Ok so one personal learning of 2025 - it might be a good idea to try to reconnect with friends you lost in your twenties due to both parts being idiots in their 20s. With some the reconnection won't work, and that's fine. But those with who it will work are absolutely worth it.

I guess it's time for my yearly books recommendations/recap of the books I read and actually liked thread. I will start with some tech non-fiction I liked, then add some other non-fic and a bunch of fiction I guess, otherwise in no particular order.

One lesson from the shocking story of this teenager's final AI interactions is how sensitive the data is. Journalists were not allowed to review the content itself. Society shouldn't have to wait for deaths & lawsuits to pile up. We need faster science. (1/n) www.washingtonpost.com/technology/2...

Google is hosting a CBP app that uses facial recognition to identify immigrants, while simultaneously removing apps that report the location of ICE officials because Google sees ICE as a vulnerable group. “Big tech has made their choice.” www.404media.co/google-has-c...

a neat (and scary) survey into AI "companions" and kids www.aura.com/reports/stat... "nearly half of the time kids and teens use AI, it is for companionship, and when they do, the themes of discussion are primarily sexual, romantic, and violent."

The surge of AI slop, deceptive synthetic videos & images flooding our feeds, is driven in large part by Agentic AI Accounts (AAAs). aiforensics.org/work/agentic... Key findings: -Over 43,000 mostly AI-made posts generated 4.5 billion views -More than 65% of accounts were created in early 2025 1/

🧵Have you ever heard of LLM-authored books being used as part of an disinformation operation? Well now you have! #AI #Disinformation open.substack.com/pub/marcowen...

My last story of the year: an update on the woman who fell in love with ChatGPT. www.nytimes.com/2025/12/22/t...

If you are an "interdisciplinary" journal that welcomes "worldwide" authorship, then either allow pdfs as first submission or make sure your system's latex compiler supports non-Latin fonts - I am about to convert 5 sentences in Cyrillic to IMAGES to embed them into the Tex, there is no other way

I wish SocSci journals either optimized their systems for LaTeX submissions or, at least, allowed submitting a pdf as the first pre-review submission. This would totally increase at least my productivity and efficiency.

ChatGPT gains local knowledge panels on the right side but where do they come from www.seroundtable.com/openai-adds-... via @nearmedia.bsky.social @gsterling.bsky.social #chatgpt #local #maps #search #openai

The ongoing radicalization of Gen Z Republicans is something media in Europe should absolutely start covering. Now.

NEW: ICE has paid hundreds of thousands of dollars to a company that makes “AI agents” to rapidly track down targets. AI Solutions 87 did not respond to a request for comment on how its AI agents work.

The sad thing is Mr Fukuyama would not even need to build some crazy database to track stuff he reads if only someone told him about Zotero... I mean he says he tracks 20k items. My Zotero library is 7k+ (just checked), excluding group libraries. No separate server or database necessary, really.

Thinking you have a good idea because the LLM said “good idea!” is identical to thinking your gas pump is grateful because it thanked you

Cloudflare cofounder and CEO Matthew Prince explains to Fast Company how the pay-per-crawl model can help fix the misinformation epidemic that is eroding the credibility of the information ecosystem we’ve all come to rely on. f-st.co/TM4ERRR

The ACM Digital Library, where a LOT of computing-related research is published (I'd say at least 75% of my own publications), is now not only providing (without consent of the authors and without opt-in by readers) AI-generated summaries of papers, but they appear as the *default* over abstracts.

“With the govt’s planned 10% cuts to the Swiss National Science Foundation, Switzerland is massively weakening its research, thereby putting the innovative strength of its economy & its competitiveness at risk,” says the President of the Research Council of the @snf-fns.ch www.srf.ch/audio/echo-d...

Students asked me yesterday for some sources to keep up with the tech news now that the semester is over and I won't spam them with those anymore. For the 🇨🇭 context I am happy to repeat my "unsponsored advertisement" here :) Follow the work of @adfichter.bsky.social @julianschmidli.com

Here we go! @lmesseri.bsky.social & I wrote about epistemic risks of synthetic participants ('AI Surrogates') in @cp-trendscognsci.bsky.social (see 🧵 here) bsky.app/profile/mjcr...

Did you know that from tomorrow, Qualtrics is offering synthetic panels (AI-generated participants)? Follow me down a rabbit hole I'm calling "doing science is tough and I'm so busy, can't we just make up participants?"

New Paper in @computationalcommunication.org The Telegram COVID-19 Protest Dataset 2020-2022 with the most wonderful co-autohors @buehling.bsky.social & @maxzehring.bsky.social >> 2020-2022 >> 715 channels & 229 public groups >> 5,641,026 messages >> Manually curated >> Open for researchers

For the latest Naked Pravda episode, I spoke to Nikolay Kononov about his new book on Pavel Durov and Telegram. Tech-bro feudalism, massive cryptocurrency schemes, free speechism, and a never-ending quest to evade regulatory oversight. meduza.io/en/episodes/...

This is fucking grim. Somebody invented a white guy, an "IT professional" named Edward Crabtree, who stopped the Bondi shooting and spread it all over the internet, which was picked up by AI agents and slop aggregation sites. The real hero is a fruit stand owner named Ahmed el Ahmed.

🇭🇺💥Thousands marched to Viktor Orbán’s office in Budapest on Saturday, led by opposition leader Péter Magyar, who called on the PM to resign over multiple child abuse scandals. Protesters carried toys and torches under the banner “Protect the children!”. 🎥Timelapse video posted by Magyar.

Made a site comparing the sizes of living things :) The great Julius Csotonyi spent 5 months painting over 60 illustrations for the site, no ai used > neal.fun/size-of-life/

🚀 We’re hiring a Postdoc! Our group is looking for a Postdoc to join the team working on computational comm research. If you’re excited about automated content analysis, large text & social-media data, open science, this might be for you. 💡 Sounds like you or someone you know? Please share/boost!

Another wrongful death lawsuit filed against OpenAI: www.cbsnews.com/news/open-ai... This time for a woman murdered by her son in Connecticut, whose conversations with ChatGPT were earlier reported by WSJ: www.wsj.com/tech/ai/chat...

New! We worked with the Public Interest Research Group to test a bunch of LLM-driven toys ahead of the Christmas. That sexual talking teddy bear wasn't a one-off; it turns out disturbing behavior is deeply built into these toys across the board.

Ah, yes. Trump receiving a worthless peace prize leads to the US dropping corruption charges against FIFA.

I posted it here before, but @favstats.eu and I wrote an entry about LLMs in the encyclopedia below 👇 Final version is out now at www.elgaronline.com/display/book... Or read the preprint: doi.org/10.31219/osf...

After 3 years and 5,322 emails (and counting) our Encyclopedia of Political Communication is finally out at @elgarpublishing.bsky.social 📚 Three volumes ⭐ 431 entries 🎓 581 wonderful authors from across the world A very short 🧵

💥 💥 💥 #PalantirFiles, Part II 💥 💥 💥 Zürich= Drehscheibe für Palantirs weltweites Business. Das Schweizer Aussendepartment führt nun eine Untersuchung gegen #Palantir. Und: Zürcher Kantonsräte sind jetzt wach geworden. Was genau mauschelt der Kanton mit Palantir? www.republik.ch/2025/12/09/w...

Via Milagros Miceli on Mastodon: "Today, DWI publishes a powerful essay by Michael Geoffrey Asia, who works impersonating an AI sex companion. This piece offers a rare glimpse into the psychological and economic realities behind one of the fastest-growing sectors: AI-assisted intimacy."

OpenAI updated its ChatGPT crawler / bot OAI-SearchBot www.seroundtable.com/openai-chatg... via Pieter Serraris #chatgpt #openai #bot

EU eröffnet Verfahren gegen Google wegen KI-Suche: Beutet Google für seine neue KI-Suche das Web aus? Die EU-Kommission hat nun offizielle Ermittlungen begonnen, auch YouTube steht im Fokus.

“AI models not only point some users to false sources but also cause problems for researchers and librarians, who end up wasting their time looking for requested nonexistent records…”

The messy implementation of Australia’s under 16 social media ban. Demanding better, responsible, compliant social media platforms rather than a messy age ban is much needed. Which many Australian colleagues like @antmandan.bsky.social have said, repeatedly. www.theguardian.com/media/2025/n...

This roundup features some of the best writings and perspectives on AI in all kinds of contexts. And thanks for mentioning the only piece i wrote this year, Eryk! I promise next year i'll be fitter happier more productive. Maybe. ;)

🧵 In early November, the troll network set up on X/Twitter to praise Kirill Dmitriev has been expanded by 24 LLM-powered bot accounts. Just a month later, Dmitriev is blaming EU today for using bots to manipulate his poll results.

noch nie ist eine technologie so schnell ins dystopische gekippt wie generative ki.

ever since i read this i’ve been convinced that the uk is basically being governed by chat www.nytimes.com/2025/12/03/m...

A Palantir billionaire just called for public hangings. “It's time to bring back masculine leadership to protect our most vulnerable,” writes Joe Lonsdale. This is where Silicon Valley authoritarianism is heading: www.thenerdreich.com/joe-lonsdale...

Let me tell you a story. Perhaps you can guess where this is going... though it does have a bit of a twist. I was poking around Google Scholar for publications about the relationship between chatbots and wellness. Oh how useful: a systematic literature review! Let's dig into the findings. 🧵

Here's the reality this example illustrates: It's not even just about people blindly trusting what ChatGPT tells them. LLMs are poisoning the entire information ecosystem. You can't even necessarily trust that the citations in a published paper are real (or a search engine's descriptions of them).

I unfortunately think you at times can't trust stuff even from people you know (in my case, at least, i did on multiple occasions receive AI stuff. Not from people I know well or people who are close to me. But from people I did know.)

Elon reposting Brendan Carr as a first reaction. But no, X is not taxed for being successful. And yes, X is fined for breaking simple DSA requirements regarding deceptive design, ad transparency, and data access for researchers