The Trump administration has proposed 40 to 60 percent budget cuts to the NIH, NSF and other science agencies while simultaneously promising a "golden age" of research — and a new dialogue published in The New Atlantis argues no one inside the administration is actually running a coherent strategy connecting the two.

The exchange between New Atlantis editor Ari Schulman and M. Anthony Mills, a senior fellow at the American Enterprise Institute's Center for Science, Technology, and Society, works through the central puzzle for R&D managers: whether the "good cop" rhetoric — the May 2025 "Restoring Gold Standard Science" executive order, promises to ease grant paperwork — and the "bad cop" crackdown on elite universities, DEI, mRNA vaccine research and pandemic preparedness funding are two halves of a master plan.

Mills offers a "steel man" version: the postwar federal research system, built after 1945 and shaped by Vannevar Bush, suffers from well-documented pathologies — publish-or-perish incentives, p-hacking, hyper-specialization, orthodoxy over heterodoxy — that decades of incremental reform have failed to fix. On this reading, disruption is the medicine: break institutions dependent on the status quo so reformers can rebuild.

He then dismantles the coherence claim. The administration, he argues, is a collection of factions with incompatible views on science: populists distrustful of experts, social conservatives focused on bioethics, the MAHA movement hostile to pharma and agriculture, the techno-optimist Tech Right, national conservatives, and a group he calls the "science reformers" — figures like NIH director Jay Bhattacharya, OSTP head Michael Kratsios and NSF TIP directorate lead Erwin Gianchandani — who genuinely want to accelerate science through metascience and the administration's Genesis Mission, not abolish agencies.

The reformers' problem, Mills argues, is that decisions on budgets and staffing are being made elsewhere — by actors like HHS secretary Robert F. Kennedy Jr., OMB director Russell Vought, and previously Elon Musk's Department of Government Efficiency — whose goals have little to do with science policy. "The destructive, punitive part of the strategy, if it really is or was a strategy, has undermined the constructive part," Mills says.

For lab managers, the historical framing matters. Mills notes the GOP has historically championed federal science funding — Congress doubled the NIH budget in the 1990s even after Newt Gingrich shuttered the Office of Technology Assessment. What has changed, he identifies in three markers: aggressive political control over grant allocation, the most serious challenge to the federal-university research link since the 1970s, and cratering Republican support for the NIH.

The dialogue also weighs whether top-down vision documents can change scientific culture. Mills contrasts Kratsios's report "Science: A New Golden Age" with Bush's 1945 "Science — The Endless Frontier": Bush succeeded because he was a proven institution-builder — co-founder of Raytheon, head of the wartime OSRD that oversaw the Manhattan Project and radar development — speaking in an era of institutional optimism. Kratsios, Mills argues, has written "a document written in the spirit of 1945 for a cultural moment that more closely resembles 1973."

Mills's sharpest line targets the underlying diagnosis: "It's a form of apocalypticism applied to science policy that rationalizes what is ultimately a much easier but far less effective mode of wielding political power."

His prescription for restoring a social consensus on American science is not presidential leadership but Congress. He points to the legislative wave of the early 1970s — the Mansfield Amendment, the Federal Advisory Committee Act, the National Research Act, the Technology Assessment Act — as proof Congress can rebuild oversight structures, and notes "glimmers" of pushback already visible in Republicans' quiet refusal to enact the administration's most radical proposed cuts to the NIH.

Whether that congressional reassertion materializes will determine whether the current disruption ends in reconstruction or simply in a smaller, less trusted federal research enterprise.