：標題：Bittensor：人工智慧基礎設施網絡

Bittensor 是一個去中心化的人工智慧網絡，允許人工智慧開發人員建構和部署機器學習模型或其他人工智慧相關的開發。

!output: Author: IOSG Ventures

!輸出：作者：IOSG Ventures

Introduction

介紹

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.

近年來，由於數據、運算能力和演算法研究的進步，人工智慧發展取得了巨大突破，特別是OpenAI GPT-4的出現，代表了基礎LLM大模型的到來，推動了生產力的提高和社會效率的轉變。

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為代表的閉源大型模型的弊端也變得明顯，即中心化模型往往對第三方集成存在限制，這損害了基於中心化模型的AI代理的可擴展性和互通性。

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.

於是，像Llama系列這樣的開源大型模式越來越受到研究者的青睞，但開源並不等於透明，也面臨許多挑戰。

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.

主要的困境是開源人工智慧開發沒有為大多數貢獻者提供經濟誘因。儘管有一些競賽獎勵存在，但通常是一次性的，後續的改進和開發工作仍然需要激情，除非達到一定規模後建立一個龐大的追隨者社區，這可能會帶來更多的收入機會和更多的貢獻者繼續提升。

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專案Bittensor嘗試利用web3代幣挖礦，讓開源AI開發更永續、可驗證、更有效率。透過Yuma 共識，旨在將資源與研究方（礦工）、驗證者（Validators）和AI 專案方（子網創建者）對接，讓整個AI 研究過程更加透明和去中心化，讓任何人都可以為AI 做出貢獻並獲得應得的收益獎勵。

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!

代幣在二級市場的表現也印證了人們的預期，價格從2023年9月的50多美元上漲到2024年12月的500多美元，實現了十倍的漲幅！

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的投資人、數位貨幣集團創辦人成立了名為Yuma的加速器，專門孵化Bittensor生態內的子網項目，並擔任CEO，展現了他對Bittensor項目的信心和潛力。

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.

當然，任何專案的成功都不可能不被懷疑。自從 Bittensor 成立以來，就出現了很多 FUD。在本文中，我們總結了許多懸而未決的問題，並試圖透過研究和分析來了解 Bittensor 在去中心化 AI 領域的未來定位和潛力。

What is Bittensor?

什麼是比特張量？

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

Bittensor 於 2021 年由來自加拿大多倫多的團隊創立，成員包括 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 是一種去中心化的人工智慧基礎設施，人工智慧開發人員使用它來建構和部署機器學習模型或其他人工智慧相關的開發。許多Web3 AI項目，無論是否擁有自己的區塊鏈，都可以連接到Bittensor的區塊鏈「子張量」並成為子網路的一部分。

What is a Subnet?

什麼是子網路？

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.

子網構成了Bittensor生態系統的核心，每個子網路都是一個獨立的基於激勵的競爭市場。任何人都可以創建一個子網，自訂它要執行的任務，並設計激勵機制（在機器學習術語中，激勵機制可以理解為目標損失函數，引導模型訓練走向理想結果）。透過支付註冊費（以 TAO 計價），人們可以建立一個子網路並接收該子網路的 netuid。請注意，子網路創建者不需要承擔子網路內的操作任務，但可以將這些任務的操作權限委託給其他人。

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).

子網路內的操作任務為其他人提供了另一種參與方式，即加入現有子網路。如果加入現有子網，有兩種參與方式：作為子網礦工或子網驗證者。除了支付註冊費（以TAO 計價，驗證者還需要質押TAO）外，只需提供一台具有足夠計算資源的計算機並將該計算機及其錢包註冊到子網，同時運行子網創建者提供的礦工模組或驗證器模組（兩個模組都是 Bittensor API 中的 Python 程式碼）。

How Does the Competitive Market of Subnets Work?

子網路競爭市場如何運作？

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.

子網路競爭的運作原理如下：假設您決定成為子網路礦工。子網路驗證器將為您指派任務來完成。子網路中的其他礦工也會收到相同類型的任務。一旦所有子網路礦工完成任務，他們就會將結果提交給子網路驗證器。

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.

隨後，子網路驗證器將對子網路礦工提交的任務品質進行評估和排名。作為子網路礦工，您將根據您的工作品質獲得獎勵（以 TAO 計價）。同樣，其他子網路礦工也會根據表現獲得相應的獎勵。同時，子網驗證者也將獲得獎勵，以確保高品質的子網路礦工獲得更好的獎勵，從而帶動子網整體品質的不斷提升。所有這些競爭過程都是基於子網路創建者編碼的激勵機制自動化的。

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

激勵機制最終評判子網礦工的表現。當激勵機制調整好後，就能形成良性循環