Written by: Haotian
Yesterday, the DeAI training platform @flock_io in the Web3 AI field officially announced a collaboration with @Alibaba_Qwen, the large language model under Alibaba Cloud.
If I'm not mistaken, this should be the first integration collaboration initiated by Web2 AI towards Web3 AI. Not only did Flock break through its circle, but it also boosted the morale of the struggling Web3 AI track. Let me explain in detail:
1) As I mentioned in my pinned tweet, Web3 AI Agents have been trying to stimulate application landing through Tokenomics and a quick deployment competitive paradigm. However, after a wave of asset issuance FOMO, everyone realized that Web3 AI is almost impossible to compete with Web2 AI in terms of practicality and innovation.
Therefore, the emergence of innovative AI technologies like Manus, MCP, and A2A from Web2 directly or indirectly burst the bubble of the Web3 AI Agent market, causing the secondary market to become bloody.
2) How to break through? The path is actually quite clear. Web3 AI urgently needs to find a complementary ecological niche with Web2 AI to solve high computing power costs, data privacy issues, and vertical scenario model fine-tuning problems that centralized Web2 AI cannot resolve.
The reason is that purely centralized AI models will ultimately face concentrated issues in computing power resource acquisition channels, costs, and data resource privacy. Web3 AI's distributed architecture can utilize idle computing resources to reduce costs, protect privacy through zero-knowledge proofs and TEE technologies, and promote model development and fine-tuning in vertical scenarios through data ownership and contribution incentive mechanisms.
Regardless of criticism, Web3 AI's decentralized architecture and flexible incentive mechanisms can have an immediate effect on solving some problems existing in Web2 AI.
3) Regarding the collaboration between Flock and Qwen. Qwen is an open-source large language model developed by Alibaba Cloud, which has become a common choice for developers and research teams due to its outstanding performance in benchmark tests and flexibility that allows local deployment and fine-tuning.
Flock is a decentralized AI training platform that integrates AI federated learning and distributed AI technology architecture. Its key feature is protecting user privacy through distributed training without data leaving local storage, transparently tracking data contributions, and solving model fine-tuning and application problems in vertical fields like education and healthcare.
Specifically, Flock has three key components:
1. AI Arena, a competitive model training platform where users can submit their models, compete to optimize effects, and vie for rewards. Its main purpose is to use a "gamified" mechanism design to encourage users to continuously fine-tune and improve their local large models, thereby screening out better benchmark models;
2. FL Alliance, which solves cross-organizational collaboration issues in traditional sensitive scenarios like healthcare, education, and finance through localized model training + distributed collaborative framework, enabling multiple parties to enhance model performance without sharing raw data;
3. Moonbase, the neural center of the Flock ecosystem, serving as a decentralized model management and optimization platform. It provides various fine-tuning tools and computing power support (computing power providers, data annotators), offering a distributed model repository, fine-tuning tools, computing resources, and data annotation support to empower users to efficiently optimize local models.
4) How should we view the collaboration between Qwen and Flock? In my opinion, the extended significance of this collaboration is even greater than its current substance.
On one hand, in the context of Web3 AI being continuously technologically suppressed by Web2 AI, Qwen, representing tech giant Alibaba, already has certain authority and influence in the AI circle. Qwen's active choice to collaborate with a Web3 AI platform fully demonstrates the recognition of the Flock technical team, and subsequent research and development between the Flock and Qwen teams will deepen the interaction between Web3 AI and Web2 AI.
On the other hand, previous Web3 AI had merely a Tokenomics shell, performing poorly in actual utility landing. Although attempts were made in various directions like AI Agents, AI Platforms, and even AI Frameworks, no concrete solutions could be provided in areas like DeFai and GameFai. This collaboration from a Web2 tech giant somewhat sets the direction and focus for future Web3 AI development.
Most critically, after experiencing a period of pure "asset issuance" FOMO, Web3 AI needs to regroup and focus on a goal that can produce real results.
In fact, Web3 AI has never been just a channel for easier and more efficient AI Agent asset deployment or a money-raising game. To compete for collaboration possibilities with Web2 AI, complementing each other's ecological needs, and truly demonstrating the indispensability of Web3 AI in this AI trend wave is crucial.
I'm glad to see more cross-border collaborations like this between Web2 AI and Web3 AI.
