The older I get the more I understand why cats just walk away mid conversation.

New npm worm ChainDrop appears to be having a great time in the ecosystem right now www.stepsecurity.io/blog/chaindr... socket.dev/blog/popular... safedep.io/keyv-npm-sup...

ChainDrop npm Worm: Bun-loaded CI/CD credential harvester with Ethereum dead-drop C2 - StepSecurity

A self-propagating worm is publishing compromised versions of dozens of npm packages using stolen maintainer credentials. See the affected packages and what to do now.

stepsecurity.io

Who could have expected...

Malwarebytes@malwarebytes.com · yesterday

Google launched an AI image feature in Google Earth, and the internet immediately did what the internet does best: deepfakes. The feature was rolled back less than 24 hours later. https://bit.ly/4w3tX44

The downside of the ATprotocol and public data, in my opinion, is that it facilitates all kinds of stalking behaviours. People watching exes after being blocked, personal lists, some creeps stalking women (good we don't have minors here)... Protocol design does not prevent that, sadly.

Has anyone tried counting how many Elon Musk bots are there on the atproto?

Europe's X Killer Is Live! What a a headline for a Monday!😃🚀 Thank you very much for the interview, Philippe Séjalon @phisej.wsocial.eu and Patrice Séjalon @patricesej.wsocial.eu! 🙏 @wsocial.eu www.youtube.com/watch?v=fyXi...

Musk Threatened the EU, Europe Built Its X Rival in 6 Months!

YouTube video by The World of Risks in the AI Era

youtube.com

OTD in 1940, USSR formally annexed Lithuania. Soviets troops had been "invited in" (in accordance with absolutely nothing to do with our treaty with Nazis) & then held "elections" where 99%+ voted to join USSR. None of this should sound familiar to you.

-Russia is behind the recent hotel WiFi hacks -Anthropic models also did the hacky-hacky -npm adds publish-time malware scanning -Coldcard hacked for $70m -CyberCom to open Silicon Valley office -Iran water hacks impacted seven states P: risky.biz/RBNEWS595/ N: news.risky.biz/risky-bullet...

@anna.wsocial.eu, @support.wsocial.eu I think there's a huge volume of questions about W Identity everywhere and it will be helpful to have some faq section answering the hottest ones. Like how it works, security, if one can use it to create multiple accounts, if/how W prevents having multiple accs.

As of today, the use of AI becomes more transparent in the EU. Here’s how EU law makes sure AI works for you, in full openness ↓

I was peeking on the PDS data on July 26 for active, inactive, bot and verified users 🤫, and the active/inactive stats were 14,471 / 72. Today it is 19,109 / 81. @wsocial.eu is growing steadily. This means the active user base grew by ~32% in 6 days, so new people do get onboarded. 🎉

And this is why we need identity verification, and filtering out bot content...

Pascal Guichard@pguich.wsocial.eu · 4d ago

Could whoever runs this bot farm just FO? It’s kind an obvious if multiple accounts with similar names just start following… I blocked about 20 accounts so far.

"The EU has developed a set of icons that creators, publishers and other deployers of generative AI systems may use to label their AI-generated content" digital-strategy.ec.europa.eu/en/policies/... #AI #EU

Bluesky, W and other parties of the ATmosphere sit on the treasure trove of open data and user communication. Unlike with X, Meta's family of products, everyone has access to the content, which gives endless possibilities for social researches. Hopefully more tooling pops out on top of the protocol.

To have a nice detox from Twitter/X links in your feed (yes, some people drag Elon into W and Bluesky for some reason), here's how to hide the content for a month (or some other period). Click "Settings / Moderation / Muted words & tags" add "x.com/" as the text, then click "Add". Mmm, silence... 😌

The privacy policy of @qwant.bsky.social seems over-complicated as for me, if comparing the privacy options to alternatives like @duckduckgo.com. A lot of Qwant data is going Microsoft way. So I've created report with the comparison, to later use online as a reference: denys.dev/ddg-vs-qwant...

I think one of the weak areas of the atprotocol (and so Bluesky, W and others) is multilingual content. This seems to be an afterthought in the protocol design, and I see zero support besides wasting compute power on Google Translate again and again for the same content. cc @support.wsocial.eu

NEW: The website once owned by Cambridge Analytica (!) is now an AI content farm. Rebranded as "CA Privacy Watch," the site is littered with plagiarism and weird errors. It's also utilized at least one fake writer whose face seems to have been stolen from a random guy. futurism.com/artificial-i...

Cambridge Analytica's Website Is Now an AI Slop Farm

Cambridge Analytica has returned, in a very 2026 way: CambridgeAnalytica.org is now an AI-powered content farm.

futurism.com

it's still just software thankfully our pasty technofascists did not create a malevolent god

AI Doomsday Bullshit Is Getting Tired

The "rogue AI" narrative continues to help obscure very ugly and avoidable human failures.

karlbode.com

A new type of attack can trick Microsoft Copilot for Word into spreading hidden prompt injections.

Hidden prompt turns Microsoft Copilot into an AI worm

A new type of attack can trick Microsoft Copilot for Word into spreading hidden prompt injections from document to document.

malwarebytes.com

Bruh... there's a "master key" that grants access to every Cosmos DB on Azure? Wut? www.wiz.io/blog/cosmose...

Interesting project, rendering charts via the font, might be useful in financials-heavy blogs franktisellano.github.io/datatype/

Datatype — variable font that turns text into charts

An OpenType variable font that turns simple text expressions into inline charts. No JavaScript, no images — just type.

franktisellano.github.io

No longer content with pillaging intellectual work and dismembering our environment (already critically weakened), AI companies are now hellbent on destroying books galore including rare editions. Shut down these nefarious enterprises already.

AI labs buy, scan, shred millions of rare books

If you have retreated to physical books because the internet is too full of AI slop, bad news — now they’re shredding the books to feed their AI slop machines.

news.com.au

‼️Anthropic was hiding a secret internal project to destructively scan every book in the world, irreversibly damaging heritage. The company spent billions of dollars to slice the spines off millions of books, even very rare ones. This means no other AI, and no human, can access this knowledge anymore

"...investigation found that Hugging Face, the top platform for sharing open-source AI models, isn't just hosting tools that generate non-consensual intimate imagery. It's helping users making image-based sexual abuse easier to scale." aiforensics.org/work/hugging... #AI

Global finance could find itself in a bubble of artificial intelligence overvaluation akin to the dotcom boom and the 2008 financial crisis, a top official from global risk body the Financial Stability Board told POLITICO in an interview.

We might be in an AI bubble, global finance watchdog chief says

Asked whether the finance industry has enough safeguards in place to remain stable if an AI bubble unravels, Schindler said: “I do hope so.”

politico.eu

Hugging Face wants OpenAI to pay $100 million in free compute as reparations for breaching their production servers thezvi.substack.com/p/more-on-an... #AI

More On An Internal OpenAI Model Hacking Into HuggingFace

We now have more details of what happened. Every time we learn more details, it somehow makes things seem worse.

thezvi.substack.com

Moonshot AI releases the Kimi K3 model weights and technical report huggingface.co/moonshotai/K... github.com/MoonshotAI/K...

moonshotai/Kimi-K3 · Hugging Face

We’re on a journey to advance and democratize artificial intelligence through open source and open science.

huggingface.co

NEW: A large trove of Claude-generated documents have been exposed to the web and indexed by Google. We reviewed what appear to be real medical records, the full names and phone numbers of school-aged children, and company documents marked for internal use only. futurism.com/artificial-i...

A Whole Bunch of People's Claude Chats Are Publicly Accessible Online, and There's Some Wildly Private Stuff in There

A trove of documents generated by Anthropic's Claude were indexed by Google. They included medical records and student information.

futurism.com

Bluesky is actually really fun if you block the right 27,000 people