI models for advanced decision-making.The Ritual Infernet comprises two main components: the Infernet SDK and Infernet Nodes.
Infernet SDK
Ritual's Infernet SDK empowers smart contract developers to request and integrate off-chain computations into their applications. These computations span a wide range of tasks, from executing automated trading models to identifying abnormal patterns on blockchain networks and generating NFT traits based on user input.
The SDK offers two distinct interfaces: Subscriptions and Callbacks, each tailored to different use cases to provide flexibility in handling off-chain computations. The Subscription interface caters to smart contracts requiring recurring, time-based requests fulfilled by Infernet Nodes at fixed intervals. An example includes running an ML model every hour for a week.
On the other hand, the Callback interface is used for one-time requests, delivering off-chain output asynchronously. An instance might involve running an ML model with specific inputs and receiving the result once.
Using the SDK, users can set up notifications to keep track of the completion of off-chain computations and initiate further actions as needed.
Infernet Nodes
Infernet Nodes are lightweight off-chain clients within the Ritual ecosystem. They're responsible for fulfilling the AI-based compute workloads requested by users through the Infernet SDK. Infernet Nodes deliver the computed workflow outputs and, if applicable, optional proofs back to the blockchain. The workflow outputs and proofs can be transmitted either through on-chain transactions or an off-chain API, accommodating different use cases and user preferences.
How does decentralized AI address existing challenges?
In confronting the hurdles that currently exist in the AI landscape, decentralized AI emerges as a compelling potential solution.
Here’s how it addresses key challenges:
Service Level Agreements (SLAs)
Challenge: Existing AI platforms lack guarantees around computational integrity, privacy, and censorship resistance.
Solution: Leveraging decentralized networks, AI models can operate with enhanced transparency, verifiability, and privacy, providing strong SLAs for users.
Centralized APIs
Challenge: Centralized infrastructure restricts native integrations and can lead to liveness issues.
Solution: Decentralized AI networks allow for a distributed infrastructure, removing the dependency on a few centralized entities. This liberates developers and users from centralized constraints, fostering seamless access and integration.
High compute costs
Challenge: Escalating costs and restricted access to AI hardware hinder developer innovation.
Solution: By tapping into a network of nodes with computational resources, decentralized AI tackles high compute costs and hardware accessibility issues, providing a more equitable environment for developers.
Oligopolistic market structure
Challenge: Organizations face a dilemma between maintaining the secrecy of closed-source models for competitive reasons and participating in open-source collaboration for the greater benefit of the AI community.
Solution: Ritual, with the aim of decentralizing AI, disrupts the market structure by incentivizing open-source contributions and making sure of transparent governance. This empowers users to participate in the contribution and governance of AI models.
How can the crypto industry benefit from AI?
From base-layer infrastructure to applications, AI models can handle intricate logic and algorithms, becoming the key to unlocking possibilities that were previously unimaginable with just smart contracts. The decentralization of AI opens doors for a broader array of projects to harness the transformative capabilities of this technology.
Here are some ways the crypto industry stands to benefit from the integration of AI:
Amplifying smart contracts: AI enhances the capabilities of smart contracts. It introduces sophisticated decision-making and adaptive functionalities, elevating the scope and versatility of smart contract applications.
Transparent governance mechanisms: AI brings transparency to decision-making. AI models can provide real-time access to up-to-date information, operate on auditable algorithms, and offer data-driven insights, fostering more comprehensive and accountable governance decisions.
Automated risk management: In financial protocols, AI becomes a strategic asset. It can autonomously manage risk parameters based on real-time market conditions, enhancing the efficiency and resilience of DeFi ecosystems.
User-friendly interfaces: By incorporating AI, crypto platforms can offer natural language interaction. Users can engage with contracts, transactions, protocols, and more using everyday language, reducing entry barriers and fostering broader adoption.
Self-improving blockchains: AI can also advance blockchain systems by autonomously adjusting core parameters. This includes balancing supply and demand, resource pricing, and optimizing the overall performance of blockchain networks.
The Ritual team
The Ritual team, led by co-founders Niraj Pant and Akilesh Potti, offers a wealth of experience. Pant brings strategic insights as a former General Partner at Polychain Capital and a researcher at the Decentralized Systems Lab. Potti, a former Partner at Polychain Capital, also draws from his background in ML at Palantir, as well as experience in high-frequency trading and quantitative trading at Goldman Sachs.
The team is complemented by a distinguished panel of advisors, including Illia Polosukhin, co-founder of NEAR and co-author of "Attention is All You Need," with a previous stint at Google Research; Arthur Hayes, Chief Investment Officer at Maelstrom and co-founder of BitMEX and 100x Group; Sreeram Kannan, founder of EigenLayer and an Associate Professor at University of Washington; Divya Gupta, previously a Partner at Sequoia, with a background in ML from roles at Airbnb, Databricks, and Palantir, and more.
Ritual’s milestones: $25 million Series A financing
Ritual announced a significant milestone in November 2023 by securing a $25 million Series A financing round, spearheaded by Archetype. This funding round received backing from prominent investors and angel contributors, including Accomplice and Robot Ventures, affirming the strong investor confidence in Ritual and propelling the company into its next phase of advancement and innovation.
The $25 million funding inflow could provide a catalyst for expanding the team, nurturing the growth of the developer network, and starting the seeding process for the network.
The final word
Ritual sees itself as more than an AI execution layer. Instead, it's aiming to establish itself as the central hub for AI within the Web3 space. The vision involves the evolution of Ritual into an AI coprocessor that's compatible across a variety of blockchains, starting with the Ritual Superchain and Infernet.
Imagine a world where AI isn't just a buzzword but the driving force behind groundbreaking applications. Through its unique approach, Ritual aims to make this a reality, heralding a new era of AI accessibility.
© 2024 OKX. Este artículo se puede reproducir o distribuir tanto en su totalidad como parcialmente en fragmentos de 100 palabras o menos, siempre que no sea con fines comerciales. Cualquier reproducción o distribución del artículo en su totalidad debe indicar de forma prominente: “Este artículo es © 2024 OKX y se utiliza con permiso”. Los fragmentos permitidos deben citar el nombre del artículo e incluir la autoría. Por ejemplo: “Nombre del artículo, [nombre del autor si corresponde], © 2024 OKX”. No se permiten trabajos derivados u otros usos de este artículo.
