AI Finance, abbreviated as AIFi, refers to the use of large language models (LLMs) to replace and augment traditional financial components. For example, using an LLM as an alternative to conventional trading models exemplifies a typical AIFi application.Among the many AIFi projects, BubbleAI, an emerging standout, has recently attracted substantial attention in both finance andweb3 circles. Having trialled the beta version myself, I have heightened expectations for the product’s official launch. BubbleAI may well serve as the nexus between traditional finance and web3 financial systems.
From my BubbleAI beta test, it is evident that the product’s logic can be divided into three segments: Data – Data Analysis – Trading tool. Accordingly, my overview of the product will encapsulate these three dimensions.
Data
The data sources for the beta version of BubbleAI cover several segments: news, Twitter, exchange announcements, and on-chain data.
The timeliness and completeness of these data sources in the beta version demonstrate an exceptional level of sophistication.Essentially, the vast majority of the data can be accessed within milliseconds and encompasses information on most tokens, including those rarely in the spotlight.
The disclosed information indicates that future data sources will incorporate more partners, including leading data providers in the industry such as Dune, Nansen, etc. From the perspective of pursuing data integrity, all products attempt to achieve perfection, but overlook the opportunities and advantages that business partnerships may bring. In this regard, BubbleAI has made a surprisedmovement.
Data Analysis
BubbleAI boasts its proprietary Large Language Model, called Bubble-GPT, a vertical model tailored for the Web3 domain.Upon acquiring raw data, BubbleAI initially matches the mentioned token within the raw data, and BubbleGPT subsequentlyconducts a sentiment-aware analysis on the specific post. For example, if the news headline and content are favorable for the token, it would be labeled “Positive”; conversely, it would be tagged “Negative.”
Furthermore, BubbleGPT can also execute a preliminary analysis on the raw data, aiding users in making trading decisions with greater speed and accuracy.
BubbleAI’s use of BubbleGPT to process data is undoubtedly a core trend. Traditional sentiment analysis tools like TestBlog havebeen proven to have significant gaps in accuracy compared to AI tool. Moreover, the accuracy of sentiment analysis often has a major impact on the profitability of a quantitative model. As a result, different financial institutions have started researching in thisdirection.
According to backtesting data from professional teams, BubbleGPT’s accuracy for sentiment analysis currently exceeds 70% – asurprising figure. OpenAI’s GPT-3.5 has also been tested for the same purpose, with results showing accuracy wavering between 60- 70%. These results demonstrate BubbleGPT’s superior capabilities in Web3 sentiment analysis as a vertical model, placing it at a high level within the industry.
Trading Tool
The current beta version of BubbleAI provides users with three trading options through both DEX and CEX channels.On the CEX side, users can link their Binance API key to BubbleAI’s Trading Tool, then seamlessly execute one-click trades on the platform, reducing the time friction. Perhaps anticipating users’security concerns, BubbleAI offers a “Trade on Binance” buttonwhere clicking automatically redirects to the user’s own trading interface, enabling assets to be traded without passing through a third party.On the DEX side, BubbleAI designed an additional interface to meet user demands.
On the Dex page, users can login via their Wallet, currently supporting Metamask, Rainbow, Coinbase Wallet and Wallet Connect. After logging in, users can still execute trades via Signal, but surprisingly, the data on the Dex page differs from the Cex page.
BubbleAI clearly tweaked and filtered the data sources on the Dex page, notably incorporating meme coin data. Pleasantly, the charting page is also functional, sparing meme coin traders the tedious process of using other charting tools like Dex Tools and providing a smoother trading experience.
BubbleAI to Launch Cold Start Campaign in Late August, Distributing $100K in Tokens to Early Participants
BubbleAI will airdrop tokens worth $10 million to early users as a reward at the end of August, according to information disclosed officially.The token is named $BBA, the abbreviation for BubbleAI. The rewards will go to the first 30,000 participants in the activity, which will last one week. Specific task details have not been revealed yet by the officials, but social media dynamics indicate it will likely involve social media and product interaction.Interested users can prepare in advance and keep an eye on the latest official Twitter updates around the end of the month, as this is likely a project with over 100x potential.
Twitter: https://twitter.com/Bubbleai_xyz
Website: https://bubbleai.xyz/
Media Contact
Company Name: Mediable
Contact Person: Jason
Email: Send Email
Country: United States
Website: bubbleai.xyz