Martin Johnsson @mrtnj.bsky.social · 06/10/2026Mitt främsta önskemål, om vi ska ha en debatt om gener och heritabilitet nu, är att vi använder de svenska termerna "anlag" och "arvbarhet" istället. 000
Martin Johnsson @mrtnj.bsky.social · 02/10/2026I'm reminded of the fallacious ethical argument summarised as "I'm only the second worse chainsaw murderer in Texas". 000
Martin Johnsson @mrtnj.bsky.social · 26/09/2026De traditionella religiösa protesterna vid Uppsala pride uteblev, kanske för att de inte tål regn. 000
Reposted by Martin JohnssonDanielle Navarro @djnavarro.net · 18/07/2026And so after a very brief hiatus, version 0.7.0 of "learning statistics with R" exists. Full rebuild in quarto, stylistic fixes, "epilogues from 2026" to comment on how the world changed since original publication, and as an added bonus, no longer misgenders the author learningstatisticswithr.comlearningstatisticswithr.comLearning Statistics with R 217146
Reposted by Martin JohnssonJulie S. Lalonde @julieslalonde.bsky.social · 25/09/2026Y'all know Bean Soup Theory? How someone can post a recipe for bean soup and the replies will be "BUT WHAT IF I DON'T LIKE BEANS?!" That is sometimes annoying trolling but it's predominantly people failing to understand how to read for context, understand intended audience, etc. 92228169
Martin Johnsson @mrtnj.bsky.social · 25/09/2026Writing a grant application with little need for equipment, a lot of need for computational infrastructure, and honestly it should probably also say "we'll request every important book on quantitative genetics on interlibrary loan". 000
Martin Johnsson @mrtnj.bsky.social · 25/09/2026I had a lot of sympathies for "skepticism" around 2010 and a lot of sympathies for "effective altruism" around perhaps 2020. I feel, incredulously, that the people involved should have the ability to formulate cogent criticisms of the AI and transhumanist nonsense I now associate with these terms. 030
Reposted by Martin JohnssonMattan S. Ben-Shachar @mattansb.msbstats.info · 24/09/2026This quote by haunts me. From Wilcox's "Introduction to Robust Estimation and Hypothesis Testing" #stats 310422
Martin Johnsson @mrtnj.bsky.social · 24/09/2026Agreed. We're an all-Canvas shop, and compared to anything else I've seen, it's really quite nice. :) 010
Reposted by Martin Johnssonhoffsbeefs.bsky.social @hoffsbeefs.bsky.social · 21/09/2026If these systems become recursive, as breeding is already setup to do; we could see self reinforcing programs where our agents go rouge and make a cow so big and beautiful it conquers the world. I'm asking congress to intervene immediately and i hope you'll do the same. 221
Reposted by Martin Johnssonhoffsbeefs.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 Martin JohnssonPeter Morrell @peterlmorrell.bsky.social · 19/08/2026QTL mapping academic.oup.com/genetics/art... Whether you are new to QTL mapping or have years of experience, you’re sure to learn something new and interesting from Shizhong Xu’s Primer. Don’t fret over the equations; there is a terrific narrative here about the past and future of QTL mapping!academic.oup.comQTL mappingAbstract. Most biologically relevant traits are quantitative, controlled by multiple genes, and can be influenced by the environment. The multiple genes, c 01810
Martin Johnsson @mrtnj.bsky.social · 21/09/2026HBIO-kollegan Erik Pelve har testkört en kul evolutionsaktivitet med post-it-lappar i fikarummet. Den som vill se den in action får gå på Scifest Uppsala! Som gammal domesticeringsforskare gillade jag särskilt den här lappen. 010
Reposted by Martin JohnssonLotta Ilona Häyrynen @lottailonah.bsky.social · 16/09/2026"Stämmer det att nya medborgare avgjorde valet" undrar SVT efter SD et als spinn. Helt bisarrt att se public service aktivt välja att stärka och bekräfta högerextrema och antidemokratiska ställningstaganden. Kvinnor "avgjorde" också valet. Skåningar avgjorde där sd tappat. Osv 19444148
Martin Johnsson @mrtnj.bsky.social · 14/09/2026Genuinely fun part of research collaborations--to see how colleagues in other fields organise data and think about data gathering and analysis. 000
Reposted by Martin JohnssonM.J. Crockett @mjcrockett.bsky.social · 13/09/2026Who is advising students to automate these messages? Why the mass delusion that professors want to read these? What makes me saddest is the erosion of trust. Even the students who take the time to write will now meet readers who are skeptical they wrote the words themselves. Everyone loses. 1039741
Reposted by Martin JohnssonJason Cordova @jasoncordova.bsky.social · 13/09/2026AI is awful for a thousand reasons, but it’s not going to cause the extinction of the human race. That is *marketing hype* from people who need to keep raising vast sums of money. They have to make it seem like it’s this god-like entity unbound by the normal laws of supply and demand economics. 3668
Martin Johnsson @mrtnj.bsky.social · 12/09/2026Theory is immensely useful to get to those rare occasions, but theory is hard work. The idea that you can get to knowledge by bouncing ideas and looking for patterns in text seemed silly to me. Now, how does an LLM encode "knowledge"? 000
Martin Johnsson @mrtnj.bsky.social · 12/09/2026If you do empirical natural science (and are honest about it), you are painfully aware that your ideas mean nothing to nature. Those rare occasions when you're able to make a specific prediction that is actually borne out by observation, it's a matter of celebration. 100
Martin Johnsson @mrtnj.bsky.social · 12/09/2026There's a segment of note taking enthusiasts interested in index card systems and digital equivalents, connecting notes and graphs of ideas. I objected to a view of knowledge that seemed implicit in their writing, that knowledge is relations between ideas, expressed in words. 101
Reposted by Martin JohnssonNick Brown @steamtraen.eu · 12/09/2026Social psychologists have some serious explaining to do. I don't recall a single article saying "around a quarter of the population of pretty much every Western country are just full-on racists", and yet here is the proof, every fucking month somewhere. www.bbc.com/news/article...bbc.comSweden election becomes tight race, as hard right eyes role in governmentPrime Minister Ulf Kristersson has promised cabinet jobs for the Sweden Democrats, but the rival left-wing bloc has a narrow lead in the polls. 141
Reposted by Martin JohnssonErik Angner @erikangner.com · 12/09/2026Separately: The PM is referencing the HC Andersen story about the Princess and the Pea to imply Wikforss is overly sensitive in a womanly way. But the point of the story is that the princess is unusually alert and perceptive. Misogyny and ignorance, what a winning combination. 19717
Reposted by Martin JohnssonErik Angner @erikangner.com · 12/09/2026Even apart from the obvious misogyny, the attack fits a pattern. The current "Blue–Yellow" nationalist–populist governing coalition stands out for its continual efforts to dismiss and delegitimize journalists, academics, and artists. 18619
Reposted by Martin JohnssonErik Angner @erikangner.com · 12/09/2026The Prime Minister of Sweden 🇸🇪 Ulf Kristersson just dismissed my colleague Åsa Wikforss – Professor of Theoretical Philosophy at Stockholm University, Member of the Swedish Academy and the Royal Academy of the Sciences, and celebrated author of numerous books – as "princess" 🧵 3219127
Reposted by Martin JohnssonKolbjörn Guwallius @kolbjorn.se · 12/09/2026Sverige från utlandet, ord och inga visor: extremhögern kan få plats i regeringen. ”country’s reputation as global champion of human rights said to be at stake” 05027
Reposted by Martin JohnssonNick Holmes @thehandlab.bsky.social · 09/09/2026f = 0.25 *every* power analysis I've read this week uses f = 0.25 It is truly the universal effect size #stats 🧪 🤦♂️ 261
Martin Johnsson @mrtnj.bsky.social · 10/09/2026Yeah, yeah, you can destroy the world and help mathematicians scoop each other, but can you turn a profit? 000
Martin Johnsson @mrtnj.bsky.social · 09/09/2026Like Brecht wrote about the virtues of the Kanzler, if your diligent unpaid work for the community is actually harmful, perhaps it would be better if you were lazy and worked less. 010
Martin Johnsson @mrtnj.bsky.social · 09/09/2026Apropå Pisa-undersökningen: läser gymnasiepojkar ens Nicanor Teratologen längre? :( 000
Martin Johnsson @mrtnj.bsky.social · 07/09/2026Like a true believer of Emacs or Vi of old, I'm asking myself why I'd ever want to leave RStudio. 010
Martin Johnsson @mrtnj.bsky.social · 07/09/2026Thank you Quarto for helping me achieve my long-standing dream of being able to do my manuscript writing in RStudio. 110
Martin Johnsson @mrtnj.bsky.social · 06/09/2026"Framework integrating"? You mean, "data analysis comparing". Much better. 010
Martin Johnsson @mrtnj.bsky.social · 04/09/2026Vårt land, så rasistiskt att det är rasistiskt mot engelsmän. 000
Reposted by Martin JohnssonPer Engzell @pengzell.bsky.social · 28/11/2025A thread of directed acyclic graphs (DAGs) that look like record covers... because that's EXACTLY what the world needs 1. Huey Lewis and the News: link.springer.com/chapter/10.1... 11319107
Martin Johnsson @mrtnj.bsky.social · 04/09/2026Give me an honest attempt to quantify something that matters to an important research question. 010
Martin Johnsson @mrtnj.bsky.social · 04/09/2026I guess what I'm saying is I've been in science long enough to not like quirky, amusing, clever stuff anymore. 110
Martin Johnsson @mrtnj.bsky.social · 04/09/2026I love this operationalisation of antisocial behaviour: someone gives you candy, and you take it. 000
Martin Johnsson @mrtnj.bsky.social · 04/09/2026Quite possible. I've not read the authors' whole oeuvre. 000
Martin Johnsson @mrtnj.bsky.social · 04/09/2026Same. I clicked on the IgNobel's website (well, their link didn't work, if memory serves, so I guess I googled from their website) and saw the retraction. Hard to tell from the outside how serious something like this is, but I think the IgNobels should at least acknowledge when they promote it. 100
Reposted by Martin Johnssonsamuel mehr @mehr.nz · 20/02/2026the new paper on bouba/kiki in chicks is utterly compelling canonical, elegant method from comparative cogsci & its partner in developmental science, ultra-simple design, ultra-clear effects, no need for fancy analyses, machine learning, or AI it appeared in an appropriately badass venue (Science) 11430162
Martin Johnsson @mrtnj.bsky.social · 04/09/2026The IgNobel winners include some typical flimsy experimental psychology (ask people hypothetical scenarios and correlate with SES, judge social class by evaluating the niceness of passing cars) and an animal study retracted for potential research animal ethics violations. 6115
Reposted by Martin JohnssonRichard McElreath 🐈⬛ @rmcelreath.bsky.social · 02/09/2026Last year, I publicly complained multiple times (see e.g. elevanth.org/blog/2025/07...) that the open access movement had been captured by publishers, and we are now worse off than before open access. Well people got mad at me for saying it. But I think things are still getting worse. 616466
Martin Johnsson @mrtnj.bsky.social · 01/09/2026"Not just this, but that" is a rhetorical argument with no substantial force, since the first part is a negation of whatever that has no bearing on what follows, but it seems persuasive on the surface. 000
Martin Johnsson @mrtnj.bsky.social · 01/09/2026The fact that LinkedIn posts look as they do suggests that prompting for business and career jargon leads to the same. 100
Martin Johnsson @mrtnj.bsky.social · 01/09/2026I suppose everyone recognises the structure "not only this, but that" as typical of LLM-generated text by now. If I prompt ChatGTP with questions about science, it gives me that in most every paragraph. I wonder what its frequency says about that kind of argument. 100
Martin Johnsson @mrtnj.bsky.social · 27/08/2026You're proposing to waste good research funding and fuel racist fantasies for what -- a poor genomic predictor of a clumsily operationalised behavioural trait that has the accuracy of the square root of zilch? 100
Martin Johnsson @mrtnj.bsky.social · 27/08/2026By the way, that's why more normal geneticists don't want you to research the genetics of racial difference and the heritability of being convicted of sex crimes etc. It's not because we can't handle the truth or whatever; it's because the research would have zero implications for anything. 111