The CEO of the AI company reveals: It is obvious that AI 'crushes' humans in stock analysis, and large funds have long adopted it?
With enough computational power and training, AI can surpass human intelligence in stock trading.
Jan Szilagyi, CEO and co-founder of the software company Reflexivity, has provided interesting insights on the use of artificial intelligence (AI) in stock trading. In an interview on CNBC's Squawk Box, Szilagyi revealed that top hedge funds are using AI software as a supplementary tool to existing tools. He also emphasized that AI has made many accurate predictions about the stock market this year.
Reflexivity, the company of Szilagyi, develops platforms that integrate large language model (LLM) analysis interfaces. According to Szilagyi, with enough computational power and training, AI can surpass human intelligence in stock trading. He stated that the ultimate goal of the company is to build an 'automated investment analyst.' Currently, Reflexivity is providing technologies that allow investors to aggregate data from various sources into a single platform, optimizing decision-making.
Szilagyi describes Reflexivity's technology as a tool capable of reducing data analysis time from two hours to just two minutes. Users can ask questions to the system, and it will automatically search for data, analyze, and provide results. This AI system also has the ability to find similar trends from past data to predict future price patterns. These 'similarities' are identified based on about 12 to 15 previous similar cases, related to specific economic conditions or parameters. From there, the AI will suggest potential stocks that could yield superior returns.
Reflexivity's clients are mainly large hedge funds, which use this system as an additional layer of intelligent analysis on top of existing data sources. Szilagyi also shared that the company's AI has accurately predicted many major market trends in 2024. For example, the AI accurately predicted the market peak in July, the subsequent bear market rally, and the following market bottom. However, in some cases, the system made predictions a bit too early, such as predicting an uptrend before the recent Federal Reserve (Fed) meeting. As a result, the Fed announced two interest rate cuts in 2025, instead of the four times initially predicted.
Szilagyi believes that this does not mean the system is wrong, as all results from AI are based on probability assessments. He explains that the system only provides a level of confidence, for example, 'there is a 70% chance this will happen, but there is still a 30% chance it won't.'
Notably, Szilagyi sparked controversy by stating that there is nothing special that makes human intelligence superior to AI in stock trading. According to him, if provided with enough computational power and continuously improving intelligence, AI can completely surpass human capabilities. He predicts that within the next 5 to 10 years, AI may replace humans in this field. However, he also acknowledges that some areas, such as the private market, will be more difficult for AI due to a lack of training data.
Although AI cannot completely replace humans yet, with the current pace of development, the time when intelligent systems excel in stock trading is no longer a distant prospect.