Vilgot Huhn @vilgothuhn.bsky.social · 30/09/2026I guess my ambivalent feeling can most clearly be described as "anti-pro-AI". More than anything, I do not relate to people who feel no sense of dread, or at least resistance to change. But that doesn't mean I think no one should use the useful little guy in the computer. 120
Vilgot Huhn @vilgothuhn.bsky.social · 30/09/2026Idea: machine learning model based on guessthecorrelation.com where human estimates are ground truthguessthecorrelation.comGuess the CorrelationGuess the Correlation. How good are you at guessing correlation coefficients from scatter plots? Test your skills! 010
Vilgot Huhn @vilgothuhn.bsky.social · 30/09/2026Your Fallible Human Eyes Are Not Better At Detecting Correlation Than OLS 001
Vilgot Huhn @vilgothuhn.bsky.social · 29/09/2026The thing about RM-ANOVA is that it seems somehow *more* difficult to use than mixed models while also being all around worse? Am I missing something here? 130
Vilgot Huhn @vilgothuhn.bsky.social · 28/09/2026man I hope the real treasure isnt the friends we made along the way because AI is getting very effective at skipping past the journey 000
Reposted by Vilgot HuhnMagnus Johansson @pgmj.bsky.social · 28/09/2026Our commentary "A Revolution in Search of an Estimand: On the Idionomic Proposal of Hayes et al. (2026)" is now available as a preprint: osf.io/preprints/ps... With @rpsychologist.com, @matti.vuorre.com, and Nils Hentati Isacsson.osf.ioOSF 03118
Vilgot Huhn @vilgothuhn.bsky.social · 27/09/2026Obviously I'm not a very expert LLM user (at least these last few months) because just based on chatting/small tasks Opus 5.5 is like Opus 5. I can't tell the difference. I've seen the demonstrations. Super intimidating. But day to day for me it's the same chatbot as yesterday. 140
Reposted by Vilgot HuhnKristoffer Magnusson @rpsychologist.com · 25/06/2026New interactive blog! "Why Adjusted Regression Coefficients Are Less Descriptive Than They Look" rpsychologist.com/descriptive-... 1520769
Vilgot Huhn @vilgothuhn.bsky.social · 24/09/2026With AI I'm constantly cycling between: 1. This tool is amazing! Claude is such a useful little guy. 2. This is monstrous! The death of art. Death of purpose. Get me of this wild ride. I hate this. 3. The fact that LLMs work this well is just deeply fascinating and bizarre and I need to know more. 131
Vilgot Huhn @vilgothuhn.bsky.social · 23/09/2026Wait, Adam Sandler’s Click (2006) was prophetic? I hate this timeline. 000
Reposted by Vilgot Huhnpaolo @paolopalma.bsky.social · 20/09/2026Working on a new taxonomy for research methods 55615
Vilgot Huhn @vilgothuhn.bsky.social · 20/09/2026They should do an obnoxiousness benchmark because the Opus models have been getting gradually worse on that. It constantly injects extra assumptions about what you mean and then counter those assumptions in an overly verbose way. Anthropomorphising. it comes of as "full of itself" 010
Vilgot Huhn @vilgothuhn.bsky.social · 20/09/2026Preparing a lecture/seminar on behavioral analysis for undergraduates. When I took the course I remember I got quite preoccupied on whether behaviorism as a philosophy/paradigm could offer a complete description of human behavior. Missing the trees for the forest: It's clinically extremely useful. 020
Reposted by Vilgot HuhnMatti Vuorre @matti.vuorre.com · 17/09/2026Papers I would like to explore: - Against SESOI: More arbitrary thresholds = More arbitrary science - Against value-driven science: open, team, or interdisciplinary science doesn't make your study more valid - Abandon "best practices" and explain your reasoning instead 3151
Vilgot Huhn @vilgothuhn.bsky.social · 15/09/2026Trumps AI/EA comments: has reaffirmed for me that "woke" now just mean "anything that includes moral concern about others". MAGA world seems to be caught up in a weird psychodrama where anyone who demands moral action are inherently bad. (bad how? morally?) 131
Reposted by Vilgot HuhnBenjamin @benjaminschneider.ch · 13/09/2026actually ratatouille was the name of the dish, you are thinking of ratatouille’s monster 21069
Vilgot Huhn @vilgothuhn.bsky.social · 10/09/2026I don’t agree with the AI doomers but I also don’t mind them being around. Almost certainly Claude will not kill us but I do sleep a bit more soundly imagining that if It Begins 1000 off-grid lesswrong sleepers will emerge to switch it off. 010
Reposted by Vilgot HuhnSE Gyges @segyges.bsky.social · 06/09/2026this neural network has accidentally been trained to naruto run 1627320
Vilgot Huhn @vilgothuhn.bsky.social · 03/09/2026Happy that someone points out that ”aphantasia” isn’t a newly discovered thing. I swear I’ve seen so many examples where it’s claimed that this was barely discussed before 2015 but I swear my 2013 textbook in cognition talked about Galton researching it a hundred years ago.youtu.beThe Aphantasia LieYouTube video by ARTexplains Science and History 010
Vilgot Huhn @vilgothuhn.bsky.social · 28/08/2026I often pass by a café that has as decoration a big screen TV with an uncanny AI video of a more idyllic Paris café and I feel dizzy trying to count the level of simulacra here 000
Vilgot Huhn @vilgothuhn.bsky.social · 28/08/2026One perk with being a toddler dad is that I finally have an excuse to be this tired. 010
Vilgot Huhn @vilgothuhn.bsky.social · 24/08/2026”The person vs situation debate” is just such a wild broad-strokes way to frame psychology that I kinda love it 130
Vilgot Huhn @vilgothuhn.bsky.social · 20/08/2026It's really cool about those London taxi drivers and violinists but I've come to hate the word "neuroplasticity" because by now my first association is superfluous science sprinkles on the well known fact that "people learn things". 121
Reposted by Vilgot HuhnJay Van Bavel, PhD @jayvanbavel.bsky.social · 19/08/2026Using AI helps people get higher grades and get their work done faster, but they learn less & suffer on actual exams I saw this in my own large Intro Psych class: -the exam average is 85% -after AI, it shot up to 95% -after I made the final in person it went to 68% www.economist.com/graphic-deta... 14391150
Vilgot Huhn @vilgothuhn.bsky.social · 18/08/2026There’s a type of academic elitism where the premise is that only niche academic experts can understand certain subjects. Obviously there’s some truth to that, but imo it sometimes becomes a kind of anti-intellectualism, weirdly. Like it’s suspect to read outside ones area. 040
Reposted by Vilgot Huhnaegisiondure @1v.lc · 13/08/2026this is gonna get me killed with hammers for saying but i like corporate memphis as an artstyle. i always did. i think it's pretty. its professional, has nice varied colors, clean forms, and is more expressive than people give it credit for. and it is just nice to be greeted with on a website. 5347
Vilgot Huhn @vilgothuhn.bsky.social · 13/08/2026By now I've repeatedly come across people on twitter/reddit using technical jargon to signal authority in debates around LLMs which turn out to just be empty gatekeeping. They don't know what these words mean either. Ironically, the thing that revealed this was innocent curious questions to Claude. 130
Vilgot Huhn @vilgothuhn.bsky.social · 10/08/2026Listening to Search Engine about the huggingface incident. I had assumed it was hyped up for headlines but actually this does sound genuinely dangerous. 010
Reposted by Vilgot Huhnpmamtraveller.bsky.social @pmamtraveller.bsky.social · 08/08/2026Bauhaus Milhaus 56103162126
Vilgot Huhn @vilgothuhn.bsky.social · 06/08/2026It’s impossible to hang around on TV-show specific subreddits and not conclude that viewers are harder on women than men. Shiv is supposedly the worst sibling in Succession. Alicent the most vile character in House of The Dragon. Cersei more evil than Joffrey. Etc. 020
Vilgot Huhn @vilgothuhn.bsky.social · 03/08/2026Trying to put the finger on why discussions about LLM consciousness bother me so much and tbh it’s not so much about the substance of the arguments as the feeling that people haven’t really grappled with the consequences of animal consciousness yet (or even fellow human consciousness). 210
Reposted by Vilgot HuhnBrawlcats @brawlcats.bsky.social · 02/08/2026airbnb I’m staying at unintentionally created the toughest punk album cover of 2027 158363876435
Vilgot Huhn @vilgothuhn.bsky.social · 25/07/2026If disgust evolved to help us avoid ingesting things that are harmful why does the idea my food has been poisoned not conjure disgust in me? (Though fear, ofc) I feel like my brain differs between chemical and biological contaminants in food? Weird. Is this just me? 010
Vilgot Huhn @vilgothuhn.bsky.social · 22/07/2026I don’t get this constant loling at ”cope”. Oh someone’s trying to find a perspective on things so they’re not crushed by the weight of their circumstances? Haha! Funny! 000
Vilgot Huhn @vilgothuhn.bsky.social · 19/07/2026Lots of negative things to say about short-form video but we really are in the golden age of comedy skits 010
Reposted by Vilgot HuhnRafe Meager (they/them) @economeager.bsky.social · 17/07/2026Hmm. Europe, Canada and Australia can once again win from this if we avoid being just absolutely bottom of the barrel stupid. 6448
Vilgot Huhn @vilgothuhn.bsky.social · 16/07/2026Love is loving someone’s bad poetry because it’s theirs and you love them 011
Vilgot Huhn @vilgothuhn.bsky.social · 11/07/2026There's something life affirming in doing things that "life is too short for". 030
Reposted by Vilgot HuhnTimothy O'Leary @timothyoleary.bsky.social · 09/07/2026Dear peer reviewers everywhere: sometimes we feel pressure to find issues with a submission as a demonstration of diligence. But we have to recognise this is a mistake: picking random and sometimes incoherent holes in research add confusion, generates unnecessary work and encourages defensive 48214
Reposted by Vilgot HuhnEris @isolyth.dev · 09/07/2026I get such a feeling of visceral disgust from this, like this is a thing that isn't human in the slightest learning how to use our bodyplan and it just looks so fucking wrong to me. It's very cool, probably some of the best hands rn? Figure's are also very good but are less uncanny to meyoutube.comNEO’s HandsYouTube video by 1X 4282
Vilgot Huhn @vilgothuhn.bsky.social · 07/07/2026I feel like I'm missing something because this seems both quite far from GNW and also didn't people already assume something like this was going on based on what models are able to do? (just based of the blogpost) 010
Reposted by Vilgot HuhnTommy Blanchard @tommyblanchard.bsky.social · 23/06/2026I wrote about blindsight. I'd always accepted the standard narrative that blindsight is a dramatic dissociation between awareness and sight. But looking at the details, I'm unsure it tells us any more about consciousness than the more typical experience of barely being able to see in a dark roomcognitivewonderland.substack.comBlinded by BlindsightCan you see without being aware? The condition "Blindsight" is often reported to show we can, but it isn't as simple as it's often portrayed 43013
Reposted by Vilgot HuhnBerna Devezer @devezer.bsky.social · 23/06/2026just thought i'd share the observation that dipping your feet in the ocean is very much like touching grass, but even better 26111
Vilgot Huhn @vilgothuhn.bsky.social · 21/06/2026Spent my childfree vacation day hungover after midsommar randomly binging The Amazing Digital Circus. It was cute. I feel I relate to it indirectly through some inner teenager. Like I could definitely see myself obsessing over this show at 15. 120