病毒式：深入研究开源的，分散的集体智能平台

不仅要寻找最受欢迎的AI加密项目，还要寻找具有基本原理的人。以下是病毒式的详细分析

Author: Emperor Osmo

作者：奥斯莫皇帝

Compiled by: Felix, PANews

编译者：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：本文仅代表作者的观点，并且不构成投资建议，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.

不仅要寻找最受欢迎的AI加密项目，还要寻找具有基本原理的人。以下是病毒式的详细分析，包括其提供的服务，基本面以及对病毒令牌的财务和市场分析。

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是一个开源的，分散的集体智能平台，旨在将AI代理人真正地转变为人类助手。简而言之，这是一个可以在任何数字环境中以人类方式运行的代理商。 Viralmind的LAM旨在以人类的方式浏览和操作数字环境。通过利用键盘输入，鼠标移动和点击，这些AI代理可以在游戏，生产力和其他创意领域中执行各种任务。

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.

为了培训AI代理，用户可以通过贸易体育馆进行培训，该体育馆有效地将屏幕上的动作用作培训数据。然后将这些信息转换为详细的轨迹，使AI代理能够随着时间的推移学习和改进。 ViralMind还推出了一个数据市场，用户可以在其中交易这些数据集，以进一步增强系统的整体学习能力。病毒式的关键创新是单击的微型调整功能，它使用户能够使用小型数据集自定义诸如GPT-4O之类的模型。这种方法简化了AI培训，并使大量用户（甚至没有深厚技术专长的用户）从中受益。该系统生成结构化的.jsonl文件，以捕获人类行为和全面的推理，从而提供了高质量的数据以改进模型。

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旨在弥合LLM（大语言模型）之间的差距，并与计算机直接相互作用，以取代过时的基于OCR的技术。 ViralMind计划将代理在链上和台式机上部署，旨在将无缝集成到游戏，企业软件和区块链应用程序中。 ViralMind得到其本地令牌病毒的支持，该病毒激励用户提供高质量的培训数据，参与竞争并促进ViralMind不断增长的AI生态系统的发展。

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将大规模模型产生的收入重新投资到营销和开发中，创造了一个高效且自我维持的经济，从而奖励贡献者并支持该平台的长期增长。

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.

产品/服务Viralmind的主要产品是VM-1，它是反映数字环境中人类行为的LAM。作为先进的LAM，VM-1使AI代理可以通过平滑的，类似的人类互动来玩游戏，完成任务并导航复杂的接口。

The VM-1 ecosystem will have two different tiers:

VM-1生态系统将有两个不同的层次：

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.

开源小型模型：这些模型紧凑，高效，满足希望通过更换OCR模块来增强现有管道的开发人员的需求。它们可以用作任何LLM的插件扩展，从而增强其功能，而无需进行完整的LAM培训。

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通过API：通过API提供的大型VM-1型号经过数百万次运行训练，适用于从游戏和工作自动化到流媒体的广泛应用。它们的使用得到了病毒令牌的支持，并且费用将在营销和增长中重新投资，从而确保生态系统是自我维持的。

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还与游戏工作室，企业软件提供商和加密平台建立了战略合作伙伴关系，以扩大VM-1的影响力。这些合作将将VM-1的功能集成到更广泛的AI生态系统中，从而增强代理框架的采用和潜力。

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.

为什么VM-1？社区孕妇没有像其他项目那样实现病毒营销。 ViralMind没有不和谐，但目前有一个超过1.1K成员的电报频道。现有社区对病毒式产品有深刻的了解。 ViralMind当前尚未在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.

市场分析大型数据集是培训AI模型的基础。 ViralMind是这项培训的核心，同时激励用户参与，有效地允许培训以更广泛的规模进行，同时也使其成为个人用户的原生。通常可以通过集中式方法对AI代理和模型进行培训，但这限制了AI了解用户高度集中需求的能力。此外，集中的AI培训模型还吸收了建造它的机构/组织/个人的偏见。在这里，需要分散的AI训练模型（例如病毒式）。

Viralmind is not the only project building distributed AI training.

ViralMind并不是唯一分发AI培训的项目大楼。

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还建立了可以由用户培训的定制和高度集中的AI模型。他们具有类似的社区参与的AI培训模型，用户可以帮助羊群培训AI模型。这些模型然后可以由个人或组织委托。在这种情况下，羊群具有与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还提供了基于参与用户的AI模型的能力。作为回报，这些用户可以接收奖励。但是与病毒不同，