Since the advent of ChatGPT, there have been people with flexible minds who have begun to delve into how to make money with it.
It is used to write copy, to generate artwork, to make online courses. Only you can't think of it, and no AI can't do it.
Of course, the closest thing to money is financial investment.
Both traders and retail investors want to "kill" a bloody road in the treacherous secondary market with the blessing of powerful AI.
Whether AI will become a Buffett for everyone or a freshly baked leek-cutting sickle, the entire industry is waiting to see.
Return over 500%?
The first to spark the craze this year was a study by the University of Florida.
The researchers provided ChatGPT with a large number of news headlines about public companies, allowing ChatGPT to use sentiment analysis to determine whether these events were favorable, unfavorable, or irrelevant to the company's stock price. The "ChatGPT index" is then made into a complex calculation formula and compared with real stock market returns.
Subsequently, the researchers advised stocks based on the "long and short strategy" trading offered by ChatGPT, where companies with good news and short those with bad news gained more than 500% over a year-long or so time interval, in stark contrast to the -12% return of buying and holding S&P 500 ETFs over the same period.
There are some views in the industry on this result.
Liang Jia (pseudonym), a quantitative private equity practitioner, said that long before this study, there were many algorithms for sentiment analysis of news headlines. In addition to news headlines, company announcements, research reports and other texts are also important data sources for many such analyses.
Compared with traditional sentiment analysis, ChatGPT has more advanced language understanding and reasoning ability, so it is more significant and more accurate in prediction.
This advantage is also reflected in the study.
When researchers asked about the impact on Oracle's stock price with the headline "Rimini Street fined $630,000 in case against Oracle," Ravenpack, a leading news analytics tool, gave a negative sentiment score of -0.52.
But ChatGPT sees it as a positive sign, arguing that "the fine on Rimini Street may increase investor confidence in Oracle's ability to protect its intellectual property and increase demand for its products and services."
It can be seen that compared with traditional tools, ChatGPT has a deeper thinking about context and its meaning when understanding natural language.
JPMorgan Chase logo photographed in New York, USA Source: Xinhua News Agency
Wall Street's AI "arms race"
The academic community has just dropped the bomb, and some people in the industry can't help but put down a real market.
In mid-May, a financial services company called Autopilot built a new ChatGPT-led investment plan in addition to the company's original portfolio and handed it $50,000 in initial capital to see if ChatGPT could beat hedge funds.
At its core, it's the same as the University of Florida research project, which lets AI analyze 10,000 news articles and reports from 100 companies, and then builds an investment fund of 20 stocks that updates positions once or twice a week.
Stocks initially selected include Berkshire Hathaway, Amazon, dr. horton and Davita Health, among others. Two weeks later, the portfolio was up about 2 percent, essentially in line with the broader market.
In this experiment, more than 25,000 traders bet a total of $15.14 million, an average of nearly $600 each, and the number is growing, which shows the enthusiasm for AI investment.
This set the fire of the AI revolution on Wall Street.
According to consulting firm Evident, about 40% of the most enthusiastic banks are hiring for AI-related positions, such as AI data engineer and quantitative roles, as well as AI governance and ethics roles.
JPMorgan Chase recently revealed that it is developing an artificial intelligence service similar to ChatGPT "IndexGPT" to select stocks for customers, and may be the first established financial institution to launch such services to customers.
In addition, several banks, including Goldman Sachs and Morgan Stanley, have begun testing ChatGPT internally; Deutsche Bank is scanning wealthy clients' portfolios with AI; Holland International is used to screen potential defaulters... Everyone is gearing up to embrace new technologies.
So, is AI really ready?
Traders work on the New York Stock Exchange in the United States, source Xinhua News Agency
"Leek sickle" or "retail light"?
As far as ChatGPT is concerned, it is still widely used in the industry as a decision-making aid.
Liang Jia believes that in terms of quantitative investment, the current limitation of ChatGPT is that it cannot generate its own strategic inspiration.
When researchers have a good investment idea, ChatGPT can accelerate the realization of this idea through efficient data collection and processing power, programming ability, text analysis ability, etc., but the idea itself is the key to making money on investment.
In addition, compared with quantitative researchers who need to process massive amounts of information, researchers in "deep" subdivisions focus on a few stocks, and they can do more in-depth research through empirical judgment, offline research, etc., and may rely less on such tools.
The "herd effect" generated by investment is also a concern for industry insiders.
The underlying logic of short-term trading is a "zero-sum game", if everyone uses similar large-model investment tools to make short-term investments in the future, it is likely that because of certain news events, they will rush to concentrate investment, resulting in strategy failure, and it is difficult for everyone to obtain good returns.
Long-term trading requires investors to have more comprehensive judgment and industry experience, which is difficult to achieve with the current level of ChatGPT.
But for common shareholders, the emergence of ChatGPT could be a powerful enabler.
Originally, many investment analysis methods had high technical thresholds and were not very friendly to retail investors. But with ChatGPT, retail investors can test and realize their good investment ideas through tool learning, which may change the existing model of securities trading.
However, it is certain that with the continuous exploration and development of large-language models tailored to the financial industry, the accuracy and efficiency of the financial decision-making process will be greatly improved.
As for who can laugh at the sea of stocks after technological iteration, this is the unpredictable charm of the secondary market.
(Source: Chao News)
For more exciting information, please download the "Extreme News" client in the application market, please do not reprint without authorization, welcome to provide news clues, once adopted, you will be paid. 24-hour reporting hotline: 027-86777777.