Olivia Guest · Ολίβια Γκεστ
@olivia.science
👁️ https://olivia.science 🍃 https://metatheory.space associate professor of computational cognitive science · she/they · cypriot/kıbrıslı/κυπραία · σὺν Ἀθηνᾷ καὶ χεῖρα κίνει
My latest pre-print is out: "Pareidolic Illusions of Meaning: ChatGPT, Pseudolaw and the Triumph of Form over Substance" I really enjoyed this one, where I juxtapose LLMs & pseudolaw to argue both create illusions of meaning that trick users. papers.ssrn.com/sol3/papers....
LLMs don't actually understand meaning or relationships bc they don't need to. The fact they can answer you with seemly relevant text to the prompt you gave them doesn't imply actual understanding. I know you need to ignore this to keep humanizing LLMs, but you're deluding yourself.
Paper recycling. I did that with my HPs long ago and one of Emoto's water non sense somebody had donated to our grad school book sale. Tip: if hard cover, you need to remove the binding before dumping them in the recycling box, that goes in the regular garbage.
"All this rests upon social psychologists treating their world of variables as more important – even more real – than the participants in their laboratory experiments. The actions of the participants are merely a means for understanding the effect of variable upon variable." (1/2)
Marc Scully@marcscully.bsky.social · 12mo agoEvery so often when I read things about people using AI ‘participants’ in psychology, I think “that can’t be just what it sounds like, though. There must be some more to this; some complex epistemology I’m missing”. But no, it seemingly is really is just “Whee!!! N= go brrrr!!!”
FWIW I don’t think that AI is an existential threat, nor on track to become one. I do however think that the leaders of the biggest AI companies, AI’s strongest boosters, and some of the folks who are most adept at using it to exploit others are actively hastening along OTHER existential threats.
Yes... Besides some of the most credentialised people are boosters because they are unhinged and speaking outside their lane and not following scholarly practice like conflict of interest declaration.
Olivia Guest · Ολίβια Γκεστ@olivia.science · 2w agoI keep hearing the argument that the 3 big pro AI academics who are usually doomers are right BECAUSE they're experts. What's important about experts is we declare our *perceived* conflicts of interest. How many AI company shares do they own? Because employees are often paid in shares in any case.
> Something seems to be lost on my peers today: it’s still easy to not use AI. [...] Eschewing AI is one thing that we can actually do to live out ethics that affirm values of human and environmental rights. It’s almost a gift! Just use a computer the same way you did three years ago!
Baldur Bjarnason@baldurbjarnason.com · 2mo ago“Potentially Useful | Hating AI in 2026” https://www.eamoncaddigan.net/posts/ai-in-2026/ > We can point out all of the harms being done by the technology right now, and we can speak to the additional harms that will be done if things continue like this, but our concerns are spurned or ignored.
I spent some time in my Applied Social Psychology tutorial this week explaining the turn towards ‘silicon sampling’ and there were genuine gasps in the room. Undergrads seem to instinctively grasp the issues with this in a way some more senior academics don’t.
Olivia Guest · Ολίβια Γκεστ@olivia.science · 12mo agoNow, we tackle in turn the issues we outlined in the abstract: First, the displacement of human participants, tracing the history of technologies to recruit real people, we end up forgetting the whole point: gathering — not fabricating! — human data. More at section 3: doi.org/10.31234/osf... 5/
✔️ AI is cognition Ban it then it harms everyone ❌ No, of course not it's a tool, silly.
Carlos Diaz Ruiz@carlosdiazruiz.com · 12mo agoThey claim that non-human objects have agency (when convenient). ✔️Vaccines cause autism ✔️Tylenol causes damage to pregnant women ✔️Video games radicalize children But do guns harm people and cause mass shootings? ❌ No, of course not, silly. People kill people; guns are objects and cannot act.
"i think i would like to stop vaping, it's not doing me any favors and i don't care for it. i was happier and healthier before i started" "you're just romanticizing the past! you can't turn back time. vapes are here to stay. can't believe you're against new technology"
something also about the way that any insufficiently new technology gets booted from the definition and relegated to the corner of "living in the past". a well made wooden chair, a wool sweater that lasts decades. to stop doing a new thing or want to do an older thing is painted as regressive. churn
it is in much the same way as they have coopted the concept of "the future" to mean their specific brand, such that if you oppose a very narrow segment of objects and behaviors, you can be deemed "anti-technology" or "against the future" (which is impossible)
i'm too tired to do the topic justice rn but this is related to something i've been thinking about lately: it's not fair that the current social understanding of "technology" has been fully coopted by ghouls. baskets are tech, silviculture is tech, even social arrangements like mentioned below are.
Olivia Guest · Ολίβια Γκεστ@olivia.science · 2d agothe grossest thing that happens in conversations on automation is the idea that more mechanisation is "good automation" and more deep structural changes (like free healthcare at the point of service, or less rampant and vile racism and sexism) is either "not even seen as automation" or somehow bad!
Case in point 🙃 bsky.app/profile/oliv...
Olivia Guest · Ολίβια Γκεστ@olivia.science · 41m agoIf I cooked food and somebody needs to know if there's gluten, I may be best placed to tell you. Not exclusively so: there's laboratory tests that can also do it. But between friends, aren't I absolutely best placed to clarify I didn't use wheat than sending it to a lab which is unrealistic?
I always love how Olivia can give everyday common sense examples to illustrate how our scientific reasoning could be improved if it were *more* like common sense, not less.
Olivia Guest · Ολίβια Γκεστ@olivia.science · 46m agoFormally, I assume you can guess why that's evidently true, but also informally is the same as well. If somebody texts me "my child needs help and in your class, can I drop off their meds in the break?" and I reply "oh, yes, are you their dad?" it doesn't preclude the same answer if it's their mum.
"Probably" is doing a lot of work here. Nobody actually knows that that's true, and there's reason to believe it's not. Here's an actual cognitive scientist's take on the matter bsky.app/profile/iris...
Iris van Rooij 💭@irisvanrooij.bsky.social · last wk."(...) no computationally tractable account exists to date for substantive and domain-general cognitive capacities (...) Moreover, there is no good reason to believe that such tractable accounts will be forthcoming via machine learning or otherwise" journals.sagepub.com/doi/10.1177/...
In any case, your original claim that what an LLM does "requires comprehensive understanding" is demonstrably false because an LLM does not and cannot understand anything.
And when Leibniz went into the understanding windmill, he found nothing but nuts and bolts but not understanding itself. I just love how a not-that simple language simulator fucks up terminology of cognition.
extra funny because to me that dumbass baby is way cuter as well
Darren Ewing@sadknob.bsky.social · 24h ago"you said it was fancy dress! you said you were dressing like a magician! you said i should wear my beaver costume! there's a special place in hell for you, waylon. there's a special place in hell."
This is, deservedly, vicious in its appraisal of Pinker. But so should all appraisals of "public intellectuals", a category that shouldn't really exist.
Dr Abeba Birhane@abeba.blacksky.app · 24h ago"For a man whose academic field was built largely on faulty studies to ignore the intellectual fraud in his circle and lash out at people like Arday for bad scholarship – well, psychologists used to call that projection" @virginiaheffernan.bsky.social www.thenerve.news/p/steven-pin...
That they simply repeat and maneuver around previous thought is a feature and not a bug here
It's also one reason they're determined to replace every researcher with chatbots, because for various reasons they can rely them to regurgitate their cant more than human scientists at this point
This applies equally well to Dawkins's gene-coding dominant/exclusive approach, which is no longer the bleeding edge of anything, but which still has a loyal audience in tech and finance for obvious reasons
An underrated factor in former academic heavyweights now leveraging their priors in non-academic circles, particularly reactionary ones, is that they can no longer throw their weight around with such abandon in 'real' scientific circles given that the foundations of their thought have become shaky
Dr Abeba Birhane@abeba.blacksky.app · 24h ago"For a man whose academic field was built largely on faulty studies to ignore the intellectual fraud in his circle and lash out at people like Arday for bad scholarship – well, psychologists used to call that projection" @virginiaheffernan.bsky.social www.thenerve.news/p/steven-pin...
If you read the replies you’ll see that Pinker was always a race science supporter, it’s just that nobody paid any attention. Maybe we should start noticing when celebrity scientists are fans of pseudoscientific ideas, and adjust our opinion of their competence accordingly
Virginia Heffernan@virginiaheffernan.bsky.social · 2d agoWhat happened to Steven Pinker? I only felt free to write this because he blocks me: www.thenerve.news/p/steven-pin...
There’s more to be said about the relationship between Pinker-style stardom and the state of academia. Not because academia is Woke, but because star systems loosen rigor at the top while solid scholarship becomes a dead end lower down. Shoddy celebrities and broke PhDs proceed from the same machine
