CVS

CVS

substack
twitter
HomeArchives

Does Web3 AI Chatbot have any prospects?

119
AI Translation
This post is translated from Chinese into English through AI.View Original
AI-generated summary
Recently, a Web3 AI Chatbot was created and the user wants to compare it with other similar applications. The current chatbot uses similar technology, which involves gathering information, saving it, generating embeddings, and using similarity queries to find relevant information. The chatbot then uses this information to generate answers for users. The use cases for the chatbot include obtaining market data, learning about individual tokens or projects, learning industry knowledge, summarizing news, and identifying trends. However, the chatbot faces challenges in terms of its advantages compared to other applications. For market data, existing platforms like CMC can be used. In terms of industry knowledge and news summarization, the chatbot may not have an advantage compared to optimized language models from large companies. For in-depth research on individual projects or tokens, the chatbot may not be cost-effective due to high data collection costs. The chatbot may be suitable for capturing short-term trends or predicting medium-term trends, but it requires high levels of refinement. Overall, the Web3 AI Chatbot's positioning is slightly awkward due to alternative platforms and the optimization of language models by large companies. The summary also provides a list of related products for reference, including AlphaRushAI, Inflation Universe, ChainGPT, Web3Chat, and Web3 Analytics AI.

Recently, I created a Web3 AI Chatbot (https://app.web3chat.info/) and wanted to see other similar applications for comparison.
Currently, chatbots use similar technologies, which involve information gathering, saving, generating embeddings, converting user queries into embeddings to query for similarity, and returning the generated answers to the user.

The main use cases are as follows:

  1. Getting market data: querying token prices, fluctuations, displaying token lists, etc.
  2. Understanding individual tokens or projects: providing introductions to individual tokens or projects, including background, direction, ecosystem projects, token economics, etc. Suitable for quickly understanding projects and conducting in-depth research.
  3. Learning industry knowledge: learning about Web3 and blockchain-related concepts.
  4. News aggregation: aggregating the latest news in the industry or specific fields.
  5. Latest trends: keeping up with the latest hot topics and general demands.

Currently, Web3 Chatbot has roughly these user demands. The problem is that chatbots in these areas may not have any advantages compared to other applications. For market data, it is sufficient to use market data websites like CMC. For industry knowledge and news aggregation, with outdated LLM training data, there may still be some practical value with new message data. However, with the optimization of language models by big companies, such as ChatGPT plugins or Google's built-in search function, these two functions may no longer have an advantage and may even be less effective. For the demand of researching individual projects or tokens, the main users are investment researchers and airdrop hunters. There is a problem of high data collection costs, and it may not be worth it to have a chatbot with in-depth analysis capabilities. For the demand of the latest trends, it can be divided into two categories. One is the demand for trend opportunities with high timeliness, such as shitcoin traders who engage in short-term speculation of low-market-cap altcoins. For these users, conversation is not a good form, and it is better to passively receive notifications, such as real-time price fluctuation alerts. The second category is the demand for medium-term trend analysis, mainly predicting possible themes for the next few weeks or months. This demand may be met by a chatbot, but it would require high cost to do it well.

Currently, Web3 AI Chatbot's positioning is somewhat awkward, with market data platforms and Web3 media applications as alternatives, and big companies optimizing their models. In a field like Web3 where data is publicly available, it is difficult to create differentiated features using LLM.

Below is a summary of the current related products for reference.

AlphaRushAI#

Website: https://alpharush.ai/

You can interact with the chatbot on Discord.
The threshold is high, requiring at least 150M RUSHAI tokens (worth more than $1000 😓) to qualify for interaction.
Due to the threshold issue, I haven't personally tested it. However, from the official documentation (https://alpharushai.gitbook.io/index/chat-with-jennaai/jenna-capability), I can see some of its features, including:

  1. Event tracking, such as providing a schedule of token unlocks for the next week.
  2. Individual token or project descriptions.
  3. Listing ecosystem projects, such as projects on StarkNet.
  4. Token prices and fluctuations.
  5. Drawing candlestick charts.

Based on the displayed effects in the documentation, it seems to collect specified Twitter messages, some project documentation, event calendars, and token price data.

Inflation Universe#

Website: https://www.infverse.io/

The feature of this chatbot is that it not only answers questions but also presents recent related news in the form of tables or summarizes the answers into tables.
However, the messages in the tables do not have links, so their accuracy cannot be verified. Sometimes, the listed messages are unrelated to the questions, and there is no continuous conversation feature.
image

ChainGPT#

Website: https://app.chaingpt.org/

It can be used for knowledge Q&A, Solidity contract writing, and review.
The interface is a bit rough, and there is no continuous conversation feature.
image

Web3Chat#

Website: https://web3chat.info/

Chatbot developed by myself.

Interaction is available after wallet login. The advantage is that it lists the message sources for verification purposes. Messages are tied to the wallet, so you can view the chat history after logging in. Messages are saved locally and not stored on the server, providing better privacy protection.
There are two modes: News Mode for getting the latest news and Project Mode for targeted understanding of individual project information.
There is a preliminary memory function and the ability for continuous conversation, but the effectiveness is average. Currently, project information is not updated in real-time.
image

Web3 Analytics AI#

Website: https://3gpt.ai/

The replies from this chatbot are annotated and displayed in a unified format, making it convenient to view the information sources.
The replies to recent news seem to need further optimization. The chat content does not follow the wallet. The functions shown in the video, such as tables, trend charts, and code replies, seem to be unfinished. The swap function has not been implemented yet.

image

Loading...
Ownership of this post data is guaranteed by blockchain and smart contracts to the creator alone.