anil oza

@aniloza

science journalist @stat covering health equity, the nih, and interaction btwn science & society | also: npr, nature, sciam, science, science news (he/him) (signal: aniloza.16)

1.650 Folgt4.988 Follower580 Posts

Bluesky

Ist das dein Profil?Übernimm es — dann gehört dir aniloza.dirty.social.Profil übernehmen

anil oza@aniloza· 11d

👋 #jsm2026

anil oza@aniloza· 14d

where have I heard this before …

Screenshot reading “Cassidy said that his approval of Kaufman comes with conditions and that any departure from these agreements by the Health and Human Services Department could hold up further health nominees. “

anil oza@aniloza· 16d

👀👀👀👀

The average biomedical Ph.D. takes five to six years to complete, meaning a majority of incoming foreign graduate students would need to apply for DHS extensions to complete their work. And while the length of postdoctoral appointments is more variable than graduate training, National Science Foundation survey data indicate that about a fifth of postdocs are likely to have worked in that role for more than six years.

anil oza@aniloza· 29d

you can also find a link to the github with details of how we did our analysis, and take a look at the data yourself:

screenshot from story reading "A note on the use of AI
This analysis was conducted in partnership with STAT by Jae Yeon Kim, assistant professor of public policy; Betsy Pleasants, a postdoctoral fellow at the Carolina Population Center; and Maria Gallo, professor and chair of the Department of Epidemiology. All are at the University of North Carolina at Chapel Hill.

The researchers used a large language model (LLM) provided by OpenAI to classify comments by whether they support or oppose the proposed rule as well as to count the number of times comments mentioned specified topics.

Classification tasks like this would not be possible to conduct in a timely manner by hand and are well-suited to the capabilities of LLMs. This sort of use of LLMs is different from more “generative” uses (like chatting with an AI bot) that are more prone to “hallucinations.”

AI tools were also used to assist in the process of writing code for this project, but the entire data analysis pipeline uses repeatable code, mostly R and some Python. None of the summary statistics or charts used in this article were generated by AI.

All work was reviewed and verified by humans.

STAT also independently reviewed all code and methodology for this project.

Per STAT’s AI usage policy, all code and underlying data are available for review on Github.

— J. Emory Parker, Data Editor at STAT"

anil oza@aniloza· 29d

The (initial) results are in: most of the publicly available comments on OMB's proposal to overhaul federal grantmaking are negative 95% of the 50k available comments were opposed, just 1% were in support Political control, peer review, and terminations were most common topics mentioned

Chart titled "commenters overwhelmingly oppose the proposed rule, percent of aalyzed comments by sentiment and type of commenter."

chart showing what commenters on OMB proposal mentioned in their comment

anil oza@aniloza· Jul 14

glad to know I am not the only procrastinator around ... 50k comments were submitted about the OMB proposal *yesterday* Since we wrote about this a week ago, 260k have been submitted. (Only 50k are available to search on the docket as of now)

anil oza@aniloza· Jul 1

Getting my slides ready for tomorrow's webinar... Day 21: 44134 total comments, 21950 mention science, 24501 mention research

anil oza@aniloza· Jun 25

Looks like groups outside the scientific community are commenting in higher numbers Day 18: 35,323 total comments, 17,458 mention science, 19,135 mention research

Graph showing increase in the number of comments on OMB proposal over time

anil oza@aniloza· Jun 24

NIH posted its updated data on Research, Condition, and Disease Categories for FY 25 and all three categories for health disparities work took hits

screenshot of NIH's website showing an increase then drop in health disparities funding

anil oza@aniloza· Jun 23

wild story from Lizzy — FDA gave compassionate use access to a 79 year old through senior NIH clinician, @muniyappar.bsky.social, could it be President Trump? Read more from Lizzy on what she found out, and what ethicists had to say about it

$screenshot of story reading "This person was a 79-year-old man at the time the request was made in April, according to three sources familiar with the matter. Those sources, who requested anonymity due to fear of reprisals, said it drew the interest of top health officials, suggesting the person receiving this drug was well connected. Senior NIH clinician Ranganath MuniyappaNIH A senior clinician at the National Institutes of Health named Ranganath Muniyappa requested the drug to treat the patient for refractory obesity with obstructive sleep apnea and pulmonary hypertension, a severe version of the disease. Pulmonary hypertension is high blood pressure in the lungs, which can be life-threatening. STAT does not know who the patient is. But given the patient demographics and the unusual nature of the application, STAT asked the White House whether the patient was President Trump, who turned 80 a week ago, is overweight, and has expressed interest in obesity drugs."$

anil oza@aniloza· Jun 22

Here is what OMB said when we reached out about concerns being raised about the clause on "disparate impacts" affecting health disparities research

Screenshot of the STAT story, reading "In response to concerns that the updated Uniform Guidance, as it is called, could be used as grounds to terminate or not fund health disparities work, an OMB spokesperson said, “In the past, unaccountable deep state bureaucrats have funded woke and wasteful projects like, ‘engaging Black sexual and gender minority youth social media influencers,’ and ‘Addressing Social Determinants of Health Among Racially Diverse Trans Women in the Deep South,’ under the guise of studying health disparities. That ends now.” The official added, “With this new rule, the Trump Administration will bring much-needed accountability to the grantmaking process and ensure taxpayer dollars are spent wisely.”"

anil oza@aniloza· Jun 18

Been a while since I checked Day 14 (I'm counting business days): 25328 total comments, 12324 mention science, 13832 mention research

line graft showing increasing number of general comments, research related comments, and science comments

anil oza@aniloza· Jun 12

As someone that has invoked Vannevar Bush more than my colleagues would like .... I appreciated this piece making the case that there are other figures we could reach for as touchstones amid all of the upheaval

screenshot of essay reading "In this chaotic time, Vannevar Bush’s Science, the Endless Frontier has emerged as a symbolically meaningful text among scientists, who frequently point to it as the basis for the government’s long-running support of university research. Certainly, Bush marked a call for a new industrial policy in the United States that would make the nation a global leader in the new world order. He would later be credited as the architect of the so-called social contract of science, whereby federal funding is allocated primarily to university researchers in pursuit of free inquiry that might later yield some economic or social benefit. But this mythologized rendering of the innovation system overlooks other key ideas, like those expressed by sociologist Robert K. Merton, that described how the United States should govern science democratically based on lessons learned from the Second World War. Rereading Merton now, even more than revisiting Bush, better exposes the vulnerabilities driving us, as social scientists, to defend a vision of science as—and for—democracy in our own era."

anil oza@aniloza· Jun 11

Here I thought I could drop off looking at the # of comments on the OMB proposal on a weekly basis ... Day 9: 15583 total comments, 8468 mention "science," new category — 9011 mention "research" at the suggestion of @dangaristo.bsky.social

anil oza@aniloza· Jun 9

We talked with Jenna and @ngfriedman.bsky.social about the report, hosted by @27unihted.bsky.social, and what the past year has meant for them

Screenshot of story reading “Working as a lab scientist, studying neuroplasticity in the visual system, had felt like the perfect fit for her, but she began to wonder, “Who will be left in the field to read my work and build on it? Will I be able to get a job as a postdoc? Will my PI fully understand the situation with funding right now?”

Because of this uncertainty, once she defends her thesis, she will start working with 27UNIHTED, the group of former NIH staffers and supporters who put the report together and have been advocating for the NIH amid changes from the Trump administration.

“I’m so excited for the day when we can just do science,” Friedman said. But for now, “I’m leaving science, and I wouldn’t be leaving science if not for what happened this year.”

anil oza@aniloza· Jun 5

Here’s the rationale the conference provided to STAT

In response, ADA said, “These attendees were escorted out by our onsite event security because they demonstrated behavior not consistent with this code of conduct. They were respectfully given the opportunity to cease this behavior and chose not to, which is why they were escorted out.”

anil oza@aniloza· Jun 5

Day 5: Thus ends my week of very scientifically checking on the OMB comments (4347/7284 mention science)

anil oza@aniloza· Jun 4

Day 4:📈📈, almost forgot to check today with the other news out of the White House last night (3439/5837 mention science)

anil oza@aniloza· Jun 3

Day 3: 📈📈📈 📈📈📈 📈📈📈 (2637/4428 mention "science")

anil oza@aniloza· Jun 3

Part of this story is also what it says about commitments to rethink the parachute model of international research, and how it is more difficult for US researchers to collaborate across borders

screenshot of the story, reading "To those involved with the program, it was an attempt to break from the longtime paradigm of researchers from wealthy countries parachuting into low- and middle-income countries to dictate research priorities without consulting the local communities. But that goal feels more elusive than ever, with the U.S. slashing foreign aid. "

anil oza@aniloza· Jun 2

Day 2: 📈📈📈 (825/1300 mention "science")

screenshot of document comments, reading 1.3k

anil oza@aniloza· Jun 1

just over the weekend, there are already several hundred comments on this 👇 211 of the 353 publicly visible comments mention the word "science"

screenshot of the regulations.gov site, showing "Document Comments" with 353 next to it

anil oza@aniloza· May 14

wow — the preprint host, arxiv, is banning authors for a year if they submit papers with hallucinated citations 🤖

Twitter thread from Thomas Dietterich, reading "Attention
@arxiv
authors: Our Code of Conduct states that by signing your name as an author of a paper, each author takes full responsibility for all its contents, irrespective of how the contents were generated. 1/
3:03 PM · May 14, 2026
·
66.7K
Views
Relevant
View quotes

Thomas G. Dietterich
@tdietterich
·
1h
If generative AI tools generate inappropriate language, plagiarized content, biased content, errors, mistakes, incorrect references, or misleading content, and that output is included in scientific works, it is the responsibility of the author(s). 2/
Thomas G. Dietterich
@tdietterich
·
1h
We have recently clarified our penalties for this. If a submission contains incontrovertible evidence that the authors did not check the results of LLM generation, this means we can't trust anything in the paper. 3/
Thomas G. Dietterich
@tdietterich
·
1h
The penalty is a 1-year ban from arXiv followed by the requirement that subsequent arXiv submissions must first be accepted at a reputable peer-reviewed venue. 4/
Thomas G. Dietterich
@tdietterich
·
1h
Examples of incontrovertible evidence: hallucinated references, meta-comments from the LLM ("here is a 200 word summary; would you like me to make any changes?"; "the data in this table is illustrative, fill it in with the real numbers from your experiments") end/"

anil oza@aniloza· May 13

@danielpayne.bsky.social & @chelseacirruzzo.bsky.social on the precarious situation Bill Cassidy finds himself in — trying to simultaneously please MAGA & MAHA, and potentially losing both in the process

screenshot of STAT story, reading "Cassidy’s fate — and political legacy — may ultimately be traced back to two votes, according to pollsters, allies, and opponents of the senator.

On Feb. 13, 2021, when some Republicans believed the events of Jan. 6 had pushed Donald Trump out of mainstream politics for good, Cassidy voted to convict Trump for inciting the insurrection. “I voted to convict President Trump because he is guilty,” he said in a statement after the vote.

Four years later to the day, with Trump back in the White House, Cassidy cast another vote, this one in line with the president’s wishes. After publicly wrestling with Robert F. Kennedy Jr.’s long history of anti-vaccine advocacy, Cassidy advanced him to become the nation’s health secretary.

“The mistake that Cassidy made, which is a political mistake, is he either should have been all in from the very beginning, or he should have stood up and said, ‘no way,’” said his opponent Fleming. “But instead, he flip-flopped, and that’s the worst thing you can do, is be indecisive or change positions.”

anil oza@aniloza· May 8

anil oza@aniloza· Apr 30

Grants that were scored very competitively by peer reviewers now amount to a coin flip. This reality is harsher for early career folks and those on soft money positions, incl @kaytesb.bsky.social Chart by @jaspar.bsky.social 👇

Bar chart showing the odds of a grant ranked in the top 10% will be funded, from 2015 to 2025. Odds fall from consistently being ~90% to 56% in 2026.

anil oza@aniloza· Apr 1

of course!!

Screenshot of a portrait of vannevar bush made into a slack emoji

anil oza@aniloza· Apr 1

guilty as charged 🫡

anil oza@aniloza· Mar 27

there's a lot of rhetoric from this administration that focuses on individual responsibility over structures, but as several folks pointed out to me, there is overlap between structural competency & the MAHA movement @ammahstarr.bsky.social @jonathanmetzl.bsky.social @ariana-tl.bsky.social

screenshot from the story reading "Metzl, who coined the term, agrees. “Structural competency is about structures, it’s not about Republicans or Democrats. I would say that certainly there’s individual choice rhetoric, but there’s nothing in what we’ve been doing that says that individual choice is not important,” he said. “My hope with structural competency was that it was something that could actually help bridge political divides around health. Because it’s community-focused and having to do with health finances and costs for medications: things that were important to our last government, and to this government.”"

anil oza@aniloza· Mar 27

Researchers we surveyed moved their international work to the US, wound down projects, and some had projects rejected Read more at the report above from @mmolteni.bsky.social @jonathanwosen.bsky.social & I