Public comment period re US science, ends Dec 26

For those in the US:  Here is a link for submitting a public comment regarding OSTP's request for information about "Accelerating the American Scientific Enterprise".   The US government does this at some rate, requesting public feedback and input on policies under consideration or development.  Unfortunately, this public comment period ends on December 26.  

Here is the slightly longer version of what I submitted (I had to trim it back b/c of character limits).  It doesn't touch on everything in the RFI, but it's a start.  Many readers will likely disagree with what I wrote, and that's fine.  Submit your own comments.  Better that OSTP hear from a variety of perspectives.

Update:  I’ve had some ask me what the point of this is, since the comments will likely not be read in a meaningful way.  I assume they will feed the comments through an AI tool and have staffers spot check ones that are flagged as interesting (by whatever criteria). I know I’m old fashioned, but I still feel like silence is acquiescence edging toward complicity. Saying something at least helped me organize my ideas a bit.

----

I appreciate the desire and need to improve the American scientific enterprise. Other rival nations are making enormous investments in basic and applied scientific research, and it is critical for US competitiveness that the US not flag or falter.

I have two overarching suggestions before delving into the specific questions posed in the RFI.  

First, it is of vital importance that the US scientific enterprise solicit advice and input from practicing scientists and engineers through advisory panels and committees.  In the last year, the US government research agencies have dissolved or disbanded a huge fraction of advisory bodies.  I think this is a mistake.  While there are always concerns about whether experts are too entrenched to be bold, across-the-board devaluing expertise is not the way to go.

Second, there needs to be coherence between priorities and financial allocations.  Complaining that the US is falling behind in critical research areas (e.g. advanced materials) while simultaneously slashing federal support of those research areas is both self-defeating and internally inconsistent.  There have been large scale cuts across many agencies in critical scientific areas, seemingly at odds with the stated policies about innovation and competitiveness.  

Regarding specific points from the RFI:

Encouraging public-private collaboration runs into the problem that existing economic incentives (the constant need for short-term returns) disfavor companies investing in research that doesn't have clear, nearly-immediate payoffs.  Norm Augustine has spoken publicly about how opening a research laboratory caused Lockheed Martin's stock to fall. (see here:  https://www.aaas.org/news/basic-research-needs-sustained-federal-investment-says-norman-augustine)  Moreover, since many very wealthy corporations (e.g. Tesla, 3M) already pay almost no federal taxes, it is challenging to encourage public-private partnerships through offering companies tax incentives.  The issue is less of a problem for small and startup companies, in part thanks to programs like SBIR and STTR.  Programs like ARPA-E and ARPA-H also are important to consider and expand.

The most financially well-off sectors, all of which depend critically on a spectrum of research from basic to the applied and near to long term, are the financial industry, energy, biomedical/pharma, and tech/semiconductors (especially tied into AI).  In the past, research consortia like International Sematech were very effective at guiding research of tremendous economic benefit.  Some state governments have had interesting and impactful programs (e.g., the Cancer Prevention Research Institute in Texas, supported by state bonds).  The US could encourage the development of industrial consortia aligned to specific aspects of research; these could be supported financially by joint investment of the constituting companies and special bond issues expressly for these topics, with the proviso that funds be used to support use-inspired basic and early-stage applied research.  It is critical that any such plan have actual oversight, to make sure that this does not become a giant give-away to the companies that have the best lobbyists.

In terms of supporting models for research that complement traditional university structures and enable projects that require vast resources, interdisciplinary coordination, or extended timelines:  There are already large-scale center/institute programs that have been deployed in the past by various federal agencies.  While not perfect, these are not a bad model - in the last year, these have suffered greatly due to budgetary uncertainties and outright cancellations.  You can't complain about failing to do research requiring extended timelines if funding is annualized and perpetually at risk.  One approach to dealing with this would be to endow such structures at the outset for multiyear support. Congress would have to be willing to do this, and given the level of impasse that seems very difficult, but with executive branch support it could be possible.  The short version: Don't complain about lack of long-term stickiness if the US government cannot be considered a reliable partner year over year.

In terms of identifying and developing scientific talent across the country:  the RFI mentions "leveraging digital tools", which sounds a lot like trying to use AI to identify people.  This is frought with risk.  Programs such as the NSF graduate research fellowship, the NDSEG, the SCGSR, and others are essential at developing talent, as are many research experience for undergraduate programs and internship programs at national laboratories.  The recent cuts to these programs are short-sighted and counterproductive.  If we want the US to develop top technical talent and have people choose technical paths over "easier" trajectories, then we need to make it financially worthwhile for students to pursue these aims.  Federally supported undergraduate scholarship, graduate fellowships, policies that reward companies for providing internships, policies that, e.g. reduce endowment taxes if university resources are spent providing such support, etc. are all worth considering.  Again, an advisory task force could likely provide quantitative information about what strategies are likely to work best.

I could go on, but these are a starting point for a discussion.  I would be happy to talk further with interested parties and engage in discussions going forward, if that would be helpful.

An open position at Rice, Assistant or Associate Professor of Advanced Computational Materials

The Rice Advanced Materials Institute has an open search in the area of computational materials.   The full position posting is here.  Please spread the word!

The brief description:  

The Rice Advanced Materials Institute (RAMI) at Rice University, located in Houston TX, seeks applications for tenure-track Assistant and/or Associate Professor position in the area of advanced computational and/or applied artificial intelligence (AI)/machine learning (ML)-informed materials research with an anticipated start date of July 1, 2026 to January 1, 2027. RAMI, working in conjunction with the Schools of Engineering and Computing and Natural Sciences and the Departments therein, seeks applicants from diverse backgrounds for potential tenure-track faculty positions in Departments to be determined by the best natural fit and input of the candidate (e.g., potential departments could include, but not be limited to, Chemical and Biomolecular Engineering, Chemistry, Electrical and Computer Engineering, Materials Science and NanoEngineering, Mechanical Engineering, or Physics and Astronomy). More experienced candidates nearing the transition to or having already recently transitioned to the Associate Professor level and conducting transformative research projects will be considered. 

 

We seek outstanding candidates with research interests in theoretical, computational, modeling, and/or simulation of materials science (and related fields) and the application of advanced AI and ML approaches to accelerate, improve, or revolutionize the same. Research spanning all of aspects of materials including soft/hard matter, inorganic/organic materials, etc. is welcomed, as long as that research maps significantly to one or more of the RAMI core research areas, including: 

 

• Next-generation Electronics/Photonics – Developing materials to enable a new paradigm in microelectronics ranging from memory/logic to communications to sensors and beyond, with a focus on low-power/voltage function.

• Energy Materials (Systems) – Developing materials innovations to transform energy storage and conversion/harvesting.

• Materials for the Environment – Developing materials innovations to assure responsible use of natural resources and long-term stewardship of our air, soil, and water resources.

This effort is being initiated through RAMI, a campus-wide institute with the goal of expanding Rice’s already strong efforts in materials research across many departments in science and engineering. RAMI includes >70 current faculty members working on a wide array of research that overlap with RAMI’s core research focus areas. Rice considers efforts to bolster these areas as a focal point of materials research in the coming years.

 

The selected candidates will be expected to teach and develop undergraduate and graduate courses within their expertise and home department; perform high-quality research in their specialized area and present findings from their research in peer-reviewed publications and conferences; establish a strong research program supported by extramural funding; be involved in service to the university and broader scientific community; and collaborate with faculty in diverse disciplines. Successful candidates will have a strong commitment to teaching, advising, and mentoring undergraduate and graduate students from diverse backgrounds. 

Applications will be reviewed in a rolling fashion but should be received by 11:59 PM (eastern time) on January 4, 2026. Applications received after this deadline will be considered as they arrive and until the position is filled.

The NSF MRSEC program - end of an era?

Are we witnessing the irreversible deconstruction of an historically successful NSF program, not because of any clear strategy or planning but instead because of the cumulative impact of many forces?

The National Science Foundation has historically supported a mixture of individual investigator (or small team) funding opportunities and large center grants.  The center programs are meant to bring together collaborative teams of researchers to tackle sets of research questions that require a larger scale approach - a hugely emphasized review criterion of those centers is always the question, "Is the proposed work really a center, in the sense of being a coherent effort larger than the sum of its parts, or does it instead read like a collection of loosely connected individual projects?"  Center programs are also a way that NSF has supported Research Experience for Undergraduates programs, providing crucial gateways into real science and engineering research and training for hundreds of students per year.  Similarly, center programs have contributed toward building up networks of facilities with specialized research infrastructure and capabilities.  Apart from topical centers that come and go (e.g., the nano centers in the early 2000s, the quantum centers now), center programs include Engineering Research Centers, Science and Technology Centers, and the less applied Physics Frontier Centers (Division of Physics), Centers for Chemical Innovation (Division of Chemistry), and the flagship center program of the Division of Materials Research, the Materials Research Science and Engineering Centers (MRSECs).

The MRSEC program has such a long history (going back to the 1960s) that it has a reasonably good wikipedia page telling the story of its origins and evolution.  In brief:  There are typically around 20 MRSECs at any one time since the program was shifted into its present form 31 years ago.  Each center award is for six years, renewable.  There is a national competition every three years, during which half of the existing centers are up for renewal (though depending on reports by visiting committees, some centers may be recommended not to submit a renewal proposal).  Research in MRSECs is organized into "Interdisciplinary Research Groups" (IRGs), as well as some seed projects.  Once upon a time, a MRSEC could have as many as five IRGs.  Over the last 15 years or so, NSF budgets have largely been flat, and DMR has tried to balance the demands of individual investigator grants vs. center grants, along with knowing that underfunding a project can be worse than not funding it at all.  As a result, the number of funded IRGs has decreased steadily, so that a typical MRSEC now has two or three IRGs (usually two).  

I won't bore you with details about how the funding competitions work.  The short version:  universities send in preproposals consisting of draft IRGs + the other center components (facilities, education, management, etc.); IRGs are reviewed in panels, and then universities are invited or not to the full proposal stage.  Suffice it to say, every three years there is a national competition that consumes many many hours of effort.  Landing a MRSEC is both a sponsored research award and a point of pride.  Typically maybe 80% of the cohort up for renewal make it, with new starts making up the balance, so that there is some rotation among institutions.   Like all NSF programs, the peer review efforts are done for free by the community.

The MRSEC program has had a big impact over the years, at minimum in the number of people trained and supported through these efforts.  The National Research Council/NAS did a study of the MRSEC program back in 2007 that is publicly available here, if you're interested.  No program is perfect, and I don't want to argue about whether the balance of, e.g., reporting requirements vs. research dollars is right, etc.  However, I think the large majority of materials researchers in the US would say that the MRSEC program is a mainstay that has been a key pipeline of people into the field, both in academia and industry.

This year, I worry that we are watching likely irreversible harm to the program, and not by the voluntary choice of anyone at NSF.   The budgetary uncertainty is crushing right now.  It is unclear when and how the agency will be making awards, and how much funding they will have (since you can't actually plan based on the possibility of a continuing resolution in the absence of actual budget bills that can pass congress).  As a result, after the preproposal phase in the current competition, NSF revised their guidance, so that instead of the typical "8-10 awards" expected (see here), now they say to expect "2-5 awards".  That's because they're doing contingency planning assuming a cut in the program from $27M to $15M.  This means that there is a real possibility that the total number of existing MRSECs could be cut by 40% at a stroke, and the next cycle of the competition will be due to start in 2028, with little reason to think that budgets will be any better or smoother by then. 

There will almost certainly never be a return to "normal", for multiple reasons, including the general evolution of all funding programs with time.  The end result of the current shakeup may also have some positive outcomes in terms of new approaches.  That said, it sure feels like paths are being set by circumstances, not considered choice.  I would say that the path of this program is a question that should be addressed by the NSF Math and Physical Sciences advisory committee, but of course that was disbanded back in April, along with 11 others.  You might imagine asking the National Academies for thoughts on this, but I gather anecdotally that is not happening much at all anymore either.  

I'm writing because I hope someone more influential than me can report on this.  At a time when "materials" are clearly of major importance to US competitiveness (e.g., they are clearly relevant to multiple priority areas of the Genesis Mission), is anyone thinking about the impact of the trajectory we are on here?

(Back to science soon, hopefully.)

Taking stock: some federal science news

Some general science news:

•  The New York Times ran an interactive article this week that shows what we all know.  This past year was a very bizarre funding environment.  The article focuses on NIH and NSF, but the major points are generalizable.  The combination of circumstances (DOGE, general administrative turmoil, uncertainty and legal cases about indirect costs, the lack of a real budget followed by a late continuing resolution, plus the government shutdown and continued lack of real budgets) has been extremely disruptive, resulting unquestionably in less science and engineering research being funded by the US government than in many years.  
• Conversations I've had with program officers at two agencies have conveyed that everyone thinks it is very likely that there will be another shutdown in January, when the present spending authority expires.  To put that another way, there is very little confidence that actual spending bills appropriating real budgets for NSF, DOE, NIH, etc. will pass the House and Senate, with some reconciled conference version getting filibuster-proof support in the latter, before then.  This uncertainty means that right now it's going to be nearly impossible for the NSF, for example, to make much in the way of awards in the meantime, since they have no budget and can't plan on a year-long continuing resolution.  
• There has been an executive order announcing the Genesis Mission, which is going to be a large federal AI+science project.  The goal is to "accelerate the AI and quantum computing revolution and to double the productivity and impact of American science and engineering within a decade", according to undersecretary of energy Dario Gil.  Broadly, the plan is to have AI/ML agents developed (presumably by private contractors or private/public partnerships) and trained on vast datasets (ones already in existence in, e.g., national labs and public repositories).  At the same time, a list of Grand Challenges will be defined (within the next 60 days), with the idea that these AI agents will be used to address these (and demonstrating application of the AI "Platform" toward at least one challenge within 270 days).  Any stated support for science and engineering research is welcome.  I hope that this ends up bearing fruit in terms of real research advances, and that university researchers can contribute effectively. (I worry about a framework for massive taxpayer-funded financial support of for-profit AI companies, privatizing financial/IP benefits from publically funded datasets.  Of course, I worry about a lot of things.  Ask anyone who knows me.).   Ideas about grand challenges would be fun to discuss in the comments.   
• We had a great physics colloquium this week from Steve Fetter at the University of Maryland about the continuing threat of nuclear weapons.  Very sobering.  One fact that I gleaned:  In terms of missile defense, the Next Generation Interceptor is likely to cost $660M per interceptor.   That is something like 50 times the cost of a Russian ICBM.  Something else to bear in mind:  The Houston Food Bank, one of the largest and most effective in the US, has an annual budget of about $64M.  The amount of resources consumed by nuclear arms since 1945 is just staggering.