Peter Thiel and Marc Andreessen want you to believe universities are the enemy of progress. In leaked text messages obtained by The Washington Post, Andreessen called for the National Science Foundation to receive "the bureaucratic death penalty" and described Stanford and MIT as "mainly political lobbying operations fighting American innovation." The Trump administration seems to have listened. Proposed budget cuts target 57 percent of NSF funding, 40 percent of the National Institutes of Health, and 24 percent of NASA. Michael Kratsios, who directs the White House Office of Science and Technology Policy and previously served as chief of staff at Thiel's venture fund, orchestrated the assault. More than 10,000 federal workers with STEM PhDs left the government last year.
The displaced researchers aren't vanishing. They're landing on platforms like ScaleAI and Mercor, where they label data and train AI models for Google, OpenAI, and Meta. The work pays roughly $30 an hour. No benefits, no stability, just gig work with a doctorate. ScaleAI has raised over $600 million, with backing from Founders Fund, the firm Thiel co-founded. Mercor counts Andreessen Horowitz among its investors. Whether intended or not, the loop is pretty blunt: fund the political push to gut academic science, then capture the cheap expert labor that results. The displacement has even spawned new approaches that automate the entire scientific research cycle.
What makes this especially galling is that Silicon Valley exists because of government-funded research. Larry Page and Sergey Brin built Google's core algorithms on NSF grants. The internet came from DARPA. Geoffrey Hinton, the Nobel laureate called the godfather of AI, relied on Canadian government money for his lab. Even the lithium-ion batteries in our phones emerged from publicly funded university labs. Thiel insists the average PhD is "99% less productive" than researchers a century ago, offering no real metric for that claim. The industry now depends on those same researchers, just at a fraction of what they'd cost with real jobs—working within autonomous agent systems.