What this covers
Sean Carroll explains what happens when the Trump administration, working through Elon Musk's Department of Government Efficiency, cuts federal science funding—particularly at the National Institutes of Health and NSF—by slashing indirect costs from roughly 60% to 15%. The core of his argument unfolds on three fronts: what indirect costs actually are, why the cuts are constitutionally illegal, and how they damage American scientific capacity. Carroll addresses the material with an ear toward skeptics of government spending, explaining both the mechanics of research funding and the legal violations involved without assuming prior knowledge of either.
Indirect costs emerge as the central point of contention. Carroll walks through the reality: universities negotiate rates to cover buildings, utilities, administrative staff, and compliance overhead—expenses that don't vanish when funding shrinks. He debunks the widely circulated claim about a million-dollar shrimp treadmill (actually $1,000 in grant support for a locomotion study), and clarifies Elon Musk's public misunderstanding of how the system works. The cuts violate Separation of Powers, he argues, since Congress controls the purse under Article 1, and specific statutes protect allocated funds; an executive order cannot unmake law. Beyond legality lies consequence: mid-stream termination of drug trials harms enrolled patients, universities face cascade failures from Fixed Costs they cannot reduce, and foreign scientists are already steering graduate students away from American positions. Carroll sets this against the Move Fast and Break Things ethos, which he argues catastrophically misfits medical research and governance itself, where distributed power and Fallibility of All Groups demand cooperation rather than unaccountable cuts.
Carroll argues that the Trump/DOGE cuts to NIH and NSF funding—especially slashing indirect cost rates from ~60% to 15%—are devastating to American science, are being implemented illegally through executive orders that usurp Congress's power of the purse, and reflect a reckless 'move fast and break things' process that ignores how the highly-scrutinized, university-based research system actually works.
- Indirect costs are real, negotiated, down-to-earth expenses (buildings, lights, administration) that cuts cannot eliminate because fixed costs remain.
- Spending money is constitutionally Congress's job; cutting allocated funds by executive order violates statute and is being challenged in court.
- The cuts damage U.S. scientific leadership and trust, deterring foreign students and postdocs from coming to the U.S.
Abrupt funding cuts will cripple ongoing research, harm patients in trials, and destroy the U.S. reputation for reliability with international scientific partners.
- At the National Science Foundation, employees have been told to expect about half the workforce will be let go, and there is a rumor (not yet realized) that the overall NSF budget will be cut by 60%, which would effectively shut down NSF-funded research.
“employees of the National Science foundation have been told to expect that about half of the workforce is going to be let go... the overall NSF budget is going to be cut by 60%”
- The funding cuts will cause a major shift in how the US is perceived by international partners and competitors, with foreign physics professors already planning to advise their students not to pursue graduate or postdoctoral positions in the United States, undermining America's ability to attract scientific talent.
“they have students who are thinking of either going to graduate school in the United States or taking up a postdoctoral position in the United States, and they are now going to strongly recommend that that not happen”
- Science takes time, and time requires trust and reliability; the US already has a slightly sketchy reputation for keeping promises—exemplified by building the Superconducting Super Collider halfway with international funding and then canceling it—and the current upheaval further erodes the trust needed for foreign scientists to commit to working in the US.
“the Superconducting Super Collider was a large particle physics experiment that was international in scope. Other countries put money into it and we built it halfway. And then we Canceled it. This makes people a little nervous.”
- The Silicon Valley 'move fast and break things' approach applied to medical research means breaking people's lives in a real way, because abruptly cutting funding forces drug, device, and therapeutic trials to stop mid-stream, which is devastating not only for the research but for the patients enrolled in those trials.
“the Silicon Valley motto is to move fast and break things. The things being broken are people's lives in a very real way.”
- Within days, two major lawsuits challenged the NIH indirect cost order—one by attorneys general of about 22 (Democratic-led) states, leading a judge to temporarily stay the cut in those states, and another nationwide suit by universities including Johns Hopkins—but litigation takes time and resources and leaves the cut active in the other 28 states.
“there have already been two major lawsuits about the NIH indirect cost order, one by a bunch of attorneys general of various states, I think 22 states... the judge quickly moved to say, okay the cut in funding cannot go through right now.”
Executive orders cutting congressionally allocated science funding violate constitutional separation of powers and existing law.
- Impoundment—the president declining to spend money Congress allocated—was historically used rarely and judiciously, but Richard Nixon abused it for personal political gain; Congress then passed a law (signed by later presidents and upheld by the Supreme Court) eliminating that power.
“guess who started being corrupt and misusing it? Richard Nixon... Congress passed a law which was signed by later presidents saying, no, there's no more ability to do that. And this was upheld by the Supreme Court.”
- The power of the purse—the ability to allocate money to be spent—is vested entirely in Congress under Article 1 of the Constitution, so the executive branch issuing an executive order to not pay congressionally allocated indirect costs is clearly illegal, especially since specific statutes mandate these allocated monies be paid.
“The ability to allocate money to be spent is a power that is entirely vested in Congress... It's in Article 1. It's the first thing that the Constitution says, spending money is Congress's job.”
- Executive orders do not have the force of law; their legitimate purpose is for the executive branch to interpret and explain in easy-to-read form what Congress meant by the laws it passed, not to invent new laws—the president is not a king and cannot make up and pass laws by himself.
“executive orders do not have the force of law. The idea of an executive order... is that Congress passes laws and the executive branch enforces the laws. And in order to do so, they need to interpret what the laws mean.”
- The Department of Government Efficiency is in a weird legal status: departments are supposed to be created by Congress, but DOGE was not approved by Congress, yet it is headed by Elon Musk who hired young people placed into government agencies with access to all their information, making cuts promulgated through executive orders.
“the Department of Governmental Efficiency, the Doge panel, which is not a department. Departments are supposed to be made by Congress and this was not.”
Indirect costs fund essential research infrastructure that universities cannot absorb; cutting them damages science more than claimed fraud savings.
- The National Institutes of Health has been ordered to cut indirect costs on grants from a typical rate of around 60% down to 15%, which amounts to roughly a 30% cut to the total grant-giving budget and is devastating for biomedical research.
“this number, something like 60% has been chopped down to 15%, which is much, much lower. So overall, a huge cut to spending on grants from the National Institute of Health.”
- Indirect cost rates are negotiated and vary by university and project: institutions in expensive areas like NYU or Columbia can justify higher rates to keep the lights on, while research at cheaper locations like the University of Nebraska, or work done at facilities elsewhere such as CERN, justifies lower rates.
“if you're negotiating with NYU or Columbia and they need a lot of money just to keep the lights on, as it were, they can make a good case for getting additional indirect cost funding. Whereas if you're at the University of Nebraska, where the research is good but the land is cheap, your indirect cost rate is going to be low.”
- Indirect costs pay for the real infrastructure needed to do research—offices, buildings, lights, heating, administrative staff—and as a matter of empirical fact typically do not even cover the universities' actual costs; universities accept losing money on grants in exchange for prestige, buildings, and research profile, so the funds are not simply enriching administrators.
“as a matter of empirical fact, it is not enough. Typically, the indirect costs do not cover all of the actual costs that the universities pay for.”
- The NIH indirect cost executive order was issued late on a Friday to take effect the following Monday and applies not only to future grants but to all existing grants, leading institutions like Penn State to tell medical researchers on NIH grants to stop spending money because incoming funds are needed to cover fixed costs.
“it doesn't go into effect on future grants, it goes into effect on all grants... Penn State... told all of its medical researchers who are on NIH grants to stop spending all of their money.”
- Cutting NIH grant income by ~30% does not reduce a university's costs by 30%, because of fixed costs—just as a household losing 30% of income still must pay rent, utilities, and loans—so the cut grinds research planning to a halt rather than producing a proportional reduction.
“If you lose 30% of your income, you don't lose 30% of your lifestyle because there are things called fixed costs”
- Elon Musk made a common mistake (in a tweet) revealing he doesn't understand the indirect cost system: if a researcher applies for $1,000 with a 60% overhead rate, the university requests $1,600 from the agency—the researcher still gets the full $1,000 and the extra $600 is over and above it, not 60% taken out of the researcher's budget nor 60% of the total.
“It is not that I apply for $1,000 and the university takes 60% of that. That is not how the math works. If I apply for $1,000, I get $1,000, but the university in total gets an extra 60%.”
The grant system is merit-driven and already heavily scrutinized, making social-justice concerns a false pretext for deep budget cuts.
- NSF grant decisions are made by peer-review panels of working scientists, not just agency bosses; collaborators are barred from reviewing each other; and in Carroll's experience panels seriously favored younger researchers doing more interesting work over famous scientists who 'phoned in' their proposals, demonstrating the system rewards merit and effort.
“there were some super famous, successful scientists who applied to grant, put in grant proposals that we looked over and we did not give them money in favor of some younger person who was frankly doing more interesting work”
- Grant proposals, especially in theoretical physics, are partly a game where researchers describe work they just did or are about to finish, because the best available way to judge how good someone's future research will be is to look at how good their recent research has been—with carve-outs for very young scientists without a track record.
“what better way do we have of judging how good research someone will do in the next three years than to look at how good the research was that they did do for the last three years?”
- The fraction of science funding devoted to DEI or social-justice goals is tiny—broader-impacts sections of proposals are typically about outreach, and review panels weight science roughly 95% versus 5% for such concerns—so cutting DEI cannot significantly reduce the budget deficit, and the 30% NIH cut is far larger than any plausible social-justice spending.
“This might be 5% of what they care about, but 95% of it is caring about your science.”
- Money spent on scientific research through the grant proposal system is some of the most carefully scrutinized spending on Earth, with very little room for fraud in the sense of misallocated money, because grants are peer-reviewed competitively—nobody proposes a grant to buy a yacht and gets funded.
“the money that is spent in scientific research through this grant proposal system is some of the most carefully scrutinized spending on Earth.”
Unaccountable executive power concentrated in a small group without expertise or transparency enables corruption and reckless action.
- In the rush to act quickly, the NSA was instructed to delete any web pages containing a forbidden list of words, an indiscriminate keyword-based approach that is not carefully thought out and will cause real damage to one of the most sensitive parts of the federal government.
“If the web pages contain any example of a certain list of forbidden words, they had to be deleted. This is not carefully thought out... This is going to cause real damage.”
- Because both political sides are susceptible to mistakes and corruption, and because every group tends to believe only it knows how to run things, democracy can only work for everyone if groups share power, cooperate, and compromise—running around unchecked breaking things will not lead to less error or corruption.
“The only way that can work for everyone is if people in all the different groups share power... we have to be able to cooperate, we have to be able to compromise, we have to be able to share power.”
- Administrative bloat in research institutions exists not because anyone likes spending on administration, but because government bureaucracy creates compliance hoops to prevent waste and fraud; cutting the administrators that enforce these would open the door to actual waste and fraud, creating an eternal bureaucratic conundrum.
“It's usually because the government bureaucracy has invented all sorts of hoops that you have to jump through in order to not have waste and fraud... that's what leads to the administrative bloat.”
- Concentrating large unaccountable power in a small number of hands without transparency or expertise is an obvious opening for corruption that is actually occurring—oversight officials overseeing Elon Musk's companies are being fired, and rules barring companies from bribing foreign officials have been openly removed under a 'level the playing field' rationale.
“Elon Musk's companies are overseen by various parts of the federal government and the people doing those oversight are being fired. There's a set of rules here in the United States that we're not allowed to, that companies are not allowed to bribe foreign officials... Those rules have now been removed”
Scientific progress drives broad human progress, and dismantling the U.S. system abandons a world-historical responsibility.
- The US system of funding scientific research was not deliberately planned but grew organically over time, with agencies like the DOE evolving from the Atomic Energy Commission and Manhattan Project, NASA for astronomy, NOAA for oceans and climate, and the NSF serving as the catch-all agency for everything else.
“what used to be funding for building nuclear weapons is now the funding for doing particle physics”
- Doing science at universities rather than standalone research institutes is cheaper and more effective because universities have a diverse portfolio of income (tuition, alumni donations, grants), so a dedicated institute funded only by grants would actually have far higher indirect costs; plus universities create constant interaction between top scientists and students that benefits both.
“If you are starting a research institute, you pay for everything... then your indirect costs, as it were, would be super duper high.”
- The importance of US scientific leadership is not nationalistic pride but that the US has demonstrated a way to be great at scientific discovery; by undermining the funding system, the country is giving up the responsibility of leading the world in science, a world-historical decision with consequences reaching deeply into the future.
“It's not that we're the leader. It's that we're showing the world a way to be great at this. And now we're giving up on that.”
- Science—proposing theories and testing them experimentally—is a huge driver of progress in a broad sense including economic, social, and cultural progress, and this scientific progress is enabled by a system of support rather than happening on its own, especially now that low-hanging scientific fruit has been picked.
“science... is a huge driver of progress and progress in a very broad sense, economic progress, social progress, cultural progress. And that scientific progress is enabled by a system of support. It doesn't happen all by itself.”
Summer salary incentives through grants are an effective policy tool that motivates continued research productivity in tenured faculty.
- Universities typically guarantee professors only nine months of salary; successful researchers can apply to have one or two months of summer salary paid through their grants, which functions as an incentive system that motivates tenured faculty to remain active in research and thereby increases total research output.
“it's basically a little incentive system for tenured faculty member to continue to be good at doing research”
- The viral claim that a million dollars was spent on a 'trampoline for shrimp' is nonsense: it was a researcher studying shrimp locomotion who initially built a device out of his own pocket, got good results, and then received about $1,000 in grant funding—the million-dollar figure was fabricated or conflated with all the university's grants.
“It was never a million dollars that was just made up... There was a grant that involved trampoline for shrimp. It was done by a researcher who was trying to understand shrimp locomotion.”