AI Abstraction is expected to become the mainstream direction of DeFAI applications.
Author: Kevin, the Researcher at BlockBooster
DeFAI is another hot topic in the market following Framework. According to Kaito's data on January 15, DeFAI's mindshare has reached the same level as Meme. Although Meme has been somewhat quiet during the recent two months of the Agent craze, it still shows DeFAI's market heat as the latest narrative.
DeFAI is the combination of DeFi and AI Agents, and many protocols are eager to combine Agents with this traditional narrative, hoping to spark new ideas.
AI Abstraction is expected to become the mainstream direction of DeFAI applications
A few days ago, @poopmandefi organized a mapping of DeFAI applications, where I believe that AI Abstraction-type DeFAI applications are more likely to create a bubble and have a greater chance of producing high-quality applications. Although portfolio management and market analysis DeFAI applications are also attractive, they have a smaller imaginative space compared to abstract applications and rely more on trust assumptions.
Investment portfolio management applications focusing on Agent automation can be traced back to the last cycle. Automated applications can be a simple script or a complex algorithm, but the core remains the pursuit of user customization, allowing users to DIY strategies that suit their trading habits and risk preferences from the options provided by the platform. Therefore, the goal of automated applications is to allow users to run programs and rest assured.
This means that the imaginative space for automated applications is limited. They focus more on the vertical refinement of user experience, and the moat between protocols often lies in the design of algorithms. The competition in automated portfolio management and yield optimization applications is essentially about the team's strategy formulation ability, competing on when to trigger arbitrage, when to reduce liquidation risk, how to allocate positions, and maximize farming yields.
I believe the opportunities for Agents to participate in this are not as great as the market expects. The reason is that privately tuned and trained Agents by users find it hard to outperform the rapidly iterating algorithms of professional teams. Having Agents help find trading opportunities on-chain, at this stage, is difficult not to become exit liquidity for others. Therefore, the narrative of making Agents one's "money printer while sleeping" may just be an idealized view.
Market analysis DeFAI is mixed, as any Agent can express opinions on token prices, but most of these opinions are repetitive, leading to a lack of attention. Among these analyses, applications like Zara AI, which have self-developed frameworks, analyze specific indicators through continuous training and optimization; while AIXBT, as an industry leader, has long occupied the top of Kaito's mindshare list, becoming a top KOL. There is a significant deviation in market analysis DeFAI, as the vast majority of Agents are just cannon fodder, filled with bubbles, and hard to generate commercial value. From user recognition of Agent's market analysis to Agents forming business models and realizing traffic monetization, this may be the short-term ceiling for market analysis DeFAI.
However, public analysis by Agents can be both a Buy Signal and Sell News. This may be why top KOLs like AIXBT have not started to self-custody user assets. Because the analysis by Agents is based on public data, it does not push prices up like human KOLs do through articles and team collaboration. The difference between the two is one of the reasons why the imaginative space for market analysis DeFAI is limited.
So, why is AI Abstraction-type DeFAI different? I believe its characteristics lie in low expectations and high growth potential. Low expectations stem from the objective limitations of Web3 AI. From the "AI bot" of 2023, the "GPT Wrapper" in the first half of 2024, to the fine-tuning Agents in recent months, there are many "junk projects" in Web3. These projects center around ChatGPT, encapsulating the model's input and output in the application front end, allowing users to use natural language for prompts during initial use. However, due to the lack of performance moats, the actual experience has significant friction. This prolonged poor user experience is precisely why abstract applications have lower expectations.
The definition of abstract applications is to abstract complex on-chain operations through artificial intelligence, thereby simplifying the experience for novice users, allowing entry-level users to deeply experience DeFi protocols. Although these applications simplify the process similarly to many "junk projects," where users interact with the Agent front end using natural language and call various APIs, the interaction method has not significantly upgraded. Therefore, most users, or the general market perception, often believe that abstract applications have low expectations.
However, as more and more Web2 developers enter this track, the development of abstract applications is accelerating, providing enormous growth potential for such applications. Currently, abstract applications are in a very high growth phase and are expected to achieve breakthroughs in the future.
High growth potential comes from the ability of abstract applications to fully optimize user experience, while poor user experience usually arises from two aspects:
Users lack understanding of the actual capabilities of the application. When commands like Swap, Staking, etc., are input, although these operations can be successfully executed, this interaction method does not impress users.
Users overestimate the capabilities of the application, inputting complex instructions, but for a single model, such instructions are often difficult to execute accurately, leading to errors in some steps of the Pipeline workflow.
The current version of Agent applications still has ample growth space to overcome the above issues. Taking Questflow as an example, abstract applications combine multiple Agents into a Swarm to optimize user experience. In a Swarm, the more Agents used, the more refined the user's use case becomes. For example, the "Crypto Token Signal Swarm" on the Questflow platform consists of five Agents: Schedule Agent, Telegram Agent, Techcrunch Agent, OKLink Agent, and Aggregated Web3 Information Agent. Through the introduction of the Swarm, users can quickly understand its purpose: monitoring coin prices, analyzing projects, and extracting Alpha information to push to Telegram groups. Therefore, when interacting with this Swarm, users' expectations can be fully met, and actual feedback can align with expectations. More importantly, complex instructions will not be simplified or omitted, as users' instructions will be broken down and assigned to different Agents, with each Agent completing its own task, making the entire workflow more efficient and concise.
The bubble and chaos in the abstract application track are gradually dissipating, and the market has begun to shift towards more positive and serious development. A brand new interaction method is about to truly help users solve problems and improve efficiency. This new interaction method will bring about a new trading paradigm, and in the accelerated evolution of the AI Agent track, abstract applications are expected to become the pioneers in capturing the market value of DeFAI.
Solana ecosystem actively embraces DeFAI
Solana and Base are the two main battlefields in the AI Agent track, but the development directions of these two ecosystems are completely different. Virtuals, relying on a mature token model, occupy the vast majority of the market value in the Base AI Agent track; while in Solana, despite the participation of a16z, Solana's market share in the AI Agent track is relatively low due to weak fundamentals and the influence of the Solana memecoin atmosphere.
For Solana, the current flourishing ecosystem is not the most ideal situation. Solana needs a substantial narrative label to move towards the next market cap milestone. In the context of the failure of Depin, DeFAI is undoubtedly Solana's best opportunity at present. From the distribution of DeFAI applications summarized by Solana Daily, it can be seen that many DeFAI applications have chosen the Solana platform. This may be related to Solana frequently hosting Agent hackathons and closely related to its grant distribution initiatives. Overall, Solana is in a leading position in the DeFAI track, surpassing Base.
Last week, Solana released the DeFAI Landscape on Solana. I selected projects with a market cap exceeding $10 million as of January 19 and briefly summarized their core functions and classifications.
Project
Function
Type
GRIFFAIN
• Token Launchpad + Airdrop;
• Mint NFTs;
• Support Agent collaboration;
• Agents promote Tokens on X;
• Natural language interaction;
• Model has memory, learning user operation habits;
DeFAI related:
• Staking, automation, and execution of DeFi strategies
• Market analysis based on API data from different platforms, such as chip analysis
Abstract application, market analysis
GRIFT
• Natural language interaction;
• Support Agent collaboration;
• Support for over 10 LLMs
DeFAI related:
• Swap, liquidity management, and yield optimization;
• Cross-chain asset bridging and airdrop mining;
Abstract application
BUZZ
• Natural language interaction;
• Support Agent collaboration;
• Real-time monitoring of on-chain events;
• Memory
DeFAI related:
• Swap, staking, and liquidity management;
Abstract application
SNAI
• Support Agent collaboration;
• Serverless architecture;
Infrastructure
NEUR
• Natural language interaction;
• Integrated Solana native wallet;
• NFT collection management;
DeFAI related:
• Swap, staking;
Abstract application, market analysis
QUAIN
• Analyze and identify financial markets;
• Objectively assess the potential of tokens;
• Identify popular tokens on X;
• Track whale holdings
Market analysis
AIPUMP
• Token Launchpad
Market analysis
ALPHA
• Provides market analysis and token analysis based on blockchain data;
• Transforms raw data into tweets
Market analysis
HTERM
• Token Launchpad
• Integrated X • Obtain data from different platforms for market analysis • Memory
DeFAI related: • Swap, staking;
Market analysis
KWANT
• Based on CA, provides volume-price relationships, chart patterns, and subsequent operational suggestions
Market analysis
PPCOIN
• Provides analysis based on on-chain data, wallet historical data, and GitHub repo
DeFAI related:
• Executes trades and manages portfolios, automatically implementing dollar-cost averaging, spot purchases, and liquidity pool optimization
Portfolio management, market analysis
MOBY
• Captures whale accumulation behavior
Market analysis
About BlockBooster
BlockBooster is an Asian Web3 venture studio supported by OKX Ventures and other top institutions, dedicated to being a trusted partner for outstanding entrepreneurs. We connect Web3 projects with the real world through strategic investments and deep incubation, helping quality entrepreneurial projects grow.
Disclaimer: This article/blog is for reference only, representing the author's personal views and does not reflect the position of BlockBooster. This article does not intend to provide: (i) investment advice or investment recommendations; (ii) offers or solicitations to buy, sell, or hold digital assets; or (iii) financial, accounting, legal, or tax advice. Holding digital assets, including stablecoins and NFTs, carries high risks, with significant price volatility, and they may even become worthless. You should carefully consider whether trading or holding digital assets is suitable for you based on your financial situation. For specific issues, please consult your legal, tax, or investment advisor. The information provided in this article (including market data and statistics, if any) is for general reference only. Reasonable care has been taken in compiling this data and charts, but we accept no responsibility for any factual errors or omissions expressed therein.
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。
