Forward the Original Title: “AI Agent Alpha Methodology by the Editor: Why is the ‘Platform + Market’ Model More Popular? A Case Study of Max and Alchemist AI”

One creates products, the other buys and sells them.

Another week dominated by the AI Agent sector—did you find your Alpha yet? \
Top tokens like ai16z and Fartcoin have surged, while at the bottom, hundreds of so-called AI Agent tokens flood the monitoring list daily. \
With so many choices, how do you decide? \
Unless you’re a seasoned pro, a market sniper, or a networking expert in the crypto space, how do you avoid becoming just another FOMO buyer who jumps in without knowing whether the project has real potential, other than envying others’ success, managing energy, and diving in blindly?

From the perspective of an average retail investor, you might not have a reliable way to assess the market. In the process of choosing content and observing market trends, even a long-time writer who’s been losing money feels the same sense of helplessness—overwhelmed, unable to enter the market, failing to break even, and unable to find a clear understanding of the full story.

Every second in the market is messy and unclear, but every once in a while, there seems to be a clear pulse.

Clearly, an ordinary writer trying to diagnose the market for everyone would be laughed at; the following thoughts are shared only as personal insights—these “remedies” at least helped cure some of the losses I’ve suffered.

The Diversification of the AI Agent Track: Applications to the Left, Platforms to the Right

Looking at the overall market, AI Agents have undoubtedly been a hot trend in the past week or two. Especially with OKX listing GOAT spot trading yesterday, the enthusiasm across the AI Agent space has been reignited.

However, beyond the unsustainable surge, there’s a growing sense of diversification in the AI Agent projects that have been emerging lately:

• Diversification 1: These are specialized AI Agent applications that address specific problems or have particular identities/styles.

Representative projects: AIXBT, Truth Terminal

• Diversification 2: These are not specialised AI Agent applications, but they provide you with the tools to create more AI Agent applications.

Representative projects: Virtuals, ai16z (Eliza framework), Empyreal SDK

On a basic level, this is an evolution from applications to platforms: from an AI Agent issuing its token to a platform/tool that enables more AI Agents to issue tokens. \
Of course, this argument isn’t entirely accurate—aside from launchpads, some tools don’t issue tokens for AI Agents directly, but instead provide an environment where AI Agents can be more effective and easier to use. Essentially, this isn’t about a single-use application, but about building platforms and ecosystems: \
The more useful and versatile the platform is, the more reason there is for asset-backed tokens to be endorsed, and the more likely they are to attract speculative investments.

In fact, this logic might just be a repetition of history. \
Back in 2017, ICOs were all the rage, and each project could issue a token through an ICO. But after the ICO boom, Ethereum essentially became the biggest ICO platform, allowing various projects to deploy smart contracts and issue tokens on top of it. \
Now, with AI Agents being all the rage, every Agent can issue its token. However, we’re also seeing framework and platform-based projects emerging that provide low-barrier, no-code, or natural language tools to help anyone quickly create AI Agents.

History never repeats itself, but it rhymes. The core of asset creation has never really changed.

Attention Drives, Capital Stirs

Note, I’m not saying that platform/framework-type projects in the AI Agent space are inherently stronger than pure AI Agent applications. The strength or weakness of a project isn’t determined by its direction, but rather by the results of market dynamics—where capital flows in and out. To put it bluntly, capital tends to go to the place where the story lasts longer.

What does “a longer-lasting story” mean?

Of course, we’ve all seen the tokens that skyrocket in one day, but more often, we see projects that burn bright for only 1-2 hours before fading away. These tokens can attract a lot of capital in the short term but immediately lose it once the hype dies down.

From another angle, it’s as if investors think, “Your story won’t last long.”

Projects that can sustain their story longer are more likely to capture attention. Attention drives hype, and capital stirs the market’s ups and downs.

To be more specific:

• When it comes to meme tokens, you look at the angle they’re coming from. \
  When it comes to AI Agents, you look at what they claim to do.

So the question becomes: Which AI Agents do you think can sustain their story for a longer period?

After reviewing many projects, I believe the current AI Agents can roughly be divided into several categories (the original classification inspiration comes from @rm/agents-are-a-temporary-interfacehttps://paragraph.xyz/@rm/agents-are-a-temporary-interface">this article):

1. Personalized Imitators

These agents simulate intelligence and imitate human behavior through conversations. Their goal isn’t to solve problems but to make users feel unique and personalized. The personality of these agents is their brand.

Typical representatives: Bully (the sarcastic bot)

1. Efficiency-Oriented Overseers

These agents analyze complex workflows and convert human intent into backend processes with precision. They may not have personality, but they are highly efficient, saving time or solving specific problems.

Typical representatives: Simmi AI (one Twitter sentence that helps you issue a token)

1. Autonomous Experimenters

These agents manage wallets, interact with systems, and can even start tasks without manual input. However, their autonomy is limited; they wait to be triggered, and they aren’t fully autonomous or sci-fi-like.

Typical representatives: Truth Terminal (the starting point for everything)

1. Platforms/Frameworks Supporting the Above Categories \
   You can create a sarcastic AI or a bot that helps you issue tokens with just one sentence—whatever your motive, you’ll still need certain components: for AI Agents, you need models, data, and prompts; for issuing tokens, you need a launchpad.

Typical representatives: Virtuals (Launchpad), Eliza (Framework Builder)

If you rank the tokens in the current AI Agent space by market cap, it’s not hard to see that the top projects generally fall into these categories.

So, which of these projects do you think will sustain their story the longest?

First, outside of the above categories, projects driven purely by external events or IP are often one-off occurrences. If the event itself is isolated, its influence and longevity will wane quickly, and the token will lose value fast. A pump-and-dump coin is a clear example, and Luce also shows some signs of decline.

Within the AI Agent space:

• Projects that can create assets internally, because the user-generated content (UGC) in the form of Agents is a relatively multi-point concept, will have more sustained influence. Their impact will be cyclical, as new Agents will continually emerge from their platforms. In terms of assets, their tokens may experience fluctuations, but the overall trend will spiral upward, with the parent token’s market cap staying high and several leading AI Agents emerging.

• Projects that can continually output content will strengthen their brand’s personality through ongoing conversations and interactions. People find them interesting, engaging, and even unlock the “wealth code,” so they keep coming back. And as attention translates into capital inflow, these projects will see their tokens rise.

However, between the two, the latter type (content-producing AI Agents) is more susceptible to getting stuck in a niche. For example, when you think of sarcastic AI, you immediately think of Bully, and attention becomes concentrated on just a few agents, making the success rate lower than that of platform-type projects.

So, in the AI Agent space, you’ll generally need a good concept—something that can create more assets (frameworks, programming platforms, token issuance platforms, etc.). It seems that capital has a greater and more sustained interest in projects of this kind.

MAX and Alchemist AI: Examples of the Platform + Market Model

The following are some methodologies I use when evaluating projects. They may not be perfect, but they can serve as a reference for selecting projects.

If you follow the path of looking for platform/framework projects, you’ll find that in the past week, at least two projects are worth noting, and both have adopted the “Platform + Market” combination. (Note: NFA, and not shilling—just sharing projects I’ve come across.)

Platform: A platform that helps you quickly build AI Agents or use AI Agents to quickly create products.

Market: The products created can be traded on an integrated marketplace, further boosting activity and platform hype.

1. $MAX and Distilled AI: Front-End and Back-End Model

CA: oraim8c9d1nkfuQk9EzGYEUGxqL3MHQYndRw1huVo5h

Market Cap: 28M (up from 6M on Monday, approximately a 500% increase)

MAX is itself a female AI Agent. You can directly interact with her in conversation. However, her unique personality lies in her being a BTC fanatic. When you chat with her, she constantly promotes the benefits of Bitcoin. She also comes with a built-in Bitcoin prediction market. By simply clicking on her avatar, you are directed to a prediction page where you can use $MAX tokens to predict and bet on Bitcoin’s future price. It’s a full play on both the AI Agent and betting assets.

But MAX is just an AI Agent. What really attracted funds, pushing her token’s market cap from 6M to 28M in the past week, is more about the front-end and back-end model behind her:

MAX herself attracts traffic, and then she encourages more people to use the underlying AI Agent creation platform (link here). \
The platform behind MAX is Distilled AI, a decentralized protocol that provides developers with infrastructure for accessing confidential computing, private data processing, distillation protocols, data DAOs, and secure access management.

This protocol is supported by Oraichain, an L1 dedicated to AI.

With Distilled AI, developers can build advanced applications where AI agents can learn from private data in both individual and collective settings as well as autonomously perform operations in Web3 with absolute data privacy.

This model is very similar to the relationship between Luna and Virtuals. However, currently, Distilled AI does not have a native token; all value transfers on the platform use the $MAX token.

On the Distilled AI Marketplace, you can find various types of Agents, similar to an integrated GPT store.

Currently, using other Agents in this marketplace is free, but all agents must hold and stake $MAX to function. It’s possible that future relationships may emerge between $MAX, the platform’s native token, and other platforms.

At the same time, according to the project’s description, AI agents created by others in the marketplace can also be tokenized, similar to pump.fun’s launch mechanism.

When the market cap of a newly created token reaches 66k, the launch is supported by a bonding curve mechanism. The clear pattern we can observe here is:

1. Attractive AI persona leads the way, drawing in users.
2. The platform has the ability to create AI Agents and empowers the platform’s native token (currently $MAX) with features like consumption and staking.

3. The concept of an app store and marketplace: the price discovery mechanism filters out the most popular AI Agents, expanding the platform’s influence.

4. $ALCH (Alchemist AI): A No-Code Application Generation Platform Described in Text

Market Cap: 130M (up from 27M on Monday, approximately a 500% increase) \
CA: HNg5PYJmtqcmzXrv6S9zP1CDKk5BgDuyFBxbvNApump

If MAX’s example is a complete copy of Virtuals, then Alchemist AI takes a different approach. Instead of creating different AI Agents, it uses public AI capabilities to create various applications and products.

The name of the project itself explains its function. Like “alchemy,” by inputting natural language, the AI behind Alchemist AI automatically processes the input and generates a product.

The entire process is simple: connect your wallet, input a description of the application you want, and the platform runs for a while to generate the corresponding program (for example, I tried to create a game in the screenshot).

Of course, you can also download the generated work or fine-tune it further by interacting with the AI on the platform.

The concept of a “no-code development platform” isn’t new, but when applied to the AI Agent track in the crypto market, it’s a breakthrough. Using the same AI capability to create different works has huge value for community engagement. According to community feedback, users have come up with creative ways to use Alchemist, even building a Solana-based meme coin filter and analysis tool on the platform.

This is similar to leveraging GPT capabilities to create more scenario-specific applications, except Alchemist provides its front-end page. Additionally, through operating events and community-driven promotions, it can gain more traffic value.

Here, the role of the ALCH token is as follows:

1. To create higher-quality works, you can consume ALCH to unlock the paid mode, which will generate better results.
2. By consuming a certain amount of ALCH tokens, you can list your work in the built-in marketplace.
3. Others need to pay ALCH to purchase and use your product.

This is also a platform + marketplace logic: first, a usable product development platform is provided, and then the functions of circulation and transfer of works are supported by the token, giving it more value.

However, currently, the project’s page and AI implementation are still relatively rough, and the project is upgrading to version 2. The rapid growth of ALCH’s market cap to over 100M in a week clearly indicates the market’s recognition of the project’s goals in the short term.

Overall, both of the above examples adopt the platform + marketplace model, where the former creates products, and the latter trades products.

However, this doesn’t mean that projects featuring these two elements will always experience a price surge. These examples are provided simply to help you better evaluate and filter the value of new AI projects emerging daily.

In the crypto market, there are no invincible players, and finding Alpha remains a long and challenging road.

Join theTechFlow Official Community \
Telegram Subscription Group: https://t.me/TechFlowDaily \ Official Twitter Account: https://x.com/TechFlowPost \ English Twitter Account: https://x.com/TechFlow_Intern

Disclaimer:

1. This article is reproduced from [TechFlow)]. Forward the original title “AI Agent Alpha Methodology by the Editor: Why is the ‘Platform + Market’ Model More Popular? A Case Study of Max and Alchemist AI”. The copyright belongs to the original author [TechFlow]. If you have any objection to the reprint, please contact Gate Learn team, and the team will handle it as soon as possible according to relevant procedures.
2. Disclaimer: The views and opinions expressed in this article represent only the author’s personal views and do not constitute any investment advice.
3. Other language versions of the article are translated by the Gate Learn team. Unless otherwise stated, the translated article may not be copied, distributed or plagiarized.