Reposted by Nicholas MancusoSasha Gusev @sashagusevposts.bsky.social · 4hA searchable mapping between NIH study sections and program officers that doesn't suck. (Claude one-shot) sashagusev.github.io/po-study-sec... 2199
Reposted by Nicholas MancusoRob Patro @robp.bsky.social · 9hblog.google/innovation-a... "Large Scale Codebase Migrations and Optimizations: Argon agents are working on migrating C/C++ codebases to Rust across Google—scaling from tens of thousands of lines ... up to 800K+ lines for the Fuchsia OS Zircon kernel." You love to see it!blog.googleGemini 4 Argon: our next era of frontier intelligenceAnnouncing Gemini 4 Argon, our frontier model for real-world coding, enterprise knowledge work, and cyber defense, rolling out soon. 071
Reposted by Nicholas MancusoGOLIKEHELLMACHINE @golikehellmachine.com · 11hthis is a man who knows he’s getting canned or impeached after the election and who is trying to set up a future political career 3637640
Reposted by Nicholas MancusoEugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 11hIn our Nature paper, we introduce the first superhuman Stratego AI, which we built using general techniques that we developed for RL & test-time compute under imperfect information. www.nature.com/articles/s41...nature.comScalable decision-making for games of imperfect information - NatureAtaraxos, an AI for the board wargame Stratego, establishes a design pattern for reinforcement learning and search that is effective under large amounts of hidden information, a longstanding desiderat... 1018446
Reposted by Nicholas MancusoArnaud Martin @evolvwing.bsky.social · 12hCell Press giving a rather biased, pro-eugenics platform to self-proclaimed "Biotech Barbie", former a Thiel Fellow who briefly married He Jiankui, and fails to properly frame the debate on germline editing. Partly bonkers, partly important, check out this thread 🤯 0118
Reposted by Nicholas MancusoJoe Hanly @hanliconius.bsky.social · 12h@cellpress.bsky.social has just released an.... interesting new 'short review' on pre-implantation gene editing, in which author Cathy Tie argues for the ethical imperative of pushing for this technology. www.sciencedirect.com/science/arti... There are, to put it mildly, a few issues here. 1/ 188
Reposted by Nicholas MancusoChris Smith @chriscrsmith.bsky.social · 12hI'm hiring a postdoc in population and evolutionary genetics at Indiana University Bloomington! We're a computational lab working on spatial demographic inference using genomic data and machine learning. You can apply here: indiana.peopleadmin.com/postings/34508indiana.peopleadmin.comIndiana University Academic Positions 03048
Reposted by Nicholas MancusoAriel Edwards-Levy @aedwardslevy.bsky.social · 12hThere is a lot of "oh dear" to go around in the results of this very thorough Pew study on the issues with synthetic samples, but this strikes me as particularly "oh dear" www.pewresearch.org/data-labs/20... 7423149
Reposted by Nicholas MancusoMasahiro Kanai @masakanai.bsky.social · 13hMulti-layered regulatory buffering reconciles the paradox and explains why chromatin captures disease risk that expression misses. At constrained genes, caQTLs colocalize with GWAS more than eQTLs (71% vs 28%). caQTLs mediate 79% of immune disease heritability, eQTLs 25%. 141
Reposted by Nicholas MancusoMasahiro Kanai @masakanai.bsky.social · 13hOur single-nucleus immune multiome atlas is now out in @nature.com! 🧬 10M PBMCs from 1,108 @finngen.bsky.social donors recruited by Finnish Blood Service, profiled at @broadinstitute.org with chromatin accessibility and gene expression in the same nuclei, to trace how disease variants act 🧵👇 12512
Reposted by Nicholas MancusoMitch Wolfe @whrwlfthrwlf85v.bsky.social · 19hNader wanted Bush to win so that he could spend the rest of his life saying “I told you so” to Democrats, and sanctimoniously claiming that Al Gore and Democrats had to earn his voters, as if they were God’s chosen that needed catered to instead of the spoiled and flaky assholes they were. 2326
Reposted by Nicholas Mancusoangela zhou @angelamczhou.bsky.social · 29/09/2026I love Jev. Jev is an example of what will be the next "frontier": compiling intelligence to cheaper and more structured outputs. Thinking of running my new AI evals class on Jev or open-source alternatives bc my institution keeps asking us to do AI stuff without additional resources to do so 171
Reposted by Nicholas MancusoData & Society @datasociety.bsky.social · 29/09/2026Speaking to @politico.com, D&S board member @geomblog.bsky.social has some advice for lawmakers: "Stop worrying about superintelligence, and start making companies pay for putting out unreliable and error-prone systems that could and have harmed people." www.politico.com/newsletters/...politico.com5 Questions for Suresh Venkatasubramanian 0153
Reposted by Nicholas MancusoPhilippe Batut @philippebatut.bsky.social · 10/09/2026🧬🔬 Why are enhancers transcribed and how does that impact gene regulation? I’m really excited to share our new paper in @science.org showing that noncoding RNAs control the timing of gene activation in embryos. With Mike Levine #ScienceResearch @columbiamed.bsky.social A few highlights below... 🧵👇science.orgNoncoding transcription controls the developmental dynamics of long-range gene regulationThe genomic regions regulating gene expression are often themselves transcribed into a variety of noncoding RNAs (ncRNAs). However, the regulatory roles of this noncoding transcription remain largely ... 5233114
Reposted by Nicholas MancusoEllie Armstrong @elliecat.bsky.social · 29/09/2026If you saw me talk at PEQG or PAG and want to try out a new method for assembling sex chromosomes, our paper and pipeline are live! Biggest shoutout to @scarey.bsky.social for pioneering this method and happy to have played a part in its application and deployment! 02011
Reposted by Nicholas MancusoThe A.V. Club @avclub.com · 29/09/2026Meta's adorable AI is, predictably, a security nightmareavclub.comMeta's adorable AI is, predictably, a security nightmareMeta's adorable AI is, predictably, a security nightmare The cutesy Muse avatar, Jolly, lulls users into a false sense of security as the AI gives their address out to Facebook users. 4164
Reposted by Nicholas MancusoJeremy Van Cleve @vancleve.theoretical.bio · 29/09/2026My group is recruiting a Ph.D student for Fall 2027! We're a friendly group of mathematically-minded and computationally-inspired biologists studying topics including social evolution, host-pathogen interactions, genetics, and demography. See vancleve.theoretical.bio for more info!vancleve.theoretical.biovan cleve research groupVan Cleve Research Group – theoretical evolution & ecology at the University of Kentucky 04965
Reposted by Nicholas MancusoTheodore Grunewald @tedgrunewald.bsky.social · 29/09/2026JCO for the win. 231447222
Reposted by Nicholas MancusoRasmus Pagh @rasmuspagh.net · 29/09/2026Very interesting report about AI and theoretical computer science, based on a recent meeting at @simonsinstitute.bsky.social simons.berkeley.edu/ai-tcs-worki... These are times of change, but I am optimistic that there will always be a valuable place for humans who seek to understand.simons.berkeley.eduAI + TCS Working Group 0123
Reposted by Nicholas MancusoStephanie Hicks @stephaniehicks.bsky.social · 29/09/2026Excited to share that @jhubiostat.bsky.social is hiring for postdoctoral fellows and tenure-track Assistant Professors! Details below and committees will start reviewing for both on Oct 1! 🎉 1810
Nicholas Mancuso @nmancuso.bsky.social · 29/09/2026As an example of this, a pet hypothesis of mine is that GxG is nonsensical for cis-regulatory variation. Cross gamete interactions dont make sense wrt local chromatin state (to me at least), and thus any interactions should be within the same gamete/haplotype, or HxH. 140
Nicholas Mancuso @nmancuso.bsky.social · 29/09/2026I thought the A24 release of Backrooms was just ok? The performances were solid and the concept really interesting but felt the message kind of tripped over itself nearing the end? 000
Reposted by Nicholas MancusoNeedhi Bhalla 💅🏽 @needhibhalla.bsky.social · 28/09/2026“I believe that science derives its authority from…interactions within a community of people trained to interrogate evidence and answer for the conclusions they draw from it. And it is time to formalize what this means, and which sorts of human judgment cannot (or should not) be offloaded to AI.” 01311
Nicholas Mancuso @nmancuso.bsky.social · 29/09/2026"And it is time to formalize what this means, and which sorts of human judgment cannot (or should not) be offloaded to AI." Excellent article and I think one that underscores much of my own thoughts on AI, as someone who uses it! 161
Reposted by Nicholas MancusoC. Brandon Ogbunu @cbo.bsky.social · 27/09/2026In my latest for @undark.org, I compare AI adoption in science to the Ship of Theseus: "If you swap out every wooden plank on a ship over time, is it still the same vessel?" I argue that science derives authority from something more than data generation + publication. undark.org/2026/09/25/o...undark.orgAI Presents Science with a ‘Ship of Theseus’ ProblemOpinion | AI can replace each part of the scientific process, leaving the enterprise recognizable but devoid of human judgment. 48243
Reposted by Nicholas MancusoCarly Goodman @carlygoodman.bsky.social · 28/09/2026Congratulations, we have created a world where teenagers cannot exist with each other in space, where we use “safety” to destroy sociality, youth, and development. 1424694
Reposted by Nicholas MancusoEd @ed3d.net · 29/09/2026I have something for thisyoutube.comAI hallucinations are a skill issue and I can prove itYouTube video by ed3d dot net 2121
Reposted by Nicholas MancusoCarl Zimmer @carlzimmer.com · 27/09/2026Four days ago, a scientist was surprised by the news that Anthropic discovered new virus genes for making DNA. He says he’s been studying them for years—and feeding his data to Anthropic’s AI as part of his own research. Coincidence? Here’s my story. nyti.ms/3VSWaycnyti.msDid Anthropic’s A.I. Really Make a Scientific Discovery on Its Own? (Gift Article)An expert at the University of Copenhagen said his team had been sharing its research with the company’s A.I. model, Claude, and that its new finding matched their work. 892116865
Reposted by Nicholas Mancusojamelle @jamellebouie.net · 28/09/2026i don’t care how much fancy math you can do, if you take open racism as a sign that someone is epistemologically sound then you are a fucking idiot 774872
Nicholas Mancuso @nmancuso.bsky.social · 28/09/2026Inyo National Forest is fantastic this time of year 061
Reposted by Nicholas MancusoArun Durvasula @arundurvasula.bsky.social · 24/09/2026new manuscript! 1133
Nicholas Mancuso @nmancuso.bsky.social · 24/09/2026Switching is trivial! I switched months ago and haven't looked back. 110
Nicholas Mancuso @nmancuso.bsky.social · 24/09/2026Time is a flat circle. If we managed to develop solutions in the past, we'll develop solutions to the current not at all new challenge 000
Reposted by Nicholas MancusoPrasad Jallepalli, MD, PhD @prasad.bsky.social · 24/09/2026I haven't looked at the press release, but FTEs and AI tokens cost $$. Open-weight models could help individual labs, but as with campaign finance laws, $$ and those who decide on their allocation have disproportionate control. P.S. Note that 1st & last author FTEs at Anthropic are Doudna alums. 1173
Reposted by Nicholas MancusoHanbin Lee @epigenci.bsky.social · 24/09/2026Love this comparison to historical anecdotes. 141
Nicholas Mancuso @nmancuso.bsky.social · 24/09/2026These chat bots have errors that are ineradicable by design! Anyways lemme fit a statistical to support my findings! 🤡 210
Reposted by Nicholas MancusoJames Pirruccello 🎃🦇🧹 @jamespirruccello.com · 23/09/2026I highly encourage people not to give AI slop presentations. Just throw something bad on some slides, it will be much more compelling than a glossy content that you didn't produce. If we want Claude to present, we can generate that at home. 3513
Reposted by Nicholas MancusoJames Pirruccello 🎃🦇🧹 @jamespirruccello.com · 23/09/2026www.anthropic.com/news/claude-...anthropic.comClaude discovers a novel enzyme system with CRISPR-like repeatsAnthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems. 153
Reposted by Nicholas MancusoPiper Below @piperbelow.bsky.social · 23/09/2026Very excited to see these three important papers published. This News and Views provides a short summary, for those interest in this new transcriptomic resource for the prefrontal cortex! www.nature.com/articles/d41...nature.comVast cellular gene-expression atlas could transform how scientists understand brain ageing and diseaseSingle-nucleus RNA sequencing captures molecular changes in the prefrontal cortex across the lifespan and in health and disease. 1113
Reposted by Nicholas MancusoSam Power @spmontecarlo.bsky.social · 23/09/2026A point which I continue revisiting in recent conversations about "what is to become of (mathematics) research?" is that there remains no scarcity of interesting things to wonder about, that many things should certainly still be written down, disseminated, and that this is "what it's all about". 1111
Reposted by Nicholas MancusoColin Carlson @colincarlson.bsky.social · 23/09/2026What's behind this tidal wave of epidemics? In the largest study ever conducted on emerging infectious diseases, we show that the popular explanation - "our relationship with nature is out of balance" - doesn't hold up. Out now in Nature, after half a decade of work: www.nature.com/articles/s41...nature.comThe anthropogenic fingerprint on emerging infectious diseases - NatureEmerging infectious disease outbreak risks are ubiquitous worldwide, and highest in landscapes where people, livestock and fragmented natural ecosystems coexist, but detection and data ... 8553276
Reposted by Nicholas MancusoLos Angeles Times @latimes.com · 22/09/2026Los Angeles needs more housing. But developers are increasingly walking away from apartment projects as the economics of building become more difficult. What’s behind the construction slowdown and what it means for the city’s housing crisis.latimes.comWhy L.A. developers aren't building more apartments, despite epic housing shortageL.A. needs more multifamily housing, but developers say it doesn't make sense to build apartments in the current market. 063
Reposted by Nicholas Mancusoex-Lethality Jane @lethalityjane.bsky.social · 21/09/2026I am once again hypothesizing that they bonded over their mutually terrible taste in musical theater. 618821
Reposted by Nicholas Mancusohoffsbeefs.bsky.social @hoffsbeefs.bsky.social · 21/09/2026Something i'm concerned about that we need to be talking about. Many in our community are using agentic AI to improve cow breeding. These models are powerful. Perhaps too powerful. I call it P(moo). 233
Reposted by Nicholas MancusoZamin Iqbal @zaminiqbal.bsky.social · 21/09/2026Careful implementation of Bayesian phylogenetics in Delphy allows huge speed gains and identical results to BEAST, but runs on your laptop browser www.nature.com/articles/s41... Amazing work from @pardissabeti.bsky.social , Patrick Varilly, @benfry.com and co. 14811
Reposted by Nicholas Mancusojamelle @jamellebouie.net · 21/09/2026one thing that is worth remember apropos this useful piece is that novel technologies do not necessarily require novel political solutions. americans have had to deal with fast-moving, disruptive technological change in the past, and they did so through collective, democratic action.nytimes.comOpinion | Bernie Sanders, Steve Case and Others on What Washington Should Do Now on A.I.How can the government reel in the technology? Here are six ideas. 372170269
Nicholas Mancuso @nmancuso.bsky.social · 21/09/2026Feels great to wake up seeing a mountain of responses from strangers calling into question one's ability to carry out research or mentor. Based on a single thread which doesn't even reflect my modal attitude about AI. Very fun. Very normal of folks. 030