Technical Interpretation: How does AO build a decentralized computing network suitable for AI Agent?

The dream of a world computer that can execute arbitrary code without trust and be shared enough to be used by the whole world is deeply rooted in the decentralized network. After Ethereum, many infrastructure projects have made attempts, and Arweave’s upcoming AO network is one of these attempts.

For a "world computer", it can be roughly divided into three parts: data calculation, access and storage. Arweave has been playing the role of "world hard drive" in the past, and AO Network (Actor Oriented) Introduces general computing capabilities and provides smart contracts.

AO: Actor-based universal computing network

The current mainstream decentralized computing platforms are divided into two categories, namely smart contract platforms and general computing platforms; smart contract platforms are represented by Ethereum. The network shares the global state memory and agrees on the operation process of changing the state. Because consensus requires a large number of repeated operations, it is only used to process high-value services at high costs; general computing networks do not agree on the operation process itself, but based on the business There is no shared state memory to verify calculation results and process request sequences, which reduces costs and allows the network to expand to more fields of calculations. This type is represented by computing power networks such as Akash.

Of course, there are also some projects that integrate general computing with smart contracts based on the security assumption of virtual machine security. That is, the consensus only processes the order of transactions and verifies the calculation results. Multiple state change calculations are processed in parallel in network nodes. The computing environment virtual machine guarantees deterministic results. Therefore, as long as the order of transactions is consistent, the final state will also be consistent.

Because this type of network does not share state memory, the expansion cost is very low, and multiple tasks can be calculated in parallel without affecting each other. Such projects are often based on the Actor programming model, represented by ICP, and AO also falls into this category. Each computing unit under Actor is regarded as a separate intelligent processing transaction independently, and computing units interact through communication (Actor is a very common architecture in traditional Web2 services). AO standardizes Actor message passing and implements a decentralized computing network.

Different from traditional passively triggered smart contracts (such as Ethereum/Solana smart contracts), AO with a general computing Actor can realize the active operation of smart contracts through the "cron" method that is triggered in a consistent fixed time cycle , such as a trading program that continuously monitors the arbitrage space.

Rapidly scalable decentralized computing capabilities, Arweave’s ultra-large data storage capabilities, Actor’s programming model, and the ability to actively trigger transactions make the AO network very suitable for hosting AI Agents. AO also supports the introduction of large AI models into smart contracts on the blockchain.

AO network characteristics

As mentioned above, the difference between AO and smart contract network is introduced. AO does not agree on the calculation process, but only on the transaction sequence, and defaults to the deterministic result of the virtual machine's operation. , thereby achieving the consistency of the final state.

AO also has a certain degree of flexibility, and the network is designed in a modular manner. There are three basic units in the network, the scheduling unit SU, the computing unit CU and the messenger unit MU.

A transaction is sent, and the messenger unit as the communication layer accepts the transaction, verifies the signature, and forwards it to the scheduling unit; the scheduling unit can be regarded as the connection point between the AO and the AR chain, helping the network to sort the transaction sequence, and Upload to the AR chain to complete the consensus. The current consensus method is POA (Proof of Authority); after the consensus on the transaction sequence is completed, the task is assigned to the computing unit, and the CU is responsible for processing the specific calculations, and the results are returned to the MU and forwarded to the user.

The CU set can be seen as a decentralized computing power network. Under a complete economic plan, CU nodes need to pledge certain assets and compete by providing computing power through computing performance, price and other factors. Income, if there is a calculation error, the assets will be confiscated. This is a standard economic guarantee.

The difference between AO and other networks

As a general computing platform, the difference between AO and smart contract platforms such as Ethereum is obvious. Filecoin, which is the "world's hard drive" like AR, has also launched its own smart contract platform FVM, but this is a state consensus machine architecture equivalent to EVM, and its experience is inferior to traditional smart contract platforms such as Ethereum.

Unlike decentralized computing networks such as Akash and io.net, AO still retains smart contract capabilities. AO ultimately maintains a global state on AR storage.

實際上，與 AO 在架構上最相似的是 ICP。 ICP 創造了非同步運算區塊鏈網路的最早範式，AO 在很大程度上延續了 ICP 的設計，例如僅對交易順序排序、相信虛擬機器確定性計算、Actor 模型非同步處理等。

最大的不同點在於，ICP 是基於容器維護狀態，即每個智能合約容器可以只維護自己的private 狀態，或者對狀態讀取設定條件；而AO 具備一個共享的狀態層，即AR，任何人都可以透過交易順序與狀態證明復原全網狀態，這一定程度上增加了網路的去中心化能力，但也喪失了ICP 中特殊隱私業務的實現可能（例如客戶有隱藏套利路徑的需求）。

在經濟與設計層面，ICP 為了確保網路效能，對參與節點作出了較高的硬體要求，這造成了較高的門檻，而AO 相對以公平發射、無准入的方式運行，質押即可參與競爭挖礦。 ICP 網路選擇了大堆疊的實現方式，為了性能犧牲了靈活性，而AO 使用了模組化的設計，MU、CU、SU 分離，用戶也可以自選虛擬機的實現方式，這也降低了一些開發者進入的成本。

當然AO 也可能存在與ICP 一樣的系統缺點，例如Actor 非同步模型下跨合約交易缺乏原子性，這會導致DeFi 類應用發展困難，AgentFi 的構想似乎很難在短時間內實現；脫離傳統智慧合約範式的新運算模式，也對開發者提出了更高的要求。但 AO 架構下 wasm 虛擬機器最高能管理 4GB 的限制，也導致部分複雜模型無法在 AO 上使用。由此看來，AO 選擇 AI Agent 的路線實為揚長避短，有趣的是，ICP 也在 2024 年年初宣布專注於 AI 領域。

當然，比較 ICP 50億美金的總市值來說，AR 目前總市值 22 億美金，仍有不小的差距。在 AI 大發展的背景下， AO 可能仍然存在較大潛力。

The above is the detailed content of Technical Interpretation: How does AO build a decentralized computing network suitable for AI Agent?. For more information, please follow other related articles on the PHP Chinese website!