| | 一场面向"小白"的炒股语言 | | --- | --- | | | 革命 | | 出品｜虎嗅科技组 | | | 作者｜陈伊凡、李一飞 | | | 编辑｜苗正卿 | | | 头图｜AI生成 | | "AI 原生 100" 是虎嗅科技组推出针对 AI 原生创新栏目，这是本系列的第「 34 」篇文章。 "做All in One的AI原生交易平台"。 创业第一天，RockFlow创始人Vakee（赖蕴琦）写下了自己的五年计划。 在 Vakee 的构想中，未来的投资世界不应只有冷冰冰的K线，更应有属于年轻人的热血与冒险。这一愿景 正在被数据验证：J.P. Morgan 报告显示，Z世代在25岁时的投资参与度已达 37%，远超前几代人——投资 不再是'大人的游戏'。作为数字原住民，这届年轻人拥抱 AI，寻找能与自己同频共振的数字理财搭子。而 RockFlow 正是为此而生。 2022 年，RockFlow 上线；2025 年 9 月，AI投资交易智能体——Bobby，作为内嵌在RockFlow的对话 Agent被推出。 就在我们聊天前，RockFlow 刚打完一场"真刀真枪"的美股实验：10 个大模型、每个大模型 ...

Core Viewpoint - The article discusses the emergence of AI trading models in the stock market, highlighting a competition involving ten AI models that trade in real-time using a set amount of capital, showcasing the potential of AI in investment strategies [1][3][12]. Group 1: AI Models and Competition - Ten AI models, including both established names like GPT and new entrants such as Doubao and Minimax, are participating in a trading competition, with Doubao currently leading [3][12]. - The competition involves each AI model managing a trading account with an initial capital of $100,000, making trading decisions every five minutes based on identical data inputs [18][24]. - The competition features three categories: Meme, AI stocks, and Classic, with a focus on AI stocks being particularly stimulating [20][15]. Group 2: Trading Strategy and Data Utilization - The AI trading agent, Bobby, provides all models with real-time market data, including K-line information, account data, and news, ensuring a level playing field [24][26]. - Each model must develop its trading strategy based on the same set of information, emphasizing the importance of independent reasoning and decision-making [26][24]. - The trading rules include a maximum leverage of 2x, no options trading, and a requirement for each trade to have a clear entry and exit plan [25][24]. Group 3: Performance and Insights - As of the latest updates, Doubao has achieved a notable profit, while other models like GPT-5 and Gemini 2.5 Pro have adopted different strategies, with GPT-5 focusing on risk management [28][29][35]. - The article highlights the distinct trading styles of the AI models, showcasing their personalities and decision-making processes, which adds an entertaining aspect to the competition [35][39]. - The overall performance of the models reflects their ability to adapt to market conditions, with some models taking more aggressive positions while others prioritize risk management [41][39].

Core Viewpoint - The integration of AI in the fintech sector is poised to revolutionize financial services, but it faces unique challenges such as data scarcity, high accuracy requirements, and the need for algorithmic transparency [2][4][21]. Group 1: Challenges in AI and Fintech Integration - Vertical data scarcity is a significant challenge as financial data is heavily regulated and not readily available [2]. - The financial sector demands extremely high accuracy, with a near-zero tolerance for errors, especially in monetary contexts [3]. - There is a critical need for algorithmic explainability in finance, requiring models to provide clear reasoning behind their conclusions [4]. Group 2: Industry Opportunities and Trends - The financial services market is vast, estimated at $36 trillion, indicating substantial opportunities for AI-driven startups in this space [8]. - Wealth transfer from older generations to younger ones is expected to create market opportunities, with 30% of global wealth shifting to the 90s and 00s generations over the next decade [9]. - The democratization of finance is a key trend, where advanced AI technologies can provide high-quality financial services to a broader audience, previously accessible only to wealthy clients [10]. Group 3: Product Case Studies - Cleo, an AI-driven personal finance assistant, targets young users and helps them make informed financial decisions [11]. - Bobby, developed by the company, serves as a 24/7 investment partner, assisting users throughout the investment process [12]. - Rogo is designed for young analysts in traditional financial institutions, showcasing the application of AI in professional settings [13]. Group 4: AI Agent Development and Functionality - The company has spent two years developing a vertical AI agent architecture, leading to the creation of Bobby AI, which aims to transform user interactions in financial services [16]. - Key features of Bobby AI include natural language interaction, precise task breakdown, and personalized user experiences [17][19][20]. - Bobby AI can facilitate complex investment actions through simple user expressions, enhancing accessibility for users [26]. Group 5: Core Challenges in AI Implementation - Technical challenges involve balancing timeliness, accuracy, and cost in the financial sector, necessitating a deep understanding of user needs [21]. - Trust is a significant concern, as users must learn to trust AI systems over traditional financial advisors, requiring time to build brand and product confidence [22]. - Regulatory compliance is complex in finance, with varying requirements across countries, making it essential for AI firms to navigate these regulations effectively [23]. Group 6: Future Outlook - The launch of Bobby AI is just the beginning, with expectations that many AI startups in finance will reshape various financial services, including digital banking and wealth management [30]. - The belief in financial and technological equity suggests that the next decade will bring significant changes to the financial landscape, driven by AI innovations [30].