David Solomon skipped the Goldman intern pep talk. His AI-era playbook is the firm’s quiet edge

(SeaPRwire) –

By: Christian Pierce

Wall Street is bleeding the very skill it cannot afford to lose. Junior hires race to churn out work faster with AI tools. They jump ship for 20% pay bumps at the first sign of a flashier role. Firms brag about cutting weeks of work down to minutes with automation. No one is slowing down to build judgment. The kind of judgment that catches a flawed S-1 assumption, reads a client’s unspoken concern, or walks away from a short-term win that will blow up three years out. Most leadership teams treat this gap as an unavoidable cost of progress. David Solomon is not.

Solomon laid out his counterintuitive framework for this year’s Goldman summer intern class. The group beat sub-1% acceptance odds for the third year running to land a spot. They walked into the town hall on the heels of Goldman’s record second-quarter results. The firm pulled in $20.34 billion in net revenues, with earnings per share of $20.98, up 92% year over year. He did not lead with promotion ladders or compensation benchmarks. He led with a lesson from his father, Jerry, who never ordered him to pick one path over another. When a young Solomon was set to turn left, his dad would force him to stop and map out the case for turning right, no lectures attached. That habit turned into his decades-old legal pad test. For every big decision, he draws a line down a sheet of paper. He lists the obvious points favoring the choice he already leans toward. Then he forces himself to fill the other side with the blind spots, the unconsidered risks, the case for the path he wants to dismiss. The payoff is not always a perfect choice. It is deeper conviction in the call he finally makes. He talked about turning down roles that promised quick, obvious wins. He spent nearly a decade at Bear Stearns. He was rejected by Goldman at least once before joining the firm in 1999. That patience put him in the CEO seat 20 years later. He gave a similar line to the 2025 intern class, urging them to take the long view. This year, he tied that advice directly to how he expects staff to use AI. He has watched models draft 95% of an S-1 IPO filing in minutes, work that once took a six-person team weeks. He cited internal Goldman research estimating AI will automate up to 25% of current U.S. working hours within a decade, hitting white-collar fields like banking, law and accounting hardest. He told a crowd at Italian Tech Week last October that most capital flooding into AI right now will fail to deliver returns, just like the dot-com bubble a generation prior. He expanded that point in a May New York Times guest essay, noting AI will disrupt labor markets but upgrade remaining jobs rather than erase them. His core instruction to interns on AI was simple. Use the tools to learn faster, to pull context quickly, to stress test your own assumptions. Never treat a model’s output as final. Argue with it. Poke holes in its conclusions. AI can parse existing data faster than any human. It cannot draw on lived experience to spot what the data misses. It cannot build the emotional connection that makes clients trust a firm with their biggest, most sensitive decisions.

This is not feel-good life advice. It is a deliberate talent strategy for an AI-first Wall Street. Every major bank is racing to roll out AI tools across every team. Most are framing the technology as a productivity hack, a way to cut costs and push junior staff to deliver more work faster. That approach will create a generation of analysts who can prompt a model to spit out a first draft, but cannot tell when that draft is wrong. It will create a generation of leaders who got to the top by chasing short-term speed, not building long-term conviction. Goldman is betting on the opposite. It is training its newest hires to slow down, to look for the gaps, to hold onto the judgment no model can replicate. That judgment is the actual product Goldman sells. The numbers on a spreadsheet, the first draft of a filing, the quick market scan—all of that will be commoditized, powered by AI, in a few years. The only edge left will be the person who can pause before turning left, who can spot the flaw in the model’s logic, who can sit across from a client and actually understand what they need. The interns who steal that habit will be the ones running the firm in 20 years. The firms that skip that training will be the ones losing clients, and market share, when the next cycle hits.

Author bio: Christian Pierce, chief financial columnist covering Wall Street leadership, capital markets, and AI’s impact on white-collar work for leading global business publications.