Before he dropped out of college and later went on to found OpenAI, Altman said he was “peer pressured and tricked” into accepting an internship at Goldman Sachs during his sophomore year.

“At the time, that was the cool thing. That was what everybody wanted to do,” Altman told the audience at the Internapalooza Conference last month. 

“I got peer pressured and kind of tricked into it,” Altman said. “I dutifully applied and got in and felt really cool.”

But Altman ended up turning down the offer. “And then I didn’t do it,” he said, and instead, he  worked on his startup Loopt before joining Y Combinator’s first batch and eventually became YC’s president. He then went on to  start frontier AI company OpenAI, which just wrapped a $7 billion share sale as it eyes a potential trillion-dollar IPO as soon as the second half of 2026. Goldman is one of OpenAI’s underwriters for its IPO.

Altman began discussing the idea of creating one’s business as opposed to working up the corporate ladder “just doing stuff.” Now in the age of AI, in part thanks to his own startup, it’s that much easier for someone to turn down a prestigious banking internship (Goldman’s intern class acceptance rate is now below 1%.)

Working at an establishment like Goldman, Altman said, meaning being around “people [who] put too much effort into trying to get taken seriously.” Because banking careers are inherently hierarchical, there’s no room to create. Startups, on the other hand, offer that freedom.

“I think this is much more true now than ever before,” he said. “You can make a whole startup, kind of by yourself in a room with a lot of AI tokens but not much else, and you don’t need anyone to take you seriously to do that.”

“You can get a long way in life and career just by doing stuff,” Altman said, adding that OpenAI gives young adults plenty of responsibility for major projects as the company believes in “identifying and betting on young, unproven talent.”

Recruiting from Wall Street

Years after Altman turned down the offer, his company is now recruiting Wall Street talent. Both junior investment bankers and senior banking veterans—including former Morgan Stanley managing director Alisha Lehr just last month—are among OpenAI’s latest recruits. It has enlisted more than 100 former investment bankers from Goldman Sachs, JPMorgan and Morgan Stanley to train AI models to build financial models and automate work traditionally handled by junior bankers, and there could be plenty more seats to fill. OpenAI reportedly plans to grow its workforce to 8,000 by the end of the year.

OpenAI itself increasingly expects finance people to be AI-native. CFO Sarah Friar said in June that she probably wouldn’t hire a finance employee today who didn’t know how to use a tool like Codex, comparing it to knowing Excel.

“I would never hire a finance person who didn’t know how to use Excel, and I probably wouldn’t hire a finance person today that doesn’t know how to use a tool like Codex,” she said.

As some bankers leave for jobs at AI labs, finance is also turning into a battleground for enterprise AI business more broadly. Anthropic launched roughly 10 pre-built AI agents for financial workflows, including pitchbooks, earnings analysis, credit memos and underwriting, with Claude is already in production at JPMorgan, Goldman and Citi.