On October 22nd, according to a document seen by BloombergOpenAI Over 100 former investment bankers were hired to assist in training their artificial intelligence systems in financial modelling skills, with the aim of replacing the industry as a wholeBankersA great deal of repetitive work has been undertaken。
The team included former employees from such institutions as JP Morgan Chase & Co., Morgan Stanley and Goldman Sachs Group Inc., who were part of a clandestine project within the original company called Mercury。
According to one person with knowledge of the project, these participants are paid $150 per hour (note: the current exchange rate is approximately RMB 1064), with the task of preparing tips and financial modelling for a variety of types of transactions, including business restructuring and the first public collection (IPO). Besides, OpenAI Access to the artificial intelligence system under development, which is intended to replace junior positions in investment banks, was also made available in advance to these contract employees。
The project highlights the fact that OpenAI, led by Sam Altman, is eager to apply its powerful artificial intelligence technology to a wider range of sectors, from counselling, finance to law and technology. Although the company ' s valuation earlier this month amounted to $500 billion (the current exchange rate is approximately 3.55 trillion yuan), making it the world ' s highest-market start-up enterprise, it has not yet earned a profit。
One speaker from OpenAI said: “We work with experts in various fields to upgrade and evaluate the capabilities of our models in different applications. These experts are recruited, managed and remunerated by third-party suppliers.”
In traditional investment banking, analysts usually work more than 80 hours a week during their participation in actual transactions, using Microsoft Excel software to build detailed M & As and leverage finance models. They are also required to respond to requests from their superiors to constantly modify PowerPoint presentations, or even to adapt their content over and over again, a high-intensity, low-autonomy culture that has given rise to Wall Street's widely known “pls fix” (changed)。
AT PRESENT, MANY START-UPS HAVE ATTEMPTED TO BECOME INVOLVED IN THIS AREA BY PROVIDING BANKS WITH AI TOOLS THAT CAN ASSIST IN THE ABOVE TASKS. DESPITE LONG-STANDING COMPLAINTS BY ANALYSTS ABOUT THE BOREDOM OF THEIR WORK, THE EMERGENCE OF ARTIFICIAL INTELLIGENCE TODAY IS RAISING DEEP CONCERNS ABOUT THEIR CAREER PROSPECTS。
ACCORDING TO SOURCES, THE APPLICATION PROCESS FOR ADMISSION TO THE MERCURY PROJECT REQUIRES ALMOST NO MANUAL INTERVENTION. THE FIRST PHASE INVOLVED INTERVIEWS OF APPROXIMATELY 20 MINUTES BY AI CHAT ROBOTS, THE QUESTIONS WERE BASED ON THE APPLICANT ' S CURRICULUM VITAE; THE SECOND PHASE TESTED CANDIDATES ' ABILITY TO UNDERSTAND THE FINANCIAL STATEMENTS; AND THE LAST PHASE WAS A FINANCIAL MODELLING EXERCISE。
The person indicated that the exercise was highly flexible and that the contractor could submit only one model per week. Specific requirements include the writing of indicative words in a language that is easy to understand and on which to base modelling tasks. Each submission will receive feedback from the reviewers and the contractor will have to correct all the issues before its results are integrated into the OpenAI system。
The paper shows that the Mercury project has attracted participants to date who have worked in many prominent Wall Street financial institutions, including Brookfield Corp., Mubadala Investment Co., Evercore Inc. and KKR & Co. In addition, MBA students from Harvard University and Massachusetts Institute of Technology were involved。
All models must be created in Excel and must follow the industry ' s usual format specifications, such as page margins, percentage italication, etc., must be strictly followed。