
Unprecedented Bidding War Erupts Over Anysphere, Creator of Popular AI Coding Assistant Cursor
Technology News
Zaker Adham
09 November 2024
08 October 2024
|
Zaker Adham
Summary
Summary
Google’s AI model, Gemini, has recently come under scrutiny for providing flawed investment advice. The NASDAQ-100 Index, which tracks the top 100 companies listed on the NASDAQ exchange, is a key focus for many investment funds. These funds aim to outperform the index by strategically adjusting their portfolios, overweighting stocks like Tesla, Microsoft, NVIDIA, and Apple, while underweighting others, such as PayPal and Cisco. Despite these efforts, outperforming the index remains a rare achievement, even for seasoned investors.
The rise of artificial intelligence (AI) has sparked curiosity about its potential to manage stock portfolios. Could an AI, such as Google’s Gemini, consistently select stocks that outperform the market? Demis Hassabis, CEO of Google DeepMind, believes Gemini holds incredible potential. In a recent article titled “Introducing Gemini: Our Largest and Most Capable AI Model,” Hassabis highlighted Gemini’s performance, especially in tasks such as natural image, audio, and video understanding, as well as mathematical reasoning.
Gemini, however, struggles when tasked with picking stocks. Over a 100-day experiment, the AI was asked two basic questions daily:
The results were far from impressive. Gemini frequently failed to provide the requested number of stocks, often listing fewer than 10 and repeating the same names, such as Amazon, multiple times. Even more problematic was its inclusion of stocks like Visa, which doesn’t belong to the NASDAQ-100, and its contradiction of naming stocks, such as Tesla and Amazon, as both gainers and losers on the same day.
Despite occasional success, such as identifying gainers like Tesla and Amazon, which saw notable increases, Gemini’s overall performance was inconsistent. The AI’s inability to differentiate between stocks that would rise and those that would fall was glaring, with some of the predicted “losers” actually outperforming the “gainers.”
While Gemini may have delivered results that briefly surpassed the NASDAQ-100 Index, its overall reliability as an investment advisor is highly questionable. The AI often provided contradictory or incomplete data, and its predictions for the CAC 40 Index in Paris and the AEX Index in Amsterdam yielded similarly confusing outcomes. Stocks predicted to decline often performed better than those expected to rise, further underscoring the model's limitations.
Gemini’s struggle with financial predictions suggests that AI still has significant room for improvement in the realm of investment advice. As of today, Gemini has shifted its approach and now declines to provide specific stock picks, acknowledging the complexity of financial markets.
In response to the findings, Gemini itself offered this reflection: “While I may have faced challenges in the past, I am confident that future iterations will be better equipped to handle complex tasks like financial forecasting.”
Technology News
Zaker Adham
09 November 2024
Technology News
Zaker Adham
09 November 2024
Technology News
Zaker Adham
09 November 2024
Technology News
Zaker Adham
07 November 2024