Interpreting the new AI agent Talus: How does the "new and old combination" create differentiation?

Author: 1912212.eth, Foresight News

The combination of AI and Crypto is a hot topic, with numerous projects emerging in various segments such as computing, data analysis, and agents, and many assets performing well in the secondary market. However, today we introduce the main character, the AI agent pioneer Talus. At the end of February this year, Talus Network completed a $3 million financing round, led by Polychain Capital, with participation from dao5 and others.

Recently, the co-founder of Delphi Labs stated that AI agents will drive the next wave of cryptocurrency adoption, becoming increasingly advanced, with more users relying on them to manage their financial lives. This trend is reminiscent of the transition from physical banks to online banking.

On the occasion of the official release of the white paper, Foresight News helps you understand the new AI agent player Talus.

What is an AI Agent?

There are many concept products in the market that ride the wave of trends, so how do we identify a true AI agent? In short, it uses AI technology to automatically perform tasks to assist users. Its main characteristics include automation capabilities, environmental perception, decision-making, learning and adaptation, and interactivity.

It is worth noting that among the aforementioned features, "proactivity" is the most critical and is the core logic for understanding AI agents. In contrast, blockchain smart contracts can only execute passively and lack proactive functions.

According to a metaphor by the co-founder of Space and Time, cryptocurrency is like cash, blockchain is like a cash register, DApps are like POS machines, while AI agents are like cashiers.

Talus: An AI Public Chain Platform Based on Move

Talus is an AI public chain platform based on the Move language. It features high performance and is optimized for AI use cases, allowing AI agents to interact and transact within the system. Participants in Talus can build new economies for AI applications from a universal protocol of accessible resources.

Talus's AI agents can provide functional support and optimization in multiple areas, such as offering wallet analysis tools, monitoring transaction statuses in real-time, and enhancing user asset management experiences. In the DeFi space, they can optimize liquidity management, automate on-chain portfolios, and improve capital efficiency and returns. Additionally, they support automated asset management and governance decisions for DAOs.

The aforementioned areas are just a part of its applications. Talus's AI agent functions can be useful for everyone, whether you are an ordinary user, a researcher, a programmer, or a project manager.

New and Old Combinations in Architecture

The layered architecture of Talus can be categorized into two aspects: selecting better components from similar systems and developing and innovating from scratch.

At the bottom protocol layer, Protochain is the POS node code name for Talus, primarily supported by Cosmos SDK and CometBFT. Cosmos SDK has proven to be flexible, powerful, and high-performing, providing a secure and scalable solution alongside CometBFT. In addition, Sui Move is utilized as the strategic decision-making smart contract language, based on its high-performance characteristics, security, and programming attributes, ensuring good reliability and flexibility in on-chain logic security, object models, and high performance.

Talus also highlights the benefits of choosing Cosmos SDK, such as IBC, which can enhance the functionality of AI agents across multiple dimensions, including interoperability, cross-chain atomicity, scalability through sharding, as well as security and isolation.

Above the protocol layer are mirror objects. The name may sound confusing at first, but as the name suggests, it is the off-chain representation of the on-chain world, akin to a mirrored projection.

Mirror objects serve as a link between resource-intensive AI operations and the on-chain environment, as direct AI computation on-chain is not feasible due to computational intensity. Generally, the main basic functions of mirror objects in operation include unique representation of resources, tradability of off-chain resources, ownership representation, and verifiability.

Above the mirror objects is the AI stack, primarily used to create smart agents for Talus using Talus's SDK, integrating off-chain components and resources, as well as oracles. The Talus SDK consists of multiple libraries and data models that can be used to build smart agents, along with ready-made smart contracts. The Talus AI Stack helps smart agent developers seamlessly access the necessary off-chain resources.

At the forefront of the entire architectural design are the smart agents, designed to perform specific tasks or make decisions on behalf of users. They are equipped with AI capabilities, enabling them to perceive the environment, analyze information, and take appropriate actions to achieve predefined goals.

Token Utility

The token that Talus will launch is TAI, with primary functions including usage by smart agents, resource purchasing and trading, network staking for security and consensus, and governance. Resource providers earn TAI for their contributions to the network in computing, data, or models. For developers, the TAI token simplifies access to these resources, facilitating project expansion.

TAI does not differ significantly in functionality from other AI-related product tokens, but specific details will need to wait for future official announcements.

Specific Use Cases

Portfolio Management

Talus can perform intelligent portfolio management for users, requiring much less cost, effort, and trust compared to humans. By guiding the smart agent to understand your investment goals and risk tolerance, it will make calculated decisions to rebalance the portfolio, optimize and diversify investments, and discover new opportunities.

Smart Oracles

Smart agents can serve as smart oracles, transmitting price data from off-chain. This is especially valuable for illiquid and hard-to-price assets (such as watches, collectible cards, or real estate). On-chain smart agents can automatically analyze the market to determine valuations in a decentralized, trustless manner.

Natural Language User Experience

AI models can use natural language processing to understand user intent and intelligently execute them based on context, market analysis, user goals, and risk aversion. These ranges can vary from simple requests like "buy $100 of Bitcoin at the best price this week" to complex tasks like "rebalance my portfolio to take advantage of the meme coin bull market."

AI-Enhanced Insurance

Smart agents can automate a large number of insurance operations. This includes managing policies in open smart contracts, collecting premiums, distributing payments from decentralized vaults, assessing claims, detecting fraud, and performing actuarial calculations.

Smart Wallets

Smart agents can be embedded in wallets, making them more powerful and autonomous. Wallets can learn activities to automate arrangements and make and request payments as needed. Smart agents can proactively and intelligently monitor for fraud, reject received scams, and identify suspicious activities.

Decentralized Credit Scoring

Smart agents with access to on-chain data can accurately and fairly assess creditworthiness. They can also adjust rates and payment schedules in real-time to respond to changing balances and financial activities. The result is a transparent and verifiable system.

On-Chain Fund Management

Using smart agent managers significantly reduces fund fees. They can allocate assets, monitor markets, perform sentiment analysis, and read white papers and reports. Smart agents can even be trained to become experts in specific industries, markets, and projects. The result is a broader access to more robust investment strategies.

Optimizing MEV

Smart agents embedded in order books and liquidity pools will reconstruct the way trades are ordered and processed. AI can intelligently predict future block values, which is far more powerful than today's simulation-based methods. This allows MEV strategies to maximize the value of the ecosystem while minimizing exploitation and negative outcomes for individuals.

Trade Optimization

Smart agents can automatically execute trade routing in a more dynamic way than existing bots. Smart agents owned by users or DEX can analyze exchange paths in real-time to optimize lower fees and higher returns for individuals and the ecosystem.

Content Moderation

Smart agents can perform tedious content moderation tasks for decentralized social networks. They can monitor AI-generated content, spam, scams, and abuse in real-time. The combination of on-chain posts, accounts, and moderation means a fairer and more open future for the read-write internet.

Continuous Auditing

Auditing is currently a poorly performing, time-consuming, and costly process prone to human error. Smart agents trained in software principles and project goals can continuously monitor smart contracts and code to detect new vulnerabilities and unexpected changes. They can even respond to new vulnerabilities in other projects to identify and fix flaws.

Game Testing

A key aspect of game development is game testing, where players identify defects by demonstrating the game. This process can be automated, speeding it up and reducing costs through smart agent players. They can even be customized to represent various skill levels, language fluency, and incentives for creating better games.

Conclusion

In the highly competitive AI space, how will Talus differentiate itself and stand out? Topics such as ecosystem, community, and technology can all be explored. The timing and intensity of incentive measures, collaborations with brands, and boundary expansion will all be key factors in Talus's influence.