"Suddenly like a spring breeze coming overnight, the iron tree can also bloom pear blossoms." How come so many DeFai projects have emerged like magic in such a short period of time? The standards and framework haven’t been fully understood yet, but another round of DeFai involution war has begun? Okay, next, let me share from a popular science perspective, what is going on in several major categories of DeFai projects?

1) In the past two days, @poopmandefi brother shared a DeFai ecosystem project distribution map, which was widely spread in the community. There are too many in the comment area Many related projects have not been included yet. Many people will be very anxious and worried about missing every wealth code, but in fact there is no need to do so.

First of all, it needs to be explained that there are certainly some good AIs in Mindshare. "New" projects, such as: $AIXBT, $BUZZ, #NEUR, #GRIFT, #Cod3x, etc., but most of them are new faces that exude an "old" flavor.

The core reason is that most of them are old projects that have been given new expectations through the new narrative of AI Agent, and some have done a lot of optimization in the DeFi field. Old projects that have experienced work but no one cares about them. Some old projects that were difficult to discover in the context of the last round of VC attracting the attention of retail investors.

2) There are four major categories in the picture below. I will try to dissect them one by one to understand my understanding:

1. AI Abstraction (AI abstraction): As the name suggests, it uses abstraction capabilities to encapsulate the information processing capabilities of large AI models in front-end product experiences that users can directly interact semantically. Users can enter some prompts. The transaction interface can be called directly in the front end of the dialog box and directly controlled by the AI The backend automates transactions.

"Purity" and the "accuracy" of the AIGC backend processing information and executing the requirements require a "fault tolerance" mechanism. Either the user feels that the instructions that can be entered and executed are too simple to compete with the current DeFi experience, or the user enters too many high-expectation instructions. , found that the program background is not fineThe Solver is executed correctly and cannot be processed.

However, this type of product can also gain the trust of a large number of users with its novel interaction mode and solutions to some basic Swap, Staking and other problems. The reason is that its stamina is highly predictable. Because users can input prompts in text, audio, etc. in a convenient way that conforms to usage habits, it will greatly reduce the threshold for use. At the same time, the AIGC background processing capabilities will gradually encapsulate more new Solver execution solutions to improve the user experience.

Anyway, this is an attempt to explore a new trading paradigm. Just like Uniswap entered the market with the AMM Swap trading pool paradigm, it was also complained about slippage at the beginning. Great friction. The AI ​​Abstraction subdivision track is indeed useless in the short term, but the big paradigm shift opportunities that have been nurtured in the long term are worthy of attention.

2. Autonomous Portfolio Management & Yield Optimization (automated portfolio management and income optimization): This type of product is the result of the previous round of DeFi market involution. A large number of projects want to get a share of the DeFi track and continue to work hard from the perspectives of personalization, customization, vertical segmentation, specialized experience, etc. However, before they can harvest the fruits of victory, the DeFi industry is almost desolate. .

Most of these DeFi revenue optimization strategies come from the team’s ability to monitor and analyze on-chain data, such as transaction depth, capital flow, APY fluctuations, and slippage predictions. estimation, price deviation, arbitrage space, risk warning, etc. Based on these real-time on-chain data analysis, a set of execution strategies is formulated, such as position fund allocation, arbitrage opportunity capture and execution, Yield income estimation, single pool or combination strategy, impermanence Loss management, liquidation risk control, and more.

Simply put, the core of this type of product is real-time on-chain data + the ability to capture trading opportunities, plus a complete set of automated analysis and execution optimized based on smart contracts. Operation experience upgrade capability. At first glance, what does it have to do with AI? The combination point is that data analysis and strategy formulation can train and fine-tune traders' strategies to create a set of possible investment opportunities that are more efficient than manual work.

Moreover, when combined with AI Agent, the space for imagination becomes even greater. Everyone can use their own strategies to fine-tune a personal trading preference.AI Agent can automatically help you find opportunities on the chain and automatically execute transactions. Letting AI Agents become people's advanced trading assistants is a long-term sexy and online narrative.

3. Market Analysis or prediction (market analysis and prediction): This type of product has captured most users of Mindshare as a super powerful single AI, such as @ aixbt_agent has indeed become a key information acquisition platform for many traders as a top KOL. However, everyone recognizes the practical application scenario capabilities of AI Agents that only provide trading strategy analysis, but there is little room for long-term imagination. For example, can my AI Agent monitor the news of AIXBT and automatically help me make decisions to buy lows and arbitrage? etc.

Theoretically, it is naturally feasible. In fact, it is entirely possible for AI Agents such as AIXBT to independently manage user assets, and at the same time help users conduct trading operations based on their own information decisions. , it’s just that this move hasn’t been made yet. At present, users of this type of products are occupying the minds of users so quickly, and with the commercial monetization capabilities behind the traffic drive, the room for imagination is actually not small.

4. DeFai infrastructure or Platform (infrastructure and platform): This type of agreement covers a wide range, except for AI Agent Native such as #ai16z and #Virtaual In addition to emerging platforms, other project business frameworks such as Bittensor, io, Atair, @hyperbolic_labs, Vana, SaharaAI, etc. that are related to AI computing power, data, fine-tuning and other businesses can be extended here.

After all, for the AI ​​Agent to function normally, data is oil, computing power is the power grid, reasoning is the transformer, AI Agent is the terminal, etc., they are all provided by upstream Service providers.

So, there is not much to say. AI Agent needs to accumulate strength in the second half of the journey, and this type of DeFi platform will definitely make great efforts. Originally, AI Agent narrative was only the earliest part of AI Narrative. Frameworks and standards, DeFai, Gamfai, MetAiverse and other core narratives are all inseparable from these AI infra platforms.

Above.

Although I have given you a clear understanding of DeFai, it does not mean that I am not optimistic about it. Starting from the current chaotic and difficult-to-judge narrative framework and standard "chaotic era" market, DeFai has at least substituted a more AI Agent application, and can better see progress and expectations step by step through experience, PMF product implementation, etc.

This is also a manifestation of the current popularity of the AI ​​Agent market turning from virtual to real. Moreover, so many old species cannot find opportunities in the old DeFi era. In the new trend Isn't this an opportunity for them to unleash their potential?