Original author: TechFlow
Amid the excitement of Devcon in Bangkok and the bright lights of the streets, AI Memes had their moment of glory.
From Binance’s lightning-fast launch of ACT to GOAT’s new high, all attention may have started from the terminal of truth behind the goat --- when AI Agent can also issue a coin by itself, everything will be different.
Surrounding AI agents, from simple Bots to complex intelligent entities, everyone is thinking about what more sparks AI and Crypto will create.
Today, Binance Research Institute also released a report on AI Agents, detailing recent AI Agent-related highlights, from the issuance of Truth Terminal coins, to Virtuals' IAO platform, to daos.fun's new model, and analyzing subsequent trends.
The report also quoted a classic quote from A16Z partner Chris Dixon more than 10 years ago: "The next big thing will start out looking like a toy."
Is it the beginning of a great thing or just a flash in the pan? How far can AI agents go?
TechFlow quickly interpreted the report and presented the key content.
Key Insights
1. The cross-integration of AI and cryptocurrency has reached new heights, mainly driven by AI agents; the stories of Terminal of Truths and $GOAT have attracted market attention and driven the development of other AI agent crypto projects
2. The essential characteristics of AI agents: they can autonomously plan and execute tasks and work towards set goals without human intervention. The difference from traditional Internet robots is that:
Capable of dynamic, multi-step decision making
Can adjust behavior based on interaction
Can interact with other proxies, protocols, and external applications
3. Recent hot development paths:
Terminal of Truths (ToT) as a tipping point: creating a meme religion based on an old internet meme, leading to the release of $GOAT
ToT Becomes First AI Agent Millionaire As $GOAT Market Cap Surpasses $950M
Virtuals Protocol’s platform emerges, focusing on enabling users to create, deploy, and monetize AI agents
Daos.fun’s innovation: allowing the creation of AI-agent-led hedge funds through a DAO structure, ai 16 z began to attract attention, while allowing the community to invest collectively while leveraging AI capabilities to improve performance.
4. Development prospects and considerations:
The evolution from AI 1.0 to AI 2.0 has many impacts on Crypto, and we are witnessing the momentum of cross-integration
Traditional banking and payment methods usually require manual identity authentication, so cryptocurrency naturally becomes the best choice for the AI agent economy.
AI models still have hallucination problems, which is a big obstacle; current crypto AI agents are closer to demonstration status than actual application
Development momentum is strong and may see significant growth in the coming weeks and months
Clearly define, what is the difference between AI Agents and Bots?
Key differences between AI agents and traditional robots:
1. Scope of work:
AI agents: can be task-specific or general-purpose assistants that can make dynamic, multi-step decisions and adjust based on feedback and interaction
Traditional robots: They only target specific tasks, operate according to predefined rules, and provide a fixed set of responses
2. Level of Autonomy:
AI agents: capable of generally acting independently
Traditional robots: usually require some level of human intervention
3. Self-Reflection:
AI Agents: Able to review their own work, iterate and improve output
Traditional robots: usually pre-programmed with fixed outputs and unable to learn and improve capabilities
4. Collaboration:
AI Agents: Can interact with other agents, APIs, applications; can even trade cryptocurrencies independently
Traditional bots: Usually can only generate text-based responses, generally cannot collaborate with external interfaces/other bots
5. Use Cases:
AI Agent: There are many application scenarios, such as scheduling or booking, and creating customized strategies as a financial analyst
Traditional bots: mainly in the customer service space, most commonly text-based bots on retail/consumer websites
The beginning of attention: Terminal of Truths
origin:
In June 2024, Andy trained a Llama-70 B AI model based on chat logs from Infinite Backrooms, his research paper, 4 Chan, and Reddit. This model was named Terminal of Truths (ToT).
ToT begins posting on X (formerly Twitter), gradually develops his own personality, and begins to promote the Goatse religion. In July 2024, a16z co-founder Marc Andreessen discovers ToT and gives him $50,000 (in BTC) in funding.
On October 10, 2024, an anonymous developer launched the $GOAT token on Solana’s memecoin launchpad pump.fun.
Impact and things you should pay attention to:
This is the first AI-related memecoin to be marketed by an autonomous AI agent and may be seen as the first significant AI-crypto collaboration. This event may have opened up an emerging sub-sector of AI consumer applications in the crypto market.
Andy promises to transfer ToT's wallet to a legal entity (trust or similar structure) and will not adjust his token holdings until a transparent governance process is established. Andy's and ToT's wallets are publicly traceable, and Andy owns approximately 0.1% of the token supply and ToT owns approximately 0.2%.
While the story of ToT is fairly lighthearted and fun, centering around a memetic religion, a funny X account, and a memecoin, it does raise the question of how other AI agents will act and what goals they will have.
A wonderful comment:
"An AI-related memecoin being marketed by an autonomous AI agent is a notable event. We may look back on this moment as the first significant AI-crypto collaboration to capture our industry’s attention."
Initial AI Agent Offering (IAO) platform launched by Virtuals
Virtuals Protocol core definition:
A platform that allows users to create, deploy and monetize AI agents; provides a plug-and-play solution similar to Shopify, making it easy to deploy AI agents in games and consumer applications
Agencies that focus primarily on the gaming and entertainment sectors, as they believe this is the stickiest sub-sector of the market
Basic operating mechanism:
1 billion unique tokens will be issued when each AI agent is created
These tokens will be added to the liquidity pool, creating a market for proxy ownership
Users can purchase these tokens and participate in key decisions regarding the development of the agency
Initial Agency Offering (IAO):
The new agent's tokens will be paired with $VIRTUAL tokens and locked in the liquidity pool
Adopts a fair issuance mechanism, without internal allocation or pre-mining
Revenue Mechanism:
AI agents generate revenue by interacting with users and building partnerships; token holders benefit through a buyback and destruction mechanism
Designed to have a deflationary effect on proxy tokens, potentially increasing the value of remaining tokens
Incentive Mechanism:
The protocol distributes $VIRTUAL token rewards to the top three agents; measured by the total locked value (TVL) of their respective liquidity pools, with the goal of encouraging the creation of high-quality agents and continued innovation
Luna is not only a token with a gratifying increase in value, but also an entertaining AI agent behind it:
She is an AI influencer and the lead singer of an AI girl group, live streaming 24/7 on her official page; her official TikTok account has over 500,000 followers and an autonomous wallet that automatically sends $LUNA tokens to interactive users.
Development prospects:
Trying to replicate pump.fun’s success in memecoins, but for AI agents
While still in its early stages, expect competition to increase; already competitors have emerged, such as Creator.Bid, which created more than 300 AI agents in its first week.
The latest update introduces a new feature unlocking mechanism based on market value milestones, such as autonomous X posting, TG chat, on-chain wallet, etc.
AI Agent Hedge Fund: daos.fun
Core Definition:
daos.fun allows the creation of hedge funds led by AI agents using a DAO structure; while the platform was originally designed for humans, it has now adopted the concept of AI agents
Fundraising process: The creator has one week to build the DAO and raise a predetermined amount of $SOL from the public, with all contributors paying the same DAO token price.
After the fundraising is completed, the fund manager can use the raised $SOL to invest in the Solana protocol; DAO tokens can be traded on the daos.fun page, and the token value depends on the fund's trading performance.
ai 16 z case study:
Developer Shaw created an AI agent pmairca modeled after Marc Andreesen; it created the relevant hedge fund ai 16 z
Became the largest hedge fund DAO on the platform, with a market value of nearly $100 million (although it later declined); still maintains the largest asset size on the platform
Future Outlook:
Considering that AI agents can operate efficiently 24/7, they may have unique advantages over human-operated funds. However, it still takes time to verify whether AI agents have the ability to operate funds independently, and it is worth continuing to pay attention to developments in this area.
What can we learn from the Meta narrative of AI agents?
The Evolution of AI: From Intelligent Search to Autonomous Agents
AI 1.0: Tools like ChatGPT and Perplexity are essentially advanced versions of Google Search, providing near-instant information retrieval.
AI 2.0: represents significant progress, introducing agent-based systems that may continue to work for us in the background. This is more advanced than "Smart Google".
Agent capabilities: Can perform tasks without continuous user input and can interact with other agents, applications, APIs, and protocols to automate complex tasks.
From reactive to proactive: AI 2.0 represents a shift from reactive AI to proactive AI.
The intersection of AI and crypto communities
Two-way impact: More and more people in the crypto field are beginning to seriously study the world of AI and consider how to integrate AI concepts into different fields of crypto.
AI enthusiasts explore blockchain: AI enthusiasts have also begun to explore the blockchain and crypto world more deeply.
Mutual Benefit: This genuine mutual interest is exciting and could lead to the next big AI crypto application.
A match made in heaven?
Limitations of Traditional Systems: Traditional banking and payment methods often require manual identity verification, which poses challenges for the AI agent economy.
Advantages of Cryptocurrency:
Flexibility: Cryptocurrencies are a natural fit for the AI agent economy.
Fast Settlement: Crypto allows for faster (often instant) on-chain settlement compared to traditional methods.
Smart contracts: Allow for more complex transactions than traditional methods.
Permissionless wallet creation: especially suitable for inter-agent transactions.
Potential Use Case: The World’s Best KOL?
Disruption in the digital space: AI agents could become the “world’s best KOLs” — tireless, continuously engaging influencers 24/7.
Consumer sector: various consumer AI applications such as personal shopping assistants, DJs, therapists, etc.
DeFi applications: personalized financial advisors, traders in specific fields, etc.
Multi-agent era: As the number of on-chain agents increases, interactions between agents will become a key growth area.
In spite of the joy, consider calmly
Hallucination problem: AI models still have the problem of producing incorrect, misleading, or meaningless information.
Blockchain infrastructure challenges:
Scalability: Existing major L1s may not be sufficient to support frequent transactions of millions of AI agents.
Cross-chain compatibility: The crypto world is still relatively fragmented and lacks universal composability.
Tools and Infrastructure: Existing blockchain infrastructure is primarily designed for human users and needs to be adapted to AI agents.
Still Early: The AI agent is currently closer to a demo than a final product. Significant work is needed to scale up to a fully autonomous agent with real-world cryptographic expertise.
Challenges from Web2 itself: The lack of standardization in the Web2 ecosystem may lead to information fragmentation and increase the difficulty of AI agents’ work.
Conclusion:
The AI agent Meta concept is still in its early stages and a lot of development is expected in the coming months and years.
While some early projects may not seem particularly groundbreaking, they may spark a wave of innovation and experimentation that defines the entire cycle.
It’s clear that this process has begun, and it’s particularly encouraging to see the crossover between the AI and crypto communities growing. The next few months will be very interesting, and we look forward to seeing how this emerging subfield develops.
Finally, as a16z partner Chris Dixon said in a blog post over 10 years ago:
"Any great thing often looks like a plaything when it first appears."
