Introduction

The AI concept is always in demand in the market.

As a narrative that persists throughout the year, regardless of the ups and downs of the crypto market, AI projects always manage to stand out. In various reviews and forecasts, you can almost always see a bullish outlook on AI.

However, AI is a vast concept, and the key question is: which specific sub-projects within AI are more likely to succeed?

A relatively safe choice is to use short-term hot events as a catalyst to find crypto projects directly related to the concept at hand, such as AI agents.

On January 11th, ChatGPT officially launched its store feature. Unlike traditional apps, the GPT store does not require users to have any programming experience. By simply entering the desired functionality in natural language, the system will create a custom AI chatbot.

In essence, this is an AI agent store, containing a multitude of robots that can handle certain tasks on your behalf following specific logic. Leveraging the appeal of GPT, it is foreseeable that the store’s applications will surge, possibly including crypto-related agent robots.

Whether these robots are useful is one question, but recognizing the popularity of the AI agent concept is another.

Almost at the same time, the well-known crypto VC Pantera explicitly stated in a recent lengthy article that they are interested in the combination of AI agents and Web3 in 2024.

(Recommended Reading: Pantera’s Key Focus Areas in 2024: AI Trends Continue, Web3 to Aid in Reasoning, Data Privacy, and Incentives)

(Related Reading:Pantera 2024 key areas of focus: AI trends continue, Web3 will aid development in reasoning, data privacy and incentives）

If we consider the GPT app store as a fuse and the attention of top VCs as a breeze, will the AI agent narrative ignite this year?

We can’t predict the future, but we can be prepared.

So, what we need to understand now is:

What exactly is the principle of AI agents, and which crypto projects might directly benefit from this?

AI agent, what does it have to do with the encryption industry?

First of all, we need to figure out how the AI ​​agent works. This will help us identify which encryption projects really fall within the scope of the AI ​​agent and which ones are just concepts.

The robots in the GPT store can at most give people a perceptual understanding of AI agents; but in principle, what kind of products are considered AI agents?

The author believes that to determine whether a product is an AI agent, we can simply grasp the key to the problem:

Programs or devices that use AI technology to automatically perform tasks or provide assistance to users.

Everything from simple chatbots to complex automated systems can be considered AI agents. But these things must at least meet the following conditions:

• Automation capabilities: The ability to perform tasks with little or no human intervention.
• Sensing environment: Can sense its operating environment through sensors or data inputs.
• Decision Making: Ability to make decisions based on its programming and the data it receives.
• Learning and Adaptation: Ability to learn and improve performance based on new data and experience.
• Interactivity: Can interact with human users or other systems to respond to requests or provide information.

But if we abstract the above characteristics, we will find that AI agents are a bit like smart contracts —- given preset conditions, the results are automatically executed.

So if we want to identify whether a project really uses AI agents or is using smart contracts to exploit AI concepts, we can actually identify it simply through the following dimensions:

Is there any initiative.

Smart contracts have no autonomy; they only react passively (Reactive) based on pre-written rules and have no ability to make autonomous decisions based on changes in the external environment. An example: you set a certain price and buy a certain token at that time.

In contrast, AI agents are generally considered active and can collect data, learn, make decisions, and initiate tasks on their own without external commands. An example: monitor market data and buy a token at a price that the AI ​​thinks is suitable for profit.

After understanding this difference, let’s take a look at the connection between AI agents and the encryption industry.

@chiefbuidl, co-founder of the well-known encryption project Space and Time, made a very classic and vivid metaphor:

Cryptocurrency is like cash, blockchain is like cash register, Dapp is like POS machine, and AI agent is like cashier.

Take a closer look at this sentence. When users trade cryptocurrencies like cash, the public chains behind them are actually responsible for recording and accounting; while Dapp acts as a transaction interface similar to a POS machine, and the AI ​​agent becomes the cashier. Let me tell you directly:

You don’t have to worry about how you use the money or how you keep the accounts; you just need to tell me roughly your intentions, and I will automatically help you spend the money and provide you with the services you want.

In this chain,Cryptocurrency, blockchain and Dapp are actually difficult to understand, but AI agents are the link that is most likely to be close to users and simplify the complex.

Therefore, AI agents are concerned with the encryption industry and can find ways to make the experience of using encryption products better (including but not limited to transaction experience).

Which crypto projects are related to AI agents?

So, which projects will be related to the concept of AI agents?

Perhaps we can divide projects into two categories. One is that the project itself has the ability to provide AI agents for others to use; the other is projects that use AI agents to improve the experience of their original products.

Type 1: Encryption projects that directly engage in AI agency business

Autonolas($OLAS): AI agent created to optimize the efficiency of encryption projects

Autonolas is a project directly related to AI agents. Its business is to design AI agents for the encryption industry to handle tasks in different scenarios.

Specifically, the Autonolas technology stack includes:

• Autonomous agency service
• Composable autonomous applications
• An on-chain protocol that protects proxy services and incentivizes development

(Related Reading:In-depth Autonolas: AI agent-driven off-chain services, full analysis of products and economic models）

The focus is autonomous agency services.

The AI ​​agents that make up these services can pull data from any AI model in the world. Every GPT, LMM or subnet is included (which means even linkages with $TAO are possible). Through service coordination, models for handling specific tasks are assigned to certain agents.

So what exactly can these agents do?

Judging from the product collection provided by Autonolas’ official website, its businesses related to the encryption industry include but are not limited to the use ofAI agents can predict the market, predict the APY income of certain protocols, act as oracles to provide more accurate off-chain data, help DAO governance, automatically manage and operate smart contracts, automatically establish DeFi capital pools, etc.

Generally speaking, AI agents can be used in all process-related operations related to encryption projects.

The other two businesses of Autonolas are based on autonomous services. Through these basic autonomous services, other developers can freely combine the functions to form their own applications; at the same time, Autonolas can also establish an application store to allow developers to register and monetize their services.

Judging from the data, the project’s AI agent service usage continues to increase. Since it doesn’t choose projects to cooperate with, Autonolas can theoretically become the standard for all encryption projects, adding a layer of AI agent capabilities to areas that need automation in your projects.

Its token OLAS was launched in the on-chain liquidity pool in July last year. The price at that time was around US$0.1. The current price is around US$4.6, and the market value has reached US$200 million.

However, considering the overall narrative of AI and the expected role of the project in business, the current market value and price are not particularly overvalued. For comparison, $TAO, which is also an AI narrative but has different business, has a market value of about US$1.5 billion.

Fetch.ai ($FET): old project, AI agent serving the entire industry

Fetch.AI (FET) also focuses on building and promoting AI agent services. These agents are designed as modular building blocks that can be programmed to perform specific tasks. These agents are able to connect, search and trade autonomously, creating dynamic markets and transforming traditional economic activities.

Compared with OLAS, Fetch is an old project. It was established in 2017 and launched on the mainnet in December 2019. IBC, which is currently connected to Cosmos, is also considered an AI project in the Cosmos ecosystem.

However, Fetch is not just an AI agent for encryption projects, but extends its services to different industries. Judging from the official examples, its business can reach a variety of services such as e-commerce, automobiles, law, Internet of Things, and weather.

In addition, another feature of Fetch is that it is developer-friendly.

Fetch also launched Agentverse, a no-code management service that simplifies the deployment of AI agents. Just like traditional no-code platforms (Replit) and services like Github’s Copilot make writing code accessible to the masses, Fetch is working to further democratize Web3 development in its own unique way.

Through Agentverse, users can easily launch their first agent, which greatly lowers the threshold for using advanced artificial intelligence technology.

However, as a project with a long history, Fetch’s actual product is still in the wish list stage and is not fully open to the public. This also makes people doubt whether the project is really doing something solid.

In terms of tokens, FET is used as network gas fee and is also used for node staking to maintain network operation. Its market value has reached 500 million US dollars. The author believes that compared to OLAS, the price/performance ratio is weaker and the room for upside is relatively limited. But it can still be treated as a relatively Beta project, and it may also see a certain increase when the AI ​​agent narrative has a catalyst.

PAAL AI ($PAAL): Focus on building AI assistants for crypto users

PAAL’s goal is to create an AI-powered platform that is accessible, user-friendly, and able to provide comprehensive knowledge, support, and tools in the ever-changing world of cryptocurrency and blockchain technology.

The project hopes to provide users with their own personal artificial intelligence assistant that they can rely on for accurate and reliable information, help with customization and scalability, and a deeper understanding of the crypto ecosystem.

Specifically, we can actually understand PAAL as an encrypted version of GPT and transaction BOT.

PAAL’s main AI tools include:

• Mypaal: The AI ​​robot with certain knowledge mentioned above answers various questions raised by users about encryption projects;
• Autopaal: Acts as a crypto expert, providing cryptocurrency-centric research, insights and analysis, monitoring market trends and informing users of specific situations;
• Autopaal

From a conceptual point of view, PAAL can indeed be regarded as an AI agent, but the business aspect is more focused than the previous two projects. Currently, it only serves the transactions and learning of crypto users. If we don’t worry about the specific differences, we can even directly understand it as an advanced trading BOT.

The current total market value of the token $PAAL is around US$100 million, which is lower than the previous two projects. However, considering the narrowness of its business and its strong correlation with the crypto market, I don’t think OLAS has the overall imagination. So high.

Type 2: Projects that use AI agents to strengthen existing businesses

The above projects are all directly engaged in AI agency, while there are also some projects whose main business is not entirely AI agency, but has integrated this function into the original business to improve their performance.

Due to space limitations, we will not introduce the principles of this type of project in more detail, but will only list them here: Root Network($ROOT): L1 specifically optimized for Metaverse, gaming and Web3 user experiences, powered byFutureverse This company provides technical support for AI and Metaverse. It is now possible to integrate AI agent capabilities into games supported by the network to enhance the gaming experience.

Parallel （$PRIME）：A card battle game with a sci-fi background funded by Paradigm. The game currently utilizes AI characters to create new in-game items, which can be stored within the AI ​​character’s own wallet, equivalent to an AI agent that creates game assets.

Oraichain ($ORAI):A company that provides AI Layer 1 data economy and Oracle services with the goal of building a trusted AI tool that powers Web3, scalable dApps, and the data economy. The project recently provided a token analysis tool called DeFi Lens, and also added AI agent capabilities to perform predictive analysis on tokens.

Due to limited energy, we have not listed all AI agent-related projects on the market.

However, whether the above-mentioned projects are directly or indirectly related to AI agents, in addition to the hot concept of AI, it is also necessary to see whether their AI agents are really implemented, that is, whether the AI ​​agent business itself can be used, rather than just a piece of paper. Empty talk.

The bubble of AI is huge, and the bubble of crypto projects catching on to the popularity of AI may be even bigger.

Focusing on projects with more solid fundamentals will be more conducive to capturing solid and sustainable value in the AI ​​narrative throughout the year.

Disclaimer：

1. This article is reprinted from [TechFlow deep tide]. All copyrights belong to the original author [Shenchao TechFlow]. If there are objections to this reprint, please contact the Gate Learn team, and they will handle it promptly.
2. Liability Disclaimer: The views and opinions expressed in this article are solely those of the author and do not constitute any investment advice.
3. Translations of the article into other languages are done by the Gate Learn team. Unless mentioned, copying, distributing, or plagiarizing the translated articles is prohibited.