The new "science golden age" looks suspiciously like a grift
Or at least that’s the vision emerging from a new 123-page “Report
to the President” from Michael Kratsios, the Director of the White House Office
of Science and Technology Policy.
In a normal administration, you could probably skip over
learning about a science policy document — but this is no normal
administration, and this is no normal science policy document. Instead, it’s
Trump’s attempt to end the actual golden age of science that was kicked off in
1945 with the publication of '“The
Endless Frontier” by Vannevar Bush.
Kratsios’s report is supposed to be the successor to Bush’s,
and even apes the format, but this isn’t really a document about science
policy. It’s one about ending science research as we know it in favor of
shoveling cash at big tech companies.
While it might be temping to wave this away as an
aspirational policy document, Trump has already begun to shift
billions in research funds to AI projects and has taken
a hatchet to science funding across the board. He’s teed up the shift
Kratsios proposes here, and since neither Congress nor the Supreme Court seem
inclined to stop him from doing whatever he wants with taxpayer money, it isn’t
doomering or hyperbole to say we are staring down the barrel of a world where
the only “science” the government intends to support is for-profit tech and
defense advancements.
Science for dummies
Part of how you can tell this isn’t a document about science
is that you are immediately faced with an absolute wall of business-speak of
the worst sort. Did you know we must “tightly couple our science and technology
enterprises to ensure that groundbreaking ideas invented in the United States
are rapidly prototyped, tested, manufactured, and scaled domestically”? You do
now!
That tone isn’t surprising, because Kratsios has no
background in science or research. He has a bachelor’s in political science and
spent most of his career working for
Peter Thiel at Thiel Capital. After a stint in the first Trump administration
as Chief Technology Officer, where he also did not do any
science, Kratsios was tasked in
2020 with overseeing the Pentagon’s research budget for a brief time, a
position that, as the Wall Street Journal put it, “carries a great deal of
power in terms of dictating investment allocations.”
He spent the Biden years at Scale AI as “head of strategy,”
which you will also note is neither a science nor a research job. But Scale AI
has contracts with
the Department of Defense and, as the Washington Post wrote in 2023, “has been
aggressively pitching itself as the company that will help the US military in
its existential battle with China.”
And now you have the necessary background for our new golden
age, which is going to be a golden age of taking money away from scientists and
universities and giving it to companies like Scale AI.
In a genuinely galling move, Katsios spends a lot of time
slamming universities because of “administrative burdens on researchers” and
the “well-documented inefficiencies” in academia, the message being that those
damn universities waste the time of amazing researchers with paperwork.
Except it’s the administration that is attempting to impose
inefficient, unnecessary burdens on those researchers with a proposed Office
of Management and Budget rule that would require senior political appointees to
approve every science grant and ensure it aligns with Trump’s personal
priorities. Researchers would also have to seek prior government approval to
attend conferences or publicize their research. And political appointees could
terminate any grant at any time if they feel like it.
But when it comes to AI, we’re going to “restore
permissionless innovation.” The report proposes to give lucky ducky innovators
a pile of cash up front for grants lasting five years or longer, saying “awards
should be fully-funded in year one, with all resources earmarked upfront, to
minimize administrative burdens and reduce pressure for researchers to generate
intermediate results to secure continued funding.”
Yeah, wouldn’t want to have to make the Very Special Tech
Bros show results or get money parceled out year by year like the plebes at
universities.
The report also proposes that we should have “fast grants”
with “simplified applications requiring just a few pages of writing” because
that is definitely how science works, just writing a bit off the dome and
getting some cash.
Oh, and we should have prizes and challenges, where multiple
teams compete. The proposed prizes include incentives like “advance procurement
commitments,” which means the federal government agrees to buy whatever is
being developed. There’s also “regulatory fast-tracking,” making sure all those
great AI thingies don’t have to wait too long to be unleashed on a market.
These aren’t incentives to do science. These are incentives
to develop things with tax dollars that you can then sell to the government for
more tax dollars.
In the suggestion that might be the most unhinged, the most
very opposite of how both government and research work, our new age will
include “golden tickets.” That’s where an individual reviewer in an agency —
yes, just one person — can “recommend unconventional proposals that may not
pass consensus-driven review panels” because somehow this will “surface
valuable ideas too divisive for committees and attract higher quality reviewers
by empowering them to exercise independent scientific judgment.”
Yes, that’s saying that a random person in an agency should
be able to shovel money at some idea even if it can’t pass a review panel, and
that the opportunity to do so will attract better employees because apparently
all the best minds in science are chomping at the bit to be allowed to
unilaterally fund AI research based on their personal vibes.
But that isn’t at all how actual science works, and actual
scientists aren’t unhappy about review panels. They work collaboratively, not
siloed in some secret race to create the next AI chatbot. A 2020 study by the
National Bureau of Economic Research examined whether peer review panels succeed in
helping agencies predict which research ideas will best advance science. Peer
review was better at finding “diamond in the rough” applicants than a single
government program official.
How about that.
Trump always wants a taste
It’s painfully clear that what Katsios is trying to do here
is turn the whole of government science research and funding into the sort of
“move fast and break things” paradigm that is the hallmark of big tech. That
method works to line the pockets of tech executives, but it isn’t science.
There’s a reason the government funds different kinds of
research than private tech companies do. What Katsios either fails to
understand or deliberately ignores is that the federal government is the main
funder of basic research precisely
because it can take a long time for it to bear fruit and is not
intended to be profitable as such.
“Basic” here does not mean rudimentary or simple, but rather
research that is designed to
advance scientific knowledge instead of achieving a specific outcome or
product. Big tech companies don’t widely fund that sort of research, because it
doesn’t make them money. Industry dollars account for only about 20 percent of
basic research funding, while the government contributes around 60 percent.
Big tech companies and startups do, however, rely on
and benefit from
that basic research. A 2019 study found that 35
percent of all patents obtained by venture-backed startups from 1976 to 2016
cited federally supported research. Basic research funded by the NIH contributed to
every single drug the FDA approved from 2010 to 2016.
Unlike inventing the next ChatGPT, actual science is slow
and iterative. Government funding of research that, on its face, would not
appear to lead to any big new invention is actually what underpins those big
new inventions.
If the government hadn’t funded research into
desert lizard venom, we wouldn’t know about semaglutide, and we wouldn’t have
Ozempic. If research scientists hadn’t studied how bees optimize nectar
foraging within a colony, we wouldn’t have the algorithm that assigns internet
traffic across computer servers.
The report assiduously ignores that a symbiotic, profitable
relationship between the federal government, universities, and private
companies already exists. Each federal research dollar generates about
$1.50 to $3.00 in long-term economic benefits. The $36 billion the
government awarded in
NIH research funding in fiscal year 2025 generated $94 billion in new economic
activity.
Trump’s proposed new golden age is intended to strip those
billions from universities, resulting in the dismantling of departments and the
shuttering of labs. But Katsios has a plan for that: Universities should just
become trade schools to churn out no-show STEM degrees!
“Require universities and community colleges to embed
practical technical training and externships into STEM curricula. Let hands-on
experience and industry credentials count toward degrees.”
It’s quite the feat to, on the one hand, strip schools of
funding for scientific research while simultaneously demanding to control what
is taught and, oh, also, colleges need to give degrees to employees at big tech
companies just for being employees at big tech companies.
The new golden age would also exacerbate the brain drain
Trump’s policies have already caused. Nature found that in the first quarter of
2025, US scientists submitted 32
percent more job applications abroad than in 2024. A March 2025 poll by Nature
of roughly 1,600 scientists found that over 1,200 said they were
considering leaving the
country. Europe is investing $565
million to make it a “magnet for researchers” and provide funding for
scientists who relocate.
The new golden age is also about embedding Trump’s
xenophobia into the fabric of American science. The report complains of an
“over-reliance on foreign students” that “creates a security challenge our
institutions are ill-equipped to address” and that a “reliance on foreign
talent sidelines American students.”
This is, to put it mildly, nonsense. International students
pay far higher tuition and
often receive no financial aid. Their tuition dollars are what make it possible
for American students to get financial aid.
In the 2023-2024 school year alone, international
students contributed $43.8
billion to the US economy and supported over 378,000 jobs. Those 1.1 million
international students accounted for only six percent of the total number of
students in higher education that year, but were such an economic engine that
for every three international students, one job was created or supported.
It’s impossible not to wonder whether part of Trump’s
loathing for universities and their leading role in scientific research is that
there is no way for him to corrupt the process and make money off it in the
same way he can by giving millions and billions to Silicon Valley types who, in
turn, give him millions in cash. There’s no way for his children to invest in
universities that the government then gives no-bid contracts to,
fattening the family’s bottom line, which is pretty much the Trump sons’
current venture capital practice.
Trump and Katsios’s view of science is narrow, cramped, and
wrong. We’re not in a crisis, we’re not a failed state. We’ve been a
science superpower for
decades, and the only thing that threatens that is Donald Trump.
Tech bros are already living high off the hog in this
administration. They don’t need any more of our tax dollars. Hopefully this new
golden age never comes to pass.
Thanks for reading Public Notice. This post is public so
feel free to share it.
