🎓 PhD Studentship: 3.5years, fully funded: Come build engaging multi-platform data collection tools for digital wellbeing interventions with me, @nballou.bsky.social, @martidisi.bsky.social at Imperial College. Application deadline Feb 26, full details here: bit.ly/phd-digital-...

Just a big THANK YOU 🙏 to @cfiesler.bsky.social for writing up how to deal with fair use images in ACM articles. As an @acm.org EiC, all I can say this is gold & should be embedded in standard instructions and forms. caseyfiesler.com/2016/10/20/f...

There is no way to safety-engineer LLMs (especially in “agentic” form) in the same way that you cannot safety-engineer general purpose computing. You intentionally afford practically infinite use cases. If anything, they are less safe because they lower the expertise barrier to doing stupid shit.

🚨Calling all new media scholars🚨 If (and I emphasize if) a 2nd edition of the New Media Reader happened to be in the works, what additions would you like to see? What do you think could be excised? @docmofo.bsky.social @n-w-f.bsky.social @psssssssss.bsky.social mitpress.mit.edu/978026223227...

Listened to @pluralistic.net.web.brid.gy's great 39C3 keynote "The Post-American Internet." He remains one of the true hopepunks: modelling to me how finding concrete opportunities for action and charging us with hope is probably the most effective activism today. pluralistic.net/2026/01/01/3...

"AI should augment not replace humans" has become a meaningless trope that is and can be happily slapped onto any AI replacing/degrading human work initiative. So what is an effective retort and/or move to a more productive stance? Just, "Can you specify what that looks like in this context"?

Spring cleaning of a former business travelling, long time Mac user: when you find yourself in possession of no less than nine now obsolete Mini-DVI to VGA/DVI dongles …

A lovely gentle and accessible dialogue on the limits of computationalism.

I used to say on Twitter, back when I was on there, “don’t retweet anything that you wouldn’t pay to promote”, as a way of trying to explain the value of algorithmic amplification. Instead, I learned a lesson about how much people really do want to subsidize and amplify things they pretend to hate.

Discworld QOTD, from Hogfather

Research is the MAGIC TIME on any project where I get to go have fun and explore. And 100% what you are looking for are the surprising and amazing details and textures that you could never make up

There isn't a single problem "solved" by edtech that couldn't be fixed with smaller classes led by well-paid teachers given real academic freedom

This kind of “automating the science pipeline” discourse, apocalyptic or utopian, takes as given the upside promise of foundation models, and doesn’t grant symmetry to the downside: if AI can exponentially automate science, it can also exponentially automate undetectable scientific fraud.

was at an event on AI for science yesterday, a panel discussion here at NeurIPS. The panelists discussed how they plan to replace humans at all levels in the scientific process. So I stood up and protested that what they are doing is evil. Full post: togelius.blogspot.com/2025/12/plea...

Quick boost: Come do a PhD with me!

I have to say, "more and more people are doing spacewalks" gotta be up there on the more-and-more-ism paper openers ranking –– dare I say, sky-high?

I fear that the hidden curriculum we put our students through is „muddling through chronic overload without being caught.“ Then again, that’s the reality we model and live as educators in HE today, so it figures it’s the reality we create for our students.

How to live and work in a world flooded with bullslop where the overwhelming majority of people is not capable nor interested in detecting bullslop nor able to do what they are asked to do without a bullslop machine.

Well fuck. You weren’t supposed to build the Stand Alone Complex. Injecting the social world with infinite undetectable impostors that increases the cost of honest signals by orders of magnitude is quickly becoming my top 2 „negative externality“ of LLMs. Class action, anyone?

inspired by this, I have done a free extract of this section of the book here. it is about understanding that *with all its problems* the BBC is as important for our intellectual health as the NHS is for our physical health: naomialderman.substack.com/p/the-bbc-an...

Research method textbooks in academia are a bit like email clients in software development: somehow, at some point in your career, everyone ends up being dissatisfied with everyone else’s and has to write their own which is 99% feature-identical to what’s already out there.

Come do your PhD with me at Imperial College in London! The LISS DTP funds 3.5y PhD tuition and stipend (+1y masters), deadline Jan 23: liss-dtp.ac.uk/applying-to-... See topics of interest below 👇

Mood.

With the 2025 @conference.aiandgames.com just around the corner, @tommy.aiandgames.com had something to get off his chest, or at least part of it. Game AI is facing an existential crisis, and we need to start talking about it. www.aiandgames.com/p/game-ais-e...

The meetings continue until morale improves.

2025: a late stage soviet bureaucrat larp you can never leave. Everything is strained and worn beyond function, everything breaks ever-faster, & the job is to make do & pretend & report numbers up & to the right, while everyone knows & copes & bonds over black humour & too much drink in private.

There is something truly remarkable about Generative AI. Several years in this revolution, half of the US GDP dumped to make these tools free, and i can't think of a single generated artifact, text, image, sound, video that had a significant cultural impact.

Feeling existential weariness today at so much academic effort promising to create AI systems and design frameworks that "augment rather than replace", support wellbeing sustainable just etc etc etc –– and yet, and yet, all it will achieve is Gutmenschen shelfware.

1. We ( @jbakcoleman.bsky.social, @cailinmeister.bsky.social, @jevinwest.bsky.social, and I) have a new preprint up on the arXiv. There we explore how social media companies and other online information technology firms are able to manipulate scientific research about the effects of their products.

This reads like what a reviewer 2 should have written, leading the handling editor to decide accordingly.

Imagine organisations took all the developer years they currently pour into AI assistants and used them to make their services more stable, accessible, performant, sustainable. The uplift in CSAT etc would be massive and long-term in comparison. The opportunity costs of AI are staggering.

Very pleased to see submissions to the special issue on Game Advertising of ACM Games: Research and Practice coming in. This is a last call! Deadline is 31 October 2025. We're looking forward to your submission, and if you need a smol extension… Slide into our emails. 😉 dl.acm.org/pb-assets/st...

on.ft.com/46M9mXI Why designers abandoned their dreams of changing the world. Free to Read.

In London, hate will never win.

Imperial opened its President’s PhD Scholarships (first deadline Nov 3): bit.ly/3JW2xuW Ping me to apply for a scholarship on: * Competence/empowerment/learning progress in games: models and DDA * The appeal of aleatoric uncertainty in games: models and tests ...

We’ll be hosting GachaCon 2026, a hybrid academic conference on gacha games with Genshin Impact as the theme, on 12–13 February 2026 at the School of Creative Media, City University of Hong Kong. 🇭🇰🇨🇳 Call for Papers: tinyurl.com/gachaconcfp

Great PhD/postdoc opportunity with a wonderful group: The Autotelic Interaction Group by @creativeendvs.bsky.social at Aalto has 1 PhD and 1 postdoc working on computational social intrinsic motivation – apply by Sep 30! www.aalto.fi/en/open-posi...

Annoyed by gross/sexist/AI/... slop in mobile game ads? Now is the time to write about it! Submit to @acmgames.bsky.social Special Issue on Game Advertising, submission deadline Oct 31. All details here: dl.acm.org/journal/game...

I find the delusion framing strange, which implies disordered perception. This is more innocent credulity of the mark of a conman? As in, the fault is in believing what the conman says, not in misperception or misconstrual.

Going back to 1969 to tell Foucault that in the future, everyone will know he was 100% right that "discourse” can produce itself without “a thinking … subject.” But also in the future we will care less about that principle than the practical consequence that students can cheat on homework.

“Between 2009 and 2012, iPhones had a built-in "Send to YouTube" button in the Photos app. Many of these uploads kept their default IMG_XXXX filenames, creating a time capsule of raw, unedited moments from random lives.” walzr.com/IMG_0001 this is beautiful 🥲

Want to work with the Smart Data Donation Service (sdruk.ukri.org/data/smart-data-donation-service/) on donated Google Play or YouTube data? UKRI has 6 fellowships for impact-driven research working with services like ours – do reach out! ukri.org/opportunity/smart-data-research-uk-fellowships/

📢The open access version of our book is available now via OUP's site: global.oup.com/academic/pro...

Who’s with me for a campaign to stop politicians wearing hi viz jackets and hard hats when talking about the economy and instead being filmed playing Total War: Warhammer

Your weekly UK HE equivalent of doomscrolling. A good sobering summary and assessment.

This is the first page I turn to when I’m showing people @molcher.bsky.social’s book about Judge Dredd and modern policing: shop.2000ad.com/catalogue/XB...

1. "'Trusting the experts is not a feature of either a science or democracy," Kennedy said." It's literally a vital feature of both science and of representative democracy. I've written a fair bit about trust in expertise as a vital mechanism in the collective epistemology of science.

Hey folks I am opening emergency commissions. My partners ebike was just stolen and they use it for work daily. I'll draw anything (Not hateful). My strengths are monsters, mice & maps. More examples below. Varied price points available. Please comment / message if you're interested.

Does cultural evolution generate more diverse and interesting work over time if fragmented into sparsely connected ecological niches? Eg, is cinema better for US/HK/JP/… cinemas having co-evolved with only sparse exchange over pre-globalisation decades?