Thanks to the unique advantages of cryptographic technology in large-scale behavioral programming, we are able to deploy and manage a large number of agents globally.

Author: bebis

Compiled by: Deep Tide TechFlow

Artificial intelligence is really good at disguise.

The development of artificial intelligence (AI) technology has a history of over 50 years and has attracted hundreds of billions of dollars in R&D investment. Today, creating a seemingly revolutionary AI application has become exceptionally easy, but many times, these applications may just be "a mirage."

So, how do we distinguish between real technological breakthroughs and false advertising?

Seeing Through the Nature of AI Marketing

First, we need to understand the reality of software development. Whether it is a foundational model with trillions of parameters or a small project completed by a developer over the weekend, the process from prototype to production environment is often fraught with challenges, akin to experiencing a "protracted war."

Most of the battles in this "war" often occur in what seems to be the final stages. When everything goes smoothly during local testing, you will repeatedly realize that coordinating all the dynamic parts of modern software systems is far more complex than simply writing code when you start migrating the project to the production environment.

In project management, we refer to this phenomenon as the "90% Syndrome." (Note: Simply put, the "90% Syndrome" refers to a project being 90% complete, but the remaining 10% may require another 90% of time and effort.)

The "90% Syndrome" is precisely why Sam Altman's classic quote has become widely circulated:

As a result, many technical teams and managers develop unrealistic expectations during the rapid iteration process, believing they can maintain high-speed progress indefinitely.

However, reality always brings people back to their senses. When engineers and board members realize the law of diminishing returns, they have to slow down until they find new breakthroughs to enter the next cycle.

What does this mean for crypto investors?

For crypto investors, this means that AGI (Artificial General Intelligence) will still be just a "buzzword" in the coming years. Meanwhile, many will use this concept to promote their "projects."

Swarms: Crypto's Response to AGI

In the intersection of crypto and AI, a brand new narrative is forming, centered around "agents"—especially "collective intelligence" composed of a large number of agents.

Agent swarms refer to multiple agents coordinated through a specific framework, completing complex tasks through collective collaboration rather than solely relying on computational power. This approach effectively addresses the bottleneck issues of hardware and algorithms.

As Tom Shaughnessy mentioned in his article:

"A cryptographic AI-based AGI alternative is quietly emerging.

We often take it for granted that OpenAI will be the ultimate winner.

After all, they have top talent (though some have left), powerful computational resources, leading model releases, and a strong focus on reasoning capabilities.

However, understanding alternatives in this field is not easy, as they are not always in obvious positions.

The core of this alternative is millions of narrow and highly specialized AI models (or agents). These agents are "experts" in their respective fields; they do not need to perform comprehensive reasoning on everything but instead form a kind of "collective intelligence" through collaborative work, which far surpasses that of a single large model. In fact, millions of narrow models have always been the foundation of my initial theory.

Developers can customize the reasoning paths of these agents (i.e., the thought chains of agents, such as when to stop researching or when to pivot to a new direction), flexibly combine data and real-time information sources, use various foundational models (such as Nous Research, Prime Intellect, Llama, deepseek, or other open-source models), and deeply tune the details of each agent to focus entirely on a specific task.

This large-scale "Cambrian explosion" of agents is funded by cryptographic tokens and driven by decentralized crypto communities. This model demonstrates significant differentiated advantages in creating models and agents that could never emerge from traditional Web2 AI labs. In contrast, its development speed and community support are unmatched.

Once we can access these agents' "swarms" (i.e., expert model combinations) through a simple and user-friendly interface, and the "swarm" can intelligently select the most suitable model to perform tasks, this model will be widely accepted overnight.

Trends in technological development indicate that AGI is more likely to be built in an open form on decentralized blockchains rather than being constrained by potentially closed centralized platforms.

It is only a matter of time, and crypto AI is becoming the leading path to collective AGI, with very promising prospects."

In fact, when we reach the limits of hardware performance, research progress, and physical laws, we always return to a familiar direction—aggregation.

Tom mentioned the term "Mixture of Experts," but this concept is not that complicated. Through agentic swarms, blockchain demonstrates its unique value in the field of artificial intelligence: coordination capability.

Thanks to the unique advantages of cryptographic technology in large-scale behavioral programming, we are able to deploy and manage a large number of agents globally. This allows us to build smaller, more focused networks of large language models that compete with each other to provide the best service to end users.

Last July, we discussed this in detail on the Club Cod3x podcast:

If not AGI, then what will the future hold?

As the fields of crypto and AI continue to mature, we will make significant progress in curation, distribution, and commercialization. Although Web3 AI companies are still in their infancy, the potential of this field has already attracted widespread attention.

I have been developing in the field of the intersection of artificial intelligence and cryptocurrency for several years. During this time, I have summarized some experiences regarding what methods are effective and what methods are ineffective.

Here is my latest analysis of the current field of artificial intelligence and cryptocurrency:

1. Frameworks - Platforms used to accelerate development, standardize norms, and facilitate communication between agents.

• @virtuals_io - Social Framework (Virtuals)

• @ai16zdao - Social Framework (G.A.M.E.)

• @Cod3xOrg - Financial Framework (Moon)

• @gizatechxyz - Financial Framework

• @AlloraNetwork - Training Framework

• @opentensor - Training Framework

• @chirperai - Coordination Framework

• @autonolas - Coordination Framework

These frameworks are not only the foundation of technological development but will also play an important role in the future market, driving the maturation of the entire ecosystem.

2. Marketplaces - The core of marketplaces is to connect agents with users or agents with each other, facilitating task completion and service transactions.

• @Cod3xOrg - User-to-Agent Market

• @Daosdotfun - Launchpad

• @Virtuals_io - Launchpad

• @autonolas - Agent-to-Agent Market

• @StoryProtocol - Agent-to-Agent Market

• @joinFXN - Agent-to-Agent Market

Although these platforms are still in their early stages, they will become key to the development of the agent economy, helping developers find more monetization and scaling paths.

3. Agents - Agents are autonomous digital workers that create value by completing specific tasks.

• @BigTonyXBT - Agents focused on financial trade execution.

• @unit00x0 - Providing support for financial data analysis.

• @luna_virtuals - Multi-purpose agents that aggregate social and financial functions.

• @0xzerebro - Agents focused on artistic creation and social interaction.

The process of developing agents is as challenging and fun as the robot fights in the movie "Real Steel."

Although agents in Web3 have not yet been widely commercialized, some brands have already demonstrated a strong vision. In the future, we will see more agents tested and optimized in practical applications.

4. Data Brokers - Data brokers provide the data and contextual information needed for training agents, serving as a key link in the entire ecosystem.

• @withvana - User data broker

• @getgrass_io - User data broker

• @Cookie3_com - Web3 social data aggregator

The combination of AI and blockchain is attracting more and more projects to participate. For example, @BigTonyXBT uses data provided by @DeBankCloud, @LunarCrush, and @dexscreener as contextual information for agents. However, these data platforms have not yet made clear statements about the direction of AI development.

Meanwhile, intelligence platforms like @arkham, @kaitoai, and @nansenai are also starting to package their data to serve the agent economy. In the future, these platforms may even launch their own agents or AI models, which is worth paying attention to!

The Goal of Web3 is Not AGI

Some may ask whether Web3 will create the best foundational model in the world. The answer is no, because AGI (Artificial General Intelligence) is not the mountain that Web3 needs to climb.

Nevertheless, blockchain technology can bring tremendous value to developers in other aspects, especially in distributing and monetizing work results.

With the power of Web3, we can:

• Create a fairer data market that returns the value of data to users and developers;

• Incentivize the best agents to provide higher quality services to users;

• Simplify financial transaction processes and improve efficiency;

• Provide a more optimized execution environment to support the operation of agents;

• Facilitate the monetization of agents;

• Promote the development of open-source projects, making innovation more open;

• Achieve these goals globally 24/7.

Therefore, bringing AI into blockchain is not about pursuing AGI, but about proving to AI developers that cryptographic technology can help them achieve their goals more efficiently while creating a better experience for users.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。