My latest for @science.org on the fraught future of NCAR as NSF's deadline nears. It can still go so many ways.
10.03.2026 21:38
👍 36
🔁 32
💬 3
📌 3
My latest for @science.org on the fraught future of NCAR as NSF's deadline nears. It can still go so many ways.
I'm late to this, but this is a very thoughtful essay on the subject from Daniel Litt
So sad to hear about the passing of Tony Leggett, a true great of ultracold physics, an Illinois legend, and a higher branch of my academic family tree. I really thought we'd be seeing him shuffle through the ICMT, thinking big thoughts about quantum, forever. physicsworld.com/a/condensed-...
Sort of a 'Don't Ask, Don't Tell' policy for ̶c̶l̶i̶m̶a̶t̶e̶ earth science at NASA.
For more on the new committee, see Adrian's reporting:
DOE's new Office of Science Advisory Committee (SCAC*), which replaces 6 terminated specialized committees, has its first meeting March 27. No agenda posted yet, but you can register.
science.osti.gov/About/Federa...
*Persis Drell has asked that SCAC be pronounced 'ess-cee-ayy-cee' and not 'skak'
Accepting this logic raises a lot of questions for science policy. Would, for example, a climate hawk government cancel grants related to research aimed toward further fossil fuel extraction? Would it terminate contracts with fossil companies? What about collaborations w/ foreign adversaries?
There are a lot of caveats to the comparison (grants ≢ contracts, the bulk of scientific research does not have what reasonable people would call ideological ends, many terminations were from guilt by association, etc.) but the fundamental logic is the same: gov. should not pay ideological enemies.
In many ways the Anthropic story is a more explicit version of what has played out across the research landscape for the past year.
The thinking goes something like this: Contractors (researchers and universities) have their own ideological agenda and use gov. contracts to pursue those goals.
Comments (15:58:15) First STABLE BEAMS of 2026!
The final #LHC run
Towards a more luminous future
Here is a long story in the Columbia Spectator on Richard Axel's involvement with Epstein
www.columbiaspectator.com/news/2026/03...
1/3
Well damn, this is cool.
Birds That Don't Exist: Niche Pre-Emption as a Constraint on Morphological Evolution in the Passeroidea | Chia et al., 2026 | Ecology Letters
onlinelibrary.wiley.com/doi/full/10....
That said, it'll need to do a lot more to outweigh the slop it's creating for academics to wade through...
AI can revolutionize science; of that there is no doubt. It has already helped us see proteins with previously unimaginable clarity. Soon AIs may dream up new molecules for batteries or find new particles hiding in data from colliders—in short, they may do many things, some of which previously seemed impossible. But they have a crucial limitation tied to something wonderful about science: its empirical dependence on the real world, which cannot be overcome by computation alone. An AI, in some respects, can be only as good as the data it’s given. It cannot, for example, use pure logic to discover the nature of dark matter, the mysterious substance that makes up at least 80 percent of matter in the universe. Instead it will have to rely on observations from an ineluctably physical detector with components perennially in need of elbow grease. To discover the real world, we will always have to contend with such corporeal hiccups. Science also needs experimenters—human experts who are driven to study the universe and will ask questions an AI can’t. As Hopfield explained in a 2018 essay, physics—science itself, really—is not a subject so much as “a point of view,” its core ethos being “that the world is understandable” in quantitative, predictive terms solely by virtue of careful experiment and observation. That real world, in its endless majesty and mystery, still exists for future scientists to study, whether aided by AI or not.
Think this still holds up and will continue to hold up. AI (meaning LLMs like Claude) is going to be a very useful tool for scientists. www.scientificamerican.com/article/dont...
This plus the (semi)-autonomous formalization of the sphere-packing result and a few other examples seem to indicate LLMs are now genuinely capable of helping out at the frontiers of math.
Anyhow, best not to post about late night court filings from the west coast when you're on the east coast and not a courts reporter. Mea culpa, and thanks to @maxkozlov.bsky.social for catching the error.
Having reviewed the parties’ positions in their joint statement (Dkt. No. 180), the following case schedule is set: 1. Plaintiffs shall file the Third Amended Complaint on the docket by February 6, 2026. 2. Administrative Record Production and Additional Discovery shall be completed by May 8, 2026. 3. Plaintiffs’ Motion for Summary Judgment and Class Certification due by June 5, 2026. 4. Defendants’ combined brief in support of their Cross-Motion for Summary Judgment and in opposition to Plaintiffs’ Motion for Summary Judgment and Class Certification due by July 2, 2026. 5. Plaintiffs’ combined reply brief in support of their Motion for Summary Judgment and Class Certification and opposition to Defendants’ Cross-Motion for Summary Judgment due by July 31, 2026. 6. Defendants’ reply in support of their Cross-Motion for Summary Judgment due by August 14, 2026. Case 3:25-cv-04737-RFL Document 183 Filed 01/30/26 Page 1 of 2 2 7. A hearing on the Cross-Motions for Summary Judgment and Motion for Class Certification is set for September 1, 2026 at 10:00 a.m.
Here is the case schedule: www.courtlistener.com/docket/70459...
The States of California, Colorado, Washington, Connecticut, Illinois, Maryland, New Jersey, New York, Oregon, Rhode Island, Vermont, and Wisconsin, the Commonwealth of Massachusetts, and the California Governor’s Office of Business and Economic Development are plaintiffs in the action titled California, et al. v. Wright, et al., No. 3: 26-cv-01417-RS (“Multistate Action”), filed on February 18, 2026, in the Northern District of California. Plaintiffs hereby submit this administrative motion to notify the Court of that action and request that it be related to the above-captioned matter titled Thakur, et al. v. Trump, et al., No. 3:25-cv4737 (the “Thakur Action” and, together, the “Actions”). The plaintiffs in the Multistate Action and the plaintiffs in the Thakur Action have consented to the relation of the Actions; the federal entities that are the defendants in those actions declined to consent. The Actions challenge some of the same funding decisions by the U.S. Department of Energy (“DOE”) and the Secretary of Energy on many of the same grounds. They should be related because they meet both requirements of Local Rule 3-12: cases are related when they “concern substantially the same parties, property, transaction, or event,” and it “appears likely that there will be an unduly burdensome duplication of labor and expense or conflicting results if the cases are conducted before different Judges.” L.R. 3-12(a).
Earlier today, a group of attorneys general in California v. Wright (a case about DOE terminations) moved to relate their case with Thakur, given the similarities. www.courtlistener.com/docket/70459...
Meanwhile, the case has continued at the district level with Judge Lin, adding other agencies like DoD, DoT, NIH, and DOE. www.courtlistener.com/docket/70459...
The government’s motion for partial stay pending appeal (Dkt. No. 7) is GRANTED in part as to the Form Termination Class and DENIED in part as to the DEI Termination Class. We express no opinion on Plaintiffs’ remaining constitutional and statutory claims, nor on whether it would be appropriate for the 19 25-4249 district court to certify other provisional classes based on those claims.
So, to recap: the 9th circuit has issued a stay of Lin's order reinstating 'Form' terminations, but has not blocked Lin's order for 'DEI' terminations.
On Feb. 23rd, plaintiffs (that is, Thakur and the other UC profs) asked to be reheard by the entire 9th circuit. Paez, Christen, and Desai (Clinton, Obama, and Biden appointees respectively) rejected that request. www.courtlistener.com/docket/70763...
Because the current record suggests that the government aimed at the suppression of speech that views DEI, DEIA, and environmental justice favorably, the government has not shown that it is likely to succeed on the merits of its claim that the district court abused its discretion when it concluded the agencies likely terminated the grants based on viewpoint. 6
However, with regard to the DEI Termination Class ("those whose research grants were terminated because of the DEI Executive Orders"), the 9th circuit did _not_ find the government was likely to succeed and did not agree to a stay.
![We are bound by NIH, which held that the APA’s limited waiver of
sovereign immunity did “not provide the District Court with jurisdiction to
adjudicate” similar APA claims challenging grant terminations. 145 S. Ct. at 2658.
Here, the Form Termination Class challenges the government’s termination of
research grants as arbitrary and capricious under the APA. NIH held that similar
claims were “based on . . . research-related grants.” Id. (citation modified).
Further, the Form Termination Class sought—and the district court awarded—
vacatur of the termination notices and reinstatement of the terminated grants. NIH
held that such relief is “designed to enforce an[] obligation to pay money pursuant
to [the] grants” at issue. Id. (citation modified). Accordingly, we conclude the
government has made a strong showing that it is likely to establish that the district
court lacks jurisdiction to review the Form Termination Class’s APA claim.5](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:jsmr7lyw5rvrpfogr2kc6c4y/bafkreihptoxaarfcb4iq6c7duzzabgblaq6osi3wwtyjcbguvoyd432kym)
We are bound by NIH, which held that the APA’s limited waiver of sovereign immunity did “not provide the District Court with jurisdiction to adjudicate” similar APA claims challenging grant terminations. 145 S. Ct. at 2658. Here, the Form Termination Class challenges the government’s termination of research grants as arbitrary and capricious under the APA. NIH held that similar claims were “based on . . . research-related grants.” Id. (citation modified). Further, the Form Termination Class sought—and the district court awarded— vacatur of the termination notices and reinstatement of the terminated grants. NIH held that such relief is “designed to enforce an[] obligation to pay money pursuant to [the] grants” at issue. Id. (citation modified). Accordingly, we conclude the government has made a strong showing that it is likely to establish that the district court lacks jurisdiction to review the Form Termination Class’s APA claim.5
On 12/23/25, the 9th circuit judges ruled that the government was likely to succeed with the 'Form Termination Class' ("those whose research grants were terminated by form letter without any grant-specific explanation") cdn.ca9.uscourts.gov/datastore/op...
Also on August 21, 2025 - SCOTUS rules in NIH v. APHA, which blocks lower courts from using the APA for relief (i.e. reinstating terminated grants)
Wrote about this with @maxkozlov.bsky.social at the time.
The government appealed, in light of the SCOTUS decision.
The panel held that the government had not shown a likelihood of success on the merits of its arguments that (1) the district court lacked jurisdiction; (2) at least some plaintiffs lack standing; and (3) plaintiffs are not likely to succeed on the merits of their claims.
August 21, 2025 - 9th circuit denies the government's request for a stay of Lin's order, ruling that the government was not likely to succeed. cdn.ca9.uscourts.gov/datastore/op...
June 4, 2025 - Neeta Thakur and five other researchers at UCs sue the Trump admin over grant terminations
June 23, 2025 - Judge Rita Lin issues an injunction, ordering NSF, EPA, and NEH to reinstate grants www.courtlistener.com/docket/70459...
Deleted post where I mixed up who was trying to do what.
Full disclosure: deleted a post from the other week where I flubbed the details of an update in a court case (mixed up the plaintiffs and the appellees). As community service, here's my effort to reconstruct where Thakur v. Trump actually stands.
PROFESSOR BUTTS FALLS ON HIS HEAD AND DOPES OUT A SIMPLIFIED CAN-OPENER WHILE HE IS STILL GROGGY. GO OUTSIDE AND CALL UP YOUR HOME. WHEN PHONE BELL RINGS, MAID (A) MISTAKES IT FOR AN ALARM CLOCK- SHE AWAKENS AND STRETCHES, PULLING CORDB)WHICH RAISES END OF LADLE C).BALLD) DROPS INTO NET E) CAUSING GOLF CLUB F) TO SWING AGAINST BALL (G), MAKING A CLEAN DRIVE AND UP-SETTING MILK CAN(H) MILK SPILLS INTO GLASS (I) AND THE WEIGHT PULLS SWITCH ON RADIO (J) WALTZING MICE(K)HEARING MUSIC AND PROCEED. TO DANCE, CAUSING REVOLVING APPARATUS (L) TO SPIN AND TURN. SPIKES(M) SCRATCH TAIL OF PET DRAGON(N)WHO IN ANGER EMITS FIRE IGNITING ACETYLENE TORCH(O) AND BURNING OFF TOP OF TOMATO CAN(P)AS IT ROTATES. WHEN NOT OPENING CANS, THE DRAGON CAN ALWAYS BE KEPT BUSY CHASING AWAY INCOME TAX INVESTIGATORS AND PROHIBITION OFFICERS.
"At a recent public meeting, Deputy Director Doug Lowy noted the NOFO approval process now requires 16 steps. He joked that it’s now more complex than a 1929 Rube Goldberg design for a can opener."
That 1929 Rube Goldberg design, for reference:
"Since Trump took office 13 months ago, NIH has posted only 84 NOFOs, down from 787 in the previous year. Many more are in limbo. NIH has 323 opportunities listed as 'forecasted.' ... many on the list were announced in 2024 and ’25 and still are not open."
Jim O'Neill has officially been nominated to be NSF Director as of yesterday.