Intent: The Starting Point of Web3 Interactive Intelligence

Original Author: Xuanrui

Content Source: AC CapitalResearch

This article will combine Anoma and TG Bot to open up the understanding of intent, and trace the AI paradigm shift behind intent, deeply understand the evolution, interaction paradigm, current situation, trends of intent concept, and finally look forward to the challenges and future of Web3 interactive intelligence.

Abstract (TL;DR)

(1) Understanding Intent: Anoma from top to bottom and TG Bot from bottom to top

(2) In-depth Intent: AI is the New UI: Complex Commands vs. Simple Intent

(3) Navigating Intent: Evolution, interaction paradigm, current situation, trends of Intent

(4) Guiding Intent: Challenges and prospects of Web3 interactive intelligence

In June, Paradigm's official website published an article "Intent-Based Architectures and Their Risks," which brought the concept of Intent to the public for the first time. The development of protocols and infrastructure related to intent has been rapid, and it has become an unavoidable topic in the crypto world at the ETHCC conference in July.

Intent is not a new concept, as DForce founder Mindao said: "The trend in the currency circle has always been to abstract and automate operations to the extreme. Aggregators, CEX, contract wallets, and cross-chain DeFi are all doing this work, as well as Chainlink and recent automated middleware, and the recent Telegram bot."

However, intent brings about new changes. In the past, the core was product interaction, but in the AGI era, the core is human-machine interaction. AI/LLM (large language models) holds more potential to enhance Crypo interaction.

Anoma from Top to Bottom and Unibot from Bottom to Top

In the past half year, protocols, projects, and infrastructure related to intent have been booming. Here are two excellent examples.

Although there are many projects related to intent, the most dazzling is the Anoma Foundation, which completed its third round of financing of $25 million at the end of May 2023. In the context of severe homogenization of Layer1, the Anoma Foundation, with its intent-centric architecture Anoma (Dapp full-stack architecture) and Namada (privacy Layer1), has received a total of $57.8 million in investment from institutions, ranking 7th in the amount of financing for unreleased Layer1/Layer2.

Looking back at Anoma's public speech at EthCC, founder Adrian Brink said the magic of Anoma lies in: all interactions start with intent, and user-expressed intent is processed through the Balck-box Architecture (Anoma black box architecture, also known as Magic Happens), which implements the core transaction process based on intent interaction:

User interaction: Users can send transparent, private, or protected intent to Anoma's Anoma black box architecture (intent gossip Layer).
Intent collection and matching: The solver of Anoma's black box architecture collects intent, performs balanced state transitions to match transaction counterparts.
Transaction processing: Matched transactions are submitted to the encrypted mempool, validators send proposer-packed blocks to the execution layer, complete execution and verify validity, and finally update the state root.

Intent: The Starting Point of Web3 Interactive Intelligence

Source: https://twitter.com/DelphiDigital/status/1696626180752056764_

Behind the new interaction form, the Anoma team has insight into the pain points of the underlying blockchain protocols. They reviewed that from the first generation of Bitcoin's scriptable settlement, the architectural limitations of the second generation Ethereum's programmable settlement have led to current application protocols having at least one Web2 component, which cannot achieve transaction counterpart discovery and resolution. They stated that the intent-centric Anoma architecture is the third generation Dapp architecture of blockchain protocols, allowing users to define the desired result state, efficiently and customizably achieve private transactions at the intent level. The intent-centric Dapp architecture will usher in a new era.

Intent: The Starting Point of Web3 Interactive Intelligence

Source: https://www.youtube.com/watch?v=mdYwfW6tMJ8

The highly financed, new technological architecture of Anoma, currently being developed by a team of 37 interdisciplinary members at Heliax, has been slow to progress over the past two years, but has already formed a complete innovative ecosystem:

Anoma: A general Dapp architecture centered on intent, which can be modularly deployed to L1, L1.5, L2, and embedded in various EVM frameworks using modular components.
Namada: The privacy settlement Layer1 of the Anoma ecosystem, using PoS mechanism and used for asset-independent cross-chain privacy, will be compatible with Ethereum and IBC chains in the future.
AnomaVM: Simplifies DApp application development on Anoma, using Juvix (intent-centric programming language) and VampIR (circuit programming language).
Typhon: Tendermint improvement of Typhon consensus based on Cosmos.
Taiga: Provides a unified execution environment for composable privacy. Taiga defines three types of intent: transparent intent, masked intent, and private intent.

Intent: The Starting Point of Web3 Interactive Intelligence

Source: https://heliax.dev/#projects

Although many institutions such as CMCC Global, Electric Capital, and Delphi Digital are enthusiastic about Anoma and express that the intent-centric architecture has unlimited potential in the future, the top-down Anoma currently only lays a good technical foundation, but has not yet seen exciting practical applications. On the other hand, TG Bot opens up the imagination of landing scenarios for intent segmentation with on-chain automated trading tools.

Unibot is a Telegram-based trading bot (TG Bot) that allows users to automate DEX trading through Telegram and provides sniper purchases, copy trading, DEX limit orders, privacy, anti-MEV, and other functions. Unibot replaces the cumbersome interaction of Uniswap with Intent and provides a convenient DeFi interaction experience. Unibot has also spawned a batch of TG Bot imitations, and the total market value of the current TG Bot track has reached nearly $200 million.

The top-down Anoma is innovation, and the bottom-up TG Bot is change; the top-down Anoma starts from the intent-based bottom-layer architecture, but the team's progress is slow; although TG Bot brings users a brand-new on-chain interaction interface, it is not intelligent and has many security risks.

Both of these divergent paths have the same goal—simplifying user interaction and introducing new programmable, customizable user interaction interfaces—User Intent Layer, allowing users to skip complex on-chain interactions and define transaction states based on user intent.

Intent: The Starting Point of Web3 Interactive Intelligence

AI is the New UI: Command Interaction vs. Intent Interaction

The above two use cases have provided a basic understanding of intent, and now it's time to explore the fundamental changes AI brings to the crypto world—changes in the user interaction interface. The AI paradigm is introducing the third user interface paradigm in computer history, shifting to a new interaction mechanism where users tell the computer what they want, rather than how to do it.

Paradigm one: Batch processing.

Paradigm two: Command-based interaction design.

Paradigm three: Result specification based on intent.

Every revolution in interaction modes in history has produced entirely new business models. Generative AI based on LLM has brought revolutionary changes to human-machine interaction, directly transforming the original human-machine interaction with various software UIs into interaction with Chatgpt chat windows, which is an unprecedented interaction experience. LLM+crypto brings a new proposition of intent-based interaction to the crypto world, and LLM has the potential to make blockchain interactions more intelligent by discovering and describing user intent.

Evolution of the Concept of Intent, Interaction Paradigm, Current Situation, Trends

After a deep understanding of the trend of AI intelligence behind intent, we further explore the evolution of the concept of intent. Intent is not a new concept, but has been present since the Web2 era. With the rise of search engines such as Google, users can input intent, and the search engine provides relevant search results to satisfy that intent.

With the rise of e-commerce platforms such as Amazon and eBay, there has been a significant shift in intent-based architecture. Users can now express the intent to purchase products, and the platform is responsible for logistics. In early 2017, Gartner released a report "Innovation Insight: Intent-Based Networking Systems," formally proposing the concept of Intent-based Networking. The key to intent-based networking is to present users with an interface where they only need to express what they want, and the platform will handle the business details.

Intent-based networking is the development trend of network automation and intelligence in the Web2 era. Gartner's research vice president Andrew Lerner pointed out in 2017 that intent-based networking will be the next milestone in the network field.

Intent: The Starting Point of Web3 Interactive Intelligence

Source: https://www.gartner.com/en/documents/3599617

The evolution from Web2 to Web3 has brought about the concept of intent in Web3, although the concept is not yet clear, there are some common understandings:

Definition given by Paradigm: Intent is a set of declarative constraints that allow users to delegate transaction creation to a specialized third-party participant network while retaining complete control over the process.
Definition given by the underlying Dapp architecture Anoma based on intent: Intent is a message sent by the user to express custom preferences. Intent is a constraint that defines what the user wants the system to do, rather than the specific execution path.
Definition given by the intent-based transaction infrastructure Propellerheads: Intent is the part of a transaction that requires assistance from others. Intent simplifies complex transactions by outsourcing difficult parts to third parties.

In real-world use cases of intent-based transactions in Web3, users create intent off-chain and outsource it to a solver off-chain, encapsulating the complexity of blockchain interaction to the solver, while retaining complete control over the process on-chain, thus lowering the barrier to on-chain interaction.

Intent: The Starting Point of Web3 Interactive Intelligence

Source: https://www.brink.trade/blog/powerful-intents-part-1

Intent also brings the concept of LLM architecture—intent-centric—where related crypto protocols and infrastructure are embedded in the interaction layer between users and Crypto, providing a better on-chain interaction experience with the help of LLM.

LLM+Crypto introduces a new human-machine interaction paradigm, where user intent can be directly translated into smart contract calls. Users only need to express their intent, and a robot/AI agent/third-party solver helps complete complex on-chain operations. The future of Web3 interactive intelligence based on intent has the potential to significantly reduce the complexity of user on-chain interactions.

AI is the New UI, and more and more Web3 projects are exploring the power of intent. Whether it's intent-based Dapps or intent-based underlying architectures, they are unleashing the potential of intent from different perspectives. We attempt to categorize this rapidly developing field from an interaction perspective into four major categories and track the latest developments and trends:

Intent: The Starting Point of Web3 Interactive Intelligence

Infrastructure: (Intent-centric intent architecture layer)

DappOS: The first intent-centric Web3 operating protocol that builds an intermediate layer between users and encrypted infrastructure such as public chains and cross-chain bridges, allowing Web3 users to interact with products seamlessly and imperceptibly. DappOS recently released V2.
Flash 2.0 SUAVE: SUAVE is an independent plug-and-play modular MEV full-chain sorting layer. The core concept of Preferences (SUAVE) is similar to the concept of Intent, both providing a way for ordinary users to customize transactions and ultimately help users achieve optimal execution. Flashbots plans to launch SUAVE Centauri in 2023Q4.
Anoma: A general Dapp architecture centered on intent, which can be modularly deployed to L1, L1.5, L2, and embedded in various EVM frameworks using modular components. The latest development in the Anoma ecosystem is the upcoming v0.22.0 testnet release of Namada, the privacy settlement Layer1 for Anoma.
Cow Protocol: CoW Protocol builds a foundational settlement network for traders and solvers, and in July released Cow Hooks, a new intent system to enable custom DeFi operations. In August, CoW Protocol processed a total transaction volume of 1.05 billion and generated a profit of 5.3 million.

Supporting Facilities: (Infrastructure related to intent and account abstract wallets)

ERC 4337: The User Operation of ERC 4337 is based on user intent, including token payment for gas, payment selection, and login methods, replacing the current transaction mempool to achieve account abstraction and improve wallet interaction experience.
Supporting facilities related to account abstraction: Bundler infrastructure (StackUp, Blocknative, AlchemyRundler, AA-Bundler, infinitism Bundler), Wallet SDKs (Safe, ZeroDev, Biconomy).
Supporting facilities related to intent: Juvix (intent-based programming language), Seaport (open-source NFT trading protocol), Gelato (Web3 automated execution layer providing smart contract automation services), okcontract (low-level intent automation tool), Delegatable (smart contract authorization proxy tool), Hyper Oracle (off-chain automation service), symmio (intent-based on-chain derivative trading framework).
Enterprise applications: (Integrated Dapp intent infrastructure, APIs, modular intent layer, domain-specific solvers)

Transaction Solver APIs: PropelleHeads;

Intent APIs: EnsoFinance;
Intent transaction infrastructure: Bob the Solver, Brink;
Composable intent tools: Essential.

Consumer Applications: (Intelligent user interaction interfaces such as Wallets, Dex, Web3 Ai Agents, intelligent search engines)

Wallets: Safe, Bitconomy, Argent, Ambire, Sequence, Versa, A3SProtocol;
Dex: CowSwap, Paraswap, Uniswap, 1inch, BananaHq, basedmarkets;
Intelligent front-end: Unibot, Dora, Web3 Analytics, kaito.ai.

Based on intent-based architecture, supporting facilities, and enterprise-side applications (due to space constraints, only listed here without further elaboration), the aim is to address the underlying challenges of intent implementation. However, infrastructure cannot be built in a day. The trend in intent consumer application layer is what truly excites, as the evolution of interaction entry points such as wallets and Dex shows the potential for intelligent interaction. Wallet intelligence: ERC-4337 introduces a new UserOps layer, where users express intent, and then the Bundler converts these intents into executable signed transactions. In recent months, the number of ERC4337 users has increased significantly, and ZeroDev, Biconomy, and Safe modular intelligent accounts are the most motivated to promote AA + intent. AA + intent has the potential to drive the arrival of wallet intelligence (SCW) in the future.

Intent: The Starting Point of Web3 Interactive Intelligence

Dex architecture transformation: Intent Based Trading aims to improve capital efficiency and user interaction experience. CoW Hooks concatenates complex actions such as trading, bridging, staking, and depositing. UniswapX already allows users to sign intent off-chain matching and on-chain settlement. The number of intent-based DEX such as BananaHq, Brink, basedmarkets is gradually increasing, and a new Intent-based RFQ narrative (SYMMIO) is unfolding. With more and more DEX and aggregators transitioning to intent-based architectures, the landscape of DEX is undergoing significant changes.

Intent: The Starting Point of Web3 Interactive Intelligence

Source: https://twitter.com/BananaHQio/status/1694013407929020740

Intelligent interaction entry points: Similar to the transformation of Web3 front-end interaction interfaces by TG Bot, the entry points focus on understanding user intent and systematically transforming it into automated, actionable tasks. Not only trading bots like TG Bot, but also Web3 Ai Agents and Web3 intelligent search engines are making Web3 interactions more intelligent.

Intent: The Starting Point of Web3 Interactive Intelligence

Source: https://twitter.com/awasunyin/status/169540582237605496

Challenges and Prospects of Web3 Interactive Intelligence

In the past, the user interaction experience of Web3 products has indeed hindered widespread adoption, and the wealth effect on-chain has masked many product issues. Now, liquidity and user attention are both scattered, and on-chain liquidity is scarce.

At the ETHCC conference in July, developers extensively discussed the future of intent-centric, how it can help users achieve more intelligent DAPP interactions in Web3. However, the application of intent-centric still faces many challenges:

Lack of intent-related programming languages (similar to Juvix)
Lack of suitable intent architecture (similar to Anoma)
Lack of domain-specific solvers (similar to Bob The Solver)
Centralized and closed intent implementation (similar to CowSwap)
Lack of front-end intent recognition solvers (similar to Unibot)
Lack of composable intent implementation (similar to Brink)
Intent security risks (as mentioned in the Paradigm article)

Challenges and hopes coexist. With the emergence of intent-centric protocols, the future of Web3 has the potential to evolve into a more user-friendly intelligent interaction interface. Users can interact with mainstream applications such as Uniswap, Blur, AAVE, etc., all in one place, making interaction as simple as hailing a ride with Uber. This presents a real opportunity for Web3 Dapps to surpass Web2 in terms of user experience. More importantly, with the combination of Anoma, Flashbots SUAVE, and DappOS, the future of intent-centric Web3 on-chain interactions will become more intelligent, customizable, and fair. By centering around intent, the power of decision-making will be placed back in the hands of users, allowing ordinary users to empower third-party robots/solvers to complete customized transactions (fees, slippage, privacy, cross-chain, MEV, or other intents). In the future, users will no longer be passive participants on-chain or be exploited by MEV searchers and validators, but will be the true masters of on-chain activities. Looking ahead to the future of intelligent Web3 interactions, whether it's account abstraction, chain abstraction, or intent, Web3 is finally focusing on user experience, and the dawn of intelligent interaction is about to illuminate the dark forest.

Acknowledgments

Special thanks to Web3 Analytics, Crypto V, Haotian, Jason Chen, Luke, Grace Deng, SixSix.eth, POOR DAO, #017, armonio.eth, Kiwibig.eth for their valuable discussions, feedback, and insights on this article.