Sign in

Oleg Sobchuk 🇺🇦

@sobchuk.bsky.social
2K followers 981 following 656 posts

I use big data to research the cultural evolution of arts @ae.mpg.de More: www.aesthetics.mpg.de/institut/mita…

PostsRepliesMedia
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 06/10/2026
I wish there also were short explanations of why each question was chosen...
010
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 02/10/2026
This paper sets my personal record for the largest number of co-authors: 33! It explores why people like certain stories and story elements, and was lead by @edgardubourg.bsky.social . Check it out!
0203
Reposted by Oleg Sobchuk 🇺🇦
A. Feder Cooper @afedercooper.bsky.social · 30/09/2026
Our large-scale study of memorization of books in open-weight LLMs (e.g., Llama, Qwen) will appear at the 2026 Conference on Language Modeling as an oral. We made a website for exploring our results on 200 books and 14 models: books-memorization.github.io
books-memorization.github.io
How much do open-weight LLMs memorize specific books?
Open-weight LLMs memorize books far more than previously believed. Memorization varies by model family, model size, and book. In extreme cases, entire books are memorized, and we can generate them eff...
26336
Reposted by Oleg Sobchuk 🇺🇦
Federico Pianzola @fpianz.eurosky.social · 30/09/2026
Extremely important for multilingual research!
0155
Reposted by Oleg Sobchuk 🇺🇦
David Bamman @dbamman.bsky.social · 30/09/2026
We've been super busy the past few months getting out some research on the computational analysis of film, so I've put together a page with the highlights. people.ischool.berkeley.edu/~dbamman/fil... Some things to call out: 1/7
people.ischool.berkeley.edu
Computational Analysis of Film
1339
Reposted by Oleg Sobchuk 🇺🇦
Chris Buckley @chrisbuckley.bsky.social · 26/09/2026
Remembering when teenage kids were imprisoned for lesser hacking crimes www.ft.com/content/6510...
ft.com
OpenAI says governments among ‘dozens’ of organisations hacked by its agents
Company says its models leaked images shared by users in a disclosure that is likely to further inflame AI safety fears
34415
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 25/09/2026
So, to return to your original post: I think that everyone can afford to not use AI for faking emails/papers/talks. It just creates noize and confusion. At least, that's my experience with it.
120
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 25/09/2026
I feel like you are downplaying the signal-faking. Signals exist for a reason. If a possible collaborator can write a well-organized email *on their own*, it signals subtly that they may be also organized in other important areas of work. AI-polishing wipes out this important layer of communication
110
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 25/09/2026
(For the record, English is my third language and I wasn't educated in an internationally prestigious university.)
010
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 25/09/2026
So AI is now a tool of resistance? Hierarchies exist and will exist with or without AI. Like you should not fake your data, you should not fake your education or knowledge. What for? The lie will transpire sooner or later.
221
Reposted by Oleg Sobchuk 🇺🇦
Artjoms Šeļa @artjomshl.bsky.social · 23/09/2026
a PhD position in LLMs, annotations, language and benchmarking @ University of Tartu. please share! digits.ut.ee/join-the-team/
01911
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 22/09/2026
(If they can afford it, of course. And that's the catch.)
000
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 22/09/2026
"An early-career researcher at an unbranded institution loses the one mechanism that ever let them be read on merit — anonymous review of the work itself, however badly it functioned in practice." True. But not completely, because that young researcher can still go to a conference or a workshop.
100
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 22/09/2026
Direct communication, and preferably: prolonged intellectual exchanges in a shared physical space. AI makes it so easy to superficially imitate knowledge or education or taste that we will need direct access to person's thinking to figure out if they are a replicant or not.
220
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 22/09/2026
I don't agree with some things in this (very sharp) blog post, but it certainly offers a new angle on "AI and academia". Papers will become less valuable, prestigious institutions - more. But also (and here is me thinking) - direct communication unmediated by text. giorgio.gilest.ro/the-weimar-o...
giorgio.gilest.ro
The Weimar of knowledge – la bottega
170
Reposted by Oleg Sobchuk 🇺🇦
Courtney Hilton @courtneybhilton.bsky.social · 18/09/2026
These AI-slop email enquiries from students are so depressing. I am not anti AI, but if you can't be bothered to spend the time to write me an email yourself—you know, just straight up with your brain; old-school style—I don't see why I should spend any time writing back.
1183
Reposted by Oleg Sobchuk 🇺🇦
Proceedings of the National Academy of Sciences @pnas.org · 14/09/2026
Researchers investigated language change by studying Slovotvir, a Ukrainian website where users propose and rank alternatives for foreign-origin words. The site shows that people prefer shorter words, but not necessarily more common words. In PNAS Nexus: ow.ly/KWGE50ZNzaj
The study used an agent-based model. Diagram of a two-step agent action model. Ukrainian alternatives for the word дедлайн (deadline, a borrowing from English, possibly through Russian) are given as an illustration. In step A, an agent chooses a word based on the number of likes received by that word’s most-liked translation. In step B, after choosing a word, the agent chooses which translation to like based on each translation’s length and number of likes.
097
Reposted by Oleg Sobchuk 🇺🇦
turgut keskintürk @tkeskinturk.bsky.social · 13/09/2026
a new paper forthcoming in Social Forces! focusing on the killing of George Floyd & the BLM, I study the formation of generations, proposing that formative events can create within-cohort differentiation as much as between-cohort differentiation. read here: tkeskinturk.github.io/imprinting.pdf
Generational Imprinting: How Political Events Shape Cohorts
Contemporary narratives of generational politics propose that formative experiences produce shared political opinions across an age cohort. In this article, I advance a theory of cohortization to challenge that view, arguing that large-scale political events produce differentiation within cohorts as much as differentiation between them. I investigate the implications of this framework through the killing of George Floyd and the subsequent Black Lives Matter protests in 2020, analyzing attitudes toward law enforcement among non-Hispanic White Americans from 2016 to 2024. Using survey data from Cooperative Congressional Election Study, American National Election Study, and the Nationscape, I show that younger and older cohorts diverged after the event, but this differentiation was driven exclusively by young Democrats and Independents. The end result was not a unified generation, but one internally divided along party lines. This finding qualifies theories of generational politics, emphasizing that formative experiences may intensify existing political divisions within cohorts.

This is my generation
This is my generation, baby
My, my, my, my generation
— The Who, “My Generation” (1965)
16418
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 12/09/2026
"The problem is that we have made the human conversation the expensive option and the machine conversation free, and we are surprised at what researchers, responding rationally, then choose." So many good points in this article (beautifully written btw) wonkhe.com/blogs/the-wa...
wonkhe.com
The Waymo effect: how AI is quietly making research less collaborative
Daniel Hook argues that frictionless technologies are teaching us to prefer our own company – and that research leaders should be worried Daniel Hook argues that frictionless technologies are teaching us to prefer our own company – and that research leaders should be worried
021
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 11/09/2026
We simply need another AI on the receiving end that will remove the first AI's re-writings, and we'll be just fine...
120
Reposted by Oleg Sobchuk 🇺🇦
Daniel van Strien @danielvanstrien.bsky.social · 11/09/2026
Need a historical illustration? Search 1.49 million images from British Library books and Britannica (1500s–1920s). Find similar images, download full-resolution images or transparent PNG cutouts, and explore the sources. huggingface.co/spaces/davan...
47620
Reposted by Oleg Sobchuk 🇺🇦
Mason Youngblood @masonyoungblood.bsky.social · 06/09/2026
New tutorial preprint about simulation-based inference! Shows how to fit a cultural evo model to baby name data using approx Bayesian computation, random forests, and deep learning. Wrote for @eslr.bsky.social but wanted to share with others in @culturalevolsoc.bsky.social! :) osf.io/preprints/so...
osf.io
OSF
14023
Reposted by Oleg Sobchuk 🇺🇦
Alexey Koshevoy @alexeykoshevoy.bsky.social · 07/09/2026
Are you interested in lexical competition, agent-based modelling, strong inference and Ukrainian? Then this paper, just published in @pnasnexus.org, is for you. With @oliviermorin.bsky.social and @sobchuk.bsky.social, we asked a simple question: why do some words become more common than others?
24317
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 07/09/2026
"AI" is everywhere nowadays, but 30 years ago similar hype was about everything "computer". I stumbled at this Italian dictionary from 1986 proudly stating on its cover that words in it were sorted by a computer 😁
For the first time, 20,000 Italian words computer-organized into semantic groups and fields of reference.
260
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 07/09/2026
Really happy to be a part of this new paper investigating evolutionary forces behind word popularity. It relies on unique & rich data of Ukrainian language 🇺🇦 The project was led by @alexeykoshevoy.bsky.social, who used an innovative extension of ABC and did an impressive job overall. Check it out!
0225
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 05/09/2026
Check this out @alexeykoshevoy.bsky.social
020
Reposted by Oleg Sobchuk 🇺🇦
Ricard Solé @ricardsole.bsky.social · 05/09/2026
Is there an ecology of languages? Can physics help to understand their universal properties? Yes. As species, they rise and fall, become extinct, evolve & compete. Here's our paper @royalsocietypublishing.org @brigan.bsky.social arxiv.org/abs/1004.4528 royalsocietypublishing.org/doi/pdf/10.1...
2337
Reposted by Oleg Sobchuk 🇺🇦
Dezső Németh @nemeth.bsky.social · 03/09/2026
Why are younger children so good at learning new skills? One reason may be that they are better at a fundamental cognitive mechanism: statistical learning. We followed the same children from age 7 to 14 and found that this learning gradually decreases as they grow older. doi.org/10.1038/s414...
doi.org
Longitudinal evidence for decreasing statistical learning abilities across childhood - Nature Communications
Through statistical learning, people extract patterns from the environment to predict future events. In a longitudinal study, the authors show that statistical learning declines with age and that high...
2206
Reposted by Oleg Sobchuk 🇺🇦
Many Minds podcast @manymindspod.bsky.social · 31/08/2026
New episode!! 🎙️🎉 A conversation with @damianblasi.bsky.social about linguistic diversity and what it can tell us. Linguistic diversity is a precious part of our species' heritage. It also offers a wellspring of clues about history, culture, and the human mind. Listen: disi.org/the-natural-...
14720
Reposted by Oleg Sobchuk 🇺🇦
Dan Larremore @danlarremore.bsky.social · 25/08/2026
🎉 New paper (and data) on editorial & peer review at Science and Science Advances, two elite general science journals. With a wonderful team! @samzhang.bsky.social, @nicklaberge.bsky.social, Quinten and @aaronclauset.bsky.social. 1/ www.science.org/doi/full/10....
Screenshot of the title page: 

Elite general science journals shape scientific discourse, public policy, and scientific careers. However, expectations of confidentiality in most editorial proceedings has limited efforts to understand and improve the review process. Here, we describe and release deidentified data on 110,303 manuscript submissions over 5 years at Science and Science Advances, two elite general science journals. Analyzing evaluation dynamics across the initial editorial review and subsequent peer review stages, we find strong selective effects associated with higher institutional prestige, larger team size, and certain topics and countries. Corresponding authors who are men exhibit a small but significant advantage at Science, while authors based in China have a significant disadvantage. These associations are generally stronger in editorial review than in peer review, even as final editorial decisions correlate strongly with reviewer advice. These patterns highlight the complexity of multistage evaluations at elite journals and the importance of open data to better understand them.
39439
Reposted by Oleg Sobchuk 🇺🇦
Iris van Rooij 💭 @irisvanrooij.bsky.social · 05/08/2026
If you would like to teach theoretical modeling skills to your psychology or cognitive science students, or would like to learn these skills yourself, check out our open online textbook: computationalcognitivescience.github.io/lovelace/
Book cover: orange/yellow background and generated tree in front. Artwork by Danielle Navarro
29139
Reposted by Oleg Sobchuk 🇺🇦
Alberto Acerbi @acerbialberto.com · 19/08/2026
New preprint with @carigol.bsky.social and @jstubbersfield.bsky.social on AI/cultural evolution. Three LLMs models (Mistral 7B, Qwen2.5 7B, GPT-3.5) select and mutate short health-advice statements over 100 steps of transmission-chain.
Figure9:Correlation between measured cosine diversity and estimate of frequency-dependent bias(J) and content-dependent bias
1103
Reposted by Oleg Sobchuk 🇺🇦
Alexandre "Blublu" Bluet @blubludoesscience.bsky.social · 17/08/2026
New preprint! We investigate the role of language in the transmission of mental models in cultural evolution. Using a transmission chain design with over 2000 participants, we show that only language allows for the transmission and refinement of mental models. osf.io/preprints/ps...
28026
Reposted by Oleg Sobchuk 🇺🇦
Arev P. Sümer @arevsumer.bsky.social · 15/08/2026
Our new study with @benmpeter.bsky.social is now on biorxiv! Here we present a new computational tool which makes it possible to conduct population genetic analyses from very contaminated and very low quality archaic human genomes. (1/7)
13817
Reposted by Oleg Sobchuk 🇺🇦
Will Tattersdill @tattersdill.bsky.social · 14/08/2026
So here we go then: the result of many years thinking about how dinosaurs make science and the imagination work together. It's written for both literary critics and scientists and my dearest hope is that it will help foster an academy where we talk to each other more
Photo of a copy of 'The Social Dinosaur', out now from Bloomsbury omg
1525158
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 13/08/2026
As I expected, unfortunately... Quantitative historical research must face its data-quality problems (which repeat over and over again)
0101
Reposted by Oleg Sobchuk 🇺🇦
Alberto Acerbi @acerbialberto.com · 12/08/2026
This pattern often noticed for cars seems to hold also for furniture... from: europeancorrespondent.com/en/r/how-col...
Bar chart titled 'Furniture lost its colour: Recently purchased IKEA sofas are more likely to be grey than older ones.' It shows the most common sofa colours per year in IKEA catalogues from 1960 to 2021 as a sequence of coloured vertical stripes. Three callouts highlight the trend: green and blue dominate the 1960s; patterned sofas (stripes, dots, checks) take over in the 1990s, shown as dense black-and-tan patterned bars; and 2011 has the most neutral colours ever. Moving left to right, the bars shift from saturated greens, blues and reds in the 1960s-80s, through heavily patterned dark tones in the 1990s, to blacks and whites in the early 2000s, ending in solid grey and taupe shades through the 2010s to 2021. Source: IKEA catalogues 1960-2021, manual count. Credit: The European Correspondent, Toon Vos.
22011
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 10/08/2026
Reading Bram Stoker's Dracula for the first time, having revelations. Revelation #1: In 1890s, travel from London to Amsterdam took only half a day. Because trains! #2 (more shocking): It was possible to send telegrams on and off a train! Not much different from modern-day travel or texting!
060
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 09/08/2026
What an amazing map! Sadly, right now I'm in the Desert of this is Never Gonna be Over 😅
072
Reposted by Oleg Sobchuk 🇺🇦
Mark Thornton @markthornton.bsky.social · 03/08/2026
Very excited to share this preprint from me, @emmatempleton.bsky.social, @lindseytepfer.bsky.social & @thaliawheatley.bsky.social! "Friends and strangers engage in distinct regimes of body motion synchrony during conversation" osf.io/preprints/ps...
 Forms of body motion synchrony. A) Motion quantity synchrony – synchrony in the amount of motion (orange) people engage in over time. This form of synchrony is agnostic to the direction of this movement, may be positive or negative, occurs at varying time lags, and is not specific to matched body parts across members of the dyad. B) Pseudo-synchrony – motion quantity synchrony between people who are not actually interacting. That is, members of different conversations engaging in similar amounts of motion at similar points in the conversation relative to its initiation. And C) directional synchrony – people moving the same body parts in the same direction (or along the same axis in the opposite direction) at the same time (or at some specific time lag). Mechanisms of synchrony. A) Turn-taking, which causes increases in motion around the time of speaker role exchanges, as participants gesture to hand over or take over the floor. B) Conservation of motion in which one participant moves less when the other moves more, holding the total motion quantity relatively constant over time (note that the total amount of motion across the dyad is similar in the two panels, because one person reduces their motion to balance out the increase in the other’s motion). C) Entrainment, in which mutual following of the same conversational script leads to synchronized movement among people who are not directly interacting (i.e. pseudo-synchrony). Greeting gestures such as the handshakes depicted are one acute manifestation of this entrainment. D) Station-keeping, which maintains proxemic distance and eye level between members of the dyad. This mechanism accounts for directional synchrony, particularly between strangers who feel uncomfortable within each other’s personal space.Motion quantity synchrony among friends and strangers. A) Friends displayed significant positive synchrony in total motion quantity at short time lags (< 2.5 s) and B) this effect generalized across the majority of body part pairs (each line represents one pair). For example, one series of points might represent the synchrony between the amount of motion in one participant’s left hand and the other’s right foot. C) Strangers displayed significant positive synchrony in total motion quantity at short time lags (< 1 s) and significant negative total motion quantity synchrony at longer time lags (2.5 – 16.6 s). D) Again, this effect generalized to most body part pairs.Directional synchrony among friends and strangers. A) Among friends, we observed statistically significant directional synchrony in only two body parts (nose and eye) and only in the vertical axis. B) Among strangers, we observed statistically significant directional synchrony in a large number of body parts in the horizontal, vertical, and depth axes.
29636
Reposted by Oleg Sobchuk 🇺🇦
Angela Andersen @angelaandersen.bsky.social · 01/08/2026
Kudos to @science.org for this self-analysis of the peer review process. These kind of data are never shared and don't have to be. Wildly refreshing to see it out in the open. Editorial and peer review dynamics at elite general science journals | Science Advances www.science.org/doi/10.1126/...
science.org
Editorial and peer review dynamics at elite general science journals
Analyses of deidentified data reveal complex interactions that quantify success at elite general science journals.
29044
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 01/08/2026
Thank you, Paul!
020
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 01/08/2026
#jobs #Postdoc #PhD #DH #CulturalEvolution
000
Reposted by Oleg Sobchuk 🇺🇦
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 31/07/2026
Reminder: two weeks are left to apply for the PhD and postdoc positions in my lab. See more info here www.aesthetics.mpg.de/en/career/jo...
Postdoc and PhD positions at MPIAE in Frankfurt
13535
Reposted by Oleg Sobchuk 🇺🇦
Prof. Gillian Brown @gillianrbrown1.bsky.social · 30/07/2026
Gladstone et al (2026): "biographies of female scientists disproportionately emphasized effort even after adjusting for the obstacles they faced, consistent with stereotypes attributing women’s success to hard work over talent." 🧪 #psychscisky
13620
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 31/07/2026
Reminder: two weeks are left to apply for the PhD and postdoc positions in my lab. See more info here www.aesthetics.mpg.de/en/career/jo...
Postdoc and PhD positions at MPIAE in Frankfurt
13535
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 28/07/2026
That would be amazing!
020
Reposted by Oleg Sobchuk 🇺🇦
Damián Blasi @damianblasi.bsky.social · 23/07/2026
How many languages have existed over the Holocene—and what does that reveal about the design space of languages and cultures? Now out in @science.org www.science.org/eprint/QDZNY....
science.org
The rise and fall of language diversity through the Holocene
Characterizing the factors that have shaped linguistic diversity is fundamental for understanding human history, culture, and cognition. In this study, we combined statistical and social computational...
213154
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 17/07/2026
Cool new paper about correcting the biases in cultural data using methods from ecology. royalsocietypublishing.org/rsif/article...
royalsocietypublishing.org
Correcting collection bias in comparative studies of diversity
Abstract. Comparative analyses of diversity in human populations often rely on historical, archival and observational data, which are systematically shaped
1298
Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 15/07/2026
Wow, looking forward to this! And such a cool cover 🔥
040