With experts at OpenAI and Google DeepMind predicting AGI could emerge before 2030, some students are leaving academia to work on AI safety, shape regulation, or launch their own startups.

"I worried I wouldn't live to graduate because of AGI," says Alice Blair, a former MIT student now working at the Center for AI Safety. "If we keep going like this, AGI will lead to human extinction." She has no intention of returning to class.

For many, the fear isn't existential—it's economic. A recent survey found that half of Harvard students believe their chosen careers could become obsolete within a few years. "If your field is going to be automated before the decade's out, every year spent in school is a year less in the workforce," says Nikola Juckovic, head of Harvard's AI safety group.

Others see a once-in-a-generation opportunity. Michael Tuell (Anysphere) and Brendan Foody (Mercor) dropped out of MIT and Georgetown, respectively—their startups are now valued at $9.9 billion and $100 million. "There's a limited window to get my hands on the steering wheel," says Jared Manter, another dropout-turned-founder.

Whether driven by fear or ambition, a growing number of top students are betting their future on the uncertain path of AGI.