Mark

Massive single-cell study by Kanai et al (www.medrxiv.org/content/10.1...):
- Once statistical power is high, constrained genes have more (though weaker) eQTLs.
- Chromatin-QTLs near constrained genes have "normal" effect sizes, colocalize more with disease, but exhibit attenuated peak-gene effects.

I don't understand how anyone can watch how blatantly Grok is manipulated to answer the way ownership desires it to and then act like the other LLM chatbots couldn't possibly be similarly but less obviously compromised to produce responses in whatever way corporate interests and priorities dictate.

Replying to a long text screenshot meme: “I ain’t reading all that. I’m happy for u tho. Or sorry that happened”

When reading a long stream-of-consciousness-type message on github or slack

Great paper! You may also be interested in Unravel (Shrestha, Barik, and Parnin UIST 2021) regarding usability of the pipe operator, and Greenberg and Buxton CHI 2008 regarding the tradeoffs of HCD and EID that you discuss.

the UI design principles and the two UI approaches

My paper "Examining the Interface Design of Tidyverse" is now published in ANZJS. This paper calls attention to the importance of UI/UX principles & approaches to statisiticians and #rstats developers.

Many thanks to reviewers who made this paper better!

🔗 onlinelibrary.wiley.com/share/author...

Many people hours, calls and messages later: OSTA is now “in (pre)print”, though the real thing lives at bioconductor.org/books/OSTA.

Check it out, get in touch. We welcome any feedback, suggestions, wishes (& contributions).

It’s been a joy working with you @estellayixingdong.bsky.social!

Do not use density-based visual encodings for non-density-preserving embeddings/dimensionality-reductions

Instagram Create an account or log in to Instagram - Share what you're into with the people who get you.

Also www.instagram.com/reel/DQ8YG60...

Chuck Schumer Helps Pull Democrats Back From Brink Of Courage

Chuck Schumer Helps Pull Democrats Back From Brink Of Courage

Trump administration demands states 'undo' full SNAP payouts as states warn of "catastrophic impact" President Donald Trump's administration is demanding that states reverse full SNAP benefits issued under recent court orders. The U.S.

Trump administration demands states "undo" efforts to provide full SNAP payouts for November amid court battle

www.instagram.com/reel/DPofE6D...

macOS 26 Tahoe: The Ars Technica review Liquid Glass brings translucent sheen to the typical batch of iterative changes.

Thanks! Are the WebGPU updates in Safari 26 only expected to work on Tahoe? Examples I try fail on Safari 26 + Sequoia + WebGPU feature flag. This article discusses it but I have not seen anything official about certain features being dependent on OS version too arstechnica.com/gadgets/2025...

TAXES FOR BILLIONAIRES

FLORIDA FOR COWARDS LEAVING NYC

The Mamdani political sign generator is kinda fun https://homes.luddy.indiana.edu/ccshan/for/for.html

Peer review seems like one of the last places AI should be used. The main args in-favor seem related to volume and bad-faith submissions of non-science content. More transparent approaches such as AbstractExplorer from @elglassman.bsky.social should be preferred doi.org/10.1145/3746059.3747773

A scatter plot titled "Fair and square? Russian federal elections, 2000-21" which visualizes the results of Russian federal elections. The x-axis represents voter turnout in percentage, ranging from 0 to 100%. The y-axis represents the percentage of votes for Putin, Medvedev, or the United Russia party, also ranging from 0 to 100%. Each dot on the plot represents a polling station. A cluster of dots appears in the upper right corner, indicating high voter turnout and a high percentage of votes for the specified candidates. A dashed rectangle highlights a grid-like pattern of dots at numbers ending in zero and five, with an annotation suggesting this pattern indicates potential foul play. The source is cited as Kobak and Shpilkin (2021).

Now this is how you detect whether an election was stolen. Humans choose rounder numbers.

by @TheEconomist

Jury acquits D.C. 'sandwich guy' charged with chucking a sub at a federal agent Sean Dunn faced single misdemeanor offense after federal grand jurors refused to indict him on the felony charge sought by prosecutors.

BREAKING: JUSTICE DEPARTMENT EATS A LOSS IN 'SANDWICH GUY' CASE AS JURORS FIND DEFENDANT NOT GUILTY www.nbcnews.com/politics/jus...

Your cat when you're choking to death in your apartment

Do we think the developers of the various 2FA apps put the "Yes" and "No" buttons on different sides on purpose

A diagram illustrating pointwise scoring with a large language model (LLM). At the top is a text box containing instructions: 'You will see the text of a political advertisement about a candidate. Rate it on a scale ranging from 1 to 9, where 1 indicates a positive view of the candidate and 9 indicates a negative view of the candidate.' Below this is a green text box containing an example ad text: 'Joe Biden is going to eat your grandchildren for dinner.' An arrow points down from this text to an illustration of a computer with 'LLM' displayed on its monitor. Finally, an arrow points from the computer down to the number '9' in large teal text, representing the LLM's scoring output. This diagram demonstrates how an LLM directly assigns a numerical score to text based on given criteria

[corrected link]

LLMs are often used for text annotation in social science. In some cases, this involves placing text items on a scale: eg, 1 for liberal and 9 for conservative

There are a few ways to handle this task. Which work best? Our new EMNLP paper has some answers🧵
arxiv.org/abs/2509.03116

It really is amazing that when a popular young Democratic politician who wants to solve real problems for people beats an old-guard institutionalist Democrat in a primary, the institutionalist Democrat just stays in the race and basically runs as a Republican.

... they are ultimately models of text. Although they can do super-human text manipulation tasks, that doesn't mean they have concepts and understanding - the chain of thought looks cute, but I am not convinced it is working at a conceptual level - rather it is working at text reporting level

Wow. Harvard nuking its PhD programs

- Science PhD admissions reduced by more than 75%
- Arts & Humanities reduced by about 60%
- Social Sciences by 50–70%
- History by 60%
- Biology by 75%
- The German department will lose all PhD seats
- Sociology from six PhD students to zero

German far right setting agenda as opponents amplify its ideas, study finds Normalisation of far-right stances likely to affect success of such parties at ballot boxes across Europe, say researchers

German study shows how far right sets agenda thanks to mainstream parties reshaping their communications to respond to what are initially fringe issues, boosting the spread of these ideas and signalling to voters that these ideas and stances were legitimate

shorturl.at/OFfig

Reaffirming the value of model organisms in training scientific minds - Nature Cell Biology As biomedical research prioritizes human models and translational promise, classic model organisms are increasingly dismissed. Here we argue that they have a lasting value, both in enabling discovery and in cultivating scientific thinking, by training researchers in systems reasoning, integrative thinking and independent inquiry.

PREACH!

www.nature.com/articles/s41...

Spatial architecture of development and disease - Nature Reviews Genetics Spatial omics has empowered the discovery of developmental and disease-associated molecular signatures, cell states and multicellular niches, as well as the evaluation of disease heterogeneity within and across organs. The authors review spatially resolved molecular changes across diseases and discuss the potential of spatial multi-omics for clinical applications, including the recent impact of artificial intelligence.

Spatial architecture of development and disease
go.nature.com/477Dmxs
#Review by Enikő Lázár & Joakim Lundeberg
Free to read here: rdcu.be/eJ9E4

Also, it is not easy to quantify the extent to which higher-quality software may result in more citations/usage because it was easier to use. And even if projects stop their software dev. at the prototype stage, we should still strive to make the code open-source, for the sake of reproducibility

I agree that papers should separate the design goals and key advancements from their implementation. One thing the post omits is how to square stopping at prototype with the fact that collaborators (e.g., expert interview participants) often expect production-quality software doi.org/10.1145/3544...

‘Completely shattered.’ Changes to NSF’s graduate student fellowship spur outcry The announcement comes months later than usual, leaving many would-be applicants stranded

Today was a hard day for Ph.D. students who found out that they can no longer apply for NSF's prestigious Graduate Research Fellowship Program. "Devastating“ was how one student described it to me. #GradSchool #NSFGRFP

www.science.org/content/arti...

Happy to share the Biodiversity Cell Atlas white paper, out today in @nature.com. We look at the possibilities, challenges, and potential impacts of molecularly mapping cells across the tree of life.
www.nature.com/articles/s41...

E-Book Contracts Are a Big Cost for Public Libraries. One Harvard Librarian Is Fighting to Change That. | News | The Harvard Crimson For the past five years, Kyle K. Courtney, who directs copyright and information policy for Harvard’s libraries, has fought to make e-books more accessible to public libraries across the United…

For the past five years, Kyle K. Courtney, who directs copyright and information policy for Harvard’s libraries, has fought to make e-books more accessible to public libraries across the United States.

Sophie Gao and Alexandra M. Kluzak report.

www.thecrimson.com/article/2025...