In 2008, Chris Anderson published “The End of Theory: The Data Deluge Makes the Scientific Method Obsolete.” Prophesying our current age of AI, it was one of those rare magazine articles that launched serious scholarly debates. Anderson’s thesis was straightforward: in an era of “big data,” scientific quests to find the ultimate causes of things were outdated. Instead, we can throw data at our problems, training machines to gradually discern what courses of action lead to better outcomes. Whether the machines (or their programmers) actually understood the dynamics behind the discernment was a distinctly secondary consideration. What Google researchers deemed the “unreasonable effectiveness of data” would rule.
Fast forward to 2025, and a new end of theory looms -- this time directed by the tech sector’s richest man. The Trump Administration has frozen research grants and fired scientists indiscriminately. Even sectors that are ostensibly Administration priorities, like AI and computer science, are being hit hard. Stanford -- one of the richest universities in the world -- has imposed a hiring freeze, and similar institutions are sure to follow. Crown jewels of health research are now in question; expect more of that inquiry to move to an already surging Chinese biotechnology industry. The US may not even make a flu vaccine for the coming winter, given Department of Health and Human Services Secretary Robert F. Kennedy’s bristling hostility to such preventive medicine. In February, the FDA suddenly cancelled a meeting of experts to choose flu strains to target.
Policy making in Trump’s second term has been, above all, fast. Alliances that took decades to build have been spurned in mere weeks. A firehose of executive orders confound courts. Mass firings have shocked the civil service. The Silicon Valley ethic of “move fast and break things” has come to the White House. A once-fringe political philosophy -- accelerationism -- permeates the halls of power. The tenets and implications of accelerationism can help us make sense of the last few weeks of disruption, while providing clues as to what is in store in the future.
The core of contemporary Silicon Valley accelerationism is what venture capitalist Marc Andreessen calls the “techno-capital spiral.” On this theory, pouring capital into technological development leads in turn to more capital accumulation, which can be invested in more technology, and so on. “We believe that there is no material problem -- whether created by nature or by technology -- that cannot be solved with more technology,” Andreessen argued in his Techno-Optimist Manifesto. Andreessen clarified and popularized an accelerationist ideology that has long cheered on the dissolution of non-market institutions that regulated business.
The devaluation of science -- manifest in massive cuts and spending freezes at the US National Science Foundation and National Institutes of Health -- may seem like an odd accelerationist priority. Isn’t scientific advance the foundation of technological development? For decades, there has been a symbiotic relationship between American science and its technology-driven industries. Free to pursue blue sky research, scholars made advances in artificial intelligence, machine vision, preventive medicine, and countless other fields. Industry capitalized on this inquiry, both commercializing these findings and hiring the bright students taught at world-leading universities. Proud of these synergies, bipartisan coalitions of Democrats and Republicans funded university research as an engine of economic innovation.
Those days now appear to be over. Both interests and ideology drive the current split between the technologists running so much of Trump’s policy, and the broader scientific community. While their long-term self-interest relies on the public goods that free scientific research creates, AI hyperscalers and platform monopolists can capitalize on the disorder now. Already burdened by a thicket of administrative rules, service-minded, government-funded scientists may finally decide to give up on the public sector and move to industry. That creates a higher supply of expert workers, likely reducing their pay. Cheaper talent boosts profits.
Such crass calculations may shock scientists. Some thought that the 2024 alliance between America's richest technologists and Team Trump portended a rational industrial policy that would protect basic research. What they failed to grasp was a long-standing and growing rift between America’s technology firms and the very institutions that have long underwritten their prosperity.
Consider, for instance, tensions between large technology firms and universities. Several tech firms have dallied with their own training and certification programs, seeking to cut out the university as a middleman education provider. Cheap “massive open online courses” and badging certifications have been a watchword of education futurists for over a decade. Now generative AI is supposed to conjure infinite artificial professors of the future. Such innovations would leave little, if any, economic basis for contemporary research universities.
Industry also wants more control over university research agendas. Not content to simply sponsor research, in 2015, Uber hired 50 faculty from Carnegie-Mellon’s robotics program. Pay gaps and stratospheric AI research costs have driven many scholars from the free-thinking atmosphere of universities to the more directed and profit-driven work of corporations. Growing pressure to obtain grants and navigate byzantine government funding rules have also made the private sector more attractive over time.
Of course, not all of the US technology sector backs what the US’ ersatz Department of Government Efficiency is now doing to undermine health, environmental, and other research. But hyperinequality narrows the range of persons who really matter in America’s political landscape. Musk and Trump can credibly threaten a well-funded primary threat against any Republican who deviates from their accelerationist orthodoxy. Even if progressive tech tycoons like Bill Gates tried to rally Democrats toward obstructionism in response, there is no convening figure like Trump on the Democratic side. So the accelerationist agenda rockets ahead, as the techno-capital spiral now overwhelms the very democratic processes that could once bring it to heel.
Despite the self-reinforcing dynamics of technological advance generating financial prowess and political dominance, much can still derail the techno-capital spiral now lofting Team Trump to new heights of power. Inflation threatens a “revenge of the real,” as vaunted advances in AI are doing little to reduce the price of eggs and other staples. They may instead be helping landlords fix higher rents via algorithmic coordination, and sellers generally to calculate new ways to extract funds from consumers. The resulting inflation will only exacerbate an anti-incumbency bias already common in post-COVID democracies.
A 2014 Google research paper called machine learning the “high-interest rate credit card of technical debt,” given the ongoing costs of maintaining the massive compute and data gathering anticipated by Anderson in his “End of Theory” article. So too may today’s accelerationist disruption of American science prove a staggering and compounding debt for years to come. Like the steroids at the core of the “Enhanced Games” recently endorsed by Donald Trump, Jr., a rapid shift of scientists from the public to the private sector may goose growth in the short run, with terrible side effects in the long run. The US could easily lose the very open-ended research critical to its most innovative products and services. There are few worse ways to lose the “AI arms race” ostensibly motivating today’s accelerationist frenzy.
Frank Pasquale
Frank Pasquale is Professor of Law at Cornell Tech and Cornell Law School. The views expressed here are the writer’s own. — Ed.
