Finance

What does Goldman Sachs want in a coder? For them to have studied philosophy.

Socrates statue in front of Academy of Athens with a blue sky and clouds in the background.
A statue of famous philosopher Socrates in front of the Academy of Athens. Hiroshi Higuchi / Getty Images
Read in app

Artificial intelligence is redefining what it takes to be a software engineer on Wall Street.

A typical software developer holds a computer science degree. They're analytical thinkers and usually more introverted, spending more time talking to machines in computer programming languages than schmoozing over business deals.

But in the age of AI, where models can generate code faster, and often better, than humans, the typical education path of a software engineer is being upended.

Goldman Sachs' chief information officer, Marco Argenti, recently encouraged his daughter, a college student, to concentrate her education on philosophy if she wants to pursue a career in engineering.

"Coming from an engineer, that might seem counterintuitive, but the ability to develop crisp mental models around the problems you want to solve and understanding the why before you start working on the how is an increasingly critical skillset, especially in the age of AI," Argenti wrote in a Harvard Business Review article this week.

Over his decades of managing engineering teams at Goldman Sachs and Amazon Web Services, Argenti said asking the right questions is one of the most important skills he's learned.

"It's not dissimilar with AI: The quality of the output of a large language model is very sensitive to the quality of the prompt," he added.

To be clear, having a technical edge, whether it's math, computer science, or data analytics, is still likely table stakes when it comes to software engineering. However, it's becoming increasingly important to train the more creative part of the brain through writing and social sciences courses, like English, philosophy, and psychology, according to some of Wall Street's top tech execs and recruiters. Doing so will prepare the coming generations of software engineers and help futureproof their jobs against AI.

With generative AI, it's possible for people, including non-tech workers, to type a command in plain English and have the models generate lines of code that execute on that ask. This emerging discipline is called prompt engineering, and it's becoming a key piece to the success or failure of AI-generated code.

"Miss the mark on the prompt, and your AI will produce code that's plausible at best, incorrect, and dangerous at worst," Argenti said.

Citi's co-chief information officer, Shadman Zafar, echoed the importance of clear and concise communication and writing skills when it comes to code.

"As AI comes in, cogent and clear communication becomes much more important, and the ability to actually phrase your questions correctly, soundly, and safely makes a lot more of a difference as well," Zafar told Business Insider. As such, communication and psychology are becoming "more valued minors or majors for students to focus on," he said.

Zafar said he's paying more attention to people with "a computer science degree and an English minor," or "a psychology major and a computer science minor."

Jensen Huang, the CEO of Nvidia, the chip-manufacturing company that's skyrocketed with the rise of AI, went as far as to advise the next generations not to learn how to code. At the World Government Summit in Dubai in March, he said that programming is no longer a critical skill with AI.

That sentiment was shared by Ken Griffin, the billionaire CEO of the hedge fund Citadel, who said the days of being "a good programmer are becoming numbered." Instead, the success of a technologist will hinge on being a really good problem solver, which involves understanding the business and the commercial aspects of technology.

While it remains to be seen whether AI will replace software engineers, those looking to futureproof their jobs should think about the parts of the job that AI could handle and zero in on tasks where there is a human in the loop, according to Deepali Vyas, global head of fintech and applied intelligence at search firm Korn Ferry.

"That software engineer might get replaced by a sort of prompt engineer," Vyas said. If that's the case, "I would make sure that I have good problem-solving, critical thinking, and writing skills," she said.

Read next

Bianca covered the intersection of finance and technology for Business Insider as a senior reporter, writing about the behind-the-scenes tech powering the country's largest financial firms. She is interested in all things cloud, data, AI and machine learning, crypto and blockchain, and cybersecurity.Her reporting has taken readers inside some of the biggest banks, hedge funds, private equity firms, and asset managers and their playbooks for spending billions every year on technology. When she's not covering finance giants, Bianca also brings readers to the bleeding edge of fintech innovation, frequently covering exciting and scrappy startups and the giant VC investors backing them. Selected works:Everything we know about how Wall Street is adopting AI, from Goldman Sachs to BlackstoneJamie Dimon says to quit if you don't like his RTO demands. Some of his tech workers might do just that.Here are 49 of the most promising fintech startups transforming how we bank, invest, and pay, according to 27 top investorsAI is fueling a culture clash inside hedge fundsInside AI's transformation of Wall Street, according to 35 insiders at banks, hedge funds, and asset managersThe secretive world of Wall Street technology is opening up like never beforeWall Street's top tech priority: building internal search engines