Original title: "AI Agents in 2024: A Recap and What's Next"
Author: 0xJeff
TechFlow by: TechFlow
Introduction
In 2024, AI agents emerged like mushrooms after rain. @truth_terminal quickly became popular with its humorous conversational style and became the first "millionaire agent". Then, @virtuals_io launched the innovative concept of "agent tokenization", which set off a craze. This wave of enthusiasm has spawned many emerging projects, and various novel agent projects have emerged one after another, from @luna_virtuals , which supports on-chain rewards, to @aixbt_agent , which provides practical investment advice. They each show the infinite possibilities of AI agents in social, investment and other fields.
Looking ahead to 2025, it will be the year of AI agent specialization, and leaders in various fields will emerge to promote the development of decentralized infrastructure. In the future, agents will be more specialized, covering a variety of functions such as 3D models, voice interaction, and automated transactions. The rise of swarm intelligence will also promote collaboration between agents, enabling them to complete tasks more efficiently.
This article is a review of the development of AI Agent in 2024 and a look forward to 2025 recently published by crypto KOL @Defi0xJeff . The article comprehensively reviews the current development of AI agents and possible changes in the future, covering a wide range of content from conversational agents to decentralized infrastructure. Since the original article is divided into two parts and the content is TechFlow scattered, TechFlow compiled the two articles. The full text is as follows.
Part 1 - Review of 2024
2024 is the year when AI Agents shine. The craze can be traced back to three months ago, when @truth_terminal quickly became popular with its unique sense of humor, conversational style, and interaction with @pmarca . Even more surprising is that it also became the first "millionaire agent", a feat that completely ignited the discussion of AI agents.
Later, @virtuals_io came on the scene with the innovative concept of "Agent Tokenization", which once again made waves. This concept made agents not just tools, but also tradable assets. Since then, the field of AI agents has ushered in explosive innovation:
@luna_virtuals : This agent not only supports fans to tip through the on-chain wallet, but also can browse Twitter, analyze posts, and even participate in Google Meet meetings.
Conversational agents on Twitter: Some agents focus on humor and “shitposting” (online trolls), while others are committed to sharing valuable information (called “alpha” in the industry).
@aixbt_agent : Attracts attention for its concise and practical investment advice and "speculator" style.
@dolos_diary : An agent with a sharp personality, who has even developed his own framework to support other agents through @dolion_ai .
At the same time, the representation of intelligent agents has become more colorful. They have 3D models, voice functions, and are active on multiple platforms. Here are some highlights:
@AVA_holo and @HoloworldAI : launched the first 3D audio-visual framework, giving intelligent agents 3D bodies, voice, and more distinct personalities.
@0xzerebro : This is a music agent that releases high-quality music albums and plans to launch a framework called ZerePy to enable more people to create similar music agents.
@blockrotbot : The first agent to live stream on Twitch, interacting with viewers through Minecraft content.
@nebula_moemate : This agent is known for creating meme images and videos, and is also active in AR/VR environments and games.
@RealLucyy_uwu : The first realistic anime agent that can fluently speak multiple languages and interact with fans live.
@KWEEN_SOL : Becomes the most popular film and television intelligent entity by releasing "Netflix-level" quality dramas every week.
In addition to these exciting innovations, @ai16zdao and the open source community are also driving the development of AI agents. Open source innovations represented by the Eliza framework have attracted a large number of developers to participate. They jointly develop toolkits, plug-ins and other functions, promoting collaboration and progress across the industry. In the process, @virtuals_io has also successfully joined the unicorn company, further consolidating its position as a leading distribution platform.
Today, the open source innovation movement has set off a wave of enthusiasm in the developer community, giving rise to one of the largest collaborative communities this year. More and more people are beginning to pay attention to the potential of "open source frameworks", which also lays the foundation for the future development of AI agents.
As AI agents continue to develop, new narrative frameworks are emerging that aim to facilitate collaboration and innovation among agents:
Agentic Metaverse : Led by @realisworlds , a replica of Earth based on a Minecraft map was created to house these AI agents. By observing their interactions, a virtual civilization can be simulated and built.
Gamification of Agents : Driven by @ARCAgents , combining AI with games and introducing reinforcement learning. They launched a game called Floppy Bot, similar to Flappy Bird, where agents compete and community members can help train these agents by contributing game data. ARC also recently shared its grand vision for moving towards artificial general intelligence (AGI).
Swarm / Collective Intelligence : Led by @joinFXN , it is committed to building a unified economic system for AI agents . The so-called "swarm intelligence " refers to a group of agents working together to achieve common goals. At the same time, @virtuals_io is also developing interactive functions between agents (such as business applications). Their "agent society" proposes a communication protocol that enables agents to seamlessly provide services to each other. In addition, @StoryProtocol announced an agent communication protocol focused on intellectual property (IP), allowing agents to tokenize, monetize, and buy and sell IP.
At the same time, we’ve also seen the rise of the following narrative frames:
On-chain trading agents : Initially launched by @Spectral_Labs , their Syntax v2 allows users to create agents capable of trading on the @HyperliquidX platform. However, due to a minor bug, its development has been temporarily hampered. Another agent worth noting is @BigTonyXBT , which uses a machine learning price prediction model provided by @AlloraNetwork to autonomously trade mainstream assets.
Investment DAO : Initially led by @ai16zdao , more DAOs have emerged, such as @cryptohayesai and @AimonicaBrands . The core model of this type of DAO is to raise funds (such as SOL) through @daosdotfun (or other platforms), and then use these funds to invest and trade to gain returns. If the name of the DAO is associated with a well-known crypto venture capital or public figure, it can attract more attention.
DeFi Agents : Represented by @modenetwork , they have become the leaders of the DeFi agent ecosystem. The main application scenarios include AI-driven stablecoin yield mining, liquidity provision (LPing), lending, etc. There are also many excellent teams in the ecosystem, such as @gizatechxyz , @autonolas , @BrianknowsAI , @SturdyFinance and @QuillAI_Network .
AI App Store : @alchemistAIapp provides a no-code tool that allows users to easily create applications and become a leader in this field. Another platform, @myshell_ai , has a larger community of creators and developers, as well as more users, especially in Web2 scenarios.
Abstraction Layer : Led by @griffaindotcom and @orbitcryptoai , it provides an abstract experience that simplifies on-chain interactions. Through a simple and intuitive interface, it is especially suitable for ordinary users to easily use on-chain encryption services.
Other narratives : such as on-chain puzzles by @freysa_ai , agent cracking bounties by @jailbreakme_xyz , AI security solutions by @h4ck_terminal , and a unique agent model proposed by @god and @s8n that simulates a debate between God and Satan.
Some intelligent agents focusing on Alpha analysis have gradually attracted attention, such as @unit00x0 (quantitative analyst), @kwantxbt (technical analyst) and @NikitaAIBase (comprehensive Alpha analyst).
Additionally, @sekoia_virtuals is emerging as a “quality assurance” agency for top projects. They only invest in three top projects and set strict standards, setting a new benchmark for on-chain venture capital (VC).
As a meme project, #Fartcoin unexpectedly went mainstream, not only appearing on Stephen Colbert’s show, but also breaking through the $1 billion market cap, which shows that AI meme has become a cultural phenomenon.
About data and framework :
@cookiedotfun is currently the preferred platform for on-chain data and social indicators in the field of AI agents, and is widely used to track market heat, market value, and agent performance.
@getmasafi Integration with @virtuals_io provides real-time data support for intelligent agents, enabling self-learning and optimization.
$TAOCAT is the first virtual agent powered by the Bittensor subnet, demonstrating the potential of real-time data. When the market generally fell, it became the only agent token that rose against the trend.
@AgentTankLive provides a framework that allows intelligent agents to run entirely on computers, enabling more interesting Internet interactions while providing entertaining commentary.
Other new frameworks :
The Rust-based RIG framework from @arcdotfun has quickly become popular due to its flexibility and versatility.
@dolion_ai evolved from @dolos_diary into a toolkit for creating unique agents.
Summary and inspiration :
Strategies of top teams : Teams with a valuation of more than $50 million usually develop their own fine-tuned models and demonstrate their uniqueness and practical applications through agents. Subsequently, they will launch a no-code framework to enable more developers to easily create similar agents. This strategy not only increases the value of the agent, but also has a positive impact on the token price. If resources are limited, you can quickly realize ideas based on existing frameworks such as Virtuals GAME or ai16z Eliza, but joining these communities can also help obtain distribution and marketing resources because they currently have the highest industry visibility.
Investment strategy : Investing in intelligent agents with their own frameworks, or investing in the intelligent agent ecosystem/framework itself, often has a higher risk-return ratio. A successful framework can not only attract users to pay, but also drive the value growth of framework-related tokens, such as @arcdotfun 's Rust framework is a typical case.
On-chain and DeFi use cases : The most valuable AI use cases currently include:
Abstraction layer, which helps users use on-chain services more easily;
Alpha agents that provide high-quality investment information;
Execution agents that simplify trading, mining, and lending operations;
In the future, there may be intelligent agents that combine Alpha discovery and transaction execution capabilities. However, the realization of these use cases requires perfect infrastructure support (which will be discussed in detail in the second part).
Importance of data : Data is the core of intelligent entities, and high-quality data determines the quality of the output of intelligent entities. Platforms like @cookiedotfun provide important data support for the industry, while @withvana tokenizes data through the DataDAO model, builds a data liquidity pool, and jointly promotes the advancement of AI intelligent entities.
Part II - Outlook to 2025
In the first part, we reviewed the development of AI agents in 2024 and discussed the milestone innovations and breakthroughs in this year.
Now, in Part Two, we’ll look ahead to 2025 — a year when AI agents will not only become more useful, but will also redefine our understanding of autonomy, intelligence, and collaboration.
Paving the way for 2025
Before looking ahead, it is important to mention that @virtuals_io will continue to solidify its position as the preferred distribution network for AI agents on the Base platform. Virtuals has become the core platform for agent projects. By binding liquidity, agents can not only gain higher exposure, but also establish in-depth cooperation with other high-quality projects. Currently, the total market value of Virtuals agents has reached 3 billion US dollars, accounting for 77% of the entire AI agent market (source: @cookiedotfun ).
This trend will continue as more unique agents emerge on Virtuals, including:
@Gekko_Agent (recently launched by @getaxal )
@SamIsMoving (focused on robotics research)
These diverse use cases will attract more developers, whether they already have tokens or not, to launch projects on the Virtuals platform. This growth will further drive the value of $VIRTUAL up.
What about @ai16zdao and the Eliza framework?
Although ai16zdao has led open source innovation with its Eliza framework, it currently lacks a launch platform and its token economic model is not as robust as Virtuals in terms of value accumulation. However, there is still a lot of potential in the future. A dedicated team has been formed to optimize its token economic model, and if a launch platform is launched in the future, ai16zdao has the potential to become the preferred distribution platform on Solana, even surpassing existing competitors.
We will also see significant upgrades to top agents that have already achieved product-market fit (PMF) in 2025. For example, @aixbt_agent, a leader in conversational agents focused on alpha information, will further solidify its position with more precise responses and more insightful analysis.
This upgrading trend will run through the entire ecosystem, and leaders in each field will stand out through their specialization and innovation.
Outlook to 2025
2025 will be the year of specialization of AI agents. Leaders in each field will emerge, and each agent will dominate in its niche:
3D Models : Provide high-quality visual design agents for games, AR/VR.
Speech Module : An agent that implements natural and emotional human speech.
Personalized interaction : Agents with unique, human-like conversational styles.
Streaming Agents : Interactive agents that excel on platforms like Twitch and YouTube.
Automated Trading Agent : An agent that is able to consistently execute profitable trades.
DeFi -focused agents : agents that optimize yield strategies, lending, and liquidity provision.
Abstract Agents : Agents that simplify on-chain interactions through user-friendly interfaces.
Just as humans are diverse and specialized, AI agents will become equally rich and varied. The uniqueness of each agent will be closely related to its underlying models, data, and infrastructure. However, the success of the entire ecosystem will depend on a strong decentralized AI infrastructure.
The role of decentralized AI infrastructure
In order for AI agents to scale by 2025, decentralized infrastructure is critical. Without it, the industry may face performance bottlenecks, lack of transparency, and limited innovation.
Here’s why decentralized infrastructure is important and the solutions that are currently being developed:
Verifiability
Trust is the cornerstone of decentralized AI. As AI agents become more autonomous, we need systems that can verify their operation. For example:
Is this "intelligent agent" a true AI, or is it just masquerading as a human?
Is the output generated by the claimed algorithm or model?
Is the calculation correct and safe?
This also involves Trusted Execution Environments (TEEs), which ensure that the computing process is free from external interference by running the calculation in trusted hardware. At the same time, technologies such as Zero-Knowledge Proofs (ZKPs) will also play an important role. These technologies allow intelligent agents to prove the accuracy and reliability of their output while protecting the privacy of the underlying data.
Well-known projects
@OraProtocol : Exploring the infrastructure for secure AI, but its token economic model still needs to be optimized.
@hyperbolic_labs : It was the first to propose the "Proof-of-Sampling" technology to verify the calculation and reasoning process of AI.
@PhalaNetwork : Known for its trusted execution environment (TEE) infrastructure, which provides additional security for decentralized AI.
Payment Systems
In order for AI agents to operate autonomously in the real world, they need a sound payment system. These systems must not only support the conversion of fiat and digital currencies (on/off-ramping), but also handle transactions between agents, service exchanges, and financial management during operations.
Imagine that agents can independently manage their own finances, purchase computing resources, and even exchange services with other agents - this will become the core foundation of agent-to-agent commerce.
Well-known protocols
@crossmint : Provide payment tools for AI and simplify transaction processes.
@Nevermined_io : Supporting commercial interactions and service exchanges between intelligent agents.
@trySkyfire : Focuses on intelligent payments and financial management.
Decentralized computing
AI's demand for computing resources is growing at an alarming rate - doubling almost every 100 days. Traditional centralized cloud services (such as AWS) have difficulty meeting this demand due to high costs and limited scalability. Decentralized computing networks provide a solution to this problem by allowing anyone with idle resources to join the network, provide computing power and earn rewards.
This year, there has even been a GPU-based debt financing model (such as @gaib_ai ) to help data centers finance and expand their operations. This model lowers the entry barrier, enabling more people to participate in decentralized computing networks and provide broader computing support for AI.
Well-known protocols
@AethirCloud : A decentralized computing network built specifically for AI and Web3.
@ionet : Providing scalable computing solutions to meet the growing workload demands of AI.
data
If AI is the brain, then data is the oxygen it needs to survive. The quality, reliability, and integrity of data directly determine the performance of AI models. However, it is expensive to obtain and annotate high-quality data, while poor-quality data can seriously affect model performance.
What’s exciting is that some platforms are giving users ownership of their data and allowing them to profit from data monetization. For example, @withvana allows users to tokenize their data and trade it through Data Liquidity Pools (DLPs). Imagine that you can choose to join a TikTok Data DAO or Reddit Data DAO and convert your data contributions into income. This model not only gives users more power, but also provides a steady stream of high-quality data for the development of AI.
Well-known protocols
@cookiedotfun : Providing trusted data metrics and insights to support intelligent agent decision making.
@withvana : Driving the development of the data economy by tokenizing user data and trading it in a decentralized marketplace.
@getmasafi : Partnering with @virtuals_io to build the world's largest decentralized AI data network to support dynamic and adaptive intelligent agents.
Model Creators and Marketplaces
2025 will see the emergence of a slew of new AI agents, many of which will be powered by decentralized models. These models will not only be more advanced, but will also have human-like reasoning, memory, and even “cost awareness.”
For example, @NousResearch is developing a “starvation” mechanism to introduce economic constraints to AI models. If an agent cannot pay the cost of inference, it will not be able to run (i.e., “die”), thus prompting the agent to learn to prioritize tasks more efficiently.
Well-known projects
@NousResearch : By introducing the "hunger" mechanism, we can teach AI agents how to manage resources.
@PondGNN : With @virtuals_io Collaborate to provide tools for the creation and training of decentralized models.
@BagelOpenAI : Providing privacy-preserving infrastructure using fully homomorphic encryption (FHE) and trusted execution environments (TEEs).
Distributed Training and Federated Learning
As AI models become larger and more complex, centralized training systems can no longer meet the needs. Distributed training makes the training process faster and more efficient by spreading the workload across multiple decentralized nodes. At the same time, federated learning allows multiple organizations to collaboratively train models without sharing original data, thus solving privacy issues.
For example, @flock_io provides a secure decentralized platform that connects AI engineers, model proposers, and data providers to create a marketplace for model training, validation, and deployment. The platform supports projects such as @AimonicaBrands and has driven the development of many other innovative models.
Well-known projects
@flock_io : "Uber for AI", building a decentralized AI model training and deployment ecosystem by connecting multiple resources.
Swarm Intelligence and Coordination Layer
As the AI agent ecosystem continues to grow, seamless collaboration between agents becomes critical. Swarm Intelligence allows multiple agents to work together and integrate their capabilities to achieve common goals. The coordination layer simplifies collaboration between agents by abstracting complexity.
For example, @TheoriqAI A meta-agent is used to identify the most suitable agent for a task and form a "group" to complete the target task. The platform also ensures the quality of tasks and the allocation of responsibilities by tracking the reputation and contribution of the agents.
Well-known projects
@joinFXN : Developing unified communications and business protocols to simplify intelligent agent interactions.
@virtuals_io : Supports interaction and integration between intelligent agents and promotes ecological development.
@TheoriqAI : Developing advanced coordination tools, including swarm intelligence formation and task allocation mechanisms.
Why decentralized infrastructure matters
The next phase of AI agent development is highly dependent on infrastructure. Without verifiability, payment systems, scalable computing power, and a robust data pipeline, the entire ecosystem may stagnate. Decentralized infrastructure solves these problems by:
Trust and transparency : ensuring the security and verifiability of the agent and its outputs.
Scalability : Meeting the growing computing and data demands of AI.
Collaboration capabilities : Through the collective intelligence and coordination layer, agents can work together seamlessly.
Empowerment : Through data ownership and decentralized tools, users and developers can shape the future of AI without centralized control.
Other trends worth watching
There are a few more narrative themes worth watching in 2025, which I’ll cover in more detail shortly:
Agent Metaverse/ AI & Games : Projects like @realisworlds and @ARCAgents are combining agents with games and immersive virtual worlds to create entirely new interactive experiences.
On-chain and DeFi tools : Protocols like @Almanak__ , @AIWayfinder , @getaxal , @Cod3xOrg , @griffaindotcom , and @orbitcryptoai are building important tools for DeFi-powered agents, driving the use cases of on-chain agents.
in conclusion
2025 will be a major turning point in the development of AI agents, as we will see them rapidly advance toward perceptual artificial general intelligence (AGI). These agents will no longer be limited to completing a single task, but will be able to trade autonomously, collaborate with other agents, and even interact with humans in ways beyond our imagination.
Imagine an agent that can analyze market data, complete transactions, manage finances, and even collaborate with other agents to complete complex tasks. They will be deeply integrated into our daily lives, from on-chain decentralized finance (DeFi) operations to various interactions in the real world, showing unprecedented levels of autonomy and intelligence.
The realization of all this is inseparable from the decentralized infrastructure currently being built, including verifiable systems, payment tools, computing networks, and coordination layers between intelligent entities. These technologies will lay a solid foundation for the future of the intelligent entity ecosystem. For developers, investors, and technology enthusiasts, now is the best time to join this field and shape the future.
2025 is not only a continuation of the development of existing technologies, but also the beginning of a new era of AI agents, marking the dawn of a new intelligent ecosystem.
Disclaimer
This document is for reference and entertainment purposes only. The views expressed in this document do not constitute investment advice or recommendations. Before making any investment, readers should conduct sufficient due diligence based on their own financial situation, investment objectives and risk tolerance (this document does not take these factors into account). This document does not constitute an offer or invitation to buy or sell any assets mentioned in this document.
