：标题：Bittensor：人工智能基础设施网络

Bittensor 是一个去中心化的人工智能网络，允许人工智能开发人员构建和部署机器学习模型或其他人工智能相关的开发。

!output: Author: IOSG Ventures

!输出：作者：IOSG Ventures

Introduction

介绍

AI development has made tremendous breakthroughs in recent years due to advancements in data, computing power, and algorithm research, especially with the emergence of OpenAI GPT-4, which represents the arrival of foundational LLM large models, driving productivity improvements and transforming social efficiency.

近年来，由于数据、计算能力和算法研究的进步，人工智能发展取得了巨大突破，特别是OpenAI GPT-4的出现，代表了基础LLM大模型的到来，推动了生产力的提高和社会效率的转变。

However, the drawbacks of closed-source large models represented by GPT-4 have also become apparent, namely, centralized models often have limitations on third-party integrations, which undermine the scalability and interoperability of AI agents based on centralized models.

然而，以GPT-4为代表的闭源大型模型的弊端也变得明显，即中心化模型往往对第三方集成存在限制，这损害了基于中心化模型的AI代理的可扩展性和互操作性。

As a result, open-source large models like the Llama series have gained increasing popularity among researchers, but open source does not equate to transparency, and it also faces many challenges.

于是，像Llama系列这样的开源大型模型越来越受到研究者的青睐，但开源并不等于透明，也面临着诸多挑战。

The main dilemma is that open-source AI development offers no economic incentives for most contributors. Even though some competition rewards exist, they are usually one-time, and subsequent improvement and development work still require passion, unless a large community of followers is built after reaching a certain scale, which could lead to more revenue opportunities and more contributors continuing to improve.

主要的困境是开源人工智能开发没有为大多数贡献者提供经济激励。尽管有一些竞赛奖励存在，但通常是一次性的，后续的改进和开发工作仍然需要激情，除非达到一定规模后建立一个庞大的追随者社区，这可能会带来更多的收入机会和更多的贡献者继续提升。

Therefore, the AI project Bittensor attempts to utilize web3 token mining to make open-source AI development more sustainable, verifiable, and efficient. Through Yuma Consensus, it aims to align resources with research parties (Miners), validators (Validators), and AI project parties (Subnet Creators), making the entire AI research process more transparent and decentralized, allowing anyone to contribute to AI and earn deserved rewards.

因此，AI项目Bittensor尝试利用web3代币挖矿，让开源AI开发更加可持续、可验证、高效。通过 Yuma 共识，旨在将资源与研究方（矿工）、验证者（Validators）和 AI 项目方（子网创建者）对接，让整个 AI 研究过程更加透明和去中心化，让任何人都可以为 AI 做出贡献并获得应得的收益奖励。

The performance of tokens in the secondary market also confirms people's expectations, with prices rising from over $50 in September 2023 to over $500 in December 2024, achieving a tenfold increase!

代币在二级市场的表现也印证了人们的预期，价格从2023年9月的50多美元上涨到2024年12月的500多美元，实现了十倍的涨幅！

Recently, Bittensor's investor and founder of Digital Currency Group established an accelerator named Yuma, specifically to incubate subnet projects within the Bittensor ecosystem, and serves as CEO, demonstrating his confidence and potential in the Bittensor project.

近日，Bittensor的投资人、数字货币集团创始人成立了名为Yuma的加速器，专门孵化Bittensor生态内的子网项目，并担任CEO，展现了他对Bittensor项目的信心和潜力。

Of course, the success of any project cannot be achieved without facing skepticism. Since the inception of Bittensor, there has been a lot of FUD. In this article, we summarize many unanswered questions and attempt to understand Bittensor's future positioning and potential in the decentralized AI space through research and analysis.

当然，任何项目的成功都不可能不受到怀疑。自从 Bittensor 诞生以来，就出现了很多 FUD。在本文中，我们总结了许多悬而未决的问题，并试图通过研究和分析来了解 Bittensor 在去中心化 AI 领域的未来定位和潜力。

What is Bittensor?

什么是比特张量？

Bittensor was founded in 2021 by a team from Toronto, Canada, including Jacob Robert Steeves, Ala Shaabana, and Garrett Oetken.

Bittensor 于 2021 年由来自加拿大多伦多的团队创立，成员包括 Jacob Robert Steeves、Ala Shaabana 和 Garrett Oetken。

Bittensor is a decentralized AI infrastructure used by AI developers to build and deploy machine learning models or other AI-related developments. Many Web3 AI projects, regardless of whether they have their own blockchain, can connect to Bittensor's blockchain "subtensor" and become part of a subnet.

Bittensor 是一种去中心化的人工智能基础设施，人工智能开发人员使用它来构建和部署机器学习模型或其他人工智能相关的开发。许多Web3 AI项目，无论是否拥有自己的区块链，都可以连接到Bittensor的区块链“子张量”并成为子网的一部分。

What is a Subnet?

什么是子网？

Subnets form the core of the Bittensor ecosystem, with each subnet being an independent incentive-based competitive market. Anyone can create a subnet, customize the tasks it will perform, and design incentive mechanisms (in machine learning terms, the incentive mechanism can be understood as the target loss function, guiding model training towards ideal outcomes). By paying a registration fee (priced in TAO), one can create a subnet and receive a netuid for that subnet. Note that a subnet creator does not need to undertake the operational tasks within the subnet but can delegate the rights to operate those tasks to others.

子网构成了Bittensor生态系统的核心，每个子网都是一个独立的基于激励的竞争市场。任何人都可以创建一个子网，定制它要执行的任务，并设计激励机制（在机器学习术语中，激励机制可以理解为目标损失函数，引导模型训练走向理想结果）。通过支付注册费（以 TAO 计价），人们可以创建一个子网并接收该子网的 netuid。请注意，子网创建者不需要承担子网内的操作任务，但可以将这些任务的操作权限委托给其他人。

Operating tasks within the subnet provides another way for others to participate, namely by joining an existing subnet. If joining an existing subnet, there are two ways to participate: as a subnet miner or a subnet validator. Besides paying a registration fee (priced in TAO, and validators also need to stake TAO), one only needs to provide a computer with sufficient computing resources and register that computer and their wallet to a subnet, while running the subnet creator's provided miner module or validator module (both modules are Python code within the Bittensor API).

子网内的操作任务为其他人提供了另一种参与方式，即加入现有子网。如果加入现有子网，有两种参与方式：作为子网矿工或子网验证者。除了支付注册费（以 TAO 计价，验证者也需要质押 TAO）外，只需提供一台具有足够计算资源的计算机并将该计算机及其钱包注册到子网，同时运行子网创建者提供的矿工模块或验证器模块（两个模块都是 Bittensor API 中的 Python 代码）。

How Does the Competitive Market of Subnets Work?

子网竞争市场如何运作？

The operation of subnet competition works as follows: suppose you decide to become a subnet miner. Subnet validators will assign tasks for you to complete. Other miners in the subnet will also receive the same type of tasks. Once all subnet miners complete their tasks, they submit the results to the subnet validators.

子网竞争的运作原理如下：假设您决定成为子网矿工。子网验证器将为您分配任务来完成。子网中的其他矿工也会收到相同类型的任务。一旦所有子网矿工完成任务，他们就会将结果提交给子网验证器。

Subsequently, subnet validators will assess and rank the quality of the tasks submitted by subnet miners. As a subnet miner, you will receive rewards (priced in TAO) based on the quality of your work. Similarly, other subnet miners will also receive corresponding rewards based on their performance. At the same time, subnet validators will also receive rewards for ensuring that high-quality subnet miners receive better rewards, thus driving the continuous improvement of the overall quality of the subnet. All these competitive processes are automated based on the incentive mechanisms coded by the subnet creator.

随后，子网验证器将对子网矿工提交的任务质量进行评估和排名。作为子网矿工，您将根据您的工作质量获得奖励（以 TAO 计价）。同样，其他子网矿工也会根据表现获得相应的奖励。同时，子网验证者也将获得奖励，以保证高质量的子网矿工获得更好的奖励，从而带动子网整体质量的不断提升。所有这些竞争过程都是基于子网创建者编码的激励机制自动化的。

The incentive mechanism ultimately judges the performance of subnet miners. When the incentive mechanism is well-calibrated, it can create a virtuous cycle

激励机制最终评判子网矿工的表现。当激励机制调整好后，就能形成良性循环