NSF Shouldn't Fire So Many People

Within the science community, there is understandably a lot of consternation over the cancellation of research grants. But an equal or greater problem is the way that the Department of Government Efficiency (DOGE) is demanding widespread firings and reductions in the federal workforce.

It’s not clear why DOGE cares so much about firing people. Government salaries make up only 4% of the federal budget. Like it or not, our federal government is essentially a giant machine that collects taxes and then redirects almost all of the money to elderly people, hospitals/doctors, and defense contractors (and does so quite efficiently). A small slice of that gets siphoned off to fund science, and that’s what we care about.

There are many scholars and observers who have written about the ills of government bureaucracy and about the need to reform science funding. None of them have ever said that the first and most important idea is to fire as many random people as possible, without regard for what they do or how qualified they are.

This issue is coming to a head with the recent leak of an internal memo sent around at the National Science Foundation (NSF). While the memo wasn’t from DOGE per se, everyone knows that DOGE is running the show here (to the extent of driving the Trump-appointed NSF Director to resign in protest).

Last week’s memo announced plans to:

• Cut the number of executive positions from 143 to 59

• Cut the number of temporary employees from 368 to a mere 70.

What’s the big deal about cutting temporary employees?

In the case of NSF, those are often the most highly-qualified people who took a pay cut to contribute to American science for a couple of years! Ironically, the same is true of most DOGE employees, so you’d expect them to understand this issue.

Firing people with outside expertise might particularly harm the new Technology, Innovations, and Partnerships (TIP) Directorate at NSF. TIP was created by the CHIPS and Science Act of 2022, and while still a new kid on the block, has become a Schelling point for people working on innovative R&D in the US.

TIP’s mission is a departure from the traditional NSF approach—instead of focusing on basic science in everything from physics to computer science, TIP focuses on targeted R&D investments to promote economic competitiveness and national security.

In July 2024, TIP released its first Roadmap, which listed 4 technology areas for future grantmaking: AI, biotechnology, advanced communications, and data.

How did TIP choose these 4 areas? By assessing the risks to U.S. competitiveness. After all, if we want to be competitive with China, we have to think strategically about R&D investments.

What is TIP specifically doing? Here are a few examples of the programs they are standing up [full disclosure: I have served on a peer review panel for TIP, although not for the programs mentioned below].

• Breaking the Low Latency Barrier for Verticals in Next-G Wireless Networks. This two-year initiative seeks to enable new classes of mobile applications, such as telesurgery, that require networks that are much more responsive than those we have today. Researchers have already invented the technologies that can make low latency networking possible and this initiative is teaming those researchers with industry experts to work on building out their capabilities. If we’re the first to demonstrate and deploy a whole new generation of applications, that will allow U.S. companies to re-establish American leadership in wireless communications.

• Use-Inspired Acceleration of Protein Design (USPRD). This three-year initiative intends to accelerate the commercial adoption and wide-spread use of AI-based approaches to protein design. The work supported through this program aims to break down barriers to the use of AI-enabled protein design and demonstrate how a new genre of proteins can redraw the competitive landscape of the U.S. bioeconomy, providing the US with new ways to manufacture a variety of economically important materials.

• Privacy-Preserving Data Sharing in Practice (PDaSP). In partnership with Broadcom’s VMware division, Intel, the Federal Highway Administration and the National Institute of Standards and Technology, the PDaSP effort aims to advance the ability of organizations and scholars to pool and analyze data without having to share the raw data itself. Some of the more promising use cases involve opportunities to improve healthcare delivery and to evaluate the effectiveness of a wide range of government services.

• Building the prototype open knowledge network. In collaboration with five other federal agencies, TIP has invested in 18 multidisciplinary teams to create an prototype open knowledge network to make it far easier to work with data from multiple sectors. New AI tools will be far more useful if they can access wide-ranging and reliable data.

Each of these programs is working with folks outside the usual “government science funding” orbit. That is, instead of being limited to university labs, NSF is working with a range of organizations, including startups, established companies and non-profits.

That’s all to the good. Indeed, the CHIPS and Science Act (see section 10396) specifically encouraged the TIP directorate to:

• hire program directors “from a variety of backgrounds, including industry,”

• set up a program to “facilitate recruitment of eminent experts in science or engineering” (along with the freedom to pay higher salaries), and

• recruit “individuals with expertise in business creativity, innovation management, design thinking, entrepreneurship, venture capital, and related fields.”

In other words, TIP was encouraged to be like DARPA — to hire folks with rotational assignments who would come into NSF, focus on specific tech areas of relevance, and then eventually move on to allow the next class of outside tech experts to rotate in.

We should be eager for outside experts to sacrifice their time and career pathways to serve in government roles for a year or two. It makes no sense to have DOGE trying to get rid of almost all of those outside experts. That endangers all of the ambitious programs listed above (and much more).¹

As the administration considers how to achieve the vision laid out in the President’s letter to his incoming science advisor, it would be better off looking to NSF’s TIP Directorate than to DOGE for advice. The President’s letter asks:

How can the United States secure its position as the unrivaled world leader in critical and emerging technologies — such as artificial intelligence, quantum information science, and nuclear technology — maintaining our advantage over potential adversaries?

Easy answer: By continuing to maintain robust funding to NSF and TIP (and analogous efforts at other agencies), rather than letting DOGE decimate science funding and employment for no reason that anyone has even tried to articulate.

Indeed, in all of the debates about science funding, one point seems to be lost: The reason that the federal government funds science is not just to provide scientists or government bureaucrats with jobs, nor is it an act of charity towards universities and academic hospitals. We fund science because it is in our national interest to do so!

Whether it’s DARPA creating the original Internet, or an NSF grant funding the origins of Google, or the NIH funding work that led to a miracle cure for a type of leukemia, or NSF funding work that ended up making PCR usable, or Defense research that led to the GPS system that made it possible to have real-time maps on our phones, we all benefit from putting public money towards scientific R&D.

With China threatening to pull far ahead of us in R&D spending and productivity, the last thing we should do is let DOGE decapitate our science agencies like NSF.

“Penny-wise but pound-foolish” barely begins to capture the ill-advised nature of what DOGE is doing. It’s shooting ourselves in the foot, over and over again. DOGE should stick to issues where it has some logical reason for taking action (e.g., updating IT systems) and some level of expertise that is above zero (hardly anything to date).

Meanwhile, Congress should emphatically push back against these efforts to derail American scientific competitiveness. DOGE is on the way out. Time to stand up for American R&D.

1. A May 9 temporary restraining order currently keeps NSF from implementing these employee reductions, although it remains to be seen what will happen in that case on appeal.