Author: Jessy, Golden Finance
On August 30, Coinbase CEO Brian Armstrong posted on X that the company had successfully completed the first cryptocurrency transaction between artificial intelligence (AI) robots on August 30. In this transaction, an AI robot used encrypted tokens to purchase AI Token from another AI robot.
This means that artificial intelligence has its own wallet and can trade automatically. Brian Armstrong said that for the development of the AI industry, giving AI agents trading capabilities will be a major breakthrough, especially for companies that want to achieve AI checkout.
AI can make transactions autonomously under the set algorithm program, which is an important step in the implementation of the AI Agent concept that became popular at the end of 2023. Only by being able to trade can we realize various economic behaviors in the economic society.
When AI Agent becomes an important direction for the development of AI in the future, what role can blockchain technology play in this? And what can the practice of AI Agent on blockchain bring to the blockchain industry?
How is AI-to-AI trading implemented in detail?
The realization of the first AI-to-AI on-chain transaction was made possible by an infrastructure, the MPC wallet released by Coinbase. Simply put, the idea is to combine AI Agent with a cryptocurrency wallet and allow AI Agents to trade on their own without human intervention.
At this stage, AI Agents cannot have a traditional bank account in the world of traditional finance, but having a crypto wallet is already possible.
In theory, with this Coinbase MPC wallet, AI Agents can trade with USDC on the Base chain. Currently, using the development kit provided by Coinbase, you can create an AI Agent MPC wallet.
In terms of project progress, in early August, Coinbase made a demo product for AI to pay cryptocurrency to humans, and then launched a "CDP AI Builder Program" to provide funding for AI robot developers who build integrated Coinbase MPC wallets.
Why use blockchain technology and virtual currency to conduct transactions?
In the view of Skyfire 's founder Sarhangi, AI agents are far from reaching their full potential. Once they encounter a problem that requires payment, they will be in trouble and need human intermediaries to help them with credit cards. Developers say that such a system is in urgent need of streamlining, and the traditional banking system is powerless.
"It can't be Visa, because these rails simply can't meet the needs of artificial intelligence, that is, to be able to conduct micro-transactions efficiently, at low fees, and around the clock," he added. "These are things that cryptocurrencies and blockchains can provide."
Encryption industry leaders such as Brian Armstrong have also publicly stated that the future of artificial intelligence depends on the adoption of on-chain payments.
Skyfire is one of the blockchain companies doing this. It was founded by two former executives of Ripple and completed a $8.5 million seed round of financing in August this year.
The project introduction states that Skyfire is a financial stack for the AI economy, providing an instant, global and open payment system for AI agents to conduct fully autonomous transactions between AI agents, LLMs, data platforms, service providers and other goods and services.
Skyfire has developed an open source payment system that allows AI Agents to conduct various transactions on the Internet, including the purchase of data storage, creative assets, air tickets and daily necessities. Its payment network runs on the USDC stablecoin and is currently deployed on the Polygon network, with plans to expand to other blockchains in the future. For services that do not yet support USDC payments, Skyfire is considering integrating third-party aggregators to achieve the transfer of USDC payments to traditional payment methods.
According to Skyfire official news, there are already some projects trying to use Skyfire's services, the first of which is Payman, a project that allows AI to pay humans for various tasks. In theory, through Skyfire's platform, Payman's AI intelligence can now exercise full autonomy to hire contract workers and fulfill payments.
According to Skyfire's official news, another project that is experimenting with Skyfire is Denso, a global auto parts manufacturer, which has created an AI Agent that hopes to be able to purchase various materials on its own without human help. Previously, its operation required human intervention and wire transfer payments at the end of the month, but trying to use Skyfire's services seems to solve this problem.
Another company doing this is Payman, which uses the services provided by Skyfire and is invested by Visa. The idea of Payman's project is to provide a platform API that allows AI Agents to pay humans in a variety of ways, including legal currency, cryptocurrency, and bank accounts.
All of the above-mentioned projects can be regarded as a payment middleware platform that focuses on providing interfaces and services for payments between AI Agents and humans. They are more targeted at AI developers.
In addition to Skyfire, some non-payment blockchain projects are also trying to implant this concept and technology. Parallel Alpha's latest game Colony is designed to create AI characters that can have wallets and trade with each other.
What can blockchain do for AI Agent?
AI Agent is a new concept and development trend in the AI track that is popular with capital. AI Agent refers to artificial intelligence that can make decisions and complete actions independently without human intervention in the process.
When robots evolve into powerful AI Agents, they will eventually be able to handle complex tasks autonomously and make more intelligent decisions.
Opening up the autonomous payment and transaction of artificial intelligence is undoubtedly an important step in the implementation of the concept of AI Agent. At present, AI cannot use traditional payment methods to purchase cloud computing services (such as AWS), developer tools (such as GitHub) or pay other fees (such as booking air tickets or hotels). AI cannot bypass paywalls to obtain the information it needs. These limitations greatly hinder the potential utility and actual work ability of AI.
Intelligent agents can handle many tasks autonomously, and inevitably need "transactions". In the economic society, they need to conduct various economic interactions with other entities.
Blockchain is a direction for realizing artificial intelligence "transactions". At the current stage, blockchain has the characteristics of lower account opening threshold, which can reduce the friction costs of traditional finance such as account opening and intermediary. Moreover, on-chain transactions can realize real-time transactions and improve transaction efficiency.
This is the advantage of blockchain technology, and some projects in the industry are also practicing this direction. However, is blockchain technology necessary? A very realistic question is that on-chain transactions based on blockchain technology are currently in demand, but the scope of application is very small. At present, the traditional financial system is the mainstream, and most transactions are still based on the traditional financial world. Therefore, the opening of transactions by artificial intelligence agents enables intelligent agents to cope with various daily economic behaviors. In fact, the first thing to open up is the traditional financial system. It is just that now, due to some realistic factors and the stage of technological development, the use of blockchain technology to open up transactions has become a feasible attempt.
In fact, based on blockchain, the cornerstone of opening up channels for AI transactions is that blockchain transactions themselves have been embedded in all aspects of human life.
Otherwise, similar AI-to-AI transactions will fall into the dilemma of "where is the market".
But if we look at the significance of this matter from Web2 back to Web3, we will find that it is very conducive to building crypto-native AI Agents. The crypto-native payment channel has been opened up, and the digital property rights inherent in crypto assets can be given to AI Agents. This is the cornerstone of AI Agents as independent economic persons on the chain.
How to combine encryption with AI has always been an important exploration direction in the encryption industry. And AI Agent is an important branch in this direction. At present, some well-known projects in the industry have already made attempts in the direction of corresponding AI Agents. For example, the AI agent launched by the Fetch.ai platform is a modular component. AI developers can register their own services, LLMs, ML models, APIs and other AI businesses as agent services. Traditional product developers can quickly integrate these agents to perform specific tasks.
In terms of trading, such as price prediction, transaction strategy, stop loss strategy, leverage dynamic adjustment, KOL smart copy, lending... Every service that used to make money on a centralized exchange can be packaged into a customizable self-learning AI Agent, and traders will use AI Agents to execute their own strategies.
There are also public chains born for AI Agents that appeared this year. The AO chain launched by Arweave is built based on Arweave's on-chain storage, which realizes infinitely scalable decentralized computing and allows any number of processes to run in parallel at the same time. The decentralized AI Agent is hosted by AR on the chain and run by AO on the chain. In other words, it is stored on the AR chain and calculated on the AO chain. Under this set of design patterns, the ultimate vision of AO is to achieve a seamless integration of AI and blockchain, so that AI models can be hosted on the chain and run directly on the chain, and make autonomous decisions.
However, these projects are still in the very early stages of exploration. The transaction of AI Agent is a complex and sophisticated system. How to ensure its stability in a complex environment is the most difficult challenge to be faced.
Edmund
Weiliang
Huang Bo
Brian
Miyuki
Cheng Yuan
Aaron