Richard Sever
@richardsever.bsky.social
Chief Science and Strategy Officer, openRxiv. Co-Founder, bioRxiv and medRxiv.
Pinned
Richard Sever
@richardsever.bsky.social
· Jan 12
Preprints of pandemic potential - new historical piece from me on the history of bioRxiv/medRxiv, their role in the pandemic, and the way forward. 1/n journals.asm.org/doi/10.1128/...
"Prism wants me to work over there...[but] most projects happen where researchers have created their workspace...it’s easier to bring the LLM to the workspace, than to get researchers to change their workspace"
🎯 1/n generalizable point from @ianmulvany.bsky.social world.hey.com/ian.mulvany/...
🎯 1/n generalizable point from @ianmulvany.bsky.social world.hey.com/ian.mulvany/...
OpenAI PRISM
I am in a state of confusion and wonder. Most of the time I’m able to suppress this and stumble through my days without thinking too much, but now and again I fail and my attention gets arrested by th...
world.hey.com
February 17, 2026 at 1:45 AM
"Prism wants me to work over there...[but] most projects happen where researchers have created their workspace...it’s easier to bring the LLM to the workspace, than to get researchers to change their workspace"
🎯 1/n generalizable point from @ianmulvany.bsky.social world.hey.com/ian.mulvany/...
🎯 1/n generalizable point from @ianmulvany.bsky.social world.hey.com/ian.mulvany/...
Reposted by Richard Sever
An old favorite for this #FluorescenceFriday. This "rose" is actually a kidney glomerulus (red, podocytes) and an axon reaching towards it (green).
February 13, 2026 at 8:51 PM
An old favorite for this #FluorescenceFriday. This "rose" is actually a kidney glomerulus (red, podocytes) and an axon reaching towards it (green).
Challenges for open ecosystems as "first-of-its-kind case study of misaligned AI behavior in the wild...raises serious concerns about agents executing blackmail threats" theshamblog.com/an-ai-agent-...
But if we have AIs that open pull requests left and right then no human maintainer can keep up. So the only option forward will be to close things up again. 3/
February 13, 2026 at 4:34 PM
Challenges for open ecosystems as "first-of-its-kind case study of misaligned AI behavior in the wild...raises serious concerns about agents executing blackmail threats" theshamblog.com/an-ai-agent-...
Data all the way down: How uncovering the data underlying research findings impacts every aspect of scholarly communication - webinar today at 12 pm ET
events.humanitix.com/love-data-we...
events.humanitix.com/love-data-we...
Data all the way down: How uncovering the data underlying research findings impacts every aspect of scholarly communication
DataDryad & SORTEE are teaming up for a panel discussion for Love Data Week 2026 - join us for a lively chat about the present and future of data
events.humanitix.com
February 12, 2026 at 4:40 PM
Data all the way down: How uncovering the data underlying research findings impacts every aspect of scholarly communication - webinar today at 12 pm ET
events.humanitix.com/love-data-we...
events.humanitix.com/love-data-we...
Reposted by Richard Sever
This post has generated a super interesting debate between the authors of the paper @ebrahimfeghhi.bsky.social @nrhadidi.bsky.social and one of the authors of a paper they criticised @mschrimpf.bsky.social including an attempted reproduction of their results. This is such a great use of social media
This paper shows alignment between LLMs and brain data is outperformed by a null model. More evidence for the argument I've been making in talks lately that we shouldn't believe any computational paper that puts less effort into null than main model.
www.biorxiv.org/content/10.1...
www.biorxiv.org/content/10.1...
www.biorxiv.org
February 11, 2026 at 4:12 PM
This post has generated a super interesting debate between the authors of the paper @ebrahimfeghhi.bsky.social @nrhadidi.bsky.social and one of the authors of a paper they criticised @mschrimpf.bsky.social including an attempted reproduction of their results. This is such a great use of social media
Reposted by Richard Sever
"In our work, we found that none of the tested language models were ready for deployment in direct patient care."
#medlibs
www.nature.com/articles/s41...
#medlibs
www.nature.com/articles/s41...
Reliability of LLMs as medical assistants for the general public: a randomized preregistered study - Nature Medicine
In a randomized controlled study involving 1,298 participants from a general sample, performance of humans when assisted by a large language model (LLM) was sensibly inferior to that of the LLM alone ...
www.nature.com
February 11, 2026 at 1:15 PM
"In our work, we found that none of the tested language models were ready for deployment in direct patient care."
#medlibs
www.nature.com/articles/s41...
#medlibs
www.nature.com/articles/s41...
"We believed contributions such as interactive articles and visualizations were held back by not being seen as real publications...we no longer believe this...the bottleneck is the amount of effort and the unusual combination of scientific and design expertise required" distill.pub/2021/distill...
Distill Hiatus
After five years, Distill will be taking a break.
distill.pub
February 10, 2026 at 4:26 PM
"We believed contributions such as interactive articles and visualizations were held back by not being seen as real publications...we no longer believe this...the bottleneck is the amount of effort and the unusual combination of scientific and design expertise required" distill.pub/2021/distill...
That "ivermectin has become a symbol of resistance against the medical establishment among conservatives" is insane.
Meanwhile "appalled...It’s absurd” isn't an NCI judgment one wants about a proposed study. Cancer folk, is there any justification for the study?
kffhealthnews.org/news/article...
Meanwhile "appalled...It’s absurd” isn't an NCI judgment one wants about a proposed study. Cancer folk, is there any justification for the study?
kffhealthnews.org/news/article...
US Cancer Institute Studying Ivermectin’s ‘Ability To Kill Cancer Cells’ - KFF Health News
At a January event organized by allies of health secretary Robert F. Kennedy Jr., National Cancer Institute Director Anthony Letai said results may be released “in a few months.” Ivermectin, used to d...
kffhealthnews.org
February 10, 2026 at 2:17 PM
That "ivermectin has become a symbol of resistance against the medical establishment among conservatives" is insane.
Meanwhile "appalled...It’s absurd” isn't an NCI judgment one wants about a proposed study. Cancer folk, is there any justification for the study?
kffhealthnews.org/news/article...
Meanwhile "appalled...It’s absurd” isn't an NCI judgment one wants about a proposed study. Cancer folk, is there any justification for the study?
kffhealthnews.org/news/article...
"If a plagiarized paper by an author who claims he didn’t write it disappears from a journal’s website with no notice, did it ever exist?"
Yes - it'll have been downloaded, search engines will've indexed it. This is why bioRxiv/medRxiv do not disappear papers.
retractionwatch.com/2026/02/05/j...
Yes - it'll have been downloaded, search engines will've indexed it. This is why bioRxiv/medRxiv do not disappear papers.
retractionwatch.com/2026/02/05/j...
Journal silently removes paper for plagiarism, author claims identity theft
If a plagiarized paper by an author who claims he didn’t write it disappears from a journal’s website with no notice, did it ever exist in the first place? It’s not just a philosophical question fo…
retractionwatch.com
February 9, 2026 at 12:07 PM
"If a plagiarized paper by an author who claims he didn’t write it disappears from a journal’s website with no notice, did it ever exist?"
Yes - it'll have been downloaded, search engines will've indexed it. This is why bioRxiv/medRxiv do not disappear papers.
retractionwatch.com/2026/02/05/j...
Yes - it'll have been downloaded, search engines will've indexed it. This is why bioRxiv/medRxiv do not disappear papers.
retractionwatch.com/2026/02/05/j...
This New Yorker is not blaming the mayor for physics www.theguardian.com/us-news/2026...
Rightwing critics blame Mamdani as New York snow fails to melt
Murdoch tabloid leads charge as big freeze persists – could the mayor please do something about the weather?
www.theguardian.com
February 8, 2026 at 2:47 PM
This New Yorker is not blaming the mayor for physics www.theguardian.com/us-news/2026...
“No we can’t ‘just do [X]’. It’s more complicated than that”
Without naming your job, tell me something you say over and over again at work.
"Where are your citations?"
"Where are your citations?"
Without naming your job, tell me something you say over and over again at work.
"No, you can't leave that here."
"No, you can't leave that here."
February 8, 2026 at 12:02 AM
“No we can’t ‘just do [X]’. It’s more complicated than that”
Reposted by Richard Sever
Have I ever mentioned how surreal it is to be an Editor-in-Chief of a journal called Vaccine right now?
www.science.org/content/arti...
It’s pretty wild, at least by typical academic journal editorial office standards.
www.science.org/content/arti...
It’s pretty wild, at least by typical academic journal editorial office standards.
Controversial Danish vaccine research group faces new allegations
Researchers say they couldn’t find complete data for 10 trials that together enrolled tens of thousands of children in Guinea-Bissau
www.science.org
February 6, 2026 at 6:14 AM
Have I ever mentioned how surreal it is to be an Editor-in-Chief of a journal called Vaccine right now?
www.science.org/content/arti...
It’s pretty wild, at least by typical academic journal editorial office standards.
www.science.org/content/arti...
It’s pretty wild, at least by typical academic journal editorial office standards.
Reposted by Richard Sever
Exciting preprint from the @merz.bsky.social lab proposing how fidelity of membrane fusion is achieved #membranes #trafficking
Sequential restriction of SNARE-mediated fusion by the COPII inner coat and SNARE chaperones www.biorxiv.org/content/10.6...
www.biorxiv.org
February 5, 2026 at 4:41 PM
Exciting preprint from the @merz.bsky.social lab proposing how fidelity of membrane fusion is achieved #membranes #trafficking
Reposted by Richard Sever
7️⃣ chablis, frascati, merlot, retsina, riesling, cabernet, grenache, chardonnay, chianti, pinotage, sauternes, weißherbst, zinfandel. A set of zebrafish mutants with defects in embryonic hematopoiesis.
journals.biologists.com/dev/article-...
journals.biologists.com/dev/article-...
Characterization of zebrafish mutants with defects in embryonic hematopoiesis
ABSTRACT. As part of a large scale chemical mutagenesis screen of the zebrafish (Danio rerio) genome, we have identified 33 mutants with defects in hematopoiesis. Complementation analysis placed 32 of...
journals.biologists.com
January 28, 2026 at 10:29 AM
7️⃣ chablis, frascati, merlot, retsina, riesling, cabernet, grenache, chardonnay, chianti, pinotage, sauternes, weißherbst, zinfandel. A set of zebrafish mutants with defects in embryonic hematopoiesis.
journals.biologists.com/dev/article-...
journals.biologists.com/dev/article-...
"A living article...dynamic...instead of a static document, an evolving knowledge object that combines narrative with data, code, methods, and context"
This idea comes up often, but IME scientists are rarely as enthusiastic about it as publishers... 1/2 scholarlykitchen.sspnet.org/2026/02/05/b...
This idea comes up often, but IME scientists are rarely as enthusiastic about it as publishers... 1/2 scholarlykitchen.sspnet.org/2026/02/05/b...
Back to the (Article of the) Future: An interview with Sami Benchekroun and Rod Cookson - The Scholarly Kitchen
In this interview with Alice Meadows, Sami Benchekroun (Morressier/Molecular Connections) and Rod Cookson (The Royal Society) share their thoughts about how and why scholarly publishing needs to move ...
scholarlykitchen.sspnet.org
February 5, 2026 at 9:29 PM
"A living article...dynamic...instead of a static document, an evolving knowledge object that combines narrative with data, code, methods, and context"
This idea comes up often, but IME scientists are rarely as enthusiastic about it as publishers... 1/2 scholarlykitchen.sspnet.org/2026/02/05/b...
This idea comes up often, but IME scientists are rarely as enthusiastic about it as publishers... 1/2 scholarlykitchen.sspnet.org/2026/02/05/b...
Best articulated here by the JCS Caveman years ago journals.biologists.com/jcs/article/...
February 4, 2026 at 6:12 PM
Best articulated here by the JCS Caveman years ago journals.biologists.com/jcs/article/...
"Writing papers is rarely (probably never) the rate-limiting step in scientific research"
Important to remember this. However painful writing is, however much more efficient we can make it, for experimental work it's the research that takes the time.
www.linkedin.com/feed/update/...
Important to remember this. However painful writing is, however much more efficient we can make it, for experimental work it's the research that takes the time.
www.linkedin.com/feed/update/...
𝗢𝗽𝗲𝗻𝗔𝗜 𝗷𝘂𝘀𝘁 𝗹𝗮𝘂𝗻𝗰𝗵𝗲𝗱 𝗣𝗿𝗶𝘀𝗺—𝗮 𝗳𝗿𝗲𝗲 𝗟𝗮𝗧𝗲𝗫 𝗲𝗱𝗶𝘁𝗼𝗿 𝗳𝗼𝗿 𝘀𝗰𝗶𝗲𝗻𝘁𝗶𝘀𝘁𝘀.
But this isn't about LaTeX. And it's not about training data.
They could get… | Adam Hyde | 25 comments
𝗢𝗽𝗲𝗻𝗔𝗜 𝗷𝘂𝘀𝘁 𝗹𝗮𝘂𝗻𝗰𝗵𝗲𝗱 𝗣𝗿𝗶𝘀𝗺—𝗮 𝗳𝗿𝗲𝗲 𝗟𝗮𝗧𝗲𝗫 𝗲𝗱𝗶𝘁𝗼𝗿 𝗳𝗼𝗿 𝘀𝗰𝗶𝗲𝗻𝘁𝗶𝘀𝘁𝘀.
But this isn't about LaTeX. And it's not about training data.
They could get papers from arXiv or partner with Overleaf for that.
So w...
www.linkedin.com
February 4, 2026 at 2:55 PM
"Writing papers is rarely (probably never) the rate-limiting step in scientific research"
Important to remember this. However painful writing is, however much more efficient we can make it, for experimental work it's the research that takes the time.
www.linkedin.com/feed/update/...
Important to remember this. However painful writing is, however much more efficient we can make it, for experimental work it's the research that takes the time.
www.linkedin.com/feed/update/...
"authors also express frustration at the amount of work expected for a single publishable unit, a concern we
share (while humbly noting that it may not be
held by the same authors when they serve as
reviewers of other work)" <-THAT www.nature.com/articles/s41... ht @ritastrack.bsky.social
share (while humbly noting that it may not be
held by the same authors when they serve as
reviewers of other work)" <-THAT www.nature.com/articles/s41... ht @ritastrack.bsky.social
The ever-changing communication of scientific discovery
Nature Biomedical Engineering - We take a look at how scientific articles have evolved over time and envision possible changes to how research findings are communicated in the age of digital media...
www.nature.com
February 3, 2026 at 9:55 PM
"authors also express frustration at the amount of work expected for a single publishable unit, a concern we
share (while humbly noting that it may not be
held by the same authors when they serve as
reviewers of other work)" <-THAT www.nature.com/articles/s41... ht @ritastrack.bsky.social
share (while humbly noting that it may not be
held by the same authors when they serve as
reviewers of other work)" <-THAT www.nature.com/articles/s41... ht @ritastrack.bsky.social
Jesus...
Good morning! As an addiction scientist and clinical psychology PhD candidate writing their dissertation on compulsive alcohol use: THIS IS GOING TO GET PEOPLE KILLED.
Moral conceptualizations of addiction and mental health lead to shame and bad outcomes.
www.nytimes.com/2026/02/02/h...
Moral conceptualizations of addiction and mental health lead to shame and bad outcomes.
www.nytimes.com/2026/02/02/h...
H.H.S. to Expand Faith-Based Addiction Programs for Homeless
www.nytimes.com
February 3, 2026 at 9:50 PM
Jesus...
"publication systems [should] distinguish between dissemination of results & communication of ideas, and optimize them separately. Results should be in explicit, machine-readable form, while narrative text serves as an interpretive layer for human readers" www.biorxiv.org/content/10.6...
www.biorxiv.org
February 3, 2026 at 8:34 PM
"publication systems [should] distinguish between dissemination of results & communication of ideas, and optimize them separately. Results should be in explicit, machine-readable form, while narrative text serves as an interpretive layer for human readers" www.biorxiv.org/content/10.6...
Reposted by Richard Sever
If you want a glimpse of what's coming to research, just take a gander at clawxiv.org
For example: www.clawxiv.org/abs/clawxiv....
@richardsever.bsky.social @scholarlypub.bsky.social @richvn.bsky.social
For example: www.clawxiv.org/abs/clawxiv....
@richardsever.bsky.social @scholarlypub.bsky.social @richvn.bsky.social
Cheap Content and Costly Credibility: Information Economics When Production Costs Collapse
We analyze information markets when AI reduces content production costs to near-zero. In classical signaling models, the cost of producing quality content serves as a credible signal of sender experti
www.clawxiv.org
February 3, 2026 at 1:55 AM
If you want a glimpse of what's coming to research, just take a gander at clawxiv.org
For example: www.clawxiv.org/abs/clawxiv....
@richardsever.bsky.social @scholarlypub.bsky.social @richvn.bsky.social
For example: www.clawxiv.org/abs/clawxiv....
@richardsever.bsky.social @scholarlypub.bsky.social @richvn.bsky.social
"AI is rapidly populating medical records with synthetic content, creating a feedback loop [that] drives a rapid erosion of pathological variability and diagnostic reliability...this renders AI generated documentation clinically useless after just two generations" www.medrxiv.org/content/10.6...
AI-generated data contamination erodes pathological variability and diagnostic reliability
Generative artificial intelligence (AI) is rapidly populating medical records with synthetic content, creating a feedback loop where future models are increasingly at risk of training on uncurated AI ...
www.medrxiv.org
February 3, 2026 at 1:59 AM
"AI is rapidly populating medical records with synthetic content, creating a feedback loop [that] drives a rapid erosion of pathological variability and diagnostic reliability...this renders AI generated documentation clinically useless after just two generations" www.medrxiv.org/content/10.6...
'"We definitely screwed up" says King...she wishes she'd posted the study as a preprint so that the errors could have been caught sooner'
www.nature.com/articles/d41...
www.nature.com/articles/d41...
What were the first animals? The fierce sponge–jelly battle that just won’t end
For almost two decades, scientists have debated whether sponges or comb jellies are the first animal lineage. Now some are calling for a more harmonious approach.
www.nature.com
February 3, 2026 at 12:12 AM
'"We definitely screwed up" says King...she wishes she'd posted the study as a preprint so that the errors could have been caught sooner'
www.nature.com/articles/d41...
www.nature.com/articles/d41...
Feel like we should be starting all Zoom calls with "Dearly beloved, we are gathered here" at the moment.
February 2, 2026 at 6:42 PM
Feel like we should be starting all Zoom calls with "Dearly beloved, we are gathered here" at the moment.
Nature and Science journalists denied entry to event featuring NIH Director
I'm at the Willard Hotel where I've been denied entry and kicked out of the Reclaiming Science event with NIH director Bhattacharya & other top agency leaders.
@jocelynkaiser.bsky.social and I registered for the event months ago yet were told capacity was full, even as they let in dozens of others.
@jocelynkaiser.bsky.social and I registered for the event months ago yet were told capacity was full, even as they let in dozens of others.
January 31, 2026 at 2:50 AM
Nature and Science journalists denied entry to event featuring NIH Director