For my sins i am on a train with nothing more vital to do except read the UK/US MoU on the Big Beautiful Tech Prosperity Deal. I know, what a charmed life I live eh?1/?

Instead of a BritGPT could we have a little corner of some supercomputer for a BritTwit? I realise Im avoiding this site to get away from the sinkhole of US politics and frankly I don’t think Im alone.

Terrified that i asked a bunch of millennials this morning if it was actually real that Macron was offering proof of his wife’s gender & instead of LOLing they launched into a detailed socio-political critique of all dramatis personae. Who says kids dont follow the news?

You're just someone that I used to know

It is time for possible #Dagstuhl proposals again! Proposals can be submitted between October 15 and November 1, 2025. For links to guidelines, more details on the process, and important dates see www.dagstuhl.de/en/institute...

Yo tambien!

From @odihq.bsky.social the best AI news story ever

I just discovered the gov replaced the boss of the CMA with the Chair of the Turing Institute because that's going so well 👀

GikII – Geek Law share.google/Dc9prjTc3DmU... kicks off tomorrow in Amsterdam and for not the first time, but only the 2nd ( I think) I'm not there 😥 Enjoy! All who are. Next year we plan to celebrate a homecoming to Edinburgh alongside launch party for new edition Law, Policy and the Internet! 🙏😃

AI agents promise to “put our brain in a jar while a bundle of AI systems does our living for us”, writes Meredith Whittaker. “But as in fairy tales, so in life: relying on magical fixes leads to trouble,” she warns

Interesting to discover that OpenAI now does have a pretty comprehensive set of transparency reports nowadays except GUESS WHAT for numbers of people requesting DP user rights eg RTBF where, tumbleweed.... @daphnek.bsky.social

Wow, so long since I've been on ScotRail to Glasgow I didn't even know they didn't have buffet trolleys anymore. Civilization falls.

“ Kyle announced a £2.7 mn gov fund to help regulators pilot AI systems across sectors,inc energy, aviation and nuclear oversight” You keep using that word “regulation”, it does not mean what you think it does.. https://www.computing.co.uk/news/2025/ai/peter-kyle-pushes-for-ai-regulation-overhaul

Now I am out of that side of things, I cannot genuinely see how we square massively increased numbers of students with less than traditional qualifications/language standards, and staff redundancies, with the laborious intervention needed to make genAI useful/check plagiarism. It's lose lose lose

Very happy to be a signatory to this letter drafted by @hutko.bsky.social countering Trumpian misinformation on DSA. Please read thread. I also note Martin's kind offer to provide DSA training to US politicians at 1% cost 🤣

I think there's a genuine case for killing the REF this time round. Not just the waste of money & distortion of hiring etc etc. It reinforces league tables which continue to encourage elite unis to trawl in all PGTs and UGs in the wild, destroying further the economics of unis further down the pole

Can u Canute? Trump can!

Cory Doctorow on Google antitrust case: "Out of misguided privacy nihilism .. Google to be forced to share the data it stole from us, so that potential competitors could tune their search tools on the monopolist's population-scale privacy violations. And that is what the court has ordered."

Honey, weve all been there.

ha thats entirely my handwriting on the right, all in caps because my father was a doctor and this is the only way I am genetically related to him.

#CfP: Special Issue “Celebrating 60 Years of ELIZA? Critical Pasts and Futures of AI” of the Weizenbaum Journal of the Digital Society (WJDS), co-edited by @strippel.bsky.social & Magnus Rust. Submission deadline: November 17, 2025. More: www.weizenbaum-institut.de/news/detail/...

When did we get to the point where the use case for future financial security of AI is fake friends you drive you crazy as opposed to, say, folding proteins in n-space? Did it happen while I was on holiday? 👀🙄

What a title! Kudos to the authors 🤣

👇

If true ( it seems horrendously anthropomorphic) this would have really bad implications for AI audit in the automated public sector. Any thoughts?

"powerful AI” encourages and permits management to replace or degrade jobs they might not otherwise have. More important than the technological change, perhaps, is the change in a social permission structure." www.bloodinthemachine.com/p/ai-killed-...

Amusing holiday anecdotes Pt 29: the WiFi password for this resort is a misspelt version of the resort name. Not sure if this is brilliance or just typical Sri Lanka ..

Jesus. If they want to teach them about loss, could they not just get them a cat/dog? ( Incidentally there is a long lit on emotional trauma from support terminated Internet of Junk - @andreamm.bsky.social ?

AI-generated errors in an Australian murder case. We'll probably see an influx of ineffective assistance of counsel petitions/appeals soon arguing AI-usage. apnews.com/article/aust...

Agree. The way the counter culture works is fascinating.

and not to bring it back to AI but seriously AI cannot do that, an AI thrown up story has no intention. you can't have a conversation about why it did anything because there's no reason behind any of the choices beyond probability.

Amusing holiday anecdotes PT 2 : 12th century Sri Lanka paintings on plaster often feature mistakes like 3 arms or extra nipples... "Clearly AI!" Said I to an uncomprehending massive..

Amen, says she from a tiny village outside Glasgow

With no malice ( it's gorgeous ), trying to get service in a nice Sri Lanka hotel is very like asking ChatGPT to do what you want and not what it thinks you want very very slowly.

I cannot believe the world we're in where the major bug with software that can't add up and makes things up is that people keep falling in love with it. Actually this says way more about modern society than it does about OpenAI TBF.

The government should just pay for a £100pm discount for anyone buying Mountjaro etc. at this point, absolute no brainer for value.

I am 60 and I remember learning British pre-decimalization currency (because our current system only came in when I was 7) and IT WAS JUST LIKE THIS AAARGH

Much of interest in this detailed investigation into Amsterdam's failed AI welfare fraud risk scoring tool. Fascinating data on how terrible *humans* are at assessing welfare fraud risk, not just AI. Defend the status quo at the peril of the powerless. www.technologyreview.com/2025/06/11/1...

“Delete your old emails to save water” isn’t bad because it’s individualising the climate crisis, it’s bad because it’s complete nonsense from top to bottom. It’s like saying “eat less lettuce to save water”, you can see the thought process but it’s that of a child who knows two facts

The battle against bot scraping doesn't just involve legal squabbles. Some website owners are fighting back with AI scraper-trapping mazes and ways to poison training data. My feature for @newscientist.com www.newscientist.com/article/2490...

AI tragedy of the commons. Can robots AI preferences be declared at this level of granularity? datatracker.ietf.org/wg/aipref/ab...

I am immensely tired of fucking paywalls. Could all the newspapers not form a consortium for a price per month. They could call it oh Newsflix? 🤣🤣

*splutter*

Important thread, read to end!

Humans are anthropomorphic machines

www.ft.com/content/dfdc... "In tests, GPT-5 was more accurate than other OpenAI models and less likely to “hallucinate”, or refuse to answer queries. " Looking forward to a tide of counter examples in 3..2..1...

There is a risk that this will get missed over the summer, so just to flag up this important call for evidence on human rights and the regulation of AI from the Joint Committee on Human Rights (and deadline for response is 5 September) - committees.parliament.uk/work/9220/hu...

MP sets up AI chatbot rather than talk to constituents. Nothing could go wrong here…. https://www.neural-voice.ai/mark-sewards Mark Sewards - AI Assistant

My take on the Shanghai World AI Conference last weekend, and the US and China AI "action plans," dueling at least in the headlines… Out now on @techpolicypress.bsky.social podcast, with a tsp of jetlag-brain but covering a lot of ground. Thx @justinhendrix.bsky.social! www.techpolicy.press/podcast/

I'd quite like to know what the official answer to this is. You'd hope it was because they're hoping to use some zero knowledge solution but I v much doubt it..