Anselm Levskaya @levskaya.bsky.social · 07/06/2025Oh gods - let us know in DMs or otherwise how we’re the source of suffering. 100
Reposted by Anselm LevskayaThomas Ptacek @sockpuppet.org · 04/06/2025My "early-career" developer feelings are complicated and alienating to SFBA-type career ladder people. 1161
Reposted by Anselm LevskayaStephen Curry @scurry.bsky.social · 23/05/2025If your test for ground-breaking discoveries can’t detect the discovery of RNAi, or of CRISPR-Cas9, or the cryoEM resolution revolution or Alphafold2, maybe it’s not a very good test. www.nature.com/articles/d41... 1518534
Reposted by Anselm LevskayaEthan Mollick @emollick.bsky.social · 20/05/2025At some point, the fact that over a billion people use this technology and that they self-report high utility has to mean something. There is lots to criticize about AI and plenty of real issues caused by AI, but the narrative that this is all a fake thing that will disappear doesn't help anyone. 4457
Reposted by Anselm LevskayaCharlie Marsh @crmarsh.com · 13/05/2025Today, we’re announcing the preview release of ty, an extremely fast type checker and language server for Python, written in Rust. In early testing, it's 10x, 50x, even 100x faster than existing type checkers. (We've seen >600x speed-ups over Mypy in some real-world projects.) 1433284
Reposted by Anselm LevskayaRyan Moulton @moultano.bsky.social · 12/05/2025This isn't true. I'm the person who ran the experiments this is BSing about. When search results are worse, people attempt fewer tasks. When they're better they attempt more. 925353
Reposted by Anselm LevskayaSam @samd.bsky.social · 23/04/2025LET’S FUCKING GOOOO THE MOST AMBITIOUS HOUSING BILL IN THE HISTORY OF THE STATE OF CALIFORNIA HAS ADVANCED OUT OF COMMITTEE 832932
Reposted by Anselm LevskayaZach Hensel @zachhensel.bsky.social · 12/04/2025@zey.bsky.social I’m not on the site with Nazis anymore — wtf are you saying about me over there? You know there’s new data on the origins question, right? 4409
Reposted by Anselm LevskayaZach Hensel @zachhensel.bsky.social · 11/04/2025The case for "lab leak" book "VIRAL" by Alina Chan and Matt Ridley leaned heavily on two facts in 2021: 1. The Wuhan Institute of Virology had sampled the virus most identical to SARS-CoV-2 2. SARS-CoV-2 lineage B, but not lineage A, was found in Huanan market By 2022, neither was true, so... 35013
Reposted by Anselm LevskayaJieyu Zheng @jieyusz.bsky.social · 12/04/2025My first outing of "The Unbearable Slowness of Being" at the Caltech Chen Neuroscience Workshop today! What does living at 10 bits/s mean for humans, flies, mice, and crows? More here: jieyusz.github.io/talks/2025_c... Thanks to Profs. @cfcamerer.bsky.social & Carlos for the invite! 183
Reposted by Anselm LevskayaDiego del Alamo @delalamo.xyz · 03/03/2025There it is again: using PLMs to predict antigen-epitope interactions from sequence alone yields a prediction accuracy of 0.65, in line with a proposed upper limit from a previous study QTed below (from doi.org/10.1101/2025.02.12.637989; I deleted an older version of this post due to typos/errors) 2163
Reposted by Anselm LevskayaThomas Wolf @thomwolf.bsky.social · 19/02/2025After 6+ months in the making and over a year of GPU compute, we're excited to release the "Ultra-Scale Playbook": hf.co/spaces/nanot... A book to learn all about 5D parallelism, ZeRO, CUDA kernels, how/why overlap compute & coms with theory, motivation, interactive plots and 4000+ experiments!hf.coThe Ultra-Scale Playbook - a Hugging Face Space by nanotronThe ultimate guide to training LLM on large GPU Clusters 217952
Reposted by Anselm LevskayaFred Scharmen @sevensixfive.bsky.social · 26/01/2025Ceding techno optimism to the right is a generational scale mistake 26494105
Reposted by Anselm LevskayaAngie Rasmussen @angierasmussen.bsky.social · 17/02/2025AI is revolutionary but it also elicits some of the dumbest LinkedIn takes of all time. I’d like this lawyer whose primary expertise is scaremongering for EA money to explain how using “future kinds of models” starts pandemics. Outline the model to pandemic pipeline for me. 36918
Reposted by Anselm LevskayaJeff Dean @jeffdean.bsky.social · 09/02/2025Well said, @carlbergstrom.com. I also feel the dismantling of our scientific institutions & funding agencies for basic science is an attack on all scientists, wherever they might be (government or corporate lab, academic institution, ...). Our collective identities are about advancing knowledge. 1747
Reposted by Anselm LevskayaJeff Dean @jeffdean.bsky.social · 09/02/2025Another example of what cutting science funding is doing to our leading university research programs in the U.S.: dismantling things like the Soybean Innovation Lab at UIUC, which have made US crop yields dramatically higher. 8432159
Reposted by Anselm LevskayaRoy Frostig @froystig.bsky.social · 04/02/2025Our online book on systems principles of LLM scaling is live at jax-ml.github.io/scaling-book/ We hope that it helps you make the most of your computing resources. Enjoy!jax-ml.github.io How To Scale Your Model Training LLMs often feels like alchemy, but understanding and optimizing the performance of your models doesn't have to. This book aims to demystify the science of scaling language models on TPUs: how... 3359
Reposted by Anselm LevskayaMichael Eisen @mbeisen.bsky.social · 04/02/2025Y'all are just noticing that scientific societies are obsequious cowards on everything except preserving their publishing cash cows? 5672
Reposted by Anselm Levskayajacobaustin123.bsky.social @jacobaustin123.bsky.social · 04/02/2025Making LLMs run efficiently can feel scary, but scaling isn’t magic, it’s math! We wanted to demystify the “systems view” of LLMs and wrote a little textbook called “How To Scale Your Model” which we’re releasing today. 1/n 39429
Reposted by Anselm LevskayaAnne Applebaum @anneapplebaum.wsocial.eu · 07/01/2025"When conspiracy theories and nonsense cures are widely accepted, the evidence-based concepts of guilt and criminality vanish quickly too." www.theatlantic.com/magazine/arc...theatlantic.comThe New RasputinsAnti-science mysticism is enabling autocracy around the globe. 12235651233
Reposted by Anselm LevskayaDerek Lowe @dereklowe.bsky.social · 06/01/2025Re-upping this reply thread from last night. Drugs don't come from nowhere, folks. And we're not ripping off the NIH, either. 57823
Reposted by Anselm LevskayaNathaniel Johnston @njohnston.ca · 20/12/2024There's been a bunch of claims (mostly on X) that ChatGPT did great on this year's Putnam math competition. Let's do a thread to talk about it! 🧵 #MathSky 24417
Reposted by Anselm LevskayaJieyu Zheng @jieyusz.bsky.social · 17/12/2024Attached is a piece of art by me for "The Unbearable Slowness of Being: Why do we live at 10 bits/s?" Hope this will inspire you to think about the brain from a new perspective! 03410
Anselm Levskaya @levskaya.bsky.social · 19/12/2024You really need long (compared to cells) metal antennae for decent radio, and biology never got that good at reducing and templating crystalline metals/alloys. 020
Reposted by Anselm LevskayaSara Hooker @sarahooker.bsky.social · 04/12/2024AI amplifying biorisk has been a major focus in AI policy & governance work. Is the spotlight merited? Our recent cross-institutional work asks: Does the available evidence match the current level of attention? 📜 arxiv.org/abs/2412.01946 25912
Reposted by Anselm LevskayaAsh Jogalekar @ashjogalekar.bsky.social · 15/12/2024In early phase drug discovery, biology and assays are make-or-break. I can remember very few programs I worked on which were hampered by the inability to make molecules, but plenty that were hampered by the inability to capture disease biology complexity on a chip or in an enzyme readout. 37913
Reposted by Anselm Levskaya🎃 scarecrowcialism 👻 @swolecialism.bsky.social · 12/12/2024All western US water issues are entirely caused by the cultivation of cash crops in places that they shouldn't be grown. Everything, literally everything else, is a rounding error 22866114
Reposted by Anselm LevskayaShae Mclaughlin @shaemcl.bsky.social · 12/12/2024Here is the model it was trained on! Preprint went up on bioRxiv last week: www.biorxiv.org/content/10.1...biorxiv.orgNucleotide GPT: Sequence-Based Deep Learning Prediction of Nuclear Subcompartment-Associated Genome ArchitectureThe spatial organization of the genome within the nucleus is partially determined by its interactions with distinct nuclear subcompartments, such as the nuclear lamina and nuclear speckles, which play... 011
Reposted by Anselm LevskayaShae Mclaughlin @shaemcl.bsky.social · 12/12/2024I trained a sparse autoencoder on the middle layer residual stream of my genome language model and found human-interpretable latent features that consistently detect specific DNA motifs! 🧵1/8 232
Reposted by Anselm LevskayaDan Roy @roydanroy.bsky.social · 23/11/2024Reviewing is dead. I just read some ICLR reviews. Absolutely junk. No insight. Written by humans, but could have been written by a chatbot or teenager who had read a few dozen reviews. Why do we think this is useful? Only 5% (if that!) of the community has worthwhile opinions. 312589
Anselm Levskaya @levskaya.bsky.social · 20/11/2024BOS, BOE have been amazing, hadn’t realized you did SFMTA as well. Thank you so much for this service over the years! 011
Reposted by Anselm LevskayaBen Recht @beenwrekt.bsky.social · 25/10/2024Following up on Kyle Cranmer's excellent and measured response to my inflammatory Higgs missive. www.argmin.net/p/toward-a-t...argmin.netToward a Transformative Hermeneutics of Standard Model PhysicsA dialogue with Kyle Cranmer 011
Reposted by Anselm LevskayaBen Recht @beenwrekt.bsky.social · 22/10/2024The Higgs Discovery did not take place. www.argmin.net/p/the-higgs-...argmin.netThe Higgs Discovery Did Not Take PlaceSome noncontroversial blogging during midterm break 342
Reposted by Anselm LevskayaJulia Evans @b0rk.jvns.ca · 22/10/2024to celebrate selling out the first print run of How Integers and Floats work, we're giving away 500 PDF copies! use code BUYONEGIVEONE at checkout to get a copy for free As usual this works with the honour system, use the code if $12 USD is a lot of money for you! wizardzines.com/zines/intege... 317754
Reposted by Anselm LevskayaMeredith Whittaker @meredithmeredith.bsky.social · 22/10/2024New Paper! w/@HeidyKhlaaf + @sarahbmyers. We put the narrative on AI risks & nat'l security under a microscope, finding that the focus on hypothetical AI bioweapons is warping policy and ignoring the real & serious harms of current AI use in surveillance, targeting, etc. 1/ 37929
Reposted by Anselm LevskayaDavid Pfau @davidpfau.com · 21/10/2024I'm starting a research nonprofit dedicated to understanding and addressing the potential wellbeing and moral patienthood of the people living on the other side of mirrors, lakes and other reflective surfaces. eleosai.orgeleosai.orgEleos AIEleos AI Research is a nonprofit organization dedicated to understanding and addressing the potential wellbeing and moral patienthood of AI systems. 4171
Reposted by Anselm LevskayaDavid Pfau @davidpfau.com · 23/10/2024My phone just autocorrected "singularitarians" to "singular Italians" and I think this is a massive improvement. 0171
Anselm Levskaya @levskaya.bsky.social · 23/09/2024Yeah but common people can’t use any of that stuff. GenAI is accessible to the public. 000
Anselm Levskaya @levskaya.bsky.social · 15/09/2024(aside: it's great to see ML people beginning to show up here.) 000
Anselm Levskaya @levskaya.bsky.social · 13/09/2024The boosters of AI safety bills like SB1047 claim that open models will enable the production of biological weapons. These claims are delusional. As a synthetic biologist and LLM engineer, I felt compelled to write why for anyone who might care: dreamofmachin.es/machine_prop...dreamofmachin.esThe backers of California’s SB1047 routinely cite AI-enabled bioweapons as a threat justifying the radical regulatory regime that places a locus of liability on general computational models, rather than on particular dangerous applications or criminal acts. 001
