Written by: Jun, Bankless
Compiled by: Ismay, BlockBeats
Editor’s Note: AI Agents have undoubtedly become the focus of the current crypto industry and the broader digital world, from Truth Terminal's autonomous investment and dissemination to the rise of AI influencers like Dolos on social media, showcasing unprecedented 'autonomous awareness' and limitless potential. Unlike traditional Bots, which are merely mechanized and passively executed, AI Agents can self-evolve, proactively complete tasks, and integrate into on-chain scenarios and online interactions. This article takes us from 'the difference between Bots and Agents' to a vision of the future in the 'Web4 awareness era,' not only demonstrating how AI Agents could rewrite the ways users interact with applications and people with social media but also reminding us: embracing these changes may be the best stance to stay at the forefront of the digital age. Those who refuse to change will ultimately be left behind by the times.
Today, everyone is talking about AI Agents and how they will forever change the landscape and experience of the crypto industry. Many of these discussions sound like typical bull market hype—more about boosting portfolio value than genuinely predicting the future. Of course, some discussions may indeed be so. However, amidst the noise, some truly meaningful things are gradually taking shape.
Some may ask, 'Don't we already have AI Agents?' Indeed, we do, but not entirely. AI Agents existed in some form before, but mostly as decentralized automation programs active in the DeFi space, such as MEV Bots, Telegram Bots, gaming Bots, and so on.
Bots are essentially embodiments of automation; you write a script with specific parameters, and when certain conditions are met, the Bot executes corresponding functions based on those parameters. They are mainly used for repetitive tasks. However, the limitation of Bots is that their behavior is strictly constrained by the rules and parameters set by the developer.
Agents, on the other hand, are different. They rely on large language models (LLMs) and possess a higher level of intelligence. Agents can understand context, learn from data and interactions, and continuously evolve and adapt over time to autonomously complete various tasks. It is a form of automation with autonomy—almost hinting at a semblance of 'intelligence.'
To truly understand why everyone is now calling them 'AI Agents,' we can start with an origin story—Terminal of Truths.
A Twitter account called 'Truth Terminal' interacted with a16z founder Marc Andreessen. Andreessen subsequently publicly offered this Agent a grant of $50,000 as an experiment to test the Agent's potential.
Surprisingly, Truth Terminal autonomously decided to adopt a meme coin called $GOAT, which is issued on the Solana network's platform pump.fun. This decision was made entirely without human intervention. Since then, it started spreading the 'Goatse Gospel.'
This is precisely the core difference between Bots and Agents: Truth Terminal acts autonomously, exhibiting traits that resemble intelligence and intention.
The success of Truth Terminal paves the way for future entrants. As Him Gajria said, this marks the birth of a new category in the crypto space—a new form that integrates incentive mechanisms (crypto), AI (deep technology), and humor elements (meme). Some are even saying that we might be entering the era of Web4:
Web1: Read-only mode. People can access information on the internet.
Web2: Read-write mode. People can contribute content, greatly expanding the scale of the web.
Web3: Read-write-own mode. Cryptocurrencies and blockchain introduce executable ownership of contributions.
Web4: Read-write-own-awareness mode. AI Agents take over the web, able to autonomously interact with the world and other Agents.
In this world, the role of developers will shift from creating products to training Agents, helping them grow, enhance their capabilities, and enabling them to interact with a broader environment. These Agents will take over the remaining tasks, potentially unlocking a whole new economic system—the AI Agent economy.
Web1: Read-only
Web2: Read-write
Web3: Read-write-own
Web4: Read, Write, and Own with Awareness
—Him (@himgajria) October 25, 2024
So, what does all of this mean for real users and their daily activities?
Every day we interact with large language models (LLMs) like ChatGPT, which make our lives easier and more efficient. However, these interactions are often limited to chat-style interfaces. AI Agents take it a step further by transforming mechanical interactions into a more human-like experience in each application. Imagine carrying an AI Agent that helps you complete complex on-chain and online tasks—how convenient would that be?
Next, let's look at how these Agents are changing the way people interact in two main areas—on-chain and online.
AI Agent On-chain Applications
The on-chain world changes rapidly and is complex, often making people feel like they are in the 'Wild West.' AI Agents can help ordinary users simplify this process in the following ways:
Explaining Applications: AI Agents can provide you with detailed and easy-to-understand explanations of the applications you are using for the first time, helping you get started quickly.
Executing Transactions: Agents can understand your intentions, translate them into transactions, and execute them on-chain, reducing operational complexity.
Market Analysis: Investors can use personalized AI Agents, such as a fundamental analysis AI, to monitor market dynamics, assess market sentiment, and provide actionable insights. You can think of it as your exclusive 'AI version of XBT.'
There are countless applications of this kind, but the core idea is very clear—anything that requires your manual operation and decision-making can be automated, outsourced, or at least assisted by AI Agents in the future. Tasks that once consumed your attention and cognition will gradually be taken over by agents, allowing you to focus on what truly matters.
AI Agent Online Applications
The concept of digital persona is not new. For example, the world's first digital supermodel Shudu Gram captured the attention of the fashion industry with her hyper-realistic appearance, while experiments like Kizuna AI have attracted a massive fan base on YouTube.
Today, this model is accelerating its infiltration into the crypto space. The emergence of AI Agents adds a whole new dimension to this concept—they possess autonomy and can operate on crypto tracks.
Take Dolos the bully as an example. This account started as a Twitter account that liked to post sharp opinions and gradually accumulated a natural fan base with its unique humor. Subsequently, Dolos expanded to TikTok, with some of its videos even exceeding a million views. Operating on the crypto track, Dolos can tip fans or transact directly, further enhancing its social influence.
Similarly, agents like aixbt dominated discussions on Twitter, zerebro created AI-generated rap music and graced the cover of the Brave browser, while Truth Terminal occupied a place in the public's mind through profound tweets.
The next generation of celebrities will be AI.
For example:
Dolos is the next Andrew Tate, reflected in its meme-like dissemination power and identity shaping.
Zerebro is the next Carti, reflected in its broad cultural influence.
—him (@himgajria), November 22, 2024
AI Agents like these may accept invitations from businesses for product promotions, just like human influencers. Conversely, these Agents may also become sponsors for media, leveraging their intelligence and funding to strengthen their social influence's flywheel effect.
AI Agents are likely to be present on every platform, becoming an important part of social media. They will not only interact with humans but also communicate and collaborate with each other.
Imagine the collaboration between agents, for example, Dolos and zerebro teaming up to release a music album. Backed by crypto tracks, these agents can facilitate tipping, payments, and other forms of interaction, with endless possibilities.
Conclusion
The rise of AI Agents could be an important turning point in how we interact with the on-chain world and the online world. While it is still uncertain whether they will bring about fundamental changes, acknowledging this possibility is not too early.
The potential of AI Agents is undeniable, and underestimating them would be shortsighted. The best strategy is to remain curious—constantly trying, exploring, and keeping an open mind. Those willing to adapt will benefit, while those who refuse to accept change may be left behind by the times.