Leaving aside AI for research at the moment, as anyone with a pulse knows at this point, we are in a period of rapid change in terms of AI tools and specifically their role in higher education.   There is an arms race among universities about what tools should be made available to students.  There are equity of access issues here - wealthy students are likely to pay out of pocket for more expensive subscriptions, regardless of what universities provide.  Princeton announced that it's providing everyone with chatGPT subscriptions (the equivalent of the $20/mo GPT 6 Astra).  Cornell is doing something similar, including Claude.   This is likely just the beginning, as the major AI firms want to hook customers for life provide a great opportunity to many users, and prestigious universities don't want to risk being perceived as inadequate relative to their peers. In the meantime, it's honestly not clear if undergraduate students even want this.  This WSJ article talks about how demand is through the roof at Harvard for a reading-intensive humanities course.  The reaction from some students at Princeton to being given GPT subscriptions is not exactly a ringing endorsement. There is no question that these tools, when used correctly, can be powerful tools to develop educational materials.  A nice example that has gone a bit viral in the last week is this site, by Nathan Liow, an undergraduate CS major at UT Austin.  He's been using Claude Opus 5.5 to create interactive visuals to go with the notes he took from his AP courses, leading to some very clean demonstrations for calculus, statistics, and physics.  Without these tools, it would be a huge investment of effort to create these kinds of visualizations.  We must figure out the ways that AI capabilities can help learning, rather than removing all of the "intellectual friction" needed to build up analytical skills, critical thinking, and knowledge. Separately:  Tomorrow is Election Day in the US.  If you are eligible, please vote!  The only way representative government can work remotely effectively is if the electorate stands up and votes.  (Corollary: Anyone who pins their political strategy on suppressing voter turnout does not believe in representative government.)

Back in the ancient mists of time, scientists and mathematicians would circulate preprints of their articles among friends and colleagues via the postal service, as a courtesy, to get feedback and to try to make sure people in the community were aware of their forthcoming work.  As the wikipedia entry says, with the advent of widespread LaTeX and the development of the web, Paul Ginsparg (then at LANL) put together an html-based site for electronic sharing of preprints, initially at xxx.lanl.gov (back before "xxx" in URLs was the kind of thing filtered and blocked by employers).   In 2001 he moved to Cornell, and by then the lab was perhaps relieved to see the site, rebranded as the arXiv, shift to Cornell's library/repository infrastructure.   PIs my age remember (fondly? maybe?) the old days of the arXiv, when Prof. Ginsparg's rather dry sense of humor pervaded the site.  The skull-and-crossbones logo.  The "help"/FAQ pages that basically said, "if you can't figure out how to .tar.gz all of your necessary LaTeX files, and you can't figure out how to make your .eps figures small, maybe you should reconsider whether you're smart enough to be sharing your ideas here".  The arXiv was for preprints, without peer review (though interesting follow-on sites like Scirate now exist for organized commenting on the articles). From these modest beginnings, the arXiv has grown enormously, including imitators/spin-offs such as chemrxiv and biorXiv and socarXiv.  The arXiv has recently become an independent nonprofit, hired a CEO, secured multiyear philanthropic support, and hosts over 3 million articles.  It's been interesting seeing some level of complaints online about some of these steps, but when the audience is so large, not everyone is going to be happy.  Rapid growth has been a major issue - see here for a graph of monthly submissions: The exponential rise (except for a slight pandemic-correlated shoulder) has been problematic, especially recently.  One hallmark of the arXiv over the years has been its ability to function with comparatively minimal need for "moderation".  Early on, one consequence of the minimalistic help and moderate technical entry barrier was that it was unusual for fringe/pseudoscience to make its way onto the server.  (Hence the establishment of viXra.)  With the ease of cranking out properly formatted readable manuscripts using AI, clearly the arXiv has been struggling.  If 10-15% of submissions need some kind of human intervention or review, the support needs are rapidly outpacing the limited count of support staff.  There can be substantial backlogs.  To help deal with this, the arXiv recently updated its policies regarding AI-generated content (and AI cannot be a co-author, because the AI tools cannot take responsibility for content), and most recently has had to limit submission rates to two papers per month per submitting author.  These moves, too, have drawn some criticism (e.g. here).  Personally, I think the operators of the arXiv face an incredibly challenging environment and are doing the best they can - the idea that they are making moves because they are establishment sticks in the mud who don't understand the New Way of Doing Science is just wrong-headed. It's completely unclear where all this is heading.  Exponential growth in nature signals instability and does not continue forever. If proponents of very heavily AI-driven research want to establish a repository specifically for that work, that's up to them.  [It is very on brand for the hard core AI advocates to argue that the arXiv is somehow morally obligated to host everything (regardless of hardware or personnel costs) so that future AI tools can read everything (a repository growing too quickly for human researchers to keep up) and summarize it.] One overarching point that should come up in any arXiv discussion: The arXiv has become a global repository for an enormous amount of human knowledge, without charging anyone publication fees.  This should make interested parties think reallllllly hard about economic models of for-profit publishers.

We are less than one month away from the end of the federal fiscal year, and traditionally there are internal deadlines for agencies to allocate their final spending by around September 9. Right now, the NSF is on track to issue about 4000 fewer (!!) awards in FY26 than it did annually back in FY21-FY24, and 2000 fewer than it did in the incredibly tumultuous FY25 (with its government shutdowns and mass cutbacks in agency personnel). This is dire, if like me you are a supporter of the agency and its vital role in the US research ecosystem.   Perhaps even more distressing, the NSF is on track to underspend its FY26 budget appropriation (congressionally approved, presidentially signed) by between $1.25-1.5B, or 15-18%. This is essentially unprecedented - in the past, the NSF has always spent ~ 99% of its appropriation in a given fiscal year. Some large portion of this is from the mid-FY clawbacks that were reported in Science and Nature, supposedly squirreled away to support an as-yet unannounced OSTP "grand challenges" program.   While technically the funds don't go away at the end of September, this kind of underspending raises the possibility of a pocket rescission. OMB and the executive branch have been pushing for massive cuts to the agency; Congress has disagreed. It sure looks like all the "see, don't worry, Congress didn't allow big cuts to the NSF" palliative statements don't hold up very well to scrutiny, if the majority party is content to just give up Article I power to the executive branch.  In this period of complete flood-the-zone craziness, the mainstream news media seemingly doesn't have the bandwidth or interest to report on this; they seem to have judged that it's too obscure, it doesn't play in Peoria, the public doesn't really care. This kind of disruption will have ripple effects that last for many years and affect US scientific and economic competitiveness, and it's happening without much notice. This week's news about an agreement between NIH and DOD to funnel NIH funds for infectious disease to DOD (or, in the official statement, to work together on projects of mutual interest), is at least getting some public attention.  Agencies agreeing to pass around at minimum hundreds of millions of dollars outside congressional oversight or what the appropriations acts say is another example of an Article I crisis, when the majority party basically hands over what are supposed to be congressional powers to executive branch. (An additional sciencey blog post coming soon!)

Yesterday I attended the 40th annual summer research colloquium of the Smalley-Curl Institute at Rice, a fun internal conference that provides a great opportunity for undergrads (including visitors), graduate students, and a few postdocs to present their work.  The keynote speaker was our EVPR, Prof. David Sholl, who gave a very informative talk about reproducibility in the chemical engineering/materials literature.  We hear a lot these days about crises of reproducibility in scientific research, and Prof. Sholl rightly points out that in some fields the expectation of reproducible results is high - no one would spend $1B on a chemical engineering plant if they weren't very sure that the catalytic processes were going to work as expected at scale.  Keys to reproducibility include, unsurprisingly, repeated results and independent replication.  One metaresult that was interesting is this paper, looking at the literature on metal-organic frameworks and how often there are published replications of syntheses; not as often as you would think or want!   A truly surprising (to me, anyway) result is this one.  The Brunauer–Emmett–Teller (BET) (yes, that Teller) method is a long-established technique that uses gas adsorption measurements to infer the surface area of porous materials.  Many research groups were given identical raw adsorption isotherms and asked to calculate the specific surface areas, resulting in a surprisingly large spread of results (Fig 1 of the paper).  Clearly not everyone had the same analysis procedures even for a technique developed in the 1930s! Some take-away lessons from this are encapsulated here, in an article titled "Five easy ways to make your research more reproducible".  Good stuff.  The talk raised a number of questions relevant to our present era of huge enthusiasm about AI-based materials research and "self-driving" labs.  If the AI models are all trained on the literature, and the literature is not representative of complete and reproducible procedures, that's a problem.   One personal anecdote about reproducibility and its challenges in materials synthesis.  Twenty years ago (!), I was working with a colleague who had a postdoc who was synthesizing Fe3O4 (magnetite) nanoparticles via wet chemistry methods (see here). We did some fun electronic transport experiments bridging very closely spaced electrodes with such nanoparticles, and we saw some very dramatic hysteretic response kick in as \(T\) was reduced below about 120 K.  That's the temperature of the Verwey transition in magnetite, where the material enters a more insulating low temperature phase.  Basically all of the devices we made with that batch of nanoparticles showed this phenomenon.  Then the postdoc took up a faculty position and a senior grad student came in and took over the synthesis, and for several months, subsequent batches of nanoparticles just didn't seem to show the effect.  The key issue is oxygen stoichiometry.  Get a little oxygen rich, and you form nanoparticles that include some \(\gamma\)-Fe2O3, which doesn't have the Verwey physics and in nanoparticle form looks really similar in x-ray diffraction to the desired magnetite.  Anyway, we started working with a collaborator who could grow epitaxial Fe3O4 films, and in those devices the electronic effect was there all the time.  All this led to this publication and subsequent papers, and I still think it's a cool set of result about a nonequilibrium transition in a correlated material.  In the end, after several months the chemistry grad student did get back to making nanoparticle batches that showed the transition. It turns out that at some point he had changed the length of a piece of tubing in the gas manifold, and unexpectedly that had altered the reaction kinetics just a little.  Changing it back got the synthesis to be reliable again.  This is an example of how finicky materials synthesis can be!

For non-US folks, feel free to skip.  For US folks:  The Office of Management and Budget, which for much of its history has been a comparatively uncontroversial element of the executive branch, has set rules and guidelines for how many executive-branch agencies conduct business and interact with, e.g., universities.  For the purposes of how the research ecosystem operates, the most relevant is OMB's "Uniform Guidance" about how grants and contracts work.  Periodically these rules are updated for various reasons, including the goals and policies of the presidential administration.  The standard way this works is that the proposed changes are published in the Federal Register; there is a public comment period; OMB makes revisions and then publishes the new rules.  In principle, Congress can act to override or prevent rule changes, but without the agreement of the President, this is an extremely challenging path. OMB has proposed sweeping changes to the Uniform Guidance, summarized here.  These proposed rule changes are huge deviations from previous practice.  For example, they would have all final grant decisions made by political appointees or hires of the executive branch (rather than, e.g., agency subject matter experts); grants could be cancelled at any time for essentially any reason (completely undefined insufficient support of the president's priorities), with no appeal process; international collaborations would be severely curtailed. That's just three for starters.  Note that this would also go beyond just the public research enterprise - it would allow the executive branch to cancel funding for things like bridges, roads, schools, agriculture, etc. for undefined political reasons.  It would be a huge transfer of power from Congress to the presidency.  Here is another summary by the AAU.  Here is an editorial essay from ars technica.   The public comment period on this runs until July 13.  Here is a link where you can make a comment.  Here is a guide for how to be effective at this from Stand Up for Science.  The APS has a tool for helping people to comment about specific aspects of the rule changes.  It is also a good idea to contact congressional delegations (representatives, senators).   It's important to have a clear public record about the proposed changes.  They may try to implement these regardless, but if so, there will be a continued fight over this in Congress and through the courts.