The DeFAI project emerges, comprehensively sorting out the four categories of the DeFAI ecosystem

Author: Haotian

"Suddenly, like a spring breeze coming overnight, even the iron tree can bloom with pear flowers." How can so many DeFai projects emerge in such a short time like magic? The standards and frameworks are still not clear, and is a new round of DeFai internal competition starting again? Alright, next, I will share from a popular science perspective what the major categories of DeFai projects are all about.

1) In the past few days, Poopman shared a distribution map of a DeFai ecosystem project, which has been widely circulated in the community. Looking at the comments, there are still many related projects that have not been included. Many people may feel anxious, worried about missing out on wealth opportunities, but there is no need to.

First of all, it should be noted that among these, there are indeed some "new" AI projects with good current Mindshare, such as: $AIXBT, $BUZZ, #NEUR, #GRIFT, #Cod3x, etc., but most of them are just new faces that exude an "old" smell.

The core reason is that most of them are old projects that have been redefined with new expectations through the new narrative of AI Agents. Some old projects that once did a lot of optimization work in the DeFi field but went unnoticed, and some old projects that were difficult to discover under the background of the last round of VC attracting retail attention.

2) The following diagram has a total of four major categories, and I will try to dissect my understanding one by one:

1. AI Abstraction: As the name suggests, this refers to some products that encapsulate the information processing capabilities of large AI models in a front-end product experience that allows direct semantic interaction by users. Users can input some prompts to directly call trading interfaces in the dialogue box, and transactions are automatically completed by the AI backend.

Because of this, these products often face skepticism at first glance, as the early product interaction experiences had significant friction. For example, the "ambiguity" of user input prompts and the "precision" required for the AIGC backend to process information and execute requests need a "tolerance" mechanism. Either users feel that the commands they can input and execute are too simple to compete with the current DeFi experience, or they input too many high-expectation commands and find that the backend program does not have a precise execution Solver to handle them.

However, this type of product can gain the trust of a large number of users due to its novel interaction mode and solutions to basic issues like Swap and Staking. The reason lies in its strong predictability for future potential. Because the way users input prompts can be text, audio, etc., which aligns with convenient usage habits, it will greatly lower the usage threshold. Meanwhile, the AIGC backend's processing capabilities will gradually encapsulate more new Solver execution solutions to improve user experience.

Anyway, this is an attempt to explore a new trading paradigm, just like when Uniswap introduced the AMM Swap trading pool paradigm to the market, it was also initially criticized for high slippage friction. The short-term view of the AI Abstraction niche track may indeed seem unpromising, but the long-term potential for a significant paradigm shift is worth paying attention to.

1. Autonomous Portfolio Management & Yield Optimization: This type of product is the result of the last round of DeFi market competition. A large number of projects trying to get a piece of the DeFi pie have been continuously working from angles like personalization, customization, vertical segmentation, and specialized experiences. However, before they could reap the fruits of victory, the DeFi industry had nearly become desolate.

Most of the yield optimization strategies in this type of DeFi come from the team's ability to monitor and analyze on-chain data, such as trading depth, capital flow, APY fluctuations, slippage estimates, price deviations, arbitrage opportunities, risk warnings, etc. Based on this real-time on-chain data analysis, a set of execution strategies can be formulated, such as position capital allocation, capturing arbitrage opportunities, yield estimation, single pool or combination strategies, impermanent loss management, liquidation risk control, etc.

In simple terms, the core of this type of product is real-time on-chain data + trading opportunity capture capability, along with a complete set of automated analysis and execution operation experience upgrade capabilities optimized based on smart contracts. At first glance, what does this have to do with AI? The connection lies in that data analysis and strategy formulation can be integrated into traders' strategy training and fine-tuning, yielding a set of potentially more efficient investment opportunities than manual efforts.

Moreover, the imaginative space becomes even larger when combined with AI Agents. Everyone can use their own strategies to fine-tune a personal trading preference AI Agent to automatically help them find opportunities on-chain and execute trades. Making AI Agents advanced trading assistants for humans is a long-term and engaging narrative.

1. Market Analysis or Prediction: This type of product, as a super-strong standalone AI, has captured a large portion of user Mindshare. For example, AIXBT has indeed become a key information acquisition platform for many traders as a top KOL. However, while people recognize the practical application capabilities of AI Agents that only provide trading strategy analysis, they lack long-term imaginative space. For instance, can my AI Agent monitor AIXBT's news and automatically help me decide to buy the dip and arbitrage?

Theoretically, this is naturally feasible. In fact, AI Agents like AIXBT could potentially autonomously manage and host user assets while making trading decisions based on their own information. However, this step has yet to be taken. Currently, the speed at which these products occupy user minds is so fast, and with the commercial monetization capabilities driven by traffic, the imaginative space is actually not small.

1. DeFai Infrastructure or Platform: The scope of this type of protocol is quite broad. In addition to emerging platforms like #ai16z and #Virtaual that are AI Agent Native, other projects related to AI computing power, data, fine-tuning, etc., such as Bittensor, io, Atheir, Hyperbolic, Vana, SaharaAI, etc., can also extend into this category.

After all, for AI Agents to operate normally, data is the oil, computing power is the power grid, reasoning is the transformer, and AI Agents are the terminals, etc. They are all upstream service providers.

Therefore, there is not much to say. The latter half of the AI Agent journey needs to gather strength, and these DeFi platforms will definitely play a significant role. Originally, the AI Agent narrative is just the earliest link in the AI Narrative, and frameworks and standards, DeFai, Gamfai, Metaverse, and other focal narratives cannot be separated from these AI infra platforms.

That's all.

Although I have clarified a clear perspective on understanding DeFai, it does not mean that I am not optimistic. Compared to the current chaotic and difficult-to-judge narrative framework and standards in the "chaotic era" market, DeFai at least introduces a more AI Agent application-oriented approach, allowing for gradual progress and expectations through experience, PMF product landing, and so on.

This also reflects the current trend of the AI Agent market moving from virtual to real. Moreover, so many old species that could not find opportunities in the old DeFi era now have the chance to unleash their potential in the face of new trends, right?