A truly stunningly bad decision by whoever is doing product management at Google for Google Earth: " If you have never verified anything in your life, here is why a less playful mapping app matters. When a photograph […] https://www.digitaldigging.org/p/how-to-plant-a-nuclear-plant-in-iran

How to plant a nuclear plant in Iran

The question is: what on earth is Google doing?

digitaldigging.org

Yet again! This keeps happening! This system built by Avon and Somerset Police turns out to have terrible results, mostly built around signals that are proxies for […] https://www.wired.com/story/british-police-built-a-sprawling-crime-prediction-machine-some-results-couldnt-be-trusted/

British Police Built a Sprawling Crime-Prediction Machine. Some Results Couldn’t Be Trusted

As UK police embrace the AI revolution, a WIRED investigation reveals the messy inside story of one region’s experiment with predictive analytics.

wired.com

"fus__ro_dah" != "fus_ro_dah": "One missing underscore in a Skyrim-themed username put an innocent Nova Scotia man in prison for 18 months. Police were looking for a man using the […] https://arstechnica.com/tech-policy/2026/07/police-missed-one-underscore-and-sent-the-wrong-man-to-prison/

A missing underscore sent innocent man to prison for 18 months

When the first step goes wrong, bad results follow.

arstechnica.com

A prompt injection attack on Microsoft's Copilot app suite, with MS painfully failing to address the vulnerability in a useful manner: "Microsoft successfully mitigated the originally submitted PoC prompt, and […] https://enklypesalt.com/posts/context-collapse-part3-ai-worming-through-word/

"Trail of Bits Claude Code skills for security research, vulnerability detection, and audit workflows" -- recommended skills for security auditing using Claude. https://github.com/trailofbits/skills

GitHub - trailofbits/skills: Trail of Bits Claude Code skills for security research, vulnerability detection, and audit workflows

Trail of Bits Claude Code skills for security research, vulnerability detection, and audit workflows - trailofbits/skills

github.com

Absolutely glorious ancient-Irish factoid from @selkies.bsky.social: "in college my History of Translation lecturer (Michael Cronin) told us that one of the first vernacular translations of The Iliad was into Irish, and the […] https://bsky.app/profile/selkies.bsky.social/post/3mrozbdp3d22p

Louise (@selkies.bsky.social)

Is this a good time to mention that in college my History of Translation lecturer (Michael Cronin) told us that one of the first vernacular translations of The Illiad was into Irish, and the translator had to give the main characters dogs because in Gaelic society a free man had a dog

bsky.app

This is grim: "But the theory that someone used AI to develop an advantage in a prediction market is entirely plausible. The Numbers experience shows us that: - We now live in a world where a movie […] https://stephenfollows.com/p/what-just-happened-to-thenumberscom-should-worry-us-all

What just happened to TheNumbers.com should worry us all

If you work in or around the film industry, there is a decent chance you have used the work of The Numbers this month, whether you realise it or not.

stephenfollows.com

"The home appliance giant LG Electronics USA said this week it plans to suspend any apps built for its smart TVs that turn one’s television into an always-on residential proxy node. The move comes less […] https://krebsonsecurity.com/2026/07/lg-to-ban-residential-proxies-from-smart-tv-apps/

a decent bufferbloat internet connection quality tester; tipped by Alexey Shipilev. I need to work on tuning mine; I'm getting a "B" ("Good under load") with a +53ms latency increase during heavy downloads. https://test.libreqos.com/

LibreQoS Internet Quality Test

Measure speed, latency, bufferbloat, and real-time app quality under load.

test.libreqos.com

Heh, I love these ghetto-tech self-hosted LLM setups. "There’s a server in my basement that has no business running a modern language model. It’s a repurposed HP StoreVirtual […] https://www.neomindlabs.com/2026/06/08/running-gemma-4-26b-at-5-tokens-sec-on-a-13-year-old-xeon-with-no-gpu/

Rachel Coldicutt: "All of this [p(doom)] myth-making and rhetorical bluster is a just a narrative trick: the hidden object is not a technology, but a bid for power. This is a plot twist familiar from Greek myths, cautionary tales, and […] https://hdsr.mitpress.mit.edu/pub/wz35dvpo/release/2

This is absolutely comical -- "Role tags were a formatting trick that became the security architecture and the cognitive scaffolding of modern LLMs": "We call the attack CoT Forgery: injecting fake reasoning into a user message or tool output. We […] https://role-confusion.github.io/

Prompt Injection as Role Confusion

LLMs can't tell who's speaking. We show they identify roles by writing style, not tags, and exploit this with CoT Forgery, injecting fake reasoning that models mistake for their own thoughts.

role-confusion.github.io

Customers of the UK Sky Broadband ISP spent most of the 23rd of June unable to access the main NHS website and app: "customers of Sky Broadband reported that they were […] https://www.ispreview.co.uk/index.php/2026/06/sky-broadband-users-accidentally-blocked-from-uk-nhs-website-and-app.html

"A live bird collage from your window" -- this is absolutely lovely. A wood-framed, colourful e-ink display which collages nearby birds (identified by their song), it's really very nicely done. Pity the e-ink displays are still so spendy :( https://github.com/Twarner491/AvianVisitors

GitHub - Twarner491/AvianVisitors: A live bird collage from your window.

A live bird collage from your window. Contribute to Twarner491/AvianVisitors development by creating an account on GitHub.

github.com

LLMs as cultural technologies, encoding language in a new way: " For some years now, I have been saying to anyone who'll listen that the best way to think about large language models and their kin is due to the great Alison […] https://bactra.org/weblog/feral-library-card-catalogs.html

This explains a phenomenon we see the world over -- empty commercial real estate staying vacant for years at a time: "The short answer is both simple and surprising: in many cases, lowering the rent on a building will * […] https://www.freerange.city/p/why-do-commercial-spaces-sit-vacant

Why do commercial spaces sit vacant?

Understanding “extend and pretend”

freerange.city

really detailed write-up of how BrightData's scraping SDK is embedded in various mobile devices and TVs running on residential broadband networks, then being resold as " […] https://blog.includesecurity.com/2026/06/the-smart-tv-in-your-livingroom-is-a-node-in-the-aiscraping-economy/

a Thoughtworks write-up of a production LLM-based system architecture in place in Bayer, where an agentic RAG system is used to improve the user experience of searching historical nonclinical study reports. Lots of fairly […] https://martinfowler.com/articles/reliable-llm-bayer.html

Building Reliable Agentic AI Systems

AI helping pharmaceutical researchers query decades of information buried in PDF reports

martinfowler.com

Very thought-provoking essay on conspiracies and the derangement of American politics: " Republicans will never lose another election for the foreseeable future. To be more specific, they will never admit defeat. Largely inspired […] https://www.liberalcurrents.com/everything-is-glonzo-now/

Everything Is Glonzo Now

Populism and conspiracism are not quite synonymous but they are inextricably linked. Yet liberals must become populist again.

liberalcurrents.com

Bert Hubert: "If we want to discuss how to improve our digital autonomy, employees from US big tech will not usefully contribute to the conversation. They can’t. And in fact, they’ll likely […] https://berthub.eu/articles/posts/do-not-invite-big-tech-to-your-digital-autonomy-discussion/

Do not invite big-tech to join your digital autonomy discussion - Bert Hubert's writings

Some time ago, I attended an event to discuss European digital autonomy and digital dependencies. I accepted the invitation without checking too carefully, and I belatedly found out that most of US big tech was invited as well, and would be presenting their thoughts. Now, even in our gloriously digitally autonomous future, we’ll still be doing business with Microsoft, Google and Amazon. And it is fine if they have a role.

berthub.eu

"For two decades, Meta had a unique, high-performance engineering org; right up until around April of this year. For the first 20 years of the company’s existence, it had a “move-fast-and-break-things” […] https://newsletter.pragmaticengineer.com/p/why-is-meta-destroying-its-engineering

Why is Meta destroying its engineering organization?

Leadership at the social media giant has been on an AI-fueled rampage through its engineering org. We report what’s happened

newsletter.pragmaticengineer.com

Entirely predictably, Microsoft's Copilot LLM could be used to steal data from their email/calendar etc. due to guardrail failure, via this exploit: " One guardrail built […] https://arstechnica.com/security/2026/06/critical-copilot-vulnerability-allowed-hackers-to-seal-2fa-code-from-users/

Critical Copilot vulnerability allowed hackers to steal 2FA code from users

SearchLeak exploit shows why the industry's approach to LLM security fails over and over.

arstechnica.com

This is the absolute epitome of 21st century data protection woes. "This might be one of the most insane scandals in game history: when Pokemon Go players scan PokeStops, theyve been unknowingly building a detailed […] https://bsky.app/profile/scherawyss.bsky.social/post/3mnyucxuets2z

Schera Wyss 💀 SLUG WORLD 2026 (@scherawyss.bsky.social)

This might be one of the most insane scandals in game history When Pokemon Go players scan PokeStops, theyve been unknowingly building a detailed visual model of the world which is being sold to a military contractor to build a no-GPS positioning for the new generation of unmanned killing machines

bsky.app

This is a major liability judgement against Google's use of AI: "A German regional court has ruled that Google is directly liable for false claims in […] https://the-decoder.com/landmark-german-ruling-declares-googles-ai-overviews-are-googles-own-words-and-makes-it-liable-for-false-answers/

Landmark German ruling declares Google's AI Overviews are Google's own words and makes it liable for false answers

A German regional court has ruled that Google is directly liable for the content of its AI search overviews. According to the court, previous limited liability protections for search engine operators don't apply to AI overviews. In this case, Google's AI had falsely linked two publishers to fraud and made claims that didn't appear in any of the linked sources. The ruling could set a precedent for AI-generated content liability worldwide.

the-decoder.com

This is a nice term from Katryna Peart to describe one of the corrupting factors LLMs introduce to documents: "narrative flattening": "When US […] https://www.publictechnology.net/2026/06/09/communities-housing-and-planning/what-uk-councils-stand-to-lose-when-ai-enters-a-reorganisation/

Omnilert, which sells an "AI gun detection" system, sued after it failed to detect a gun prior to a January 2025 school shooting. Turns out accuracy matters! […] https://arstechnica.com/tech-policy/2026/06/school-shooting-survivor-sues-ai-gun-detection-firm-after-system-failed-to-spot-weapon/

School shooting survivor sues AI gun detection firm after system failed to spot weapon

How accurate does an AI system need to be?

arstechnica.com

this is pretty much the plot of Charlie Stross' "Halting State" (2006): teenagers are being recruited online by the FSB, in online forums and in-game chats, then assigned alternate-reality-game-style "tasks" in the real world which […] https://mastodon.social/@adrianhon/116696642949969646

Adrian Hon (@adrianhon@mastodon.social)

Attached: 1 image Distributed terrorism as pervasive game/ARG isn't new, but it looks like it's here to stay. This sort of thing was a staple of near-future sci-fi back when "collective intelligence" was cool – think Global Frequency – but it was usually for good. Oh well! Archive: https://archive.is/L16IJ

mastodon.social

Interesting Kafka trick: "Kafka Interceptors have quietly existed since 2016, yet most teams overlook them as an observability superpower. This article shows how Kafka Interceptors, combined with […] https://thej3.com/kafkas-quiet-observability-superpower-kafka-interceptors-aca88c33867e

"fully offline, human-powered local AI" -- an LLM and a voice model, running on a Raspberry Pi 5, driven off hand-cranked electricity generation! I was very sceptical, but they've put in the work to optimise the platform and choose models very […] https://squeezlabs.github.io/handcrank/

I spotted Pitt Bros using this charcoal at a recent event, so that's a plug for me https://fairbrotherecofuelsandbarbecues.com/product/grill-fanatics-restaurant-grade-marabu-charcoal-10kg/