Goldman CEO Solomon: His dad's advice shapes how interns should think about AI

David Solomon, CEO of Goldman Sachs, recently met with the firm's 2026 summer intern class during a town hall to discuss career strategy and artificial intelligence. The interns had been selected from an applicant pool so competitive that Goldman maintained an acceptance rate below 1% for a third consecutive year. The timing coincided with the bank's announcement of record second-quarter results: net revenues of $20.34 billion and earnings per share of $20.98, representing a 92% increase year-over-year. Solomon's core message rested on a decision-making habit he traces to his father, Jerry Solomon. The habit is simple but deliberate: when facing a big decision, Solomon takes out a piece of paper, draws a line down the middle, and lists considerations on both sides. The crucial step comes next—even when one side appears overwhelmingly obvious, he forces himself to think more deeply about the other side, to identify what he might be missing. Solomon explained that his father never told him what to do directly but instead had "an incredible way" of making him wrestle with issues. "If I was running down the road to turn left, he had a great ability to get me to pause and really think about why left was a better choice than right," Solomon recounted. "He never said right was better—even when he thought so—he wanted me to spend more time thinking about right before I went left." The payoff, Solomon told the interns, is not necessarily a better decision every time; it is more conviction in the decision you ultimately make. Solomon then connected this habit to career advice. He emphasized patience and the value of saying no to opportunities that looked attractive short-term. "Some of the most important decisions I made were saying no to things where the grass looked greener in the short term," he said. "I stayed in places longer and that therefore opened up bigger opportunities than I would have imagined." His own career exemplified this approach: he spent nearly a decade at Bear Stearns before joining Goldman in 1999, reportedly after being turned down by the firm at least once earlier. That patience eventually led to his CEO role two decades later. Solomon extended this logic to artificial intelligence. He encouraged interns to "really try to understand what these things do and don't do. Where are they powerful? How do they help you? How do they make you smarter and better?" But his next instruction pivoted from adoption to friction. "Ask questions, challenge the conventional wisdom," he said. "You have the ability to pick any topic and, in a much shorter period of time, get much smarter on it. But you have to challenge what the models give you." Solomon's own practice with AI models is to use them as a sparring partner rather than an oracle. He asks them to verify his read on events, to explain implications, and to explore what would change his assessment. The core of his skepticism rests on what he believes AI cannot do. "The technology has no ability to do what we do as human beings—to think about whether something makes sense," he said. "There is a form of intelligence which is analyzing data, but there's also a form of intelligence which is taking life experience and using it to think about or infer things that data wouldn't." He acknowledged AI's power to accelerate routine work: the technology can now draft roughly 95% of an S-1 IPO filing in minutes, a task that once required a six-person team weeks to complete. Yet the judgment call at the end still belongs to a person. This view extends to Goldman's core business. "Across our business, a lot of the relationship is about clients wanting to talk to someone, be heard, be understood, and have an emotional connection. That's not going away," he said. Solomon's public position on AI has shifted over the past year. Last October at Italian Tech Week in Turin, he said he was "not smart enough" to know whether AI is a bubble but was certain that "it's not different this time"—meaning that a significant portion of capital being deployed now would not deliver returns, much as happened during the dot-com run. By May, in a guest essay for the New York Times, he had moved to a more definitive stance. He conceded the premise of disruption—"Will A.I. disrupt the labor market? Absolutely"—and cited Goldman's own research estimating that AI could automate up to 25% of current U.S. working hours within a decade, concentrated in white-collar fields like accounting, banking, and law that form his own client base. But he argued that this automation would free workers for more complex tasks, upgrade the jobs that remain rather than eliminate them, and create new roles managing AI systems. His message to interns, tying his father's method to AI and career patience, was essentially the same instruction passed down. Go slow. Look at the other side. Don't turn left until you know why.