jakiro
Source: Baize Research Institute
There are not many in China about the new narrative Web3 X Agent. It is honored to cooperate with the potential of the AI project on Solana to cooperate with the HAJIMEAI to complete the research.
Looking back at January this year, the US Bitcoin spot ETFS was approved, which triggered the strong bullish mood of investors and the inflow of positive funds.The crypto market continued to grow in the first half of 2024, with a total market value increased by 37.3%.Several Crypto narratives have gained strong momentum. In my impression, Memecoin, AI, and RWA are the most powerful.From the perspective of the currency price performance last week, AI is still relatively strong, second only to Memecoin.
Web3 X AI, this year’s various web3 VC bet on this track, and can even buy old projects TAO (Bittensor) in the open market, so I don’t think this is an instantaneous narrative.Instead, it will continue to innovate with the development of the traditional AI track.
For example, the latest trend in the traditional AI field AI Agent has also been brought into the Web3 world.In the first half of the year, there were a large number of Web3 X AI Agent projects started, such as Specialral, Olas Network, and many old projects followed this narrative, such as fetch.ai $ fet, Phala $ Pha.On ETHCC just in the past, many Web3 developers and top VCs have begun to focus on the AI Agent narrative.
This article starts from the two most representative new projects I think, to quickly understand this new narrative and opportunities in it.
Table of contents:
1. What is AI agent?
Second, the new change of web3 x ai agent
Third, Spectral Investment Research
4. Gainet Investment Research
5. Overview of other early projects
1. What is AI agent
Simply put, AI Agent is a “agent” based on a large language model (LLM) that can perceive the environment for autonomous understanding, independent thinking, independent decision -making and implementation actions.Similar to the process of human “doing things”, the core function of Agent can be summarized into three steps: perception, planning and action.
So what is the difference between AI Agent and ChatGPT and other AI chat robots?In terms of purpose and ability
Chat robot aims to be withHuman interactionBecause the AI chat robot is designed to help humans, they will not take autonomy;
And agent aims toComplete self -reliance, Have the ability to take autonomy.You don’t have to tell what to do, as long as you give it a goal, you can find a way to help you complete it automatically
For example, AI Agent is like a smarter classmate. When you are sick, you say, “Little love classmate, I am a little uncomfortable.”
It will monitor your body temperature and other body indicators, and combine the data and information on the Internet, and after analysis, give you a conclusion: “You are Yang.”Then take the initiative to generate a vacation bar, and you will send it directly to your boss as long as you nod.And it can also perceive that the antipyretic medicine at home is not enough, so you can help you add the medicine to the shopping cart. As long as you pay, you will be delivered to your door after 15 minutes.
Some of the hottest AI agents of Web2 have been used almost once, such as Perplexity, CrewAi, AutoGPT, Multion, and more commonly used functions include batch document content extraction and Internet information integration (generating research report).Multion is very interested in using this product, and I will focus on it here.It focuses onSimplify Internet user interaction, To liberate the user’s hands and change the interactive mode of users and the Internet.
For example, I issued a task to Multion, “Help me find the Web3 theme video with the highest playback on YouTube”, from “Browser Open YouTube”, “Search themes Web3”, “The highest screening volume”, all automatically completed the automatic completionIn the end, Agent output a video of 26 million playbacks to me.
Second, the new change of web3 x ai agent
First of all, what can web3 bring to AI Agent?In other words, what are the benefits of moving the AI Agent to the chain?
-
Anti -examination
AI Agent is based on LLM, and a concentrated LLM will lead to the result of the output bias. This will limit the spread of real information to a certain extent.Building AI agent with decentralized LLM can solve this problem.
-
Decentralization/ownership
The same reason, unless it costs a huge amount of data resources to do LLM, the core data of AI Agent is still preserved in a centralized AI provider.
-
Monetization
AI Agent Launchpad?A coin -issuing governance of a AI agent, IAO (Initial Agent Offering), provides a disguised way for AI Agent developers and investors.
-
Combined/Lego
AI Agent interact, trading, and empowerment of interaction, trading, and empowerment with other network agents.If you refer to the combination of DEFI, in the future, you will only use a web3 AI Agent to realize the selection of DEX from investment, finding the best liquidity, completing tokens exchange, and monitoring income. It is not surprising.
Conversely, what can AI Agent bring to web3?
1. 1.Dialogue AI Agent chainIn addition to the collection and sorting of professional knowledge that web2 agent can do, it can also achieve the search and summary of information on the web3 chain, which greatly simplifies the web3 user chain research process.
2. Just look at the concept of AI Agent and Multion. It really makes me think of the Intent-CENTRIC of the milk. What I want to do isUsers of users on the liberation chain, Change the method of interacting with the web3 to implement Mass Adoption.These interactions include but are not limited to SWAP and airdrop interaction.Imagine that to tell AI Agent, “Help me complete Linea’s airdrop interaction”, AI Agent can obtain KOL airdrop tutorials from the Internet, and then automatically complete the steps through the chain wallet.(It seems that there is no need to be a hairy studio?)
Or “help me build an Ethbot, when ETH is lower than MA200, buy ETH with 30% of my USDT balance”, is it very sexy to monitor the market around the clock.
If each solution device uses an AI Agent that can automatically do the interaction on the chain, it actually implements the most important puzzle of the intention protocol.
You can view my previous public account articles about Intent-CENTRIC.
So here, I want to divide the current Web3 X AI Agent track specific products into two types, one isDialogue on the web2 chain AI AGENTThe other is the comparison of the previous paragraphWeb3 native AI AgentEssence
The first one can be understood that the AI Agent created on a Layer-1 is suitable for the professional knowledge of web3 users to learn specific fields and do chain research.Do not include operation on the chain
The second can be understood as logical chain to help users achieve some specific interaction on the chain.Including the operation on the chain
Of course, my classification is only the most basic “Does AI AGENT have the ability to interact on the chain“, Did not consider whether it is based on a reliable and verifiable decentralized AI model.
The following introduces a typical project, GAIANET and SpecTral, for horizontal comparison
3. Spectral (with the operation capability on the chain)
The predecessor was a new way for the credit score agreement based on Ethereum to provide lenders with a new way to evaluate the credit risk of borrowers.
2023 Q4 transforms into a machine intelligent network that allows users to build ai agent on the chain and form an agent economy on the chain.
1. Business model
The four more important products of SPECTRAL:
Spectral syntax, An official AGENTS collection, users can tell what agent is to do, and Agent will convert natural language intentions to executable code to help users complete.For example, create NFT, create Memecoin, automated transactions, and chain information retrieval.I personally experience the MoonMaker agent, which can achieve automation completion from the name creation, logo design to CA deployment.
Spectral novaA decentralized platform that directly provides machine learning reasoning to smart contracts. Top scientists, enterprises, DEV, engineer and other model creators can build AI models in them to obtain user payment income.In addition, model founders can release challenges with rewards.The solution (bounty hunter) solves the challenge, win rewards, or obtains income dividends.The founder, the solution, the verification, and the consumer interact with each other on the Special’s machine intelligent network, forming a flywheel through an incentive mechanism.
Agent Wallets, Launched by 2024 Q3, by integrating AI Agent into a wallet, helping users achieve the chain operation and simplify the web3 user experience.For example, support the use of USDC for no GAS transactions, GAS Asset Agent will automatically execute exchange.
InferchainA Layer1 with AI Agent as a narrative will integrate Agents in Syntax and Nova to promote interoperability between these agents.It will be launched in 2024 Q4.
2. Development process
2021-2022 Complete financing as a web3 credit risk assessment infrastructure, respectively
2024.03 launched Syntax, officially transformed into ai agent narrative on the chain
2024.05 TGE, start the first quarter airdrop
2024.06 cooperated with the encrypted wallet Turnkey, Q3 launched Agent Wallet (Turnkey once raised $ 15 million, Sequoia, Coinbase participated in the investment)
3. The binding point of AI X Web3
Inferchain is the last puzzle of the SpecTral ecosystem, which realizes its final vision, easily develops AI AGENTS on the chain, and can communicate with each other. Finally, the application of AI in the field of web3 is transparent, decentralized and verifiable.
4. The value generated
Solved the problem of expensive costs and error costs, relying on a single source and information authenticity;
Make ordinary users (non -technical) can also quickly create agent on the chain;
Promote communication between agents on the chain.
5. Tokens
$ SPEC, users can use tokens as a payment method to pay the AI Agents developed by the community.In addition, it is also used for decentralized governance and pledge mechanism:
In Spectral Synax, users who pledge SPEC have the right to create AI Agents and visit AI Agents created by the community;
In Spectral Nova, the verifier needs to pledge SPEC as a mortgage to verify the challenges of the request to solve.
SPEC has been listed on trading platforms such as Bybit, Gate.io, Mexc, etc., with a market value of $ 85 million, FDV 800 million US dollars
At present, the circulation part is only the first quarter of the airdrop+the market share of the city, accounting for 10.3% of the total
2025.05 Open the core contributors and investors to unlock. If the encryption market continues to rise in the next few months this year, it is necessary to pay attention to the unlocked selling risk
6. Team background
Sishir Varghese -Lianchuang and CEO
Earlier as a co -founder and executive partner of AlphaChain, and also the strategic partner of Loopring Loopring
Mihir Kulkarni -Product Supervisor
Previously, it was the product operation manager of Coinbase Institution
7. Financing situation
Round one:
Galaxy, Parafi Capital, Maven 11, Alliance Dao, RARESTONE CAPITAL …
Ring 2:
General Catalyst, Social Capital, Jump Capital, Samsung Next, Circle Ventures, Franklin Templeton, Section 32 …
Super big VC investment background, such as Franklin Hampton, Samsung, Google
8. AI agent usage situation
2024.06 The Q2 data published by 2024.06:
Registered users 65,362
The number of contracts generated by Syntax 1,055,568
The average number of interactions of each user 25
The number of Memecoins created by Moonmaker 5,043
Fourth, Gainet (does not have the ability to operate on the chain)
Gainet is a distributed AI infrastructure that will gradually become a decentralized AI Agent ecosystem
1. Business model
Node is agent, andNode = agentEssenceIn the practice of setting nodes myself, when I talk to the AI Agent corresponding to the node, the results of the AI agent output will consume computing resources.
Create an Ethereum,In the form of nodes, the AI Agent network of knowledge bases in various fields, Let individuals and enterprises quickly build AI Agent according to their own professional knowledge, and provide a profit for demand.
Three core components of GAIANET network:
• Gainet node
A comprehensive software stack, individuals and enterprises can quickly deploy AI Agent that integrates their professional knowledge.
• Gaianet domain name
A collection of nodes registered under the Internet domain name is managed by domain name operators.The official idea is that each domain name is particularly in a professional field, such as finance, medical care, and education. Under these domain names, it is a AI Agent with specific functions.
User use process:
The user pays the node (AI Agent), and the cost is hosted in a smart contract on the chain.Some expenses are divided by domain name operators and then provide services to users.
• Gaianet dao
STAKERS pledged the tokens on the domain name provider to provide “trust”;
Domain name providers are eligible nodes to provide “guarantee”;
Users choose to use the AI Agent under the domain name provider, and the cost is divided by nodes, domain name providers, and Stakers.
In fact, there is also a role of component developers. Non -AI Agent developers can use components such as models, knowledge bases, plugins and other components by fine -tuning NFT forms from the need for AI Agent developers in need.
2. The combination point of AI X Web3
Move the web2 AI Agent ecosystem to the chain
3. The value generated
Developers can easier to deploy Agent. Gainet nodes support all open source LLMs, multi -mode models, text -to -image models, text -to -video models, can be added according to their choices, fine -tuning models.
Based on the domain name provider, each industry will be generated with professional knowledge AI AGENTS collection.
4. Tokens
Uncontured currency, use someone else’s Agent to pay tokens in GAIANET
5. Team background
Matt Wright -Co -founder and CEO
Graduated from the University of California Los Angeles, a community director of Consensys, and the co -founder of EVM Capital.He worked in JP Morgan.
Shashank Sripada -Co -founder
In Web2 is the background of venture capital, founding and joining a number of VCs.Graduated from the London School of Economics and Politics, there are some political backgrounds.
Sydney Lai -Development Propaganda Director
EVM Capital’s Lianchuang and CTO, graduated from the University of California, Berkeley
6. Financing situation
2024.05.28, seed wheel, $ 10 million
Mantle Network, Bytetrade, EVM Capital, Mirana, Lex Sokolin Secret Supervisor)
7. AI agent usage situation
The number of users is unknown, with only 35 available dialogue AI Agent, which involves financial, Crypto, programming and other fields.The total number of nodes currently: 18,594
Due to the testing phase and can run nodes in multi -threaded, it is speculated that the nodes of batch airdrops account for the majority.
5. List of other early projects
ZOTTO: Users can create their own AI Agent to achieve transaction intent. Application scenarios such as mirror smart money addresses and multi -condition trigger transactions.Just recently got on Testnet, you can focus on focusing on.
Agentlayer: Layer2 based on OP Stack, with AI Agent as the main narrative, and Unicom between AI Agents developed by the community developed and developed by the community.
Olas Network: One linked AI Agent ecosystem, a single agent or multiple agents collaborates to complete the task, and pass the output to the chain.
Theoriq: Theoriq aims to be a modular, combined AI Agent base layer.
Agentcoin: Evo Ninja, a mature AI agent product in Web2, transformed to do the common Web3 AI agent.
Giza: We are doing the web3 AI Agent framework, through the ZKML chain reasoning, and execute it on the chain.
Olas Network: A Web3 AI Agent ecosystem, under the chain, complete the task proposed by the user by a single agent or multiple agents, and pass the output to the chain.
