Lightnews — Scholar-powered news

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 3d

If the OP author truly believed that then I challenge them to think about why it is they decided to defend LLMs rather than agreeing with and extending the prior utterance they responded to (like an LLM would).

Tim Price @timprice.bsky.social · 3d

I have pondered this sentence and am now dumber for having done so

That's a very dismissive construct of LLMS, IMHO. Yes they are stochastic and yes they are 'robotic' but here's a sentence for you to ponder:

Humans are just advanced language output algorithms that spit out returns based on input and their reference library.

2

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 3d

This is why computer scientists (and computer science fetishists in silicon valley) need to take a gorram sociolinguistics class.

1

Reposted by The Grumpy Data Scientist

Olivia Guest · Ολίβια Γκεστ @olivia.science · 5d

We end on an appeal to our fellow psychologists:

"To sit idly by while deskilling and displacing of [our]selves is normalised [...] serves [solely] the technology sector, which avoids criticism and self-reflection and prefers pseudoscience and misinformation."

10/

8 Do not Embrace AI
In this paper, we unpacked why we think psychologists need
to be on high alert — not just to avoid another replication crisis,
but to avoid the total collapse of our science. What we signpost
in Table 1 may have been novel to readers until this point, but
the deeper problems are absolutely known. Also, as Crystal Steltenpohl et al. (2023, pp. 9–10) state: “Intentions alone are not
enough to move science forward. Creating responsible, considered processes for rigorously transparent open science requires
involving interested parties from a wide range of backgrounds,
perspectives, research areas, and training paradigms.”
Indeed, because many such warnings go unheeded — such as
the need for a cultivation of shared values and especially the principles of impartiality of researchers and academic freedom from
corporate influence — we find ourselves in polycrises that affect
our universities, political systems, planet, and ultimately all humanity. “When historians of science look back on the 2010s in
social and personality psychology, the decade will likely stand out
as a period of exceptional doubt and self-scrutiny in the field.”
(Schiavone & Vazire, 2023, p. 710) Why did we ever stop? Should
we ever stop?

Importantly, Hazel Rose Markus (2005, p. 180, emphasis
added) explains that: “Social psychology is often defined as the
study of how people respond to and are influenced by other people”
Algorithms, chatbots, LLMs, machines, models, inanimate objects are not people — they are the products of people (Guest,
2024, 2025). And to paraphrase Rae Carlson (1984): What’s social about chatbots? Where’s the person in an LLM?
We must sure up our subfields from the slow but certain corrosive power wielded by the harmful nonsense that is modern
displacement AI. To sit idly by while deskilling and displacing of
our students, participants, and selves is normalised — or worse
still to profit from it — serves not science but the technology sector, which avoids criticism and self-reflection and prefers pseudoscience and misinformation.

1 14 35

Reposted by The Grumpy Data Scientist

Autistic Self Advocacy Network (ASAN) @autisticadvocacy.org · Jul 29

Generative AI changes what things mean because it focuses on words, not the meaning. It adds words, changes meaning, and makes "plain language" that is not accessible. We know, because we tested it. Generative AI has no place in plain language. autisticadvocacy.org/2025/07/asan...

Dark grey textured background. There’s a lighter grey box with text that reads: Artificial intelligence is when a computer program does things that normally need to be done by humans. We call artificial intelligence “AI” for short. There are lots of different kinds of AI.
The way AI works is by using data from people. Data is information a computer can read. AI programs read data made by people to understand what to do.
Generative AI is a specific kind of AI. Generative AI can use data to make new things.
Generative AI can create many kinds of new things, like:
text
images
music/movies
and more
People have already started using generative AI to write plain language.

Grey textured background. Text reads: In our statement, we explain the reasons why people should not use generative AI to write in plain language. Plain language is an important part of making things accessible for people with disabilities. Using generative AI makes “plain language” that is not actually accessible. We hope anyone who writes in plain language will not use generative AI.
Here are the reasons we talk about in our statement:
Generative AI changes what things mean
Plain language is an idea that is too new for generative AI to understand
Generative AI focuses on words, not ideas
Generative AI has discrimination built-in
Plain language is by and for disabled people
Read our full statement to learn more about each of these reasons.

5 100 210

Reposted by The Grumpy Data Scientist

Frames @oxinabox.bsky.social · 4d

as a girl with a PhD in natural language processing and machine learning it's actually offensive to me when you say "we don't know how LLMs work so they might be conscious"

I didn't spend 10 years in mines of academia to be told ignorance is morally equal knowledge.

We know exactly how LLMs work.

66 1.6K 5.7K

Reposted by The Grumpy Data Scientist

Dr. Damien P. Williams Wants A Better Sociotechnical Paradigm @wolvendamien.bsky.social · 4d

"If it is inaccessible to the [marginalized/minoritized/oppressed], it is neither radical nor revolutionary."

5 160 440

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 8d

!!!! @annleckie.com’s Ancillary series id one of my favorite SciFi series, period. Heck. I’ve loved all of their books :)))

Ann Leckie @annleckie.com · 8d

www.amazon.com/gp/aw/d/B0FP...

Radiant Star

Radiant Star - Kindle edition by Leckie, Ann. Download it once and read it on your Kindle device, PC, phones or tablets. Use features like bookmarks, note taking and highlighting while reading Radiant Star.

www.amazon.com

3

Reposted by The Grumpy Data Scientist

Darien @boy-rat.bsky.social · 12d

I’m so fed up of seeing Ai slapped on everything!

I spent 15 years building and running an Ai fraud detection company

LLMs are NOT Ai 🤖 not even remotely related and as for Sam Altmans claim that they don’t fully understand how their system works… I’m happy to explain it to him using crayons 🖍️

2 21

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 11d

Upsetting as the details are, I was thinking about this one :( www.npr.org/sections/sho...

Their teenage sons died by suicide. Now, they are sounding an alarm about AI chatbots

Grieving parents and online safety advocates at a congressional hearing called for new laws to regulate AI companion apps to protect the mental health of minors.

www.npr.org

1 3

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 12d

@olivia.science @stevierea.bsky.social @irisvanrooij.bsky.social? Know anyone who might?

1 1

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 12d

I’m curious: I have a hunch about the information theoretic reasons behind why LLM outputs might trigger the positive assessment that it does in some users/what it is about the model that exhibits “sycophantic” behavior. Anyone we know have access to the transcripts openAI released in court?

1 1

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 14d

Remember kids, the next time he forgets his login credentials it’s CENSORSHIP BY thE TERRible LeFt

dr. caitlin m. green @caitlinmoriah.bsky.social · 14d

it was his own staffer 💀

Nick Sortor & @nicksortor • 16h
* BREAKING: The Secret Service has launched an investigation into the escalator situation at the United Nations today, per Karoline Leavitt
If someone did this intentionally, they should be
CRIMINALLY charged.
It put President Trump is a VERY compromising position.
"If we find that these were UN staffers who were purposefully trying to trip up President Trump and FLOTUS well, there better be accountability for those people," Leavitt said

Steve Cortes @CortesSteve
Follow S *
Considering the recent trend of left wing violence, including Pres Trump taking a bullet to the head, crap stunts like this one cannot be tolerated.
Get the UN the hell out of America!

2

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 15d

Found footage of the meeting to prep for this press conference.

Idiocracy Electrolytes GIF

ALT: Idiocracy Electrolytes GIF

media.tenor.com

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 15d

I’m not convinced the brain worm is dead.

Science Magazine @science.org · 15d

Federal officials said the FDA would approve a drug called leucovorin as “the first FDA-recognized treatment pathway for autism”—an unusual move for a pill only tested in a few small studies. https://scim.ag/48xOJRj

Trump’s autism initiative embraces little-tested vitamin as a treatment

FDA to approve leucovorin despite questions about whom it might help

scim.ag

1 2

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 19d

:( crunch time blues.

We salute you and your effort 🫡

1

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 20d

This. Not all peer reviewers treat authors with respect, but a majority of them do. And I will 100% own up to the fact that peer review has not only shored up my writing and occasionally my methods, but even got me to think about extensions and improvements for subsequent work.

Meagan Phelan @meagan-g-phelan.bsky.social · 20d

During #PeerReviewWeek, @science.org
heard from individual authors who talked about how peer review strengthened their work, making it functionally richer, more accessible, more pointed regarding limitations--sometimes in collaboration with preprint review. See author posts in thread below. 🧵

2

Reposted by The Grumpy Data Scientist

J. Hrebinka (Shanahan) 🇺🇦🇯🇴 @enceladosaurus.bsky.social · 23d

I made the mistake of devoting myself to AI4SG early in my career because I wanted to believe I could use my skills to make the world a better place. Instead, I found exactly what Dr. Birhane describes - particularly with respect to accountability laundering. I'll share one example🧵

Dr Abeba Birhane @abeba.bsky.social · 23d

AI is the wrong tool to tackle complex societal & systemic problems. AI4SG is more about PR victories, boosting AI adoption (regardless of merit/usefulness) & laundering accountability for harmful tech, extractive practices, abetting atrocities. yours truly
www.project-syndicate.org/magazine/ai-...

The False Promise of “AI for Social Good”

Abeba Birhane refutes industry claims about the technology's potential to solve complex social problems.

www.project-syndicate.org

4 28 59

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 24d

Grabbing this too!

1 2

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 24d

Dunno if this would interest you, but you might like the book “Street Data” by Shane Safir. It was one of my favs last year and a pretty harsh look at how folks metricize learning (I’m unashamedly trying to get more folks to talk about that book with)

1 2

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 29d

🎉🎊

1

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 29d

Oh man. I hope you realize that I’m going to devour this paper. And I may want to hit you up about a project some folks and I are doing at UCLA that’s pretty clearly related based on the abstract if you’d be interested in that.

1 1

The Grumpy Data Scientist @zaqdelinguist.bsky.social · 29d

Gaddammit. I just sent a paper out for review on a related topic and am embarrassed I missed this paper 🙈

1 1

The Grumpy Data Scientist @zaqdelinguist.bsky.social · Sep 8

Your paper is hitting a really important chord with folks on top of being well written and well argued? 🤓

1 5

The Grumpy Data Scientist @zaqdelinguist.bsky.social · Sep 8

Exactly. You show up to these things and it’s basically a sales pitch/product demo.

Not (A.) I. @aerialeverything.cryptoanarchy.network · Sep 7

This is why I have a deep suspicion of the "AI literacy" workshops being floated to profs by schools mainlining AI. Ed tech has co-opted "AI literacy" to mean "learn how to use our products" rather than "learn how AI works."

Fabio Chiusi @fabiochiusi.bsky.social · Sep 7

“people with lower AI literacy are typically more receptive to AI,"

"people with lower AI literacy are more likely to perceive AI as magical and experience feelings of awe in the face of AI's execution of tasks that seem to require uniquely human attributes."

futurism.com/more-people-...

1 1

The Grumpy Data Scientist @zaqdelinguist.bsky.social · Sep 6

Freaking wish the UC board of regents would read this . . .

Olivia Guest · Ολίβια Γκεστ @olivia.science · Sep 6

Finally! 🤩 Our position piece: Against the Uncritical Adoption of 'AI' Technologies in Academia:
doi.org/10.5281/zeno...

We unpick the tech industry’s marketing, hype, & harm; and we argue for safeguarding higher education, critical
thinking, expertise, academic freedom, & scientific integrity.
1/n

Abstract: Under the banner of progress, products have been uncritically adopted or
even imposed on users — in past centuries with tobacco and combustion engines, and in
the 21st with social media. For these collective blunders, we now regret our involvement or
apathy as scientists, and society struggles to put the genie back in the bottle. Currently, we
are similarly entangled with artificial intelligence (AI) technology. For example, software updates are rolled out seamlessly and non-consensually, Microsoft Office is bundled with chatbots, and we, our students, and our employers have had no say, as it is not
considered a valid position to reject AI technologies in our teaching and research. This
is why in June 2025, we co-authored an Open Letter calling on our employers to reverse
and rethink their stance on uncritically adopting AI technologies. In this position piece,
we expound on why universities must take their role seriously toa) counter the technology
industry’s marketing, hype, and harm; and to b) safeguard higher education, critical
thinking, expertise, academic freedom, and scientific integrity. We include pointers to
relevant work to further inform our colleagues.

Figure 1. A cartoon set theoretic view on various terms (see Table 1) used when discussing the superset AI
(black outline, hatched background): LLMs are in orange; ANNs are in magenta; generative models are
in blue; and finally, chatbots are in green. Where these intersect, the colours reflect that, e.g. generative adversarial network (GAN) and Boltzmann machine (BM) models are in the purple subset because they are
both generative and ANNs. In the case of proprietary closed source models, e.g. OpenAI’s ChatGPT and
Apple’s Siri, we cannot verify their implementation and so academics can only make educated guesses (cf.
Dingemanse 2025). Undefined terms used above: BERT (Devlin et al. 2019); AlexNet (Krizhevsky et al.
2017); A.L.I.C.E. (Wallace 2009); ELIZA (Weizenbaum 1966); Jabberwacky (Twist 2003); linear discriminant analysis (LDA); quadratic discriminant analysis (QDA).

Table 1. Below some of the typical terminological disarray is untangled. Importantly, none of these terms
are orthogonal nor do they exclusively pick out the types of products we may wish to critique or proscribe.

Protecting the Ecosystem of Human Knowledge: Five Principles

4 10