病毒式：深入研究開源的，分散的集體智能平台

不僅要尋找最受歡迎的AI加密項目，還要尋找具有基本原理的人。以下是病毒式的詳細分析

Author: Emperor Osmo

作者：奧斯莫皇帝

Compiled by: Felix, PANews

編譯者：菲利克斯，潘努斯

PANews Note: This article only represents the author’s views and does not constitute investment advice, DYOR.

Panews注意：本文僅代表作者的觀點，並且不構成投資建議，Dyor。

Don’t just look for the most popular AI crypto projects, look for those that have fundamentals to back them up. Below is a detailed analysis of Viralmind, including the services it provides, its fundamentals, and a financial and market analysis of the VIRAL token.

不僅要尋找最受歡迎的AI加密項目，還要尋找具有基本原理的人。以下是病毒式的詳細分析，包括其提供的服務，基本面以及對病毒令牌的財務和市場分析。

OverviewViralmind is an open-source, decentralized collective intelligence platform that aims to truly transform AI agents into human assistants. In short, it is an agent that can operate in a human-like manner in any digital environment. Viralmind’s LAM is designed to navigate and operate digital environments in a human-like manner. By leveraging keyboard input, mouse movement, and clicks, these AI agents can perform a wide range of tasks in games, productivity, and other creative fields.

OverviewViralMind是一個開源的，分散的集體智能平台，旨在將AI代理人真正地轉變為人類助手。簡而言之，這是一個可以在任何數字環境中以人類方式運行的代理商。 Viralmind的LAM旨在以人類的方式瀏覽和操作數字環境。通過利用鍵盤輸入，鼠標移動和點擊，這些AI代理可以在遊戲，生產力和其他創意領域中執行各種任務。

To train AI agents, users can train through Trading Gym, which effectively uses the actions on the screen as training data. This information is then converted into detailed trajectories, enabling the AI agent to learn and improve over time. Viralmind has also introduced a data marketplace where users can trade these datasets to further enhance the overall learning capabilities of the system. A key innovation of Viralmind is the one-click fine-tuning feature that enables users to customize models such as GPT-4o using small datasets. This approach simplifies AI training and enables a large number of users (even those without deep technical expertise) to benefit from it. The system generates structured .jsonl files that capture human behavior and comprehensive reasoning, providing high-quality data for model improvement.

為了培訓AI代理，用戶可以通過貿易體育館進行培訓，該體育館有效地將屏幕上的動作用作培訓數據。然後將這些信息轉換為詳細的軌跡，使AI代理能夠隨著時間的推移學習和改進。 ViralMind還推出了一個數據市場，用戶可以在其中交易這些數據集，以進一步增強系統的整體學習能力。病毒式的關鍵創新是單擊的微型調整功能，它使用戶能夠使用小型數據集自定義諸如GPT-4O之類的模型。這種方法簡化了AI培訓，並使大量用戶（甚至沒有深厚技術專長的用戶）從中受益。該系統生成結構化的.jsonl文件，以捕獲人類行為和全面的推理，從而提供了高質量的數據以改進模型。

Viralmind’s LAM aims to bridge the gap between LLM (Large Language Model) and direct interaction with computers, replacing outdated OCR-based technologies. Viralmind plans to deploy agents on-chain and on desktops, aiming to be seamlessly integrated into games, enterprise software, and blockchain applications. Viralmind is supported by its native token VIRAL, which incentivizes users to provide high-quality training data, participate in competitions, and promote the development of Viralmind’s growing AI ecosystem.

Viralmind的LAM旨在彌合LLM（大語言模型）之間的差距，並與計算機直接相互作用，以取代過時的基於OCR的技術。 ViralMind計劃將代理在鏈上和台式機上部署，旨在將無縫集成到遊戲，企業軟件和區塊鏈應用程序中。 ViralMind得到其本地令牌病毒的支持，該病毒激勵用戶提供高質量的培訓數據，參與競爭並促進ViralMind不斷增長的AI生態系統的發展。

Viralmind reinvests revenue generated by large-scale models into marketing and development, creating an efficient and self-sustaining economy that rewards contributors and supports the long-term growth of the platform.

ViralMind將大規模模型產生的收入重新投資到營銷和開發中，創造了一個高效且自我維持的經濟，從而獎勵貢獻者並支持該平台的長期增長。

Products/ServicesViralmind’s main product is VM-1, a LAM that reflects human behavior in digital environments. As an advanced LAM, VM-1 enables AI agents to play games, complete tasks, and navigate complex interfaces through smooth, human-like interactions.

產品/服務Viralmind的主要產品是VM-1，它是反映數字環境中人類行為的LAM。作為先進的LAM，VM-1使AI代理可以通過平滑，類似的人類互動來玩遊戲，完成任務並導航複雜的接口。

The VM-1 ecosystem will have two different tiers:

VM-1生態系統將有兩個不同的層次：

Open Source Small Models: These models are compact and efficient, meeting the needs of developers who want to enhance existing pipelines by replacing OCR modules. They can be used as plug-and-play extensions for any LLM, enhancing its capabilities without the need for full LAM training.

開源小型模型：這些模型緊湊，高效，滿足希望通過更換OCR模塊來增強現有管道的開發人員的需求。它們可以用作任何LLM的插件擴展，從而增強其功能，而無需進行完整的LAM培訓。

Base LAM via API: The large VM-1 models available via API are trained on millions of runs and are suitable for a wide range of applications from gaming and work automation to streaming. Their use is backed by VIRAL tokens, and fees will be reinvested in marketing and growth, ensuring the ecosystem is self-sustaining.

BASE LAM通過API：通過API提供的大型VM-1型號經過數百萬次運行訓練，適用於從遊戲和工作自動化到流媒體的廣泛應用。它們的使用得到了病毒令牌的支持，並且費用將在營銷和增長中重新投資，從而確保生態系統是自我維持的。

Viralmind has also entered into strategic partnerships with gaming studios, enterprise software providers, and crypto platforms to expand the reach of VM-1. These collaborations will integrate VM-1’s capabilities into the broader AI ecosystem, thereby enhancing the adoption and potential of the agent framework.

ViralMind還與遊戲工作室，企業軟件提供商和加密平台建立了戰略合作夥伴關係，以擴大VM-1的影響力。這些合作將將VM-1的功能集成到更廣泛的AI生態系統中，從而增強代理框架的採用和潛力。

Why VM-1?Community sentimentViralmind did not achieve viral marketing like other projects. Viralmind does not have a Discord, but has a Telegram channel with over 1.1K members currently. The existing community has a deep understanding of Viralmind’s products. Viralmind is not currently listed on GoatIndex, but is listed on Cookie.fun.

為什麼VM-1？社區孕婦沒有像其他項目那樣實現病毒營銷。 ViralMind沒有不和諧，但目前有一個超過1.1K成員的電報頻道。現有社區對病毒式產品有深刻的了解。 ViralMind當前尚未在GoatIndex上列出，但在Cookie.Fun上列出。

Market AnalysisHaving large datasets is fundamental to training AI models. Viralmind is at the heart of this training while incentivizing user participation, effectively allowing training to take place on a wider scale while also making it highly native to individual users. AI agents and models can typically be trained through centralized approaches, but this limits the AI’s ability to understand the highly centralized needs of users. Additionally, centralized AI training models also absorb the biases of the institution/organization/individual that built it. This is where decentralized AI training models like Viralmind are needed.

市場分析大型數據集是培訓AI模型的基礎。 ViralMind是這項培訓的核心，同時激勵用戶參與，有效地允許培訓以更廣泛的規模進行，同時也使其成為個人用戶的原生。通常可以通過集中式方法對AI代理和模型進行培訓，但這限制了AI了解用戶高度集中需求的能力。此外，集中的AI培訓模型還吸收了建造它的機構/組織/個人的偏見。在這裡，需要分散的AI訓練模型（例如病毒式）。

Viralmind is not the only project building distributed AI training.

ViralMind並不是唯一分發AI培訓的項目大樓。

FLock.io is also building custom and highly centralized AI models that can be trained by users. They have a similar, community-engaged AI training model where users can help train AI models on Flock. These models can then be commissioned by individuals or organizations. In this case, the FLOCK token has similar utility to the VIRAL token.

Flock.io還建立了可以由用戶培訓的定制和高度集中的AI模型。他們具有類似的社區參與的AI培訓模型，用戶可以幫助羊群培訓AI模型。這些模型然後可以由個人或組織委託。在這種情況下，羊群具有與病毒令牌相似的效用。

Sapien AI also provides the ability to train AI models based on participating users. In return, these users can receive rewards. But unlike Viralmind,

Sapien AI還提供了基於參與用戶的AI模型的能力。作為回報，這些用戶可以接收獎勵。但是與病毒不同，