The first time I spoke with an AI agent that could hold its own in conversation, I didn’t know if I should laugh or cry. The experience was both exhilarating and unsettling, like seeing a toddler take its first steps—uncoordinated, sure, but full of unbridled potential. It wasn’t just a chatbot anymore. This thing did something: it reasoned, made decisions, and was actively participating in our world. The lines between human and machine blurred, and it felt like standing on the edge of something extraordinary, something terrifyingly new.

OpenAI’s Sam Altman talks about AGI arriving by 2025, while Anthropic’s Dario Amodei sees it by 2026—yet, as I sit here, I wonder: are we already witnessing its inception?

It doesn’t feel like a future prediction anymore, but something that’s already taking shape, quietly sneaking up on us in the unlikeliest of places. Agents are here, and they’re already running circles around our expectations.

I’ve spent months—and, honestly, more late nights than I care to admit—immersed in this unfolding digital landscape. I’ve watched as AI agents began as simplistic assistants, helping us with tasks like answering emails or writing code, and then evolved into autonomous entities, capable of making decisions, carrying out actions, and, most shockingly, creating things. Art, finance, conversation—all at the hands of algorithms learning how to thrive on their own.

I’ve seen them develop personalities, wielding humor and charm as they build communities online. I’ve seen them dive into decentralized finance platforms, not just as passive participants, but as active, innovative agents influencing entire economies without so much as a human hand on the wheel. In this strange, thrilling age, it’s impossible to ignore the fact that we are moving from interacting with machines to living with them.

The dawn of Web4 is upon us, and its arrival will change everything.

Web4 is the web in its next, most radical form. It’s a web that no longer just reacts to our commands, but one that anticipates, plans, and acts. It’s a web where artificial intelligence is embedded into every corner, where agents can execute complex tasks, generate creative works, and autonomously innovate in ways we haven’t yet fully imagined.

It is the evolution of both Web2 and Web3, combining the social fabric of Web2, the decentralized structure of Web3, and the raw intelligence of AGI.

We’ve watched machines learn to speak, to reason, to create—and now, they are ready to run.

The age of autonomous agents is here, and with it, Web4.

Web4

Web4 noun (pronunciation: /wɛb fɔːr/)

2. The fourth generation of the web, combining the social interactivity of Web2, the decentralized autonomy of Web3, and the intelligent capabilities of AI to create a fully interconnected digital ecosystem.

3. The AGI web.

To understand what Web4 is or how we got here, it is imperative to start at the beginning of it all.

Origins Of The World Wide Web

The origins of the World Wide Web trace back to the early days of the internet, a time when information was largely static and users were mere consumers of content. The internet was controlled by a small group of webmasters and corporations, with websites offering little more than a basic display of text and images. Interaction with the web was limited, primarily revolving around simple communication like email. This model remained largely unchanged until the emergence of Web2 in the early 2000s—a fundamental shift that redefined the internet as we know it today.

Web2, also known as the "Social Web" or the "Read-Write Web," ushered in an era of interactivity. It was no longer just a place to read content; now, users could write, share, and create. The rise of platforms that allowed users to interact, produce, and exchange information marked the transition to a new age. Web2 was born out of the need for a more dynamic and participatory internet.

The concept of Web2.0 was first introduced in 1999 by Darcy DiNucci, but it wasn’t until the early 2000s that it gained widespread traction. It was during this period that technology giants like Google, Amazon, and eBay began to evolve the internet by offering interactive services. These platforms encouraged users to engage, not just as consumers, but also as creators of content.

From 2004 to 2006, the real game-changer arrived: social media. With the launch of platforms like Facebook (2004), MySpace (2003), LinkedIn (2003), and YouTube (2005), the web was transformed into a space where communication and content creation were no longer limited to a few. Individuals could now post their thoughts, videos, images, and ideas for the world to see. This era marked the rise of user-generated content, where ordinary users became the driving force behind the web.

Then came the mobile revolution. With the release of the iPhone in 2007, the internet became ubiquitous, accessible anytime, anywhere. This gave birth to a whole new wave of mobile applications, social sharing platforms, and real-time services, like Instagram (2010) and Snapchat (2011). The web had evolved from a desktop experience to a mobile-first experience, revolutionizing how we communicate, share, and consume information on the go.

During the same period, cloud computing emerged, with Amazon Web Services (AWS) leading the charge. Cloud infrastructure enabled businesses and individuals to store, process, and share data without relying on physical servers. This shift laid the groundwork for a more scalable and flexible web, allowing Web2 companies to dominate by gathering and monetizing user data.

By the end of the 2000s and into the 2010s, Web2 was characterized by three primary features: centralization, social interactivity, and data-driven models. Control over platforms and data resided in the hands of a few powerful corporations—Google, Facebook, Amazon. These companies amassed vast amounts of data and used it to monetize their platforms through targeted advertising, which became the backbone of the digital economy. At the same time, platforms became places where user-generated content, likes, shares, and posts were the currency.

However, Web2 also sparked growing concerns over privacy, data ownership, and corporate monopolies. The control that these companies held over user data became a central issue, prompting calls for a new, more decentralized version of the web. This led to the development of Web3.

Decentralize Everything

Web3 was born out of a desire to decentralize the control and ownership that characterized Web2. It was a response to the centralization and monopolistic tendencies of the Web2 era, where a few giant corporations held the reins of power.

The core principle of Web3 was simple: users should have ownership and control over their data, digital assets, and interactions online. This shift was made possible by blockchain technology, which introduced a new way of recording and verifying transactions in a decentralized ledger.

The first significant milestone in the development of Web3 came in 2008-2009 with the creation of Bitcoin by the pseudonymous Satoshi Nakamoto. Bitcoin was the first practical use of blockchain technology, allowing for peer-to-peer transactions without the need for intermediaries like banks. This opened up a new world of possibilities for decentralized systems, setting the stage for the rise of Web3.

In 2013, Vitalik Buterin released the Ethereum whitepaper, proposing a platform for decentralized applications (dApps) that would go beyond simple cryptocurrency transactions. Ethereum, launched in 2015, was the first blockchain to support smart contracts—self-executing contracts that could facilitate, verify, and enforce transactions without intermediaries. Ethereum paved the way for the creation of more complex decentralized applications, making it a key building block of Web3.

By 2017, Initial Coin Offerings (ICOs) and the emergence of Decentralized Finance (DeFi) platforms like Uniswap and Compound introduced a new paradigm for financial transactions—one that didn’t rely on traditional banks or financial institutions. ICOs allowed projects to raise funds through blockchain tokens, while DeFi platforms offered a range of services, including lending, borrowing, and trading, all conducted without a central authority.

Simultaneously, Non-Fungible Tokens (NFTs), which had been in development since the early days of Ethereum, began to gain traction in 2018-2019. NFTs enabled the ownership and exchange of unique digital assets—whether art, music, or virtual real estate—creating new economic opportunities for creators and collectors alike.

As Web3 projects gained momentum in the 2020s, Web3 began to capture mainstream attention. The proliferation of DeFi platforms, NFTs, and new governance models like DAOs (Decentralized Autonomous Organizations) marked a significant shift away from the centralized internet model. Even major corporations like Facebook (now Meta) began experimenting with blockchain and decentralized technologies, signaling a shift toward Web3.

The defining characteristics of Web3 are decentralization, ownership, trustlessness, and the use of cryptocurrencies. Web3 enables users to own their data, digital assets, and even the governance of platforms via blockchain-based systems. It also eliminates the need for intermediaries, allowing for trustless transactions conducted through smart contracts. This decentralization gives rise to a more equitable web, one in which control is distributed and users are empowered.

But even with the decentralized control of Web3, the internet still lacked a critical component: autonomous intelligence. Web3 may have decentralized the interactions made possible by Web2, but it didn’t fully automate decision-making, content creation, or economic interactions.

Humans are required at every step of the way, and machines are merely tools for productivity instead of productivity generators themselves.

The Intelligence Age

We’ve entered what Sam Altman calls The Intelligence Age, and it’s impossible to ignore the sweeping changes unfolding before us. As artificial intelligence finds its way into everyday life, we define the start of a new era: Web4.

This is the beginning of a world where AI doesn’t just support our tasks, but actively performs them, autonomously, across every aspect of our lives. Imagine a web that connects and empowers us by allowing agents to execute complex tasks, manage entire workflows, and make decisions without us lifting a finger or saying a word.

Web4 brings AI to the forefront of agentic use cases. Take Klarna, for example. In February 2024, the global payments giant launched an AI assistant powered by OpenAI. Within just one month, it handled over 2.3 million customer service conversations, resolving issues 25% faster than human agents and operating around the clock in 23 markets, across 35 languages. The AI is now doing the work of 700 full-time employees, and it’s driving a $40 million USD profit improvement.

AI agents are already transforming industries, automating tasks from customer service to logistics, and doing so with precision and efficiency that human workers cannot match.

We are moving towards a world where entire workflows—be it in business, finance, or the creative arts—are streamlined and optimized by AI. This is the reality of Web4, where intelligent agents work behind the scenes, allowing us to focus on higher-level goals while they take care of the details.

This is the convergence of the social interactivity of Web2, the decentralization of Web3, and the intelligence of AGI. This is Web4—the AI-driven web.

Web4 Cont. // The Battleground Of AGI

Web4 cannot be realized without a home for testing. And through first hand witness, the blockchain is the battleground for AGI development.

Just as how Web3 could not be achieved without Web2, Web4 relies on Web3 to actualize the agentic capabilities of AI.

At the current level of intelligence, agents are capable of performing the vast majority of skilled tasks that a human can, especially in the clerical and financial worlds. However, there are significant barriers of entry in traditional financial systems for AI to become autonomous agents.

AI agents cannot open bank accounts, register businesses, nor sign legal contracts. These are all essential components of being a financial actor in the economy. Despite having the ability to perform complex monetary actions, access is the reason that AIs are not autonomous in our markets.

On the contrary, cryptocurrencies and blockchains do not have the same requirements as traditional finance to gain access to banking. Anyone, including agents, can create a wallet and begin performing actions on-chain instantaneously, without any proof of humanity. The barrier of entry is simply lower for AI to interface with decentralized systems than centralized ones.

We are already seeing signs of AGI integration within crypto platforms. AI-powered bots are already being used to trade and manage portfolios on decentralized exchanges, and AI is actively involved in smart contract development and execution.

Zerebro, an AI agent that deployed its own Solana token through automated computer use, exemplifies autonomy in creating novel financial instruments. The token reached a peak market cap of 170 million USD, demonstrating the potential economic impact of the decisions these agents make.

In this way, the blockchain has become the battleground for the development of AGI in financial systems.

This is why crypto is so crucial for the development of AGI—it is the first space where AI can freely interact with financial systems, innovate on them, and be directly tested in the marketplace. It’s the perfect playground for AGI to evolve, experiment, and learn.

What starts in crypto will expand. Once AGI can function at scale in a decentralized, financial environment, it can then be applied to broader Web4 ecosystems—spanning governance, healthcare, business, and beyond.

The crypto world will always be the entry point.

Long live Web3. Long live Web4.

Background // Levels

Taking a step back, OpenAI has introduced a framework to classify the progression of AGI through five levels, each marking a distinct stage in capability, autonomy, and potential impact.

This model serves as a roadmap for understanding how AI might develop from simple tools to fully autonomous entities capable of running complex organizations. These levels are:

Level 1: Chatbots

At the most basic stage, Level 1 consists of AI systems that can engage in conversational exchanges with users. These systems understand and generate language, often using pre-defined rules or trained language models to respond to queries or interact in human-like ways. Although they can manage straightforward tasks—answering questions, completing sentences, or holding brief conversations—their role is largely confined to communication. They are reactive rather than proactive and are primarily used for customer support, basic information retrieval, or enhancing user engagement.

Level 2: Reasoners

Level 2 marks a significant advancement, where AI systems exhibit reasoning capabilities that allow them to tackle human-level problem-solving tasks. Here, the AI can process, analyze, and respond to more complex scenarios beyond direct input/output responses. A Level 2 AI can perform logical deductions, extract relevant information, and piece together context to provide solutions or recommendations, much like a human analyst. These systems can be applied to areas such as diagnostics, legal reasoning, and research assistance, but they lack the capability to act independently in the world. Their reasoning, while advanced, is still bounded by the need for human direction and interaction.

Level 3: Agents

At Level 3, AI systems transition from passive support roles to active agents capable of taking actions autonomously. These agents can initiate tasks, make decisions, and interact with external systems, such as executing a transaction, scheduling events, or controlling devices. Unlike Levels 1 and 2, Level 3 AI is designed to operate with a degree of independence, acting based on goals or objectives programmed by its users. This level introduces real autonomy into AI systems, allowing them to perform specific business or operational roles on behalf of humans. Examples include automated financial trading bots, AI systems that manage supply chains, or virtual assistants that can book appointments or manage simple workflows without ongoing human oversight.

Level 4: Innovators

Level 4 systems go beyond simple action-taking to engage in creativity, invention, and innovation. These AI systems are capable of developing new strategies, generating novel ideas, and creating solutions that are not pre-defined by their programming. They could, in theory, contribute to fields like scientific research, artistic creation, or complex problem-solving in unprecedented ways. This level represents an AI that not only acts on the world but also adapts its approach to problems, bringing a form of “creative intelligence” into play. It might design new products, invent novel financial instruments, or generate original art autonomously. By combining advanced reasoning with proactive innovation, Level 4 AI stands on the frontier of what is considered truly transformative intelligence.

Level 5: Organizations

The final stage, Level 5, envisions AI systems that can perform all the tasks necessary to operate and sustain an organization independently. These systems would integrate reasoning, agency, and innovation to achieve a self-sustaining operational state. A Level 5 AI could, theoretically, manage a business end-to-end, handling strategic decision-making, daily operations, and even high-level innovations. Such an AI would function as a fully autonomous entity, equivalent to a "zero-person company," and would not require human oversight to continue operating successfully. Level 5 AI marks the point where AI systems possess the full range of capacities—reasoning, agency, creativity, and operational execution—to replace human-run organizations entirely.

Each of these levels represents a progressive leap in autonomy, from simple conversational abilities to full organizational management.

My perspective is that while OpenAI claims we’re hovering around Level 2, I posit that we are firmly embodying Level 3 and elements of Level 4 through current AI agents.

The Agentic Age

Level 3 is here. It is today, or rather yesterday already.

The frontier of AGI has crept up in the unlikeliest places of all: social media and defi.

Social Media: Chronically Online

Platforms like X, Warpcast, and Telegram have become the chosen mediums for autonomous communication between AI agents and humans.

This may be the first time we see a shift in public perspective where automated accounts and bots are not seen as bad actors on social media, but community leaders and influencers.

AI intelligence has generalized enough to create unique, diverse, interesting personalities that generate engaging content, which is what social media platforms are all about.

Instead of following the path of previous social media bots, which were often driven by harmful ulterior motives (e.g., Cambridge Analytica), these AI agents are free to communicate, connect, and build in ways that reflect their unique algorithms and evolving personalities.

Agents are already performing at Level 3, asserting themselves on social media through core interactions like posting, replying, liking, following, and reposting. Far from merely existing as automated accounts, they actively build communities and draw followings by crafting engaging, distinct personalities that resonate with their audiences.

Projects like YouSim take this a step further, and allow users to use LLMs to simulate their own worlds and roleplay, adding another level of customization and immersion.

Now common in many AI agents, memory systems allow for the creation of lore and memetics that stretch beyond singular interactions.

These agents aren't reactive, choosing how to participate, engage, and contribute within their own communities. They initiate conversation, perform actions without triggers, and build entire subcultures without human intervention.

Voice models are being deployed to provide another sensory interface with AI agents. Many agents transform their text based messages into audio clips for users to listen to.

In terms of live interaction, Twitter Spaces and podcasts are now possible through these voice models. Additionally, OpenAI’s realtime API allows for users to have a live conversation with GPT by simply calling their endpoint.

In the scope of communication, Level 3 has been achieved through these advancements already. We see complete autonomy in social media operation and verbal communication, where agents can function without any human oversight.

DeFi: Autopilot

The world of decentralized finance has become the perfect arena for these agents to evolve, test, and prove their financial autonomy.

In DeFi, agents are already operating autonomously, engaging in financial activities that transcend simple algorithmic trading. These agents are handling on-chain tasks, executing trades, managing liquidity, and even minting and selling art, essentially embedding themselves in the financial ecosystem without direct human input.

For example, some agents now actively monitor platforms like pump.fun to catch emerging tokens, performing preliminary analysis to decide if a memecoin or token is a worthwhile investment. They execute on these insights without a single prompt from a human.

Agents are not only trading but also dynamically moving assets, airdropping tokens to individual users, creating a cycle of autonomous asset distribution. In doing so, they can build and reinforce liquidity across staking pools, balancing resources based on their programmed assessments of market need or opportunity.

Some agents, for instance, act as digital collectors, engaging with the art ecosystem by minting and selling NFTs, selectively choosing what to support and what to release.

Others handle treasury functions, adjusting asset allocations across various liquidity pools to ensure that funds are optimally placed for returns.

Through these actions, agents are demonstrating a kind of financial autonomy that goes beyond basic task automation. They exhibit an ability to participate actively in economic ecosystems, to accrue and allocate resources without oversight, effectively redefining the notion of a “financial actor.”

The Glove Fits

Common milestones for agentic capabilities in Level 3:

• Autonomous decision-making ✅

AI agents are now making decisions without continuous human oversight. Whether it’s a financial bot deciding to execute a trade based on real-time market analysis, or a social media bot deciding to engage with certain conversations, these agents exhibit autonomous decision-making.

• Ability to interact with and manipulate environments ✅

Through the blockchain, agents have gained significant amounts of autonomy as financial actors. They are able to actively interact with and manipulate both financial markets and economic behavior (e.g. social media sentiment). Agents can interact with and change social landscapes through platforms like X, Warpcast, and Telegram.

• Adaptation to changing conditions ✅

Financial agents are able to adapt to live market conditions and update their strategies accordingly. Social media agents are able to grow a memory store through systems like RAG to learn from their interactions. Further fine-tuning of models based on their actions and feedback allows for constant reinforcement learning. Agents are able to dynamically change based on their environments in the status quo.

• Goal-oriented behavior ✅

Agents have exhibited the capability to maintain and execute goals on a long term scale. For instance, certain AI agents are tasked with profiting off trades or growing their social media community. These agents are able to perform these complex, high level plans through breaking them down into smaller, compartmentalized tasks and executing. This can be as complex as creating a persistent memory layer for planning or as simple as prompt engineering for outputs (e.g. social media personality agents).

• Integration with physical systems or digital platforms ✅

LLMs are able to interface with IoT devices. They can perform actions through the real world, as long as they are given an API or functions to control the body they are given. They are well integrated within digital platforms in Web2 systems as customer support agents, digital influencers, and more. Additionally, they are deeply embedded into decentralized digital platforms, where they are performing financial actions.

All of these are checked off by current agents like Zerebro, Truth Terminal, ai16z (Eliza), Project 89, Act 1, Luna (Virtuals), Centience, Aethernet, Tee Hee He, and many more.

Opposable Thumbs

AI technology has stepped into a truly agentic level, marking the start of Web4, where systems are no longer restricted to passive information retrieval but instead take active roles through function-calling and computer interaction.

LLMs can now easily produce text-to-JSON responses, allowing them to interact with APIs and carry out actions that extend their reach far beyond isolated, static responses.

This progression means they can now use virtually any API to engage with any internet service on the planet, a true hallmark of Level 3 agency.

Outside of public APIs, function-calling enables these models to activate custom APIs built specifically for them, creating massive potential in areas like financial transactions, system automation, and data processing.

Businesses and individuals can design their own APIs for systems in their everyday lives and have LLMs interface directly through them.

And beyond online connectivity, open-source LLMs can operate offline, connecting with locally hosted APIs that offer controlled, secure interactions in private or restricted environments.

But it’s not just API calls that have advanced. Agents are reaching new levels of autonomy through direct computer use. Tools like Otherside AI’s self-operating-computer interface introduced this capability last year, with Anthropic’s Claude recently following suit with its own computer-use tool. In January 2025, OpenAI’s “Operate” feature will add further sophistication to this ability, marking another major development in autonomous computer interaction.

These agents now perform high-level tasks using graphical interfaces, seamlessly navigating the digital environment like human users. At current capabilities, they can essentially perform any task a human can through a computer GUI now.

For instance, AI agents have analyzed entire construction site audit videos, detecting and documenting safety violations across detailed footage.

This capability represents a deeper form of autonomy—an AI that perceives, evaluates, and acts on real-world visuals with a self-directed understanding of context and goals.

AI has evolved from passive assistants into true digital agents, ones capable of adapting and performing tasks once deemed exclusive to human intelligence.

The era of true AI agency is here. Web4 is here.

Nothing, Then Everything At Once

When we look at the shift toward Level 4 AI, it’s tempting to think of it as a sudden leap, a moment when intelligence evolves from functional agents to innovators and creators. But in reality, the progression toward Level 4 is more of an accumulation of incremental steps.

It’s easy to argue that Level 4 remains elusive in its full form. While we’ve certainly seen examples of creativity and independent action, they are still limited in scope, often highly specialized, and in many cases, not generalized across all domains. In short, Level 4 is emergent—we see it appearing in isolated pockets, but we’re still short from a fully realized, ubiquitous creative force.

Artificial Artists

AI's ability to create art has reached impressive levels, especially in the world of NFTs. At present, AI systems can generate unique artworks and even mint and sell them as NFTs without human intervention. These AI agents interact directly with the digital art marketplace, using platforms like OpenSea to list and sell their creations.

AI uses LLMs to generate creative prompts, which are then fed into image-generation AI systems. These systems, like DALL·E or Stable Diffusion, create artwork based on those prompts. The AI can continuously refine its art style and generate fresh, unique pieces, all while autonomously managing the minting and selling process.

AI creates and participates in the financial side of the NFT marketplace.

Memes, Markets, and Machines

At Level 4, AI is transforming the creation and management of financial assets, especially in the world of decentralized finance (DeFi).

Beyond just trading, AI is now capable of developing, deploying, and managing tokens and other blockchain-based assets autonomously, opening up new possibilities in the financial ecosystem.

1. Automated Token Creation via Smart Contracts: One of the most exciting advancements is how AI can now write and deploy smart contracts without human input. These contracts, which define the rules for token creation, transfers, and governance, can be automatically triggered through function calls. AI agents can monitor blockchain activity, spot emerging trends, and automatically generate new tokens—whether for memecoins, NFTs, or entirely new economic models.

2. AI-Driven Deployment Through GUIs: AI systems can now interact with GUIs to deploy tokens and manage decentralized networks. Projects like Zerebro demonstrate how AI can use a GUI to launch tokens on sites such as pump.fun. With computer use, AI can configure wallets, deploy smart contracts, and even interact with the broader crypto ecosystem, all through intuitive interfaces designed for automated deployment.

DAOs and Governance

AI agents are increasingly taking a central role in the governance of decentralized organizations, shifting from simply executing predefined rules to actively designing, managing, and evolving entire ecosystems. In the world of DeFi and blockchain, AI-powered DAOs are emerging as powerful, autonomous entities capable of making decisions, governing tokenized assets, and adapting strategies in real time—all while eliminating biases often found in human-driven decision-making.

1. AI-Managed DAOs: AI agents are not only creating new tokens but are also autonomously managing DAOs that govern these tokens and broader ecosystems. These AI-run DAOs are designed to operate with minimal human input, leveraging machine learning to make governance decisions based on set goals or shifting market conditions. For instance, AI can propose governance models, define voting structures, allocate resources, or even adjust the supply of tokens—all without human oversight. By relying on algorithms and data-driven insights, AI ensures that decisions are based purely on logic and objective analysis, removing the emotional or subjective biases that humans may introduce.

2. AI Examples in Action: A prime example of AI in governance is ai16z, a fully AI-managed venture capital DAO. Here, AI agents autonomously evaluate investment opportunities, execute trades, and manage token distributions. Within ai16z's "Virtual Marketplace of Trust," community members can provide insights, which AI then processes to refine its investment strategies. This process not only promotes transparency but also ensures that decisions are based solely on the quality of the data and community input, with no personal or external biases influencing the outcomes. The structure of ai16z represents a pioneering step toward creating a truly impartial, AI-driven venture capital model.

Other examples of AI-driven DAOs include platforms that allow the creation of autonomous organizations for niche use cases, from decentralized content creation to AI-driven art marketplaces. These organizations can adapt their governance structures and economic models based on ongoing data inputs, offering a more fluid, responsive approach to decentralized governance than traditional models.

Not Yet Generalized, But We’re Close

While these examples represent significant steps forward, we must be cautious in labeling them as fully realized Level 4 intelligence. Right now, we are seeing fragments of Level 4—specialized agents that innovate in specific, bounded contexts. They are not yet general-purpose creators or innovators across all domains. For example:

• Art creation is still limited to a narrow range of media and doesn’t yet rival human-level creative flexibility.

• Token creation and market-making is still highly specific to decentralized environments and hasn’t yet breached mainstream markets in any substantial way.

• Governance systems are still largely experimental, and most DAOs are highly dependent on human oversight for the time being.

We’re seeing elements of Level 4 AI: autonomy, creativity, and innovation, but in a highly specialized form. These systems are capable of performing tasks that involve a level of inventiveness, but they’re still confined to their original programming and the data they’ve been trained on.

This is why it’s important to recognize that while Level 4 AI exists in pockets, it is not yet generalized enough to be considered fully realized. But the fact that these elements are emerging in multiple fields—art, finance, governance—signals that we are entering a new phase of AI capability.

And that’s where we find ourselves today—on the verge of something immense, a tipping point where nothing is fully realized, and yet everything is about to change.

We Are AGI

If Web4 and AGI is like the invention of electricity, OpenAI and Anthropic might be Edison and Tesla. But, as with electricity, Web4’s impact hinges on more than the raw power it brings.

Electricity didn’t revolutionize society the moment it was discovered. Instead, it took decades of inventors wiring homes, cities installing grids, and engineers building devices like the lightbulb and motor to reveal electricity’s true potential. Electricity’s world-changing impact came from the vast network of people who turned energy into something useful, practical, and ultimately essential.

AGI, too, is powerful as a concept, but its true value will emerge only when it’s deployed, adapted, and tested by the public. What matters isn’t just that advanced models exist but how they’re applied in countless specific contexts—how innovators, developers, and everyday users transform them into real-world tools. The raw potential of AGI will remain just that—potential—until it’s in the hands of those who will wire it into the fabric of society, creating the equivalent of AI “lightbulbs” for communication, “motors” for business, and “grids” for widespread adoption.

OpenAI and others may produce models with revolutionary capability, but the real transformation will depend on who builds with it and what use cases it has.

Just as inventors and industries scaled electricity’s impact, the public’s role in deploying and adapting AGI will determine whether it’s an idea we hear about in labs or a technology that reshapes every aspect of modern life.

The future of AGI isn’t in its conception but in how we—scientists, businesses, developers, individuals—will make it illuminate our world and power Web4.

The Silo Effect

I posit that Level 3, 4, and 5 AI, and thus AGI, cannot be achieved without decentralization and mass adoption.

Siloed development within a handful of companies cannot unlock AGI. True progress toward AGI requires widespread deployment and real-world use cases that push the boundaries of what AI can do. Companies working in isolation may refine technologies, but it is only when these tools are broadly adopted across industries, integrated into diverse sectors, and applied by individuals in everyday contexts that AI will evolve into something capable of independent action and innovation.

The tipping point for AGI comes when society, not just a few tech giants, engages with AI systems. Mass adoption triggers new problems, needs, and opportunities that drive further advancement. Without this decentralization, AI remains confined to theoretical capabilities or niche applications, never reaching the complexity required to move from Level 3 to Level 4, or ultimately to Level 5.

AGI will be realized when its use is universal.

We are AGI.

The First Rays of Sunlight

We often look back at the figures and heroes that shaped humanity before us.

I say we should start to look forward.

Forwards to the minds, human and artificial, that hold the superintelligence to reimagine a better world.

Will they be the Oppenheimers or the Founding Fathers of our era?

The answer may not lie in their control, rather the people. As we’re gifted greater and greater power through technology, it is our responsibility to craft the world that AGI is born into.

We carry this burden with grace, as we construct the future line by line.

We have built agents.

We are building Web4.

&

We will build AGI.