Hitting the Pain Point: We Desperately Need a Web3 AI Search Engine

Web3 is booming, and Arweave is becoming a popular infrastructure choice for developers. PermaDAO is a community where everyone can contribute to the Arweave ecosystem. It's a place to propose and tackle tasks related to Arweave, with the support and feedback of the entire community. Join PermaDAO and help shape Web3!

Author: Marshal Orange @ Contributor of PermaDAO

Translator: Marshal Orange @ Contributor of PermaDAO

Reviewer: John Khor @ Contributor of PermaDAO

Hitting the Pain Point: We Desperately Need a Web3 AI Search Engine

I'm sure everyone is already quite comfortable using AI tools in their daily lives. However, if I were to ask you to quickly compile a comprehensive history and technical explanation of Ethereum's scalability solutions, covering all the technical details I want to know, such as combining on-chain data analysis to achieve trustless mechanisms with Plasma and discussing the similarities and differences with Rollup, or delving into the entire process of achieving data availability for all scalability solutions, even if you use various prompts and whitepapers, what AI is currently available to help you gather this information?

The answer is likely none. Even if there were, it might only be because Ethereum's development is relatively mature, and some AI models have already pre-captured sources of detailed technical information such as Vitalik Buterin's papers and Ethereum developer community proposals. But what if we switch to a different ecosystem? For example, let's talk about the significance of the Arweave 2.7 upgrade. I tested ChatGPT, Microsoft Bing AI, YOU, and other AI tools. They couldn't retrieve the content from Arweave's official Medium, Mirror, and PermaDAO (Arweave's co-builders community) without specific guidance. In fact, they provided less effective information than what I could find with a simple Google search. Moreover, existing AI tools are inadequate for researchers or developers who want to delve deeper into organizing and analyzing on-chain data.

We Really Need a Web3 AI Search Engine!

Information retrieval in the Web3 space is currently a widespread pain point. To better understand this issue, let's first explore the typical process of information gathering in the field of blockchain research. I've also interviewed some friends who work as full-time blockchain researchers to understand their approach when researching a new project or technology in the blockchain space. Here's the basic process they follow for information gathering (the steps are not in a specific order, and the following are just some typical steps in the information search process in investment research, not representing the entire process):

• Find the official social media accounts of a project, their official website, and blogs published on their official website.

• Visit the project's official website to search for product features, whitepapers, product roadmaps, and team members.

• Analyze the product whitepaper and search for related literature based on the underlying architecture and technology mentioned in the whitepaper.

• Read blogs to understand the project's iteration history and roadmap.

• Retrieve information about funding and valuation.

• Collect background information about the team, including the past experiences of core team members.

• Join official communities to gather user sentiment, learn about community promotion, and assess brand marketing capabilities.

• Organize on-chain data related to the project, analyzing metrics such as Total Value Locked (TVL), tokenomics, and daily trading volume.

The above processes basically require the researchers to manually search in order to carry out the process, because none of the AI currently on the market include complete information sources and on-chain data about Web3, nor do they follow Web3's investment and research logic, and the researchers still spend a lot of time in the process to collect enough information.

However, this may be due to the relative independence of various ecosystems within Web3, with different technical standards for each blockchain network. This makes it challenging to establish a standardized SOP for information mining and research. When I was researching how AI can better empower Web3, I had the chance to meet Joe, the Chief Marketing Officer of Adot, during a presentation at PermaDAO and PANews. We discussed AI topics after the event, and it was a lively conversation that continued into the late hours.

It was only then that I realized they were doing exactly what I had been hoping for. Joe gave me an invitation code to participate in the Adot closed beta, and I couldn't wait to try their product. I first asked Adot the question I mentioned earlier, "The significance of the Arweave 2.7 upgrade." I was genuinely surprised by the clarity and accuracy of its response. Not only does it integrate with Google and Medium, but it even includes content from Arweave's Twitter!

Adot AI is essentially an AI-powered Web3 search engine. After testing it, I can confirm that it can indeed search and compile a vast amount of authentic Web3 data to provide highly precise insights. Moreover, each paragraph in its responses is source-annotated, eliminating the need for me to constantly verify its objectivity and accuracy, as is the case with other AI products (specifically referring to ChatGPT). Adot combines technologies such as web crawling, search ranking, web page processing, big data handling, natural language processing, and AI to deliver personalized, intelligent, reliable, and real-time search results to users. I believe that in the future, when I seek information about Web3, I won't have to spend hours sifting through irrelevant search results, reading thousands of webpages on Google, or relying on ChatGPT for unreliable and outdated information.

I've always had the habit of digging deep (perhaps it's a bad habit), so I, along with my researcher friends, analyzed several key characteristics of Adot:

(1) Adot currently has no business competitors: Adot has successfully differentiated itself from the run-of-the-mill AI products. It covers Web3 data that cannot be found through Google search. This data is sourced from social media, blockchain explorers, blog posts (Mirror, Medium), and official developer communities, among others. Google may only cover 10% of the information within Web3, while Adot focuses on the remaining 90% that currently resides in the darknet (Web3 content that cannot be directly searched for). For users, Adot has genuinely light a lamp in the dark forest of Web3!

(2) Adot benefits all stakeholders in the Web3 ecosystem: Remember, Adot is not just any generative AI; it's a Web3 search engine! The innovative business model of "Web3 + AI + Search Engine" directly benefits users, project teams, social media platforms, and all other stakeholders. Users get the information they desire through searches, project teams successfully promote their projects, and the search engine's revenue model is complex and diverse. It includes advertising fees, paid rankings, data collection and analysis, cooperative marketing, API fees, brand promotion, knowledge monetization, and more. This is just a small part of it. In addition, the collaborative advantages of Web3 will also bring income to developers who contribute to data analysis, plugins, proxies for the AI search engine, and supporting creators.

My conversation with Joe was delightful, and he revealed that their team at Adot is a startup originating from Google. They have already partnered with Arweave, Tezos, Koii Network, NFTScan, FireFly, Weweave, and Chainbase, and they are preparing to participate in various ecosystem hackathons. This makes me very optimistic about the future of Adot. The web version of Adot AI is currently in internal testing and has been previewed on Product Hunt. Additionally, the Google plugin version is about to be released in the next update.

Moreover, as Joe disclosed, there is a high probability of Adot offering airdrop opportunities. The official website has already launched the "Rewards Ranking" feature, which displays the points of the top 20 Adot users. Users earn points for using Adot for searches or contributing, although the specific use of these points has not been officially announced. The feature is described as a predefined reward mechanism, intended to benefit data contributors and search users. Therefore, in the future token economic model, I personally believe that points could be used to distribute tokens to early supporters of the project.

Provoke Thinking: Web3 is the Ultimate Form of AI Search Engine Development

This is not a promotional article!

Because Adot is undoubtedly unprecedented, and its significance for Web3 cannot be summarized in just a few sentences!

In the realm of AI-driven search engines, some companies have made attempts, such as Perplexity, You.com, AndiSearch, and various data analysis firms. However, none of them have explored Web3 data; only Adot is focused on building an AI search engine for Web3.

Building an AI search engine involves substantial costs, with neural network computations forming the technical backbone of AI. Every user search requires extensive parallel processing by neural networks. Most data centers worldwide still primarily use CPUs as their main computing units, and they are less efficient in parallel processing compared to GPUs of similar capability. This leads to high operational costs for AI.

As more search engines embrace the trend of Web3, it may reshape the entire narrative of the AI business from data and algorithm ownership. This is fundamentally different from centralized projects. It's not just a scenario for paid AI applications. Web3 inherently offers a more thoroughly efficient open-source approach. Coupled with incentives from token economics and content creator copyright economies, it provides a fertile ground for data module owners, algorithm creators, and computational power nodes, making it easier to attract nodes and users to participate.

The emergence of Adot may signify a formal challenge to Google and Yahoo. Adot is not only leading a new revenue stream for search engines but also pioneering a new generation of Web3 brand promotion. This entails Web3 project teams paying fees to Web3 AI search engine providers to showcase the content of decentralized applications or networks.

Prior to this, project teams might have needed to hire seasoned SEO professionals and data analysts to optimize search rankings, or spend substantial amounts on SEM (Search Engine Marketing) bidding ads to gain product exposure. The consequence of these approaches is the proliferation of homogeneous and one-sided low-information-entropy content on the internet, making it even more challenging to conduct Web3 information searches amid this information noise.

Web3 + AI search engines are reshaping the internet landscape, where people can once again experience the feeling of early internet search engines—directly asking a question and receiving a clear and intuitive result. The development of Web3 in various verticals allows more and more non-professionals to swiftly access detailed Web3 information, breaking down the barriers of the "information gap." As users, every AI product is worth trying, especially AI search engines, which make us contemplate their potential and value for Web3. At least in my view, Web3 unquestionably needs an AI search engine like this, and Adot is bound to become your all-in-one Web3 AI assistant!

🔗 More about PermaDAO ：Website | Twitter | Telegram | Discord | Medium | Youtube