高中生用ChatGPT“全自动”炒股，一个月竟躺赚25%

Core Insights - The article discusses an experiment conducted by a 17-year-old high school student, Nathan Smith, to test the effectiveness of AI in stock trading, specifically using GPT-4o and DeepSeek as competitors [2][4][14] - The experiment aimed to determine if AI could outperform the stock market by making investment decisions without human intervention [2][4] Group 1: Experiment Setup - Smith allocated $100 to each AI, with the goal of maximizing returns over six months, focusing on micro-cap stocks with a market cap under $300 million [4][5] - DeepSeek invested all funds immediately across three companies, while GPT-4o adopted a more cautious approach, reserving some funds for future opportunities [5][6] Group 2: Performance Comparison - In the first week, DeepSeek's portfolio decreased by 18.06%, while GPT-4o achieved a 6.72% return, outperforming the Russell 2000 Index [5][6] - Over the course of the experiment, GPT-4o's portfolio increased by 25.2% from June 30 to August 15, significantly surpassing the S&P 500's 4.5% increase during the same period [9][11] Group 3: AI Decision-Making - GPT-4o demonstrated a mix of bold and cautious investment strategies, including a notable decision to invest 32% of its assets in aTyr Pharmaceuticals based on anticipated positive clinical trial results [12][14] - Despite some questionable decisions, such as continued investment in underperforming stocks, GPT-4o's overall strategy led to substantial gains [12][14] Group 4: Broader Implications - The article references academic studies indicating that AI, particularly ChatGPT, can predict stock returns based on sentiment analysis of news articles, suggesting a potential for positive investment strategies [13][14] - The rise of AI in stock trading could significantly alter market dynamics, indicating a shift towards AI-driven investment strategies [14]