IOSG：從AI x Web3技術堆疊展開基礎設施新敘事

欧易(OKX)最新版本

【遇到注册下载问题请加文章最下面的客服微信】永久享受返佣20%手续费！

APP下载   全球官网 大陆官网

币安(Binance)最新版本

币安交易所app【遇到注册下载问题请加文章最下面的客服微信】永久享受返佣20%手续费！

APP下载   官网地址

火币HTX最新版本

火币老牌交易所【遇到注册下载问题请加文章最下面的客服微信】永久享受返佣20%手续费！

APP下载   官网地址

最近大型語言模型（LLMs）的快速發展引發了人們對利用人工智慧（AI）改造各行業的興趣。區塊鏈產業也未能倖免， AI x Crypto敘事的出現使其備受矚目。本文探討了融合AI和加密的三種主要方式，並探討了區塊鏈技術在解決AI產業問題上的獨特機會。

The recent rapid development of the large language model (LLMMs) has given rise to an interest in using artificial intelligence (AI) to transform industries. Nor has the chain industry in the sector been spared, and the emergence of the AI x Crypto

AIxCrypto 的三種途徑包括：

AixCrypto has three paths:

• 1. 將AI融入現有產品：像Dune這樣的公司正在利用AI增強其產品，例如引入SQL copilot來幫助用戶編寫複雜查詢。

  1. Integration of AI into existing products: Companies such as Dune are using AI to enhance their products, for example by introducing SQL copilot to help users compile complex queries.

• 2. 為加密生態系統建立AI基礎設施：Ritual和Autonolas等新創公司專注於開發AI驅動的基礎設施，專為加密生態系統需求量身定制。

  2. Establishment of AI-based infrastructure for encryption ecosystems: new start-ups such as Ritual and Autonolas focus on the development of AI-driven infrastructure, tailored to the need for encryption systems.

• 3. 利用區塊鏈解決AI產業問題：Gensyn、EZKL和io.net等專案正在探討區塊鏈技術如何解決AI產業面臨的挑戰，例如資料隱私、安全性和透明度。

  3. Using sector chains to solve AI industry problems: projects such as Gensyn, EZKL and Io.net are exploring how sector chain technologies can address the challenges facing AI industries, such as data privacy, security and transparency.

AI x Crypto的獨特之處在於區塊鏈技術可望解決AI產業內在問題。這項獨特交匯點為創新解決方案開啟了新的可能性，有益於AI和區塊鏈社群。

AI x Crypto’s unique feature is that the sector chain technology is expected to solve the internal problems of the AI industry. This unique point of contact opens up new possibilities for a new solution that would benefit AI and the sector chain community.

在深入探討AI x Crypto領域時，我們旨在識別和展示區塊鏈技術在解決AI產業挑戰方面最有前景的應用。透過與AI產業專家和加密建造者合作，我們致力於促進充分利用兩種技術優勢的尖端解決方案的發展。

As we go further into the AI x Crypto field, we aim to identify and demonstrate the most promising applications of block-link technology in addressing the AI industry challenge. By working with AI industry specialists and encryption builders, we aim to promote the development of cutting-edge solutions that take full advantage of two technological advantages.

AI x Crypto領域可分為基礎設施和應用兩大類。儘管一些現有基礎設施持續為AI用例提供支持，但新的參與者正在市場上推出全新的AI原生架構。

The AI x Crypto domain can be divided into two main types of infrastructure and applications. Although some of the existing infrastructure continues to support the AI example, new participants are introducing a new original AI structure on the market.

1. 1計算網絡

在AIxCrypto領域，運算網路對於提供AI應用所需的基礎設施起著至關重要的作用。這些網路根據其支援的任務，可以分為兩種類型：通用計算網路和專用計算網路。

In the AixCrypto domain, the network plays a central role in providing the infrastructure needed for AI applications. These networks can be divided into two types: a universal network and an ad hoc network, depending on the mission they support.

1.1.1通用計算網絡

1.1 Generic Calculator Network

通用運算網路（例如IO.net 和Akash）為使用者提供透過SSH存取機器的機會，並提供命令列介面（CLI），使用戶能夠建立自己的應用程式。這些網路類似虛擬專用伺服器（VPS），在雲端提供個人運算環境。

Universal computing networks (such as IO.net and Akash) provide users with access to machines through SSH and command-line interfaces (CLI) that enable users to build their own applications. These networks are similar to virtual ad hoc servers (VPS) that provide personal computing environments on cloud ends.

IO.net基於Solana生態系統，著重於GPU租賃和運算集群，而基於Cosmos生態系統的Akash主要提供CPU雲端伺服器和各種應用模板。

IO.net is based on the Solana ecosystem and focuses on the GPU leasing and operation cluster, while Akash, based on the Cosmos ecosystem system, provides mainly CPU cloud servers and various applications templates.

IOSG Ventures 的看法：

IOSGVentures comment:

與成熟的Web2雲端市場相比，運算網路仍處於早期階段。 Web3運算網路不如Web2那種「樂高」建置模組，例如基於主要雲端服務供應商（如AWS、Azure和Google Cloud）的無伺服器函數、VPS和資料庫雲端專案。

The Web3 network is less developed than the Web2 " Lego" model, such as the non-server function, VPS, and database cloud projects based on major cloud service providers (such as AWS, Azure and Google Cloud) than the mature Web2 cloud market.

計算網路的優點包括：

The advantages of calculating the network include:

• 區塊鏈技術可以利用未使用的運算資源和個人電腦，使網路更具可持續性。
• 點對點（P2P）設計允許個體將未使用的計算資源貨幣化，並提供成本更低的計算，從而潛在地降低75%-90%的成本。
• 但由於以下挑戰，運算網路很難投入實際生產並取代Web2雲端服務：
• 定價固然是通用運算網路的主要優勢，但要與成熟的Web2雲端公司在功能、安全性和穩定性等方面競爭仍具挑戰性。
• 點對點風格可能會限制這些網路快速交付成熟和穩健產品的能力。去中心化的特性會額外增加開發和維護成本。

1.1.2專用計算網絡

1.1.2 special computing network

專用計算網路在通用計算網路基礎上增加了一個額外的層，使用戶可以透過設定檔部署特定的應用程式。這些網路旨在滿足特定的用例，例如3D渲染或AI推理和訓練。

Selective computing networks add an additional layer to the general computing network base to enable users to deploy specific applications through configuration files. These networks are designed to satisfy specific examples, such as 3D rendering or AI reasoning and training.

Render是一家專注於3D渲染的專業運算網路。在AI領域，像Bittensor、Hyperbolic、Ritual和fetch.ai這樣的新玩家專注於AI推理，而Flock和Gensyn主要專注於AI訓練。

Render is a professional computing network focused on 3D rendering. In the AI domain, new players like Bittensor, Hyperbolic, Ritual and Fetch.ai are focused on AI reasoning, while Flock and Gensyn are mainly focused on AI training.

IOSG Ventures 的看法：

IOSGVentures comment:

專用運算網路的優點：

Select the advantages of running the network:

• 去中心化和crypto 特性解決了AI 產業普遍存在的中心化和透明性問題。
• 無許可計算網路和驗證方案確保了推理和訓練過程的有效性。
• 隱私保護技術，如Flock採用的聯邦學習，允許個人為模型訓練貢獻數據，同時保持其數據在本地和私密。
• 透過支援智能合約與下游區塊鏈應用集成，使得AI推理可以直接在區塊鏈上使用。

 Source: IOSG Ventures

雖然專門的AI推理和訓練運算網路仍處於早期階段，但我們預期Web3 AI應用將優先使用Web3 AI基礎架構。這種趨勢已經在Story Protocol和Ritual與MyShell合作引入AI模型作為智慧財產權等合作中明顯。

Although the specialized AI reasoning and training network is still in its early stages, we expect Web3AI to use the Web3 AI infrastructure first and foremost. This trend is already evident in the collaboration of Stanley Protocol and Ritual with MyShell to introduce AI models as intellectual property rights.

儘管基於這些新興AI x Web3基礎設施構建的殺手級應用程式尚未出現，但成長潛力巨大。隨著生態系統的成熟，我們預計會看到更多利用去中心化AI運算網路獨特能力的創新應用程式。

Although the killer-level applications based on these new AI x Web3 infrastructures have not yet emerged, they have great potential for growth. As the ecosystem matures, we expect to see more innovative applications using the ability to decentralize AI’s network unique capabilities.

數據在AI模型中起著至關重要的作用，開發AI 模型的各個階段都涉及數據，包括數據收集、訓練數據集存儲和模型存儲。

The data play a critical role in the AI model, and each stage of the development of the AI model involves data collection, training data storage and model storage.

2.1資料存儲

去中心化儲存AI模型對於以去中心化方式提供推理API至關重要。推理節點應該能夠隨時從任何地方檢索這些模型。隨著AI模型可能達到數百GB的大小，需要一個強大的去中心化儲存網路。去中心化儲存領域的領導者，如Filecoin和Arweave，可能可以提供這項功能。

Going to centralize the ALI model is essential for decentralizing the reasoning API. The reasoning nodes should be able to retrieve the models from anywhere at any time. As the AI model may reach hundreds of GB sizes, it requires a strong decentralization of the network. The leaders of going centralize the storage domain, such as Filecoin and Arweave, may be able to provide this function.

IOSG Ventures 的看法：

IOSGVentures comment:

• 這個領域存在著巨大的機會。
• 針對AI模型優化的去中心化資料儲存網絡，提供版本控制、儲存不同的低精度模型量化以及大型模型快速下載等功能。
• 去中心化向量資料庫，因為它們經常與模型捆綁在一起，透過插入必要的與問題相關知識提供更準確的答案。現有的SQL資料庫也可以增加向量搜尋支援。

2.2資料收集和標記

收集高品質數據對於AI訓練至關重要。基於區塊鏈的項目如Grass利用眾包收集資料進行AI訓練，利用個人的網路。透過適當的激勵和機制，AI訓練者可以以較低的成本獲得高品質資料。 Tai-da和Saipen等項目則專注於數據標記。

The collection of high quality data is important for AI training. Projects based on sector chains, such as Grass, use public packages to collect data for AI training and personal networking. Through appropriate incentives and mechanisms, AI trainers can obtain high quality data at lower cost.

IOSG Ventures 的看法：

IOSGVentures comment:

我們對這個市場的一些觀察：

Some of our observations on the market:

• 大多數數據標記項目受到GameFi的啟發，吸引用戶以「標記賺取」概念和開發者以承諾為高品質標記數據降低成本。
• 目前在這一領域尚未出現明顯的領導者，而Scale AI主導著Web2資料標記市場。

2.3區塊鏈數據

在訓練專門針對區塊鏈的AI模型時，開發人員需要高品質的區塊鏈數據，希望能夠直接在其訓練過程中使用。 Spice AI和Space and Time提供具有SDK的高品質區塊鏈數據，使開發人員能夠輕鬆將數據整合到他們的訓練數據管道中。

Spice AI and Space and Time provide high quality chain data with SDK, allowing developers to easily integrate data into their training data pipelines.

IOSG Ventures 的看法：

IOSGVentures comment:

• 隨著對與區塊鏈相關的AI模型需求的增長，對高品質區塊鏈數據的需求將激增。然而，大多數數據分析工具目前只提供以CSV格式匯出數據，這對AI訓練目的並不理想。
• 為促進專門針對區塊鏈的AI模型的發展，至關重要的是透過提供更多與區塊鏈相關的機器學習運維（MLOP）功能來提升開發人員體驗。這些功能應該使開發人員能夠無縫地將區塊鏈資料直接整合到他們基於Python的AI訓練管線中。

由於有動機使用較不複雜的模型以減少計算成本，中心化的AI提供者面臨信任問題。例如，去年有時用戶認為ChatGPT表現不佳。後來這被歸因於OpenAI的更新旨在提升模型效能。

Centralized AI providers face trust problems because of the motivation to use less complicated models to reduce costing costs. For example, last year some users thought that ChatGPT was not doing well. This was later attributed to OpenAI's update aimed at improving model effectiveness.

此外，內容創作者對AI公司提出了版權擔憂。這些公司很難證明特定數據未包含在其訓練過程中。

In addition, content creators have raised concerns about the copyright of AI companies. These companies have difficulty proving that certain data are not included in their training processes.

零知識機器學習（ZKML）是一種創新方法，解決了與中心化人工智慧提供者相關的信任問題。透過利用零知識證明，ZKML使開發人員能夠證明其人工智慧訓練和推理過程的正確性，而無需洩露敏感資料或模型細節。

ZKML is a creative approach that addresses trust issues associated with centralized artificial intelligence providers. Through the use of zero knowledge, ZKML enables developers to prove the correctness of their AI training and reasoning processes without revealing sensitive data or model details.

3.1 訓練

開發人員可以在零知識虛擬機器（ZKVM）中執行訓練任務，例如由Risc Zero 提供的虛擬機器。該過程產生一個證明，驗證訓練是否正確進行，且僅使用了經授權的資料。該證明作為開發人員遵守適當訓練規範和資料使用權限的證據。

Developers can perform training tasks in ZKVM, such as those provided by Risc Zero. This process should result in a demonstration that the training is being conducted correctly and that only the authorized data are being used. It should serve as evidence that developers are complying with proper training and access to data.

IOSG Ventures 的看法：

IOSGVentures comment:

• ZKML為證明模型訓練中授權資料使用提供了獨特解決方案，而這在人工智慧模型的黑盒子特性下通常很難實現。
• 這項技術仍處於早期階段。計算開銷巨大。社群正在積極探索ZK 訓練更多用例。

3.2推理

與其訓練對應物相比，ZKML 用於推理的時間要長得多。這個領域已經有幾家知名公司湧現，它們各自採用獨特方法使機器學習推理變得不信任和透明。

ZKML takes much longer to reason than its training counterparts. This field has already seen several well-known companies that have used their own unique methods to make machine-learning reasoning untrustworthy and transparent.

Giza專注於建立全面的機器學習營運（MLOP）平台，並在其周圍打造一個充滿活力的社群。他們的目標是為開發人員提供整合ZKML 到推理工作流程的工具和資源。

Giza focuses on building a comprehensive machine learning platform (MLOP) and building a vibrant community around it. Their goal is to provide developers with tools and resources to integrate ZKML to reasoning work processes.

另一方面，EZKL 透過創建用戶友好的ZKML 框架以提供良好的效能，優先考慮開發體驗。他們的解決方案旨在簡化實現ZKML 推理的過程，使更多開發人員能夠輕鬆使用。

On the other hand, EZKL provides good efficiency through a user-friendly ZKML framework, giving priority to developing experiences. Their solution is to simplify the process of implementing ZKML reasoning and make it easier for more developers to use.

Modulus Labs 採用了不同的方法，他們開發了自己的證明系統。他們的主要目標是顯著減少與ZKML 推理相關的計算開銷。透過將開銷降低10倍，Modulus Labs 試圖使ZKML 推理在實際應用上更具實用性和效率。

Modulus Labs uses different methods, and they develop their own systems of proof. Their main goal is to significantly reduce the calculation of the costs associated with ZKML reasoning. By reducing the costs by 10 times, Modulus Labs is trying to make ZKML reasoning more practical and efficient in practice.

IOSG Ventures 的看法：

IOSGVentures comment:

• ZKML 特別適用於GameFi 和DeFi 場景，其中無需信任至關重要。
• ZKML 引入的運算開銷使大型人工智慧模型難以有效運作。
• 該行業仍在尋找在其產品中大量使用ZKML 的DeFi 和GameFi 先驅，以展示其實際應用情境。

4.1 代理網絡

代理網絡由配備執行特定任務的工具和知識的眾多人工智慧代理組成，例如協助進行鏈上交易。這些代理可以相互協作以實現更複雜的目標。幾家知名公司正在積極開發類似聊天機器人的代理和代理網絡。

The agency network is made up of a wide range of artificial intelligence agents that are equipped with tools and knowledge to carry out specific tasks, such as facilitating chain transactions. These agents can work together to achieve more complex goals. Several well-known companies are developing proxies and proxies like chat robots.

Sleepless、Siya、Myshell、characterX 和Delysium是正在建立聊天機器人代理的重要參與者。 Autonolas 和ChainML 正在為更強大的用例建立代理網路。

Sleepless, Siya, Myshell, CharacterX, and Delysium are important participants in the ongoing establishment of chat machine agents. Autonolas and ChainML are creating proxy networks for stronger examples.

IOSG Ventures 的看法：

IOSGVentures comment:

• 代理對於現實世界的應用至關重要。它們可以比通用人工智慧更好地執行特定任務。區塊鏈為人工智慧代理提供了幾種獨特機會。
• 擁有激勵機制：區塊鏈透過諸如非同質化代幣（NFT）等技術提供激勵機制。有了明確的所有權和激勵結構，創作者受到激勵，在鏈上開發更有趣和創新的代理商。
• 智能合約的可組合性：區塊鏈上的智能合約高度可組合，像樂高積木一樣運作。智慧合約提供的開放API使代理商能夠執行在傳統金融系統中難以實現的複雜任務。這種可組合性使代理能夠與各種去中心化應用程式（dApps）進行互動並利用其功能。
• 內在的開放性：透過在區塊鏈上建立代理，它們繼承了這些網路的內在開放性和透明性。這為不同代理之間的可組合性創造了重大機遇，使它們能夠合作並結合各自的能力來解決更複雜的任務。

4.2 其他應用

除了前面討論過的主要類別外，在Web3 領域中還有幾個有趣的人工智慧應用正在受到關注，儘管它們可能還不夠龐大以形成獨立的類別。這些應用跨越各種領域，展示了人工智慧在區塊鏈生態系統中的多樣性和潛力。

In addition to the major categories discussed earlier, interesting applications of artificial intelligence in the Web3 domain are receiving attention, even though they may not be large enough to form an independent category. These applications span a variety of domains, demonstrating the multiplicity and potential of artificial intelligence in a regional chain system.

影像生成：ImgnAI

Image generation: ImgnAI

影像提示變現：NFPrompt

Image hint change: NFPrompt

社區訓練的人工智慧圖像生成：Botto

Community-trained artificial intelligence image generation: Botto

聊天機器人：Kaito、Supersight、Galaxy、Knn3、Awesome QA、Qna3

Chat robots: Kaito, Supersight, Galaxy, Knn3, Awesome QA, Qna3

金融：Numer AI

Finance: Numer AI

錢包：Dawn_wallet

Wallet: Dawn_wallet

遊戲：Parallel TCG

Game: Parallel TPG

教育：Hooked

Education: Hooked

安全：Forta

Security: Forta

DID：Worldcoin

創作者工具：Plai Lab

Creative tool: Plai Lab

AI x Crypto 之所以獨一無二，是因為它可以解決人工智慧領域最困難的問題。儘管目前的AIxCrypto 產品與Web2 AI 產品之間存在差距，並且對Web2 用戶缺乏吸引力，但AIxCrypto 仍具備一些獨特功能，只有AIxCrypto 才能提供。

AI x Crypto is unique because it can solve the most difficult problems in an artificial intelligence domain. Although there is a gap between the current AixCrypto and Web2AI products and is not attractive to Web2 users, AixCrypto still has a number of unique features that only AixCrypto can provide.

5.1高性價比的計算資源:

AIxCrypto 的一個主要優勢在於提供高性價比的運算資源。隨著對LLM 的需求增加，市場上開發者增多，GPU 的可用性和價格變得更具挑戰性。 GPU 價格大幅上漲，且短缺。

One of the main advantages of AixCrypto is to provide a high-priced operating resource. As demand for LLM increases, market developers increase, and GPU availability and prices become more challenging.

DePIN 專案等去中心化運算網路可以透過利用閒置運算力、小型資料中心的GPU和個人運算設備來幫助緩解這個問題。雖然去中心化運算功率的穩定性可能不如集中式雲端服務，但這些網路提供了多樣化地域的高性價比運算設備。這種去中心化方法最小化了邊緣延遲，確保了更分散和更有彈性的基礎設施。

DePIN projects, etc., can help alleviate the problem by using idle computing, small data center GPUs, and personal computing devices. While decentralizing power may not be as stable as centralized cloud services, these networks provide multi-dimensional high-cost computing facilities. This decentralisation method minimizes marginalization and ensures more decentralized and resilient infrastructure.

透過利用去中心化運算網路的力量，AIxCrypto 可以為Web2 使用者提供價格實惠、易得的運算資源。這種成本優勢對於吸引Web2 用戶採納AIxCrypto 解決方案具有吸引力，並且尤其在對AI計算的需求持續增長的情況下。

Through the use of decentralised networks, AixCrypto can provide Web2 users with affordable and easy-to-reach resources. This cost advantage is attractive for attracting Web2 users to adopt the AixCrypto solution, especially in the context of growing demand for AI calculations.

5.2賦予創作者所有權:

AI x Crypto 的另一個重要優點在於保護創作者的所有權權利。在當前的人工智慧領域，有些代理容易被複製。透過簡單編寫類似提示，就可以輕鬆複製這些代理程式。此外，GPT 商店中的代理商通常由中心化公司擁有，而不是由創作者擁有，限制了創作者對作品的控制以及有效實現盈利的能力。

Another important advantage of AI x Crypto is that it protects all rights of creators. In the current artificial intelligence domain, some agents are easily copied. It is easy to copy these agents through simple writing-like tips. Moreover, agents in GPT shops are usually owned by centralized companies rather than by creators, limiting the ability of creators to control their works and to realize their profits effectively.

AI x Crypto 利用加密領域普遍存在的成熟NFT 技術來解決這個問題。透過將代理人表示為NFT，創作者可以真正擁有他們的作品，並從中獲得實際收益。每次用戶與代理商互動，創作者都可以獲得激勵，確保對他們努力的公平回報。基於NFT 所有權的概念不僅適用於代理，還可用於保護人工智慧領域中的其他重要資產，如知識庫和提示。

AI x Crypto uses mature NFT techniques that are prevalent in encryption domains to solve this problem. By presenting the agent as NFT, the creator can really own their work and derive real benefits from it. Each time users and agents interact, the creator can be motivated to ensure fair rewards for their efforts. The concept of NFT ownership is not only suitable for agency, but also for protecting other important assets in artificial intelligence fields, such as knowledge banks and tips.

5.3保護隱私並重建信任:

用戶和創作者對於中心化人工智慧公司存在隱私擔憂。使用者擔心自己的資料被濫用用於訓練未來模型，而創作者則擔心自己的作品被使用但卻缺乏適當的歸因或補償。此外，中心化人工智慧公司可能會犧牲服務品質以降低基礎設施成本。

Users and creators have privacy concerns about centralized AI companies. Users fear that their data will be misused to train future models, while creators fear that their work will be used without proper attribution or compensation.

這些問題難以透過Web2 技術解決，而AIxCrypto 則利用先進的Web3 解決方案。零知識訓練和推理可透過證明所使用的數據和確保應用正確模型，從而提供透明度。諸如受信任執行環境(TEE)、聯邦學習和完全同態加密(FHE)等技術實現安全、保護隱私的人工智慧訓練和推理。

It is difficult to solve these problems through Web2 technology, while AixCrypto uses advanced Web3 solutions. Zero knowledge training and reasoning can provide transparency by proving the data and ensuring the application of correct models.

透過優先考慮隱私和透明度，AIxCrypto 使人工智慧公司能夠重新獲得公眾信任，並提供尊重用戶權利的人工智慧服務，使其區別於傳統的Web2 解決方案。

Through preferential consideration of privacy and transparency, AixCrypto enables AIJ to regain public trust and provide artificial intelligence services that respect user rights, distinguishing it from the traditional Web2 solution.

5.3保護隱私並重建信任:

用戶和創作者對於中心化人工智慧公司存在隱私擔憂。使用者擔心自己的資料被濫用用於訓練未來模型，而創作者則擔心自己的作品被使用但卻缺乏適當的歸因或補償。此外，中心化人工智慧公司可能會犧牲服務品質以降低基礎設施成本。

Users and creators have privacy concerns about centralized AI companies. Users fear that their data will be misused to train future models, while creators fear that their work will be used without proper attribution or compensation.

這些問題難以透過Web2 技術解決，而AIxCrypto 則利用先進的Web3 解決方案。零知識訓練和推理可透過證明所使用的數據和確保應用正確模型，從而提供透明度。諸如受信任執行環境(TEE)、聯邦學習和完全同態加密(FHE)等技術實現安全、保護隱私的人工智慧訓練和推理。

It is difficult to solve these problems through Web2 technology, while AixCrypto uses advanced Web3 solutions. Zero knowledge training and reasoning can provide transparency by proving the data and ensuring the application of correct models.

透過優先考慮隱私和透明度，AIxCrypto 使人工智慧公司能夠重新獲得公眾信任，並提供尊重用戶權利的人工智慧服務，使其區別於傳統的Web2 解決方案。

Through preferential consideration of privacy and transparency, AixCrypto enables AIJ to regain public trust and provide artificial intelligence services that respect user rights, distinguishing it from the traditional Web2 solution.

5.4追蹤內容來源

隨著人工智慧生成的內容日益精密，區分人類創作和人工智慧生成的文字、圖像或影片變得更加困難。為防止濫用人工智慧產生的內容，人們需要一種可靠的方式來確定內容的來源。

As the content generated by artificial intelligence becomes more sophisticated, it becomes more difficult to distinguish between words, images or videos produced by human creation and artificial intelligence. To prevent the misuse of the content generated by artificial intelligence, people need a reliable way to determine the origin of the content.

區塊鏈在追蹤內容來源方面表現出色，就像在供應鏈管理和NFT 中取得的成功一樣。在供應鏈產業，區塊鏈追蹤產品的整個生命週期，用戶可以識別生產商和關鍵里程碑。同樣地，區塊鏈追蹤創作者，並在NFT 的情況下防止盜版，由於其公開性，NFT 尤其容易受盜版的影響。儘管存在這種脆弱性，利用區塊鏈可最大程度減少假NFT 所導致的損失，因為用戶可輕鬆區分真假代幣。

Block chains perform well in tracking content sources, as they do in supply chain management and NFT. In the supply chain industry, the chain tracks products throughout their life cycle, users can identify producers and key milestones. Similarly, block chains track creators and, in the case of NFTs, prevent pirated versions from becoming open, NFTs are particularly vulnerable to pirated versions. Despite this vulnerability, the use of the chain can minimize losses caused by fake NFTs, as users can easily divide counterfeit currency.

透過應用區塊鏈技術追蹤人工智慧生成內容的來源，AIxCrypto 可為用戶提供驗證內容創作者是人工智慧還是人類的能力，從而減少濫用可能性，增加對內容真實性的信任。

AixCrypto can provide users with the ability to verify whether the creator of the content is artificial intelligence or human, by tracking the source of the content generated by artificial intelligence through the application chain technology, thereby reducing the likelihood of misuse and increasing confidence in the authenticity of the content.

5.5利用加密貨幣開發模型

設計和訓練模型，特別是大型模型，是一個昂貴且耗時的過程。新模型還存在不確定性，開發人員無法預測其性能。

Designing and training models, especially large models, is an expensive and time-consuming process. The new models are uncertain and developers cannot predict their performance.

加密貨幣提供了一個對開發人員友好的方式，可以收集預訓練資料、收集強化學習回饋以及從感興趣的方進行籌款。這個過程類似於典型加密貨幣專案的生命週期：透過私人投資或起飛台籌資，並在啟動時向活躍貢獻者投放代幣。

Encrypted currency provides a friendly way for developers to collect pre-training data, collect back-to-back chemical studies, and raise funds from interested parties. This process is similar to the life cycle of a typical encrypted currency project: to raise funds through private investment or flight platforms and to send money to life donors at start-up.

模型可以採用類似方法，透過出售代幣籌集資金用於訓練，並向數據和回饋的貢獻者空投代幣。透過精心設計的代幣經濟模型，這個工作流程可幫助個人開發人員比以往更輕鬆地訓練新模型。

Models can be used in a similar way to raise money through the sale of tokens for training purposes, and to air-drop money to contributors of data and feedback. Through well-designed proxy economic models, this process can help individuals develop new models more easily than ever before.

AI x Crypto計畫開始瞄準Web2開發者作為潛在客戶，因為加密有獨特的價值主張，且Web2人工智慧產業市場規模可觀。然而，對於不熟悉代幣且不願涉足基於代幣系統的Web2開發者來說，代幣可能成為一道障礙。

AI x Crypto started targeting Web2 developers as potential clients, because encryption has a unique value profile and the Web2 artificial intelligence market is impressive. However, for Web2 developers who are not familiar with and do not want to be involved in the system, it may become an obstacle.

為了迎合Web2開發者，減少或去除代幣的實用性可能對於Web3愛好者造成困擾，因為這可能會改變AIxCrypto專案的根本立場。在努力將有價值的代幣整合到人工智慧SaaS平台時，找到吸引Web2開發者並保持代幣實用性之間的平衡是一個具有挑戰性的任務。

To cater to Web2 developers, reducing or wiping out the use of the money may cause trouble for Web3 lovers, as it may change the fundamentals of the AixCrypto project. Finding a balance between attracting Web2 developers and maintaining the use of the money is a challenging task as efforts are made to integrate valuable coins into artificial intelligence SaaS platforms.

為了彌合Web2和Web3商業模式之間的差距，並同時保持代幣價值，可以考慮以下幾種潛在方法：

In order to bridge the gap between the Web2 and Web3 business models and at the same time maintain the value of the intergenerational currency, the following potential approaches could be considered:

• 在專案的分散式基礎設施網路中利用代幣。實施質押、獎勵和懲罰機制來保護基礎網路。
• 將代幣作為支付方式，同時為Web2用戶提供使用入口
• 實施基於代幣的治理
• 與代幣持有者分享收益
• 利用收益回購或銷毀代幣
• 為專案提供的服務，為代幣持有者提供折扣和額外功能
• 透過精心設計符合Web2和Web3利益的代幣經濟模型，AIxCrypto專案可以成功吸引Web2開發者，同時保持其代幣的價值和實用性。

我們最愛的AIxCrypto場景利用了用戶協作的力量，借助區塊鏈技術在人工智慧領域完成任務。一些具體的例子包括：

Our favorite AixCrypto scene uses the power of user collaboration to accomplish its mission in an artificial intelligence field, with the help of block-link technology. Some specific examples include:

1. 集體進行AI訓練、Alignment 和基準測試的資料貢獻（例如Chatbot Arena）

1. Collective data contributions to AI training, Alimentation and baseline testing (e.g. Chatbot Arena)

2. 合作建構一個大型共享知識庫，可供各種代理人使用（例如，Sahara）

2. Collaboratively build a large shared knowledge bank for use by a variety of agents (e.g. Sahara)

3. 利用個人資源，進行網路資料抓取（例如，Grass）

3. Use of personal resources for online data capture (e.g. Grass)

透過利用基於區塊鏈激勵和協調的用戶集體努力，這些模型展示了去中心化、社群驅動的方法對AI開發和部署的潛力。

These models demonstrate the potential of decentralized, community-driven approaches to AI development and deployment through the use of user-based efforts based on block-links to stimulate and coordinate.

我們正處於AI和Web3的黎明階段，與其他產業相比，人工智慧與區塊鏈領域的整合仍處於早期階段。在排名前50名的Gen AI產品中，並沒有與Web3相關的產品。頂尖的LLM工具與內容創作和編輯相關，主要針對銷售、會議和筆記/知識庫。考慮到Web3生態系統中大量的研究、文件、銷售和社群工作，為客製化的LLM工具的開發提供了巨大的潛力。

We are in the dawn phase of AI and Web3 and, compared to other industries, the integration of artificial intelligence and sector chains is still in the early stages. Among the top 50 Gen AI products, there is no web3-related product. The top LLM tools are related to content creation and editing, focusing on sales, conferences and notes/knowledge banks. Considering the extensive research, documentation, marketing and community work in the Web3 system, they offer enormous potential for the development of customized LLM tools.

目前，開發者正專注於建立基礎設施，將先進的AI模型引入鏈上，雖然我們尚未達到目標。隨著我們繼續發展這項基礎設施，我們也在探索最佳用戶場景，以安全和無需信任的方式在鏈上進行AI推理，這為區塊鏈領域提供了獨特機會。其他產業可以直接使用現有的LLM基礎設施進行推理和微調。只有區塊鏈產業需要自己的原生AI基礎設施。

Presently, developers are focusing on building infrastructure to introduce advanced AI models into the chain, although we have not yet reached our goal. As we continue to develop this foundation, we are also exploring the best user scenes, using AI reasoning on the chain in a safe and untrustworthy manner, which offers unique opportunities for the regional chain. Other industries can directly use the existing LLM infrastructure for reasoning and fine-tuning.

在不久的將來，我們預計區塊鏈技術將利用其點對點的優勢來解決人工智慧行業中最具挑戰性的問題，使AI模型對每個人都更加負擔得起、易於訪問和盈利。我們也期待加密領域將跟隨AI產業的敘事，儘管略有延遲。在過去一年中，我們見證了開發者將Crypto，代理和LLM 模型結合。在接下來的幾個月內，我們可能會看到更多多模態模型、文字視訊生成和3D生成影響Crypto 領域。

In the near future, we anticipate that the sector chain technology will use its point-to-point advantage to address the most challenging issues in the artificial intelligence industry, making the AI model more affordable, accessible and profitable for everyone. We also expect the encryption domain to follow the AI industry, albeit with a slight delay. In the past year, we have witnessed developers combining the Crypto, Agent, and LLM models. In the coming months, we may see more multi-model models, text video generation and 3D production affecting the Crypto domain.

整個AI和Web3產業目前並未得到充分的重視，我們迫切期待AI在Web3 中的引爆時刻，一個CryptoxAI 的殺手級應用。

The entire AI and Web3 industry is currently not receiving sufficient attention, and we are eagerly looking forward to the use of a CryptoxAI killer at the time of the AI detonation in Web3.

欧易(OKX)最新版本

【遇到注册下载问题请加文章最下面的客服微信】永久享受返佣20%手续费！

APP下载   全球官网 大陆官网

币安(Binance)最新版本

币安交易所app【遇到注册下载问题请加文章最下面的客服微信】永久享受返佣20%手续费！

APP下载   官网地址

火币HTX最新版本

火币老牌交易所【遇到注册下载问题请加文章最下面的客服微信】永久享受返佣20%手续费！

APP下载   官网地址