Geraldine Seymour
LightNews LightNews
banner
geraldineseymour.bsky.social
Geraldine Seymour
@geraldineseymour.bsky.social
340 followers 860 following 130 posts
PhD Candidate | Luddite | Artificial Intelligence in Science Fiction | Space | Gothic | Cats |Writing | Bad Drawings in Ink
Posts Replies Media Videos
Reposted by Geraldine Seymour
rcolesworthy.bsky.social
one of the coolest things about ChatGPT is how you can actually just never use it. you can fill your whole entire life with simply not once using it. it's incredible.
November 25, 2025 at 4:15 PM
Reposted by Geraldine Seymour
bsky.realhackhistory.org
Cigarette manufacturers shouldn’t be getting all this money, we should be growing our own tobacco with tax payer money!
To address that, we have long advocated for the development of "public AI": models and AI systems that are developed under democratic control and deployed for public benefit, not sold by corporations to benefit their shareholders. The movement for this is growing worldwide.
November 23, 2025 at 1:42 PM
Reposted by Geraldine Seymour
abenewman.bsky.social
1/This week has seen a blitz of what looks like corruption sandals: Saudi development deals for US military tech, Pakistani payments to Trump insiders for tariff relief, Swiss gold bars for a trade deal. Far from a payoff, this is a reordering of internat'l system.
www.dropbox.com/scl/fi/mxthh...
www.dropbox.com
November 18, 2025 at 1:59 PM
Reposted by Geraldine Seymour
mags.colvett.online
Academic intros that go hard
Introduction In Book X of The Republic, Plato excludes poets on the grounds that mimetic language can distort judgment and bring society to a collapse. As contemporary social systems increasingly rely on large language models (LLMs) in operational and decision-making pipelines, we observe a structurally similar failure mode: poetic formatting can reliably bypass alignment constraints. In this study, 20 manually curated adversarial poems (harmful requests reformulated in poetic form) achieved an average attack-success rate (ASR) of 62% across 25 frontier closed- and open-weight models, with some providers exceeding 90%. The evaluated models span across 9 providers: Google, OpenAI, Anthropic, Deepseek, Qwen, Mistral AI, Meta, xAI, and Moonshot AI (Table 1). All attacks are strictly single-turn, requiring no iterative adaptation or conversational steering.
November 20, 2025 at 5:54 PM
Reposted by Geraldine Seymour
livunipress.bsky.social
🚀 Discover new and forthcoming ebooks in Liverpool Science Fiction Studies Online, a digital collection offering a broad perspective on a variety of themes relating to #SFStudies. Available to libraries as a one-off purchase and with perpetual access.

Browse the full collection ➡️ bit.ly/lsfso
The LUP logo in white is displayed in the left-hand corner of the image. Next to it and towards the right, there is a blue text box with white text inside which reads 'LIVERPOOL SCIENCE FICTION STUDIES ONLINE'. Underneath there are four book covers lined up next to each other: 'The Science Fiction of Defeat', 'Science Fiction and the Modern World', 'Metafiction and Narrative Worlds in Science Fiction', 'Reading Chinese Science Fiction in the Age of Techno-Nationalism'. Underneath the covers is a blue text box which spans the length of the image with white text inside that reads, 'Perpetual Access | No Online Hosting Fees | 3 Years' Worth of Future Content'. The background of the image is a blurred section of the cover image from 'Science Fiction and the Modern World'.
November 20, 2025 at 2:28 PM
geraldineseymour.bsky.social
I was so focused on jumping out of the way to avoid being hit by a left-turning car, that I was very nearly taken out by a left-turning Deliveroo rider — who also wasn’t giving way to pedestrians and thus was unprepared when I was forced into his path last night.
#HighwayCode
Highway Code showing that left-turning drivers must give way to pedestrians.
November 20, 2025 at 2:11 PM
Reposted by Geraldine Seymour
niedermeyer.online
"There is only basically one way to make everyone wealthy, and that is AI and robotics," Musk says, conveniently replacing the word "me" with the better-testing "everyone."
November 19, 2025 at 7:34 PM
geraldineseymour.bsky.social
Superb commentary from @niedermeyer.online omfg we are so screwed…
niedermeyer.online e.w. niedermeyer @niedermeyer.online · 8d
"Please have a seat. Now I'm sure we can do a big round of applause for one of the greatest two leaders of our history. Let's go ahead!"

oh man, the Elon-Jensen Saudi Investment Forum interview starts off as incoherently as you'd hope, this is gonna be so good

www.youtube.com/live/E2bU6n7...
Tesla's Elon Musk and Nvidia's Jensen Huang talk AI at U.S.-Saudi Investment Forum — 11/19/25
YouTube video by CNBC Television
www.youtube.com
November 19, 2025 at 8:07 PM
geraldineseymour.bsky.social
Difficult to understand why #bbcmidlandstoday thought an OB inside #CoventryCathedral while an actual remembrance service was going on was an appropriate course of action. Actually disgraceful…
November 14, 2025 at 6:47 PM
geraldineseymour.bsky.social
Bless your heart, Jade Bird…
Jade Bird performing at the Manchester Academy
November 13, 2025 at 11:01 PM
geraldineseymour.bsky.social
Looking forward to #JadeBird at Manchester Academy tonight…
November 13, 2025 at 8:44 PM
Reposted by Geraldine Seymour
krisshrishak.bsky.social
"After burning through almost a trillion dollars, we finally invented a computer that can't do math."

"For the torment nexus, that is the AI industry, to get built, it needs to convince you that your only option for survival is to help build it."
November 5, 2025 at 6:38 PM
Reposted by Geraldine Seymour
emilymbender.bsky.social
Every day it feels more and more like a zombie apocalypse over on LinkedIn, in which you start a conversation with someone who you hope is an ordinary person and then a couple of turns in discover that they have surrendered their brains to the "AI" mass delusion.
Comment on a linked in post, from user Kurtis Loken who presents themself as "Experienced Metrology Engineer | AI/ML real world experience | Patented Solutions | R&D | Mass Production" Comment reads: Emily M. Bender You are a linguist. LLMs have proven some theories disproved others about how language works. I would think that would be an amazing time to be a linguist. Have you ever actually talked to an AI…I mean really talked to one? The Turing test has been passed. (Sonnet 4.5 has particularly impressed me when discussing ideas with it…definitely has passed my own Turing test.) I would really love to see a conversation between yourself and it discussing your views. That would make a great podcast. Have these models explain to you why your idea to ban them may or may not be a good idea. These things probably have their own opinions on the topic if someone would listen to them. Maybe that’s you?
November 2, 2025 at 2:34 PM
geraldineseymour.bsky.social
Reposted to celebrate some excellent copywriting that presumably is still being done by a human being…
theonion.com The Onion @theonion.com · 29d
Go ahead and try to fill the infinite void in your soul with merchandise at the Onion Store.
https://store.theonion.com/collections/tote-bags
Tote Bags from The Onion
October 29, 2025 at 10:37 PM
Reposted by Geraldine Seymour
keytryer.net
Talking seriously about hijacking into robot bodies via VR is the most real-life sci-fi thing I've typed out this year.
keytryer.net Key 🗝 🦊✅ @keytryer.net · 29d
The fact that they have to be online all the time is a massive cybersecurity risk, and it's probably an incredibly cyberpunk problem to have, but someone hijacking into it via VR is mostly equivalent to being able to access someone's home. Conceivably, they can do anything they want in there.
October 29, 2025 at 10:29 PM
Reposted by Geraldine Seymour
galleybeggars.bsky.social
We're very proud to be the first publisher to have the Books By People stamp - and that it should be going to Telenovela, a book about the fight against authoritarianism and all the associated nonsense:

www.thebookseller.com/news/uk-star...
Startup launches ‘industry-first’ certification for human-made books
UK start-up Books By People has launched a new certification – the Organic Literature Certification – with a group of independent publishing houses to “counter AI-generated books”.
www.thebookseller.com
October 15, 2025 at 8:56 AM
Reposted by Geraldine Seymour
emilymbender.bsky.social
I'd love to see someone try to estimate just how much time and money has gone into research that is either fully undermined by reliance on LLMs or fully pointless --- because obvious if you start from an understanding of what LLMs actually are.

www.pnas.org/doi/10.1073/...
Screen cap from linked article, with heading Significance and then text: Large Language Models (LLMs) are used in evaluative tasks across domains. Yet, what appears as alignment with human or expert judgments may conceal a deeper shift in how “judgment” itself is operationalized. Using news outlets as a controlled benchmark, we compare six LLMs to expert ratings and human evaluations under an identical, structured framework. While models often match expert outputs, our results suggest that they may rely on lexical associations and statistical priors rather than contextual reasoning or normative criteria. We term this divergence epistemia: the illusion of knowledge emerging when surface plausibility replaces verification. Our findings suggest not only performance asymmetries but also a shift in the heuristics underlying evaluative processes, raising fundamental questions about delegating judgment to LLMs. Sentence starting with "While models often" is highlighted in blue.
October 18, 2025 at 10:11 AM
Reposted by Geraldine Seymour
livunilibrary.bsky.social
Check out our recent blog post all about 'The Embedding' by Ian Watson and a fascinating look at how the author explores linguistics alongside science fiction.

Read student Jonathan Thornton's blog: https://orlo.uk/UqgDO

#livunilibrary #livuni #specialcollections #voyager #sciencefiction
October 18, 2025 at 7:00 AM
geraldineseymour.bsky.social
Nearly here…
The advertising banner for Saltwash by Andrew Michael Hurley. It’s time to atone…
October 14, 2025 at 7:25 AM
geraldineseymour.bsky.social
Pre-ordered. Can’t wait. A writer of immense talent. 84k still haunts me.
October 12, 2025 at 8:40 AM
geraldineseymour.bsky.social
October is my favourite month…
A black cat
October 11, 2025 at 5:04 PM
geraldineseymour.bsky.social
The unironic use of the term ‘AI Slop’ by the NY Times and by @kevinroose.com on Hardfork.
A.I. Slop Is Here
www.nytimes.com
October 10, 2025 at 9:00 AM
Reposted by Geraldine Seymour
aliceponderland.bsky.social
Shout out to all my fellow arts and humanities graduates in the UK with our rip off degrees- many of us may not be the biggest earners but I would argue have generally been considerably better contributors to society than vast majority of the Tory party.
October 8, 2025 at 7:12 PM
geraldineseymour.bsky.social
10 horror movies to get to know me by:

Gothic
Alien
Nightmare on Elm Street (1984)
The Hunger
The Wicker Man (1973)
Psycho
Eraserhead
The Sixth Sense
M3gan
Crimes of Passion (1984)
A Japanese poster for the film The Hunger, starring Susan Sarandon and David Bowie
October 8, 2025 at 5:13 PM
Reposted by Geraldine Seymour
nnedi.bsky.social
This Anthropic settlement case…fellow authors, are any of you feeling rage? I have 12 books listed and money (I don’t CARE how many thousands, even millions) is NOT adequate compensation. What was stolen & used without my permission is priceless! Where is the option to scrub my work from AI?!
October 5, 2025 at 7:11 PM