The report OSTP Director Michael Kratsios delivered to the president this week is a serious document, and treating it as a press release would be a mistake. Science: A New Golden Age runs 122 pages, is explicitly positioned as the first comprehensive rethinking of the US research enterprise since Vannevar Bush's Science: The Endless Frontier in 1945, and arrives with something reports usually lack: a mechanism.

That mechanism is the news. Alongside the report, Kratsios and OMB Director Russ Vought issued a Fiscal Year 2028 R&D Priorities Memorandum directing federal agencies to submit implementation action plans within 90 days and fold the guidance into their FY2028 budget submissions. Reports do not move money. Budget memoranda do, and this one reaches across roughly $200 billion a year in federal science and technology spending.

The reaction from researchers has been genuinely split, and the split is more interesting than either the celebration or the alarm, because a good deal of what the report proposes has been advocated for years by people with no connection to this administration.

The diagnosis is not a partisan invention

Start with what the report gets right, because dismissing it makes the analysis worse.

Its central empirical claim is that investment in research has grown while scientific productivity has slowed. That is not a talking point. It reflects a substantial body of academic work, including research on declining returns to R&D spending and on the falling disruptiveness of published papers relative to the volume produced.

The report's specific complaints are similarly grounded. It cites a finding that investigators spend nearly half of their federally funded research time on paperwork rather than research or teaching, which matches what almost any working scientist will tell you unprompted. It notes that effective indirect cost rates at NIH-funded institutions average more than 40%, while those same institutions frequently accept 10 to 15% overhead from private funders. And it argues that incentive structures reward conformity over bold inquiry, which is the standard critique of consensus peer review, made for years by metascience researchers who would not describe themselves as administration allies.

Several of the proposed remedies come from that same reform literature: faster grants, golden tickets that let a single enthusiastic reviewer fund a proposal others would reject, long-horizon awards, ARPA-style program managers with real authority, and dedicated metascience units to study what works.

The structural problem with funding individuals

The report's signature proposal is to refocus support on individual scientists rather than legacy institutions. The intellectual case is strong. Funding a person for a long period, rather than a specific project on a short cycle, lets researchers change direction when the evidence points somewhere unexpected, which is precisely what project-based grants discourage. The Howard Hughes Medical Institute's investigator model is the standard example, and studies comparing it to conventional NIH funding have found more risk-taking and more unusual results.

Here is the complication that the individual-funding proposal and the indirect-cost critique create together, and it deserves more attention than it is getting.

An individual scientist is not a laboratory. The lab is a building with ventilation and power redundancy. It has instruments that require service contracts, an animal facility with veterinary staff, an institutional review board, a biosafety office, a grants administration team, a compliance function, and technicians who keep all of it running. None of that belongs to the scientist. It belongs to the institution, and indirect cost recovery is how the federal government pays for it.

Which means the two headline reforms interact badly. Move money from institutions to individuals while also compressing indirect rates, and you have funded a person who still needs a place to work, without funding the place. The HHMI comparison actually illustrates the problem rather than solving it: HHMI investigators are embedded in universities whose infrastructure is paid for by federal grants with full indirect recovery. The model works as a supplement layered on a stable base, not as a replacement for it.

The same logic explains the 40% versus 10 to 15% comparison, which is real and also incomplete. Private funders can accept low overhead precisely because federal grants cover the underlying infrastructure. A foundation paying 12% is not demonstrating that research costs 12% to support. It is benefiting from a base someone else built. If federal rates fell to private levels, the base itself would not be funded, and the private rate would stop working too.

Why the reaction is split, and why that is not hypocrisy

The puzzle for an outside observer is why researchers would object to reforms many of them have requested. The answer is about context rather than content.

At the same hearing week, House Science Committee Democrats pressed Kratsios on reports that nearly $1 billion in congressional appropriations to NSF was being redirected toward a $1.5 billion OSTP-backed project, along with questions about slowed grantmaking. Kratsios responded that the White House is not involved in NSF funding decisions and emphasized the administration's desire to direct spending toward grand challenges. At a separate hearing, members of both parties raised concerns that agency restructuring, staff reductions, and funding cuts have damaged the research base.

That is the context in which the reform agenda lands. A scientist can believe that peer review is too slow and too conservative, and simultaneously be unwilling to hand discretion over their funding to an executive branch they believe is destabilizing their institutions. Those are compatible positions.

There is also a durable structural issue underneath, and it is worth stating in non-partisan terms because it applies to every administration. Nearly every proposal here, replacing consensus peer review with program-manager discretion, funding individuals selected by agencies, and directing money toward designated national missions, concentrates more decision-making authority in fewer, more politically proximate hands. That is exactly what makes ARPA-style models fast and effective, and it is also what makes them dependent on the judgment of whoever holds the authority.

What the Bush comparison actually invites

Positioning this as the successor to Science: The Endless Frontier is a large claim, and it sets up a fair test.

Bush's framework established a specific compact. The federal government would fund basic research without dictating its direction, universities would perform it, investigator curiosity would drive topic selection, and applications would emerge unpredictably downstream. That arrangement built the postwar American research system, and the report is correct that the world it was designed for has changed, particularly as private-sector R&D has grown to dwarf federal spending in several fields.

The alternative on offer combines mission-directed national initiatives, individual rather than institutional support, and AI-accelerated discovery, with the Genesis Mission as the flagship: more than $5 billion committed across 278 AI-for-science projects selected from over 5,000 applications. That is a coherent vision, and it is worth noticing that in one respect it is less a departure than a return. Directed, mission-oriented, patron-selected research is closer to how science was organized before 1945 than to anything unprecedented. Bush's innovation was the deliberate separation of funding from direction, on the argument that nobody can predict which basic questions will matter. Reversing that separation may well be right for current conditions. It should be argued on those terms rather than as pure novelty.

The test that will actually settle it

The useful question is not whether the ideas are good. Many are. It is whether this becomes a different way of doing science or a new label on existing machinery, and three things will show which within a year or two.

Whether the 90-day agency action plans produce structural change or reclassified existing programs. Agencies are practiced at relabeling current work to match new priorities, and the plans will be readable.

Whether new funding mechanisms arrive alongside stable base funding or instead of it. Golden tickets and fast grants layered on a solid foundation are an improvement. The same mechanisms replacing a shrinking base are a reallocation with better branding, and the FY2028 budget submissions will show which is happening.

And whether the discretion these reforms create is exercised in ways that survive a change of administration. The strongest version of this agenda would build durable institutions that outlast their authors. The weakest would concentrate authority that whoever comes next inherits and uses differently.

The report identifies real problems that working scientists have complained about for years, and it proposes remedies with genuine intellectual pedigree. It also arrives from an executive branch in an active dispute with Congress over control of appropriated research funds, which is why the same document reads as overdue renewal to some scientists and as leverage to others. Both readings are responding to something real, and the documents due in 90 days will tell you considerably more than the 122 pages did.

Further reading