On the weekend, as most students were stumbling back from the bars, Songyee Yoon was rushing across her South Korean campus. Around dinner time, she would run some programs on her college’s supercomputer and then, as the computer chugged through her program, she’d wait sleepless in her dorm. “I woke up at like 2 a.m., 3 a.m. to walk down the campus because I was so curious to see the result,” she said.
She was such an oddity on campus that a writer used her as inspiration for a TV show about her college.
“The intention was not to create characters based off of a real character,” she said. But as the writer talked to students to get material, “she kind of kept hearing about this weird girl.”
And so Yoon became the inspiration behind “Genius Girl” on the Korean TV show KAIST.
Today, if writers were going to make a show about Yoon’s life, it’d look more like HBO’s “Silicon Valley.” After completing a PhD from MIT, she climbed her way to becoming president of South Korean video game developer NCSoft, and today, she’s announcing Principal Venture Partners (PVP), a $100 million fund to back AI startups. The fund will write early-stage checks anywhere from $100,000 to “single-digit millions” and has already invested in six startups, including model maker Liquid AI.
Her fellow partners include a who’s-who of AI academia: there’s Daniela Rus, a renowned researcher that Yoon met through Yoon’s work on MIT’s board; Dawn Song, a MacArthur Fellow who’s published extensively on computer security; and Jeremy Nixon, the founder of the AGI House, a AI hacker house that’s made headlines for attracting young talented founders.
PVP is one of the few investment firms with such a deep bench of academic powerhouses — something that Yoon sees as an advantage when the firm is trying to win deals.
“I think founders would like to have a diversified set of advisors who can bring different perspectives,” she said. Yoon believes that the research backgrounds of the PVP team give them a profound understanding of how AI “has evolved over time” and where it might be headed.
The team is betting that the next generation of unicorns will be AI-native companies, meaning they were built with AI in mind from the start, not with AI applications jerry-rigged onto the platform after the fact. Yoon’s not worried that they may have missed the boat on investing in foundational companies like OpenAI or Anthropic. “If you look at the top 10 NASDAQ companies, more than half of them are digital-native companies who started after the introduction of broadband,” she said.
Yoon said the firm will invest across sectors. She’s particularly excited about the potential for AI to transform the insurance industry, whether that means using AI to help people understand what their plans cover, or insurance companies that specialize in underwriting autonomous robots.
Yoon is also worried about the issue of AI’s potential to exacerbate cultural colonialism, a topic she wrote about last year. She gave the example of large model makers proclaiming, “oh, we train this AI using all the data in the world.”
“But if you think about it, 35% of the world’s population do not even have access to broadband,” Yoon said. “And they cannot be authors of the data that has been used for training this AI. So it’s inevitable that those kinds of cultures and viewpoints cannot be reflected.”
She admits it’s a complicated problem that can only begin to be solved through continuous conversations and increased representation through the industry — like, say, an AI-focused fund with three female partners.
“We don’t say it’s a female fund, but I think I see a lot of the female founders come to us because they know that we’ll have more sympathy,” Yoon said. “And that we can see their real strength and real super power.”
