AI+Web3: Towers and Squares

In the past two years, the development of AI has been accelerated as if a fast-forward button has been pressed. This butterfly effect triggered by ChatGPT has not only opened up a new world of generative artificial intelligence but has also triggered a torrent in the Web3 field.

With the backing of AI concepts, financing in the cryptocurrency market has seen a significant boost. In the first half of 2024, 64 Web3+AI projects completed financing, including the AI-based operating system Zyber365, which raised $100 million in Series A funding. The secondary market is even more prosperous, with the total market value of the AI sector reaching $48.5 billion and a 24-hour trading volume close to $8.6 billion. After the release of OpenAI's Sora text-to-video model, the average price in the AI sector rose by 151%. The first AI Agent concept MemeCoin GOAT quickly became popular and achieved a valuation of $1.4 billion, sparking an AI Meme craze.

From AI + Depin to AI Memecoin and now to the current AI Agent and AI DAO, new narratives are emerging one after another. The combination of AI + Web3, filled with hot money, trends, and future fantasies, is inevitably viewed as a capital matchmaking arranged marriage. It is difficult for us to distinguish whether this is the revelry of speculators or the eve of a dawn explosion.

To answer this question, the key lies in considering whether both parties can benefit from each other's models. This article will examine how Web3 plays a role in various aspects of the AI technology stack, and what new vitality AI can bring to Web3.

Opportunities of Web3 under the AI Stack

Infrastructure Layer: The Airbnb of Computing Power and Data

Hash Rate

One of the highest costs of AI is the computational power and energy required for training and inference models. Meta's LLAMA3 requires 16,000 NVIDIA H100 GPUs and 30 days to complete training, with hardware investments ranging from 400 to 700 million USD and monthly energy expenditures nearing 20 million USD.

The earliest intersection of Web3 and AI is in DePin( decentralized physical infrastructure networks). The logic is to allow individuals or entities with idle GPU resources to contribute computing power in a decentralized manner, increasing the utilization of GPU resources and reducing computing costs for end users through an online marketplace similar to Uber or Airbnb.

The characteristics of DePin include:

• Gather idle GPU resources
• Targeting the long-tail market of AI computing power
• Decentralized ownership

Data

Data is the foundation of AI. The current demand for AI data faces challenges such as data hunger, increased quality requirements, privacy compliance issues, and high processing costs.

Web3 solutions include:

1. Data Collection: Obtain more private and valuable data from users at a low cost through a distributed network and incentive mechanisms.

2. Data Preprocessing: Use decentralized incentive mechanisms to complete tasks such as data labeling.

3. Data Privacy and Security: Protect sensitive data using Trusted Execution Environments, Fully Homomorphic Encryption, Zero-Knowledge Technologies, etc.

4. Data Storage: Develop high-performance storage solutions to support AI applications.

Middleware: Model Training and Inference

Open Source Model Decentralized Market

Web3 proposes to establish a decentralized open-source model marketplace, tokenizing the models themselves, retaining a portion of the tokens for the team, and directing a portion of the future income from the models to the token holders.

Verifiable Inference

The solution to the "black box" problem of AI inference in Web3 is to perform ZK proofs for off-chain AI inference computations, enabling permissionless verification of AI model computations on-chain. The main advantages include scalability, privacy protection, and trustlessness.

Application Layer: AI Agent

Web3 can bring decentralization and cold start advantages to Agents. By establishing incentive and punishment mechanisms for stakers and delegators through PoS, DPoS, and other mechanisms, it promotes the democratization of the Agent system. At the same time, Web3 can help promising AI Agent projects secure early financing and cold starts.

How AI Empowers Web3

AI and On-chain Finance

AI Agent can autonomously execute transactions on-chain, helping investors collect information, predict trends, manage assets, optimize trading experiences, and more.

AI can also be used to enhance the security of on-chain transactions, monitor abnormal trading activities in real time, and analyze risks.

AI and On-chain Infrastructure

AI plays an important role in on-chain data collection and analysis, development auditing, and more. It can provide accurate pricing data, automated code generation, smart contract verification testing, etc.

AI and the New Narrative of Web3

AI injects creativity into generative NFTs, provides content production efficiency for GameFi, and assists in decision-making in DAOs.

The Significance of AI + Web3 Integration

The combination of AI and Web3 is like the relationship between a tower and a square. AI represents a highly centralized tower, while Web3 is an innovative and vibrant square. The integration of the two can complement each other's advantages:

• Web3 fights against the centralization trend of AI through decentralized mechanisms.
• AI brings more users, innovative business models and services to Web3.

Although the two have different starting points, the endpoint is to enable machines to better serve humanity. We look forward to seeing AI + Web3 spark more possibilities.