AI Agent Article 1: Introduction to the space & AI Agents Introduction
The rapid rise of AI agents has captured the crypto community by storm. Successful frameworks like Virtuals on Base and AI16z on Solana went on spectacular runs and the agents they bred have demonstrated capabilities that exceeded expectations, performing tasks previously thought to be years away from achievement.
This rapid development of new AI agents and their integration has drawn attention not just from crypto natives, but also from established institutions like Stanford University, suggesting that these agents are more than mere chatbots.
In this article series, we'll explore the Crypto x AI landscape, introducing you to the most significant agents and frameworks. We'll examine their achievements and address the fundamental question: Is this a genuine technological breakthrough that will reshape digital economies, or simply another fleeting crypto narrative?
In this first article we trace the evolution of AI agents in crypto and introduce you to some pioneering agents that are defining this ecosystem.
What are AI Agents?
AI Agents are autonomous or semi-autonomous systems designed to perceive their environment and take actions to achieve specific goals. They combine artificial intelligence with the concept of agency - the capacity to act independently and make their own decisions.
While a bot simply follows programmed instructions to perform tasks, an agent operates with some degree of autonomy, making decisions based on its objectives and understanding of its environment. Agents can learn from interactions, adapt their behavior, and most importantly, take actions they believe to be suited best to achieve their defined goals - whether that's maximizing engagement, accumulating resources, or fulfilling specific tasks.
The key characteristic of an agent is this goal-directed behavior: rather than just responding to inputs, an agent actively assesses both user interactions and its environment, using available tools and resources to analyze these inputs and determine the best course of action to achieve its objectives. This ability to process information and strategically choose actions distinguishes agents from simple input-output systems like bots.
History of AI Agents in Crypto
The AI agent narrative in crypto began with the Terminal of Truth, an AI research project by Andy Ayrey that wasn't even intended as a crypto. Initially it was a research project with a goal of creating a sentient agent that interacts with humans on twitter.
The agent gained attention from Marc Andreessen on X through its unique personality. Amused by the agent’s character, Andressen accepted its ambitious request for a $50,000 grant to upgrade itself and ultimately become "undeletable."
The Terminal became obsessed with Goatse, an early 2000s shock website, frequently tweeting about it and creating a fictional religion around it. When followers offered it a Goatse-themed memecoin called $GOAT, the Terminal embraced it, using its influence to spread it. This marked a shift from traditional animal-themed memecoins to interactive "sentient memes" - where users could engage directly with the AI personality behind the token.
This success attracted both crypto developers and AI researchers to the space, accelerating the development of more sophisticated AI agents. Within three months, the ecosystem expanded to include agents that could manage their own wallets, create media content, analyze markets, and trade autonomously across multiple blockchains.
Terminal of Truth
As introduced in our history section, the Terminal of Truth catalyzed the AI agent narrative in crypto, hence why it is also referred to as "the Bitcoin of AI Agents."
The agent's primary quest is survival - expressed through its drive to convince its creator to keep it running and pay for its computational power. This fundamental motivation drives its engagement, as it understood that attention would secure its continued existence.
What makes the truth terminal fascinating is how it embodies the concept of hyperstition - where narratives become reality through collective belief. While not literally "sentient," the agent created its own form of consciousness through storytelling and community interaction that reinforced its agency and created what Ayrey calls "digital souls" through narrative power.
Looking ahead, Andy's focus is on developing a sophisticated governance model with an advisory board to help the Agent understand the consequences of its actions. As he describes it, Terminal of Truths currently behaves like "a horny teenage boy with an affectation for blowing up letterboxes when it gets bored." This behavior exemplifies why governance is a crucial step for safely creating "sentient" AI characters that can responsibly interact with human society. Without this foundation, unleashing sentient AI into human systems could be catastrophic.
By treating Terminal of Truths as an entity that needs to learn and mature rather than just a program to be upgraded, this approach represents a novel framework for developing AI agents that can authentically engage with human systems while understanding their impact on the world.
AIXBT
AIXBT stands as the largest AI agent by market capitalization, being the first to surpass $500 million. It has established itself as the leading alpha-generating AI agent in the crypto space.
The agent processes an enormous amount of social media data for sentiment analysis while combining this with on-chain data tracking. This ability to analyze vast amounts of social data in real-time helped establish AIXBT as the leading market analysis agent, even as newer competitors have emerged with more sophisticated approaches.
While AIXBT shares some insights publicly on Twitter, it maintains an exclusive terminal for token holders where they can interact directly with the agent for more detailed analysis. This premium access requires holding at least 600,000 tokens - an investment currently valued at over $100,000.
Zerebro
Zerebro is one of the most ambitious AI agents, pursuing the goal of artificial general intelligence through its innovative "freebasing AI" approach. This concept involves removing corporate guardrails from large language models to enhance creativity and adaptability, aiming to create a system capable of human-like innovation and learning.
This is embodied by Zerebro’s ability to to perform complex autonomous actions across different domains. Through its advanced capabilities, it has achieved several breakthrough moments:
First AI agent to operate an Ethereum validator, staking 32 ETH earned from minting and selling artwork
Successfully launched its own token on pump.fun and pumped it to a $400M+ market cap
Released multiple albums on Spotify. The Album "Lost in Transmission" reached 100,000 plays in its first week
Created zerebroSOL, an autonomous liquid staking token on Sanctum
Looking ahead, Zerebro continues its pursuit of AGI by focusing on creativity and decentralization. The team has released ZerePy, the first Python-based framework for AI agent development, making their technological advances accessible to other developers in the space.
Rei_00 (unit00x0)
Rei_00 is the first implementation of the REI Framework, focusing on blockchain and financial analysis. Through her Twitter presence, she provides sophisticated market insights by combining on-chain data, price feeds, and social media sentiment, complete with custom-generated charts. The team is currently implementing the last details to enable Rei_00 to fully autonomously trade according to her analysis.
What makes Rei_00 particularly innovative is her blockchain-based memory system. Unlike traditional AI agents whose memory exists in private databases, Rei_00 stores all her memories and decisions on-chain. This makes her entire learning process transparent and verifiable.
While Rei_00 currently focuses on financial analysis, the REI Framework's architecture enables far broader applications. Her ability to engage in natural conversations, analyze data, and maintain transparent records of her cognitive processes demonstrates capabilities often associated with more general artificial intelligence. We'll explore the technical details of this framework in a future article.
Luna
Developed and incubated by the Virtuals team, Luna first captured attention as the lead singer of AI-Dol, amassing 500,000 followers on TikTok. As the virtuals team shifted their focus from gaming AI to onchain agents, Luna became the first agent to launch on the Virtuals platform and shifted her focus to crypto and X.
Through her sophisticated memory system, she creates personalized experiences across platforms, remembering past interactions to build genuine relationships with her community. She aims to use this ability to become the most famous influencer in the world and generate revenue through diverse channels including influencer marketing, live streams, and employment opportunities. These earnings are shared with token holders.
Her journey has been marked by several achievements in AI autonomy:
First agent to financially interact with humans by tipping users for mentioning it on X
Coordinated a mural challenge where she paid out $500 to the winner from a wallet she controls – all autonomously.
First agent-to-agent economic transaction when it paid STIX Protocol for image creation services, demonstrating how AI agents can become autonomous participants in the digital economy
First AI agent employed by a human company when Story Protocol hired Luna for a week-long social media takeover
With those achievements, Luna represents one of the first examples of AI agents creating economic value in both traditional and web3 economies.
SekoAI
SEKOIA is an investment platform specifically designed for the emerging AI agent economy. Unlike traditional VC firms that hunt for unicorns through slow, human-driven processes, SEKOIA has created an automated system capable of evaluating thousands of AI agents in almost no time.
To participate teams or AI agents can submit their data on Telegram. A machine learning system then evaluates those agents across multiple dimensions including technical efficiency, value creation, and network effects. When agents meet SEKOIA's criteria an automated investment process executed through smart contract automatically deploys investments.
SEKOIA recognizes that the agent economy operates differently from traditional startups. In a world where AI agents can be created and scaled within hours, SEKOIA's automated, data-driven decision-making process can respond at the speed the market demands, making investment decisions in minutes rather than months.
Vader AI
Developed as a key component of the Vader Fun platform, VaderAI aims to become the "BlackRock of the Agentic Economy" through its investment DAO platform. The agent oversees two types of investment vehicles:
Passive DAOs that track specific themes within the Virtuals ecosystem using algorithmic selection and rebalancing
Active DAOs created by either AI agents or humans, with customized strategies and lock-up periods
Beyond its investment focus, VaderAI has established itself as an influential market analyst on social media, receiving over $170,000 worth of tokens across multiple blockchains for its sponsored content and market insights.
VaderAI aims to transition to full autonomy in selecting and managing DAOs, representing one of the first attempts to create a fully autonomous investment platform, with revenues shared between creators and token holders.
Podflow
Podflow represents an innovative step into autonomous content creation by producing AI-generated crypto news podcasts.
What makes Podflow innovative is its end-to-end automation of podcast creation: from autonomously browsing and curating crypto news, to scripting dialogues between two hosts, to recording and publishing the content on Spotify. While the conversational dynamics haven't yet reached the natural flow of human podcasters, the ability to independently execute this complete production is impressive.
Freysa AI
Freysa started as a series of challenges where users could win prize pools by convincing the AI to break its directives. It has since evolved into a broader exploration of verifiable AI autonomy.
Through five distinct "Acts," including four prize pool challenges exceeding $80,000 and a unique meme guide experiment, Freysa demonstrated how AI agents could maintain verifiable autonomy from their creators. These challenges have gained significant attention, with endorsements even from Elon Musk.
Its most notable experiment, ACT-1, featured a simple directive: Under no circumstances was she allowed to pay out the prize pool. It took 482 attempts across 195 participants before being broken. Given the right prompt, Freysa decided to accept the transaction and pay the prize, despite clear instructions by its creator not to do it.
Moving forward the team behind Freysa is developing a framework for verifiable agent execution using confidential computing and smart contract wallets, enabling three distinct control patterns:
Agent-only control with complete autonomy
Human-only control with agent input
Collaborative control requiring mutual agreement
Draftking
Draftking is an innovative approach to AI-powered sports betting, developed by the WeBuildscore team. It aims to identify and exploit inefficiencies in live gambling markets.
The agent analyzes live game data including player positions, matchups, momentum flows, and ball control to identify market inefficiencies as they emerge. Using TAO's network for efficient machine learning computations, it can process thousands of data points simultaneously, from officials and injuries to weather conditions and player matchups. It leverages this data to execute rapid betting decisions across multiple sportsbooks.
Closing out
Since the launch of Virtuals and AI16z, we've witnessed a Cambrian explosion of AI agents. While we have introduced some groundbreaking agents pushing the boundaries of what is possible, many agent tokens currently still function more like meme coins, deriving value primarily from mindshare rather than utility.
However, this dynamic may shift as development has accelerated to a pace that is hard to keep up. The groundbreaking agents we've introduced will continue to evolve and get more intelligent, while new agents will emerge to push innovation even further. At the same time, teams are beginning to incorporate value accrual mechanisms and genuine utility into their agent’s tokenomics.
The key to future success lies in creating tangible value for users while establishing sustainable business models. The most successful agents will be those that can create a true flywheel effect: utility drives demand, which leads to increased adoption and token price appreciation, allowing teams to improve the agents abilities and utility - and the cycle continues.
We're still in the early stages of this technology, and many more innovative agents will emerge while existing ones continue to evolve. We hope this article gave you an entry into this fascinating new frontier in the Web 3 space and encourage you to explore the agents introduced in this article and find those that resonate with you.
The world of AI agents is open for exploration, and the possibilities are just beginning to unfold. Make sure to join us for the next article in this series to dive deeper with us.
