At the beginning of February 2024, OpenAI released the first generative video model, Sora, turning a new page in the development of artificial intelligence. On February 13, Sam, the founder of the renowned blockchain storage project Arweave, announced on Twitter that Arweave had officially launched a super parallel computing machine, AO, which supports the integration of large AI models into blockchain smart contracts. This announcement quickly sparked intense discussion among AI and blockchain professionals.
https://twitter.com/samecwilliams/status/1757161860028150159
Arweave is a new type of blockchain storage network designed to solve the problem of permanent data storage and access. Arweave features a method called “permanent storage,” which ensures that information is never lost by storing data on the blockchain. Arweave’s crypto-economic model aims to incentivize users to store and transfer data, ensuring the security and reliability of the network. This allows users to safely store important data, such as documents, pictures, videos, etc., and access them at any time without worrying about data loss.
Notably, Arweave has demonstrated its unique advantages in the NFT domain. NFTs often require off-chain data storage, and if this data is lost or altered, the NFT loses its meaning. Arweave provides a solution that enables the permanent storage of NFT-related data, thus ensuring the long-term value and significance of NFTs. Platforms like Opensea and Mintbase have already formed strategic partnerships with Arweave, and the Metaplex NFT project on the Solana blockchain also uses Arweave as the default tool for metadata storage.
Arweave is also committed to the project of permanently preserving human knowledge and history, earning it the reputation of the “Eternal Alexandria Library on the Blockchain.”
AO is a distributed, decentralized, Actor Oriented computing system based on Arweave.
An Actor is a basic unit of a concurrent computing model in computer science. Actor Oriented refers to methods based on the Actor model, where each Actor can modify its assigned private state. However, to modify another Actor’s state, it must do so indirectly by sending messages. This is suitable for building highly concurrent, distributed, and fault-tolerant systems, which is why the project is named AO.
AO consists of three units:
AO leverages blockchain orchestration, with each unit functioning as a horizontally scalable subnet, executing a large number of transactions simultaneously, thus achieving high-performance computing with theoretically nearly unlimited computing performance.
The core goal of AO is to provide trustless and collaborative computing services without any actual scale limitation, offering a new paradigm for applications combined with blockchain technology. Compared to other high-performance blockchains like Solana, Aptos, and Sui, AO can support the storage of vast amounts of data, such as AI models. Moreover, unlike Ethereum, which can only use a single shared memory space, AO allows an arbitrary number of parallel processes to operate within compute units and collaborate with other units through open message passing, without relying on centralized memory space.
AOS is a decentralized operating system built by the Arweave official team based on AO. Its function is similar to that of smart contracts, but the contract development language is Lua. Lua is a concise, lightweight, and extensible scripting language, which is quite different from the common smart contract languages such as Solidity, Rust, and Move. AOS allows developers to initiate command-line processes and start issuing commands, providing a development experience similar to creating a new server instance on Alibaba Cloud and connecting to it via SSH, except that this command-line process has the properties of smart contracts. This means that these commands can achieve seamless user interaction across networks, and possess decentralized and trustless computing as key advantages.
The essence of AO is a decentralized cloud service based on the SCP (Storage-based Consensus Paradigm). There are three progressive relationships: “cloud services,” “decentralized cloud services,” and “SCP-based decentralized cloud services.” Cloud services are already very familiar to everyone, and their importance is self-evident. They have become an important infrastructure of the big data era, but they are almost all built and controlled by centralized giants, such as Amazon Web Services (AWS), Google Cloud Platform (GCP), Microsoft Azure internationally, and Alibaba Cloud, Tencent Cloud, Huawei Cloud domestically.
Decentralized cloud services refer to the sharing and exchange of cloud resources within a decentralized network, facilitated by blockchain economic incentives. Users can earn digital currency or tokens by providing their processing power, storage space, or bandwidth to the network. They can then use these digital currencies or tokens to access services.
Decentralized cloud services rely on market forces to determine the value and distribution of computing power, storage space, and bandwidth. This achieves a more efficient and direct allocation of resources. This economic structure encourages user participation in the network and fair competition, strengthening cloud infrastructure while breaking the monopoly held by giants. Since all underlying processing power, storage space, and bandwidth are provided by users, if the economic structure of a decentralized cloud service network is no longer viable, users can choose to leave the network or join other decentralized cloud service networks, limiting the potential for malicious behavior by service providers.
Notable projects in the decentralized cloud services space include Dfinity, Ankr, Akash, etc. It’s worth noting that Arweave can also be categorized within the decentralized cloud services space, though it provides solely cloud storage services.
One challenge of decentralized cloud services is the potential complexity and difficulty of implementing consensus mechanisms, as they require nodes in the network to agree on data storage and retrieval. SCP (Storage-based Consensus Paradigm) is one effective way to address this challenge. SCP is a consensus paradigm based on storage, its core idea being that as long as storage is immutable and transactions on it are traceable, the same results will be obtained regardless of where the computing application is performed.
AO addresses this by uploading all states before and after computation, along with inputs and outputs, to Arweave. Any third party can download all the data, run the execution environment (e.g., a virtual machine), and sequentially execute the inputs to obtain a final consistent result. This achieves permissionless verifiability, achieving a trustless consensus.
This means AO can realize distributed verifiable computation, which is AO’s biggest advantage compared to other decentralized cloud services.
The evolution of AI necessitates the combined drive of three key elements: algorithms, computing power, and data. For emerging AI startups, partnering with cloud service giants is often the only option to reduce costs and access more computing power, as not all companies have the capability to build their own computing resources. Furthermore, developing AI applications increases the demand for cloud services. Meanwhile, cloud service giants are developing their own AI systems, and it is now evident that centralized cloud service providers have almost all developed their own generative large models.
In this environment of both cooperation and competition, cloud service giants leverage their monopoly on computing power and their dominant position in cloud services to place emerging AI companies at a severe disadvantage. This competition, which prioritizes financial power over innovation, ultimately harms the interests of users.
Decentralized cloud services can balance this unfair competition and have attracted users to contribute computing power and data to the network, promising to become a new player in the AI era.
Arweave’s introduction of AO marks its entry from the niche of decentralized storage into the broader arena of decentralized cloud services. Its perpetual chain-based storage is no longer just about storing user data but serves as a permanent host for cloud computing, emphasizing infinitely scalable, verifiable computing. This significantly enhances its potential, aiming to become a leading player in the AI era.
At the beginning of February 2024, OpenAI released the first generative video model, Sora, turning a new page in the development of artificial intelligence. On February 13, Sam, the founder of the renowned blockchain storage project Arweave, announced on Twitter that Arweave had officially launched a super parallel computing machine, AO, which supports the integration of large AI models into blockchain smart contracts. This announcement quickly sparked intense discussion among AI and blockchain professionals.
https://twitter.com/samecwilliams/status/1757161860028150159
Arweave is a new type of blockchain storage network designed to solve the problem of permanent data storage and access. Arweave features a method called “permanent storage,” which ensures that information is never lost by storing data on the blockchain. Arweave’s crypto-economic model aims to incentivize users to store and transfer data, ensuring the security and reliability of the network. This allows users to safely store important data, such as documents, pictures, videos, etc., and access them at any time without worrying about data loss.
Notably, Arweave has demonstrated its unique advantages in the NFT domain. NFTs often require off-chain data storage, and if this data is lost or altered, the NFT loses its meaning. Arweave provides a solution that enables the permanent storage of NFT-related data, thus ensuring the long-term value and significance of NFTs. Platforms like Opensea and Mintbase have already formed strategic partnerships with Arweave, and the Metaplex NFT project on the Solana blockchain also uses Arweave as the default tool for metadata storage.
Arweave is also committed to the project of permanently preserving human knowledge and history, earning it the reputation of the “Eternal Alexandria Library on the Blockchain.”
AO is a distributed, decentralized, Actor Oriented computing system based on Arweave.
An Actor is a basic unit of a concurrent computing model in computer science. Actor Oriented refers to methods based on the Actor model, where each Actor can modify its assigned private state. However, to modify another Actor’s state, it must do so indirectly by sending messages. This is suitable for building highly concurrent, distributed, and fault-tolerant systems, which is why the project is named AO.
AO consists of three units:
AO leverages blockchain orchestration, with each unit functioning as a horizontally scalable subnet, executing a large number of transactions simultaneously, thus achieving high-performance computing with theoretically nearly unlimited computing performance.
The core goal of AO is to provide trustless and collaborative computing services without any actual scale limitation, offering a new paradigm for applications combined with blockchain technology. Compared to other high-performance blockchains like Solana, Aptos, and Sui, AO can support the storage of vast amounts of data, such as AI models. Moreover, unlike Ethereum, which can only use a single shared memory space, AO allows an arbitrary number of parallel processes to operate within compute units and collaborate with other units through open message passing, without relying on centralized memory space.
AOS is a decentralized operating system built by the Arweave official team based on AO. Its function is similar to that of smart contracts, but the contract development language is Lua. Lua is a concise, lightweight, and extensible scripting language, which is quite different from the common smart contract languages such as Solidity, Rust, and Move. AOS allows developers to initiate command-line processes and start issuing commands, providing a development experience similar to creating a new server instance on Alibaba Cloud and connecting to it via SSH, except that this command-line process has the properties of smart contracts. This means that these commands can achieve seamless user interaction across networks, and possess decentralized and trustless computing as key advantages.
The essence of AO is a decentralized cloud service based on the SCP (Storage-based Consensus Paradigm). There are three progressive relationships: “cloud services,” “decentralized cloud services,” and “SCP-based decentralized cloud services.” Cloud services are already very familiar to everyone, and their importance is self-evident. They have become an important infrastructure of the big data era, but they are almost all built and controlled by centralized giants, such as Amazon Web Services (AWS), Google Cloud Platform (GCP), Microsoft Azure internationally, and Alibaba Cloud, Tencent Cloud, Huawei Cloud domestically.
Decentralized cloud services refer to the sharing and exchange of cloud resources within a decentralized network, facilitated by blockchain economic incentives. Users can earn digital currency or tokens by providing their processing power, storage space, or bandwidth to the network. They can then use these digital currencies or tokens to access services.
Decentralized cloud services rely on market forces to determine the value and distribution of computing power, storage space, and bandwidth. This achieves a more efficient and direct allocation of resources. This economic structure encourages user participation in the network and fair competition, strengthening cloud infrastructure while breaking the monopoly held by giants. Since all underlying processing power, storage space, and bandwidth are provided by users, if the economic structure of a decentralized cloud service network is no longer viable, users can choose to leave the network or join other decentralized cloud service networks, limiting the potential for malicious behavior by service providers.
Notable projects in the decentralized cloud services space include Dfinity, Ankr, Akash, etc. It’s worth noting that Arweave can also be categorized within the decentralized cloud services space, though it provides solely cloud storage services.
One challenge of decentralized cloud services is the potential complexity and difficulty of implementing consensus mechanisms, as they require nodes in the network to agree on data storage and retrieval. SCP (Storage-based Consensus Paradigm) is one effective way to address this challenge. SCP is a consensus paradigm based on storage, its core idea being that as long as storage is immutable and transactions on it are traceable, the same results will be obtained regardless of where the computing application is performed.
AO addresses this by uploading all states before and after computation, along with inputs and outputs, to Arweave. Any third party can download all the data, run the execution environment (e.g., a virtual machine), and sequentially execute the inputs to obtain a final consistent result. This achieves permissionless verifiability, achieving a trustless consensus.
This means AO can realize distributed verifiable computation, which is AO’s biggest advantage compared to other decentralized cloud services.
The evolution of AI necessitates the combined drive of three key elements: algorithms, computing power, and data. For emerging AI startups, partnering with cloud service giants is often the only option to reduce costs and access more computing power, as not all companies have the capability to build their own computing resources. Furthermore, developing AI applications increases the demand for cloud services. Meanwhile, cloud service giants are developing their own AI systems, and it is now evident that centralized cloud service providers have almost all developed their own generative large models.
In this environment of both cooperation and competition, cloud service giants leverage their monopoly on computing power and their dominant position in cloud services to place emerging AI companies at a severe disadvantage. This competition, which prioritizes financial power over innovation, ultimately harms the interests of users.
Decentralized cloud services can balance this unfair competition and have attracted users to contribute computing power and data to the network, promising to become a new player in the AI era.
Arweave’s introduction of AO marks its entry from the niche of decentralized storage into the broader arena of decentralized cloud services. Its perpetual chain-based storage is no longer just about storing user data but serves as a permanent host for cloud computing, emphasizing infinitely scalable, verifiable computing. This significantly enhances its potential, aiming to become a leading player in the AI era.