Arweave, AO, AI—Modular framework and flexible security

Author: Arweave Oasis, Source: Twitter @ArweaveOasis

Recently, Messari, a well-known research organization in the industry, published an article titled “Arweave, AO, and AI – Modular Framework and Flexible Security”, written by seth bloomberg.The content is very rich and has great reference value for understanding AO and AR architecture and future development prospects.However, due to its copyright issues, all the full text cannot be translated and displayed to readers.Therefore, the author hopes to cut the clips from them and show them to you through interpretation.The original content can be obtained at Messari (pay required).

On May 30, Seth Bloomberg, the author of this article, posted some summary of the article on X.The details are as follows:

• Arweave has historically relied on external applications and ecosystems to transfer data to its network.AO is a new network built on Arweave that will now create consistent demand for Arweave.AO will become Arweave’s growth catalyst and new platform for application development.

• I think one of the greatest technical values ​​of AO is that it separates the consensus mechanism from the computations required by the application.Splitting these to AO provides a modular architecture and allows developers to extend the security and computing power of their applications.

• Applications in AO incentivize computing power providers to process status updates and other messages.This creates a new market for application and computing power providers.It ensures that the application gets the level of resources it should have on demand.This is very different from most smart contract platforms.

• AO is a virtual machine-independent platform that opens up potential for deploying compute-intensive applications.Teams like @autonomous_af are already developing DeFi automation.However, on-chain LLM (big language model) would be a completely different species.

After reading the full text through, you will find interest in AO, a distributed world computer that is completely implemented in different forms.The main points summarized in Insight are:

• AO is a brand new protocol built on Arweave, a permanent data storage network.One of the core value propositions of AO is its ability to run applications in full parallel without scalability restrictions (in AO terms such applications are called process Process) and maintain relevant state locally without the need toShare the global status across the entire network.

• All applications on AO communicate through defined messages (becoming Message in AO terms) standards, providing processes with ready-to-eat read and write data solutions.

• A unique feature of AO is its flexibility to scale process security.Process developers can configure the security required by their applications and extend security by effectively paying additional validators (called computing unit CUs in AO terms).

• Due to the scalability characteristics of AO and the virtual machine-independent architecture (here, AO can support multiple virtual machines, allowing developers to select different virtual machines to run their applications according to their specific needs, without being limited to a single one.) and native cron job functions (referring to the system’s native support for timing task scheduling without relying on external tools or third-party services), many teams want to build automation tools and AI-powered products on the network.

Modularity and flexibility

The author summarizes two major infrastructures that have emerged over the past few years: modular frameworks and flexible security.

Modular framework: enables developers to select and combine common blockchain modular components (e.g., execution, data availability, settlement, and consensus).The Rollup and Celestia data availability layers of Base are used as representative examples.

Flexible and scalable security: refers to the ability of some networks to protect their services more efficiently by renting security rather than starting a very heavy validator network on their own.The author uses Eigenlayer as a case for corresponding explanations.

The goal of a modular framework is to encourage the optionality and specialization of each component.For example, developers are free to choose the execution environment that best suits their needs.Flexible security providers help networks better manage and fine-tune the economic security of their systems.

AO is a typical example of using both infrastructure models at the same time.This new system built on Arweave provides developers with the flexibility to choose execution environments and security models.

Unlike chains like Ethereum and Solana that have a single global concept of state (e.g., user account balances, smart contract data, etc.), AO localizes state to each application (called processes in AO terms).Localization state makes it easier for applications to compute in parallel, completely frees up overall performance limitations compared to non-parallel environments, and can customize security for computing.

Unlike other Rollup ecosystems, AO defines a unified global messaging standard (Message) for all applications.The authors believe that this approach is conceptually similar to the Cosmos ecosystem, which utilizes IBC for chain-to-chain communication.As a result, AO can maintain its modular framework and as its ecosystem grows, applications will benefit from this native communication standard.In the long run, AO has separated from the traditional smart contract platform model and formed its own unique architecture to promote a prosperous application development ecosystem.

AO architecture

The author believes that the relationship between AO and Arweave is roughly similar to the relationship between sovereign Rollup and the data availability layer.But AO provides a common framework similar to a smart contract platform, with the main goal being to achieve trust interoperability of these different applications through scalable computing services.

Interoperability between applications stems from AO’s messaging standards.Ecosystems such as Optimism, Polygon, Arbitrum, and zkSync usually develop economic activities on the chain before developing interoperability solutions to solve fragmented user experience problems.And AO will start its journey with native interoperability.

We have introduced the architecture of AO in various articles, and we also have some interpretations from the author’s perspective in this article:

Process

From the end user’s perspective, a process can be considered an application.If the consumer is using a product built on AO, it usually comes in the form of a process.

A process can also be considered as a series of ordered logs (i.e. messages) written in Arweave, representing its state at any given point in time.

Each process runs independently from other processes on AO, allowing them to operate in parallel without affecting each other.Processes interact with each other through messages.AO is actually a message delivery protocol, so the concept of message is the core structure.

Message and Message Unit (MU)

Interaction with a process, whether initiated by the end user or initiated by another process, is expressed as a message.Each message in AO complies with the specific format of ANS-104 for Arweave Ecological Read and Write Data.As for what ANS-104 is, you can check this link for details.

The author compares the direct difference between AO and Ethereum.In AO, a process completes interaction between processes by requesting information from another process through a message and waiting for data to return.But on Ethereum, applications (i.e., smart contracts) can directly access the state of any other application due to the globality of EVM.

There is a fundamental difference between the two.From a modular perspective, it is advantageous to pre-standardize the interoperability of different processes; most modular networks (such as Optimism’s Superchain) are developing similar standards.

Scheduling unit (SU)

The author simply compares the scheduling unit to the sorters in many Rollup systems.Since the sorter is responsible for a series of operations in many Rollups (e.g., transaction processing, transaction sorting, zero-knowledge proof generation, etc.), the scheduling unit is more like a subset of a typical sorter.

The scheduling unit has two main processes associated with it:

Make sure each message is unique and in order.This is conceptually similar to the nonce increment in other blockchain environments such as Ethereum.This is crucial for the normal operation of a process.

Make sure every message is written to Arweave.This enables processes to access each other’s data.

Each AO process will have an associated scheduling unit.

Computing Unit (CU)

The computing unit provides computing power for updating the AO process.The message unit notifies the computing unit of its service needs.

A market is formed between the computing unit (supplier) and the user (requirer) who needs calculations for a specific process.Once again, this architecture is different from the traditional blockchain model.Nodes on traditional platforms such as Ethereum need to process transactions, while computing units can selectively bid for their updated processes.

Once the calculation is completed, the calculation unit returns the signed calculation output proof to the original message unit.

AO Security Model

In the security model part, there is currently little relevant information, and more details need to be found after the AO white paper is released.However, the author of the article gave his own understanding.

He believes that AO has also taken a completely different path from Ethereum in this post.In the Ethereum ecosystem, security is guaranteed uniformly by the Ethereum PoS mechanism, so whether it is a simple transfer operation or a complex DeFi interaction, they share the same security level, which often leads to waste of resources.

In the AO security framework, although all data uses the security guarantee of Arweave’s SPoRes consensus mechanism, at the AO level, there is more flexibility to customize security levels according to different needs and goals.

At present, although there is no definite information, it is generally believed in the ecosystem that PoS mechanisms are likely to be used in the security mechanism of AO (for example, staking and penalizing AO tokens).Arweave, a fully decentralized permanent storage network, coupled with the AO computing platform that improves scalability and flexibility, the PoS mechanism is obviously in line with development needs.

Therefore, AO can propose a certain different staking scheme for each component role and set up a corresponding penalty mechanism.

• Computing unit– The calculation unit pledges the output proof of its signature.Anyone can question the output of the computing unit, and if it proves that there is a bad behavior, its collateral can be slashed.

• Message unit– The message unit pledges the messages it passes in the system.If it is found that it passes and signs an invalid message, its stake can be slashed.If the invalid signature is caused by inappropriate behavior of the computing unit, the message unit may initiate a Slash event for the computing unit.

• Scheduling unit– If the dispatching unit fails to sort the messages correctly or does not upload the messages to Arweave, it can be slashed.The latter Slash events are similar in design to data availability guarantees.

Ultimately, a process can design its own security model in a sense.During code execution of a process, for example, it may be decided to ignore a certain computing unit or message unit that is considered untrusted.

The picture shows the security consistency of the traditional smart contract platform and is safe and customizable with the AO platform.This allows AO to customize different levels of security for different businesses, such as small transfers between friends do not need to be consistent with the security of B2B bulk transactions.

The future of AO and AI

The author further expresses his point of view on the future combination of AO and AI.He believes that AI can be classified into two types:

Completely deterministic and fully parameterized, such as a robot with configurable settings;

Nondeterministic, adaptable, such as ChatGPT or LLM applications.

In the development of AO, the author believes that the starting point will be the former, such as DeFi automation tools.

DeFi automation

An early DeFi automation project on AO was @autonomous_af.The team built a “DCA agent” that enables users to buy specified tokens in the USD Cost Average Method (DCA).

DCA Agent Products Follow Pattern:

Users define the tokens they want DCA to buy, along with other parameters such as slippage tolerance, specific DeFi pools, frequency of DCA transactions, and amount of each DCA transaction.

The DCA agent responds to the received notification (i.e., a timed task) and executes the DCA transaction when a predefined condition is met.

The user can eventually pause the DCA agent or permanently deactivate it.

It should be clear that this kind of proxy performs operations in a rule-based way and effectively follows the instructions defined in its underlying scripts.This is the former in classification – completely deterministic and parameterized AI.

In this regard, Arweave founder @samecwilliams has actually expressed a similar view that in the current mainstream financial system, a large number of transactions are not operated by investors, but are automatically traded by various robots.So this is also naturally applicable to DeFi interaction scenarios.In fact, the safest way to achieve certain goals is to set strict rules and actions for the agent.This actually brings these products closer to the same level as traditional financial products and features (e.g., setting a stop loss, setting a DCA product from a bank account, etc.), which is a good thing from a user experience point of view.

Beyond DeFi Automation

In addition to the simpler AI mentioned above, the mainstream AI routines in the technology industry are currently focused on non-deterministic and adaptable AI, such as Chatgpt and LLM models.

Obviously, this type of system will be more advanced than a “DCA agent”.But it is also very expensive.Generally speaking, LLM products require GPUs to provide the necessary computing power, and GPU computing is more expensive than typical CPU computing.The cost of self-hosting an LLM base model can easily reach about $20,000 per month.The author gives a one-of-a16z data, where some generative AI startups spend 80% of their money on AI calculations alone.

Therefore, if you want to build applications using LLM on AO, economic cost considerations are inevitable.But compared to other smart contract platforms, AO’s architecture enables developers to scale and fine-tune the security level of their processes.This structure will be of great benefit to AO developers, as most LLMs generate less valuable messages.

The final thought

At the end, the author gives his thoughts on the AO organization:

AO’s unique architecture provides an attractive platform for application development, from DeFi to AI-powered applications.

Asynchronous messaging combined with parallel computing makes applications richer and more complex than typical smart contract applications.

The scalability and flexible security of supporting processes are also unique to AO, especially LLM-driven products, which will take advantage of this property.