Silicon Valley is increasingly building a labor model that depends on scientists who have been pushed out of stable academic and public-sector jobs. The result is a growing pool of highly trained researchers doing short-term, lower-paid gig work for AI companies that need their expertise but do not want the cost of hiring them full time.
That shift is not happening by accident. It is tied to deep federal cuts to science funding, attacks on universities, and a tech industry that now profits from the very research system it helped weaken.
How public science built Silicon Valley
Much of modern tech grew out of government-backed research, not pure private entrepreneurship. Semiconductors, the Internet, lithium-ion batteries, touchscreens, and even the research foundations behind generative AI all came from labs and programs supported by public money.
The article notes that Google’s origins also depended on National Science Foundation funding at Stanford, where Larry Page and Sergey Brin were graduate students. It also points out that Geoffrey Hinton, widely known as the “Godfather of AI,” relied on Canadian public funding for his lab after leaving a U.S. academic post to avoid Pentagon contracts.
That history matters because it shows a basic contradiction. Silicon Valley now presents itself as the engine of innovation while its leaders work to weaken the institutions that made that innovation possible.
A political push against universities and research
A central part of the story is the alliance between wealthy tech figures and the Trump administration’s science agenda. Conservative venture capitalists Peter Thiel and Marc Andreessen have pushed hard against universities, federal science agencies, and public research infrastructure.
Leaked text messages cited in the article show Andreessen calling universities “at Ground Zero of the counterattack.” He also described Stanford and MIT as “mainly political lobbying operations” and called for the National Science Foundation to receive “the bureaucratic death penalty.”
The administration’s proposed budget reflected that hostility. It included cuts of 40 percent for the National Institutes of Health, 57 percent for the National Science Foundation, and 24 percent for NASA, according to the article.
What the cuts did to scientists
The effect on the scientific workforce has been severe. More than 10,000 federal workers with STEM PhDs left the federal workforce last year, and university labs were forced to fire researchers, cancel studies, or shut down work entirely.
Some academics left for Europe, while others retired early. The article describes a broader chill across scientific institutions, one that could last well beyond any single administration.
This environment created a new labor market for Silicon Valley. Researchers who once expected postdoctoral work, grants, or faculty jobs were suddenly facing fewer options and weaker bargaining power.
Why AI firms want PhDs on gig platforms
AI companies often market their models as if they can perform at “PhD-level” across many fields. But those systems still need real experts to create prompts, generate training data, and check outputs.
That need has opened the door to platforms such as Mercor and ScaleAI, both of which received venture funding from Thiel. Mercor later reached a $10 billion valuation, while Meta bought a 49 percent stake in ScaleAI at a $29 billion valuation, according to the article.
The business model resembles ride-hailing platforms. Workers are sold flexibility, remote access, and the promise of income in a difficult market. In practice, the arrangement often shifts risk onto the worker while the company captures the value.
How the work is structured
Researchers interviewed in the article described assignments that sounded attractive at first but became harder to justify once unpaid labor was counted. One doctoral student in applied mathematics was paid about $90 per hour to solve advanced math problems, but the company covered only two and a half hours per question.
If a response was incomplete or wrong, the worker received nothing. That meant the actual pay could drop sharply once hidden labor was included.
A recent graduate of an MIT engineering program said she initially viewed the compensation as fair. After tracking unlogged time, she found the effective rate was much lower than advertised and “not being super worth it.”
Key features of the emerging gig model
- High skill requirements that depend on advanced academic training.
- Short-term contracts with unstable earnings and no real career path.
- Hidden unpaid labor from preparation, revisions, and corrections.
- No guarantee of future work, even for highly credentialed researchers.
- Heavy dependence on the weak academic job market created by funding cuts.
This structure makes the work look flexible while making the worker absorb more of the risk.
From researchers to replaceable labor
The article argues that what looks like a free market is often the product of policy choices. Federal cuts to science funding and pressure on universities reduced opportunities, and that made it easier for tech platforms to offer gig work as a fallback.
One doctoral student described the situation as feeling “akin to being farmed.” That phrase captures the imbalance at the center of the system: the industry depends on researchers’ expertise but keeps the labor cheap and fragmented.
The ads also sell a romantic version of this work. They show academics hiking, reading in hammocks, or spending time with friends while earning money between jobs. The message is clear: if academia no longer has room, the platform will let scientists keep working “in their field” without the stability of a real position.
Who profits from the shift
The beneficiaries are clear. Venture-backed firms such as Mercor and ScaleAI stand between desperate workers and AI labs that need expert human input. The article notes that Mercor’s founder, Brendan Foody, became a paper billionaire in a matter of months by supplying AI companies with researchers trying to stay afloat.
At the same time, major tech companies continue to invest heavily in model development while relying on the labor of workers who are often underpaid and insecure. That creates a system where wealth concentrates at the top, but the people producing the underlying value remain exposed.
The broader pattern is not just about AI. It also reflects a deeper move in Silicon Valley to redefine progress as whatever produces fast returns for investors, even when it weakens the public institutions that generate long-term scientific breakthroughs.
The long-term risk for science
Basic research rarely makes quick profits, which is why it has historically depended on public support. The problem now is that the tech industry is treating that public system like a resource to strip down and replace with private platforms.
That approach may generate short-term gains for a small set of investors and executives. It also risks leaving a generation of scientists with fewer pathways into stable work, weaker research institutions, and a labor market designed to treat expertise as a flexible input rather than a profession.
The result is a scientific workforce increasingly pushed into contractor-style jobs that look modern but often reproduce old forms of exploitation. Silicon Valley helped build the public science system that powered its rise, and it is now profiting from the instability created as that system is pulled apart.
Read more at: www.thenation.com
