How do you view the cooperation between Flock and Qwen?

Author: Haotian

Yesterday, Flock.io, the DeAi training platform in the Web3AI field, and Qwen, a large language model under Alibaba Cloud, officially announced their cooperation. If I remember correctly, this should be regarded as the first integration cooperation initiated by web2 AI to web3 AI. Not only did Flock achieve a real breakthrough, but it also boosted the morale of the web3AI track under the heavy pressure of the downturn. Let me talk about it in detail:

1) I have explained in the top tweet that web3 AI Agent has been trying to stimulate the landing of Agent applications through Tokenomics, and also engaged in the rapid deployment of the competitive paradigm, but after the Fomo boom of asset issuance, everyone found that web3 AI has almost no chance of winning compared with web2AI in terms of practicality and innovation.

Therefore, the birth of innovative web2 AI technologies such as Manus, MCP, and A2A directly or indirectly punctured the bubble in the Web3 AI Agent market, causing bloodshed in the secondary market.

2) How to break the situation? The path is actually very clear. Web3 AI urgently needs to find an ecological niche that complements web2 AI to solve the high cost of computing power, data privacy, and vertical scene model fine-tuning problems that web2 centralized AI cannot solve.

The reason is nothing more than that the pure centralized AI model will eventually lead to concentrated outbreaks in the channels and costs of obtaining computing power resources, data resource privacy issues, etc., while the distributed architecture attempted by web3 AI can use idle computing power resources to reduce costs, and will also protect privacy based on software and hardware technologies such as zero-knowledge proof and TEE, while promoting model development and fine-tuning of vertical scenes through data ownership and incentive contribution mechanisms. No matter how much it is criticized, the decentralized architecture and flexible incentive mechanism of web3 AI can play an immediate effect in solving some problems existing in web2 AI.

3) Speaking of the cooperation between Flock and Qwen. Qwen is an open source large language model developed by Alibaba Cloud. With its excellent performance in benchmark tests and the flexibility and freedom that allows developers to deploy and fine-tune locally, it has become a common choice for some developers and research teams.

Flock is a decentralized AI training platform that integrates AI federated learning and AI distributed technology architecture. Its biggest feature is that it protects user privacy through distributed training without "data leaving the local area", transparently traces data contributions, and then solves the fine-tuning and application problems of AI models in vertical fields such as education and medical care. Specifically, Flock has three key components:

1. AI Arena, which is a competitive model training platform where users can submit their own models to compete with other participants for optimization effects and rewards. Its main purpose is to motivate users to continuously fine-tune and improve their local large models through the design of "game-like" mechanisms, and then select better benchmark models; 2. FL Alliance (Federated Learning Alliance), in order to solve the cross-organizational collaboration problems in vertical sensitive scenarios such as traditional medical care, education, and finance, the Federal Learning Alliance has achieved this through localized model training + distributed collaboration framework, and multiple parties have jointly enhanced model performance without sharing original data; 3. Moonbase (Moonbase), it is the nerve center of the Flock ecosystem, equivalent to a decentralized model management and optimization platform, providing various fine-tuning tools and computing power support (computing power providers, data annotators), it not only provides a distributed model repository, but also integrates fine-tuning tools, computing power resources and data annotation support, enabling users to efficiently optimize local models. 4) So, how do you view the cooperation between Qwen and Flock? In my opinion, the extended significance of their cooperation is even greater than the current essence of cooperation. On the one hand, in the context of web3 AI being generally crushed by web2 AI, Qwen, representing the technology giant Alibaba, has a certain authority and influence in the AI ​​circle. Qwen's active choice to cooperate with a web3 AI platform fully proves web2 AI's recognition of the Flock technical team. At the same time, a series of subsequent research and development by the Flock team and the Qwen team will deepen the linkage between web3AI and web2AI. On the other hand, web3 AI once had only the shell of Tokenomics, and its performance in actual utility was very unsatisfactory. Although it tried many directions such as various AI Agents, AI Platforms, and even AI Frameworks, it was unable to come up with a real solution to the problem when it came to DeFai, Gamfai, etc. This announcement from a web2 tech giant has, to a certain extent, set the tone for the future development path and focus of web3 AI;

The most important thing is that after experiencing a period of pure "asset issuance" Fomo craze, web3 AI needs to regroup and focus on a goal that can produce real results. In fact, web3 AI has never been only a channel for deploying AI Agents to issue assets more easily and efficiently, nor is it a game of issuing assets to make money. It is necessary to strive for cooperation with web2 AI, complement each other's ecological niches, and truly play the indispensable role of web3 AI in this wave of AI trends.

I am very happy to see more cross-border cooperation between web2AI and web3AI.