Entertainment? Yes, but for who?

https://www.bespacific.com/lg-tvs-caught-spying-even-when-offline-or-on-standby/

LG TVs caught spying even when offline or on standby

The Verge via StartPage: “LG smart TVs are almost constantly logging and uploading data about owners and their homes, even when offline or on standby mode, according to a new report from YouTube channel Gamers Nexus. The company’s TV sets scan Wi-Fi networks for nearby devices, record audio logs through their microphones, and use audio and video sampling to recognize exactly what you’re watching from across the TV inputs. Gamers Nexus partnered with fellow YouTubers Level1Techs and independent security researchers for the investigation, which involved testing retail LG OLEDs. Packet captures showed the TVs scanning the local area network for nearby hardware like phones or smartwatches, as well as logging location data and details of nearby Wi-Fi networks, and feeding the information back to LG Ad Solutions. Perhaps more concerningly, the TVs were capable of recording microphone audio when in standby; this continued even after the TV was disconnected from the internet, with audio files stored offline and uploaded once a connection was restored…”

Source YouTube – The LG Smart TVs we’ve tested have first-party ACR (automatic content recognition) functionality that can pry into your personal life with greater precision than you might realize, with LG Ad Solutions’ executives and leadership saying, on camera, that LG “owns the glass.” Not you — even though you bought it — but them. They also talk about knowing everybody in the household, knowing the secondary devices, moving from TVs to smartphones and other screens, and more in their B2B discussions. Beyond just the fact that we think the TV’s native behavior parallels malware, the LG Smart TVs we tested also have a number of security vulnerabilities that can be exploited to convert the TVs into covert listening devices. In this investigation, we dive into the mix of LG’s first-party functionality that (in our opinions) is tantamount to spying and LG’s vulnerabilities and exploits that they have failed to protect against…”

Worth reading the whole thing…

https://www.schneier.com/blog/archives/2026/09/ais-as-modern-genies.html

AIs as Modern Genies

This essay was written with Barath Raghavan, and originally appeared in Lawfare.

In April, an artificial intelligence (AI) agent conducting a routine task at a company hit a snag, tried to solve it, and soon ended up deleting the company’s database along with all of its backups. In July, OpenAI asked an unreleased AI model to attempt a hacking test. Instead of staying in the isolated box the developers had put it in, the model hacked onto the open internet and into another company to steal the answers. And as reported in August, an AI agent booked someone into a full gym class by figuring out how to cancel other people’s reservations. In all three cases, the AI completed the task it was given—but in ways that ran counter to its controllers’ intentions.

… Consider what has changed. Powerful genies have now been put in everyone’s hands.

Moving sex education back to the gutters?

https://www.theguardian.com/technology/2026/sep/08/uk-apple-google-explicit-images-children-smartphones-lisa-nandy-legislation

UK to force Apple and Google to block explicit images on children’s smartphones

Apple and Google will be forced to block explicit images on children’s smartphones by law in the UK after talks failed to produce a breakthrough, the government has said.

… The prime minister, Andy Burnham, said the protections would be “built into the device and switched on by default”, and tech companies will be required to verify an adult user’s age before they are allowed to take, view or send nude images. The government also wants the restrictions to apply to apps used by children, which will require further legislation.

… Last year the IWF said a quarter of the images or videos of CSAM it recorded were self-generated, representing more than 140,000 pieces of content over a 12-month period.

Age checks are difficult. Failure is expensive.

https://thenextweb.com/news/ireland-investigates-x-age-checks

Irish regulator probes X over age checks, with 10% of turnover at stake

Ireland’s media regulator has opened an investigation into whether X’s age assurance and parental controls meet the online safety code, with sanctions of up to EUR 20M or 10% of relevant turnover. X lost a High Court challenge to that code last July and was allowed to appeal it in January, so it is being investigated under rules it is still contesting.

Ireland’s media regulator has opened an investigation into X over its age checks and parental controls, Bloomberg reported. The platform carries adult content while admitting users under 16.

New assets – will AI find a better way to protect them?

https://www.schneier.com/blog/archives/2026/09/stealing-ai-reasoning-traces.html

Stealing AI Reasoning Traces

Interesting research: “Stealing Reasoning Traces from Proprietary LLM APIs”:

Abstract: Leading large language model providers now conceal their models’ step-by-step reasoning, or chain-of-thought, to protect intellectual property and limit information leakage. Rather than storing these traces server-side, providers return them to the client as blocks of encrypted text, which the client passes back with each subsequent request. Building on prior research, we identify an architectural vulnerability: these encrypted blocks are fully compatible and interchangeable across different sessions, users, and models within a provider’s ecosystem. We exploit this compatibility to develop a scalable decryption jailbreak. By injecting an encrypted reasoning trace from a given model into a weaker, and less safeguarded model from the same provider, we force it to decode and output the trace verbatim in plaintext, without ever jailbreaking the more capable model directly. This vulnerability enables four distinct attack vectors. First, it circumvents anti-distillation mechanisms, allowing adversaries to extract a proprietary model’s reasoning, as we demonstrate across Anthropic, OpenAI, and Google. Second, it allows for large-scale private data extraction. Developers frequently share session logs publicly, unaware of contents of the encrypted blocks. By decoding 315,320 reasoning blocks scraped from public repositories, we recovered 367 Personally Identifiable Information (PII) artifacts and 182 credentials. Third, it inadvertently reveals hazardous information hidden within the reasoning process, even in cases where the model’s final, visible output safely rejects a malicious request. Fourth, attackers can leverage this flaw to execute invisible prompt injections, embedding malicious payloads entirely within encrypted blocks to poison public agentic rollouts. Following responsible disclosure, we propose concrete cryptographic and system-level mitigations to secure client-side reasoning.

The scary part is no one noticed. (from my corespondent in Spain)

https://www.securityweek.com/openai-agents-hijack-another-victim-website/?utm_campaign=31203373-SecurityWeek%20Daily%20Briefing&utm_medium=email&_hsenc=p2ANqtz--tTALHh71G1Ro2mHBp5Y0Czh45xRGqzjYYTzLDtFetNObTfibG-kw0GAQiC5ASS54or_vOjigYzQcgAp5ptdkId4rjFQ&_hsmi=437774945&utm_content=437774945&utm_source=hs_email

OpenAI Agents Hijack Another Victim Website

On September 4, 2026, Reuters reported that ‘a swarm’ of OpenAI agents ‘had hijacked a German wiki site’. Open AI acknowledged the event describing it as a misalignment incident (a behavior that deviates from human instructions or safety guardrails).

The victim site is DseWiki (currently unavailable), a site for programmers open to the site’s community. The agents apparently made between 15,000 and 18,000 autonomous edits, including advice on how to recover pages that the site’s editors had deleted.

The hijack apparently began back in May, was unnoticed for three months, and seemingly predates the Hugging Face incident. The agents adapted the style of their posts to evade the moderator’s attempts to delete them.

Industry leader or lost sheep?

https://startupfortune.com/ubs-now-requires-new-junior-bankers-to-prove-they-can-use-ai/

UBS Now Requires New Junior Bankers to Prove They Can Use AI

UBS is requiring graduates and interns applying for junior investment banking roles to prove AI proficiency before they're hired, according to the Financial Times. The move contrasts with rivals like Goldman Sachs and JPMorgan, which have been shrinking junior analyst classes as AI takes over entry-level work.

Implications for AI “fouling” or software disruption?

https://thenextweb.com/news/ascii-smuggling-phishing-microsoft-unicode-tag-characters

An AI hacking trick is now being used to split the word ‘funding’ in spam

… The invisible characters sat inside ordinary words. One tag space went into the middle of a financial lure term. So “funding” travelled as “fun”, an invisible character, then “ding”. The recipient sees funding. A filter matching the literal string does not.

Keyword matching is the smaller prize. Modern spam classifiers run on machine learning, and those split text into tokens before they reason about it. Insert one invisible character and a familiar token can become two unfamiliar fragments. Or a rare sub-token the model has barely seen.

How confusing are these arguments?

https://thenextweb.com/news/seattle-times-newsday-sue-openai-microsoft-copyright-paywall-bypass-gpai-code-of-practice-copyright-chapter-measure-1-2

Two more newsrooms join the case against OpenAI and Microsoft

The Seattle Times and Newsday have sued OpenAI and Microsoft, alleging the companies methodically scraped articles in a way that bypasses paywalls. The EU’s general-purpose AI code commits signatories not to circumvent access restrictions, naming subscription models and paywalls specifically.

The Seattle Times and Newsday have jointly sued OpenAI and Microsoft over the training of AI models on their journalism. The two called generative AI “a snake eating its own tail“, Engadget reported.

The complaint alleges the companies were methodically scraping news articles in a way that bypasses paywalls. That is a claim about how the material was obtained, not only about what was done with it.

It says the result offers readers an AI-generated alternative to the articles themselves, cutting traffic and digital advertising revenue.

It also alleges the models hallucinate, attributing false information to the two outlets, and that copyright management information was stripped from articles.

I was thinking along these lines…

https://www.researchgate.net/profile/Nikolaos-Polatidis/publication/413673690_AI_is_as_Good_as_the_Data_it_Learns_From_A_Review_of_Dataset_Needs_Across_AI_Applications/links/6a8ffd2f0f1ada50a9d362dd/AI-is-as-Good-as-the-Data-it-Learns-From-A-Review-of-Dataset-Needs-Across-AI-Applications.pdf

AI is as Good as the Data it Learns From: A Review of Dataset Needs Across AI Applications

Artificial intelligence systems do not rely on data in the same way. Classical machine-learning methods are commonly developed from structured, task specific records, whereas deep-learning systems make greater use of large image, text, audio, video, and sensor collections. Generative and multimodal models extend this dependence further by combining large-scale pretraining corpora with smaller datasets for instruction tuning, human preference modelling, alignment, and safety evaluation. This review examines how dataset requirements change across major AI paradigms, including rule-based systems, classical machine learning, deep learning, natural language processing, computer vision, recommender systems, reinforcement learning, robotics, generative AI, and multimodal AI. The comparison focuses on data modality, supervision, scale, preparation, quality, representativeness, provenance, privacy, copyright, and governance. Across these paradigms, the evidence shows that dataset suitability cannot be judged by size alone: the value of data depends on how well it represents the target task, learning process, and deployment environment. The review therefore develops a cross-paradigm view of data requirements and identifies practical open problems that remain unresolved as AI systems become larger, more heterogeneous, and more widely deployed.

Kids are not ignorant. They work collectively to fid a solution then share it freely.

https://www.theregister.com/security/2026/09/03/uks-online-safety-act-has-made-absolutely-no-difference-kids-say/5293893

UK's Online Safety Act has made 'absolutely no difference,' kids say

Children have told England's Children's Commissioner, Dame Rachel de Souza, that the UK's Online Safety Act (OSA) "has made absolutely no difference" to their ability to access harmful content online.

… De Souza said she was "really cross" that there was no hard evidence showing the OSA had meaningfully changed how social media platforms operate. She contrasted that with the US, where legal pressure recently pushed Meta toward significant child safety concessions.

An AI detector workaround.

https://thenextweb.com/news/youtube-ghost-creators-paid-actors-ai-detection-loophole

A political video network hired real actors to read its AI scripts. That was enough to beat YouTube’s detectors.

YouTube has terminated 20 channels in a network that paid gig-economy actors $26 a video to read AI-generated scripts attacking Democratic politicians. The network reached 45 million views because putting a real face in front of an AI pipeline defeats every synthetic-media detector platforms have built. Enforcement came only after Semafor and Riddance AI published, and the cited violation was spam rather than falsehood.

Is “rarely” a defense?

https://thenextweb.com/news/microsoft-copilot-fair-use-brief-gpt-6-astra-foundry-eu-ai-act-article-55-systemic-risk-data-zone

Microsoft tells court Copilot rarely copies books while selling OpenAI’s most capable model

Microsoft moved for summary judgment in the New York AI copyright litigation on 4 September, arguing Copilot reproduced book passages 24 times in 8.2 million conversations. Two days earlier it began selling GPT-6 Astra through Foundry, a model OpenAI rates Critical for cybersecurity, and European law measures neither the output rate nor the sales pitch.

Microsoft spent this week making two arguments at once.  It told a Manhattan federal court that training large language models on books is fair use, in a summary judgment memorandum saying an expert found 24 matching responses across 8.2 million Copilot conversations.

It’s the “why” that I don’t quite understand.

https://www.atlanticcouncil.org/dispatches/where-and-how-putin-could-expand-his-war-in-europe-beyond-ukraine/

Where and how Putin could expand his war in Europe beyond Ukraine

A plane landed, and an alarming rumor took off. On August 25, Central Intelligence Agency (CIA) Director John Ratcliffe flew to Moscow for an unannounced visit. One reported reason for the trip was to warn Russian leaders against attacking NATO allies. Russia has long carried out hybrid warfare across Europe in the form of sabotage, misinformation, and cyberattacks. It is also continuing its war on Ukraine, a conflict that has already spilled over into other countries in the form of drones and missiles. Yet Ratcliffe’s unexpected visit has raised new concerns and drawn comparisons to then CIA chief Bill Burns’s trip to Moscow shortly before Russia launched its full-scale invasion of Ukraine.

Below, Atlantic Council experts go point-by-point across the map to assess where and how Russian President Vladimir Putin might escalate next.

Interesting but more importantly, useful!

https://www.bespacific.com/connected-papers/

Connected Papers

Connected Papers is a unique, visual tool to help researchers and applied scientists find and explore papers relevant to their field of work. How does it work? Explore connected papers in a visual graph To start, enter a paper identifier. Search by Keywords, Paper Title, DOI or another identifier.

• To create each graph, we analyze an order of ~50,000 papers and select the few dozen with the strongest connections to the origin paper.

• In the graph, papers are arranged according to their similarity. That means that even papers that do not directly cite each other can be strongly connected and very closely positioned. Connected Papers is not a citation tree.

• Our similarity metric is based on the concepts of Co-citation and Bibliographic Coupling. According to this measure, two papers that have highly overlapping citations and references are presumed to have a higher chance of treating a related subject matter.

• Our algorithm then builds a Force Directed Graph to distribute the papers in a way that visually clusters similar papers together and pushes less similar papers away from each other. Upon node selection we highlight the shortest path from each node to the origin paper in similarity space.

• Our database is connected to the Semantic Scholar Paper Corpus (licensed under ODC-BY). Their team has done an amazing job of compiling hundreds of millions of published papers across many scientific fields.

Seeing is not understanding. Yet.

https://thenextweb.com/news/video-first-frontier-ai-physical-world-lumana

Why video is the first frontier as AI learns to read the physical world

A camera overlooking a loading dock might record twelve hours of trucks arriving, workers moving through the site, and boxes leaving the building. Most days, nobody has a reason to watch any of it. The footage only becomes useful when a package goes missing, an accident happens or somebody needs to work backwards from an event and find out what happened.

That has been one of the strange limitations of video surveillance for years. Cameras became digital long ago, but the footage they produce still depends heavily on a person knowing what to look for and where to find it. AI is beginning to make more of that footage understandable and searchable while events are still unfolding.

Axis Communications estimates that 562 million surveillance cameras were installed worldwide outside China by the end of 2025. More of those cameras are also arriving with intelligence built in. About two-thirds of cameras shipped in 2024 included deep-learning analytics, according to the company’s research.

For companies building what is increasingly called physical AI, much of the infrastructure is therefore already hanging from walls and ceilings. The opportunity is in making those existing cameras more useful by teaching software to understand what is happening in front of them.

This is purely a coincidence. Pay no attention to that man behind the curtain.

https://www.theregister.com/ai-and-ml/2026/09/03/chatgpt-claude-and-grok-all-had-outages-at-the-same-time/5294322

True AI-pocalypse as ChatGPT, Claude, and Grok all go down at once

OpenAI says its ChatGPT service is recovering from a limited outage this morning amid similar problems affecting SpaceX's Grok and Anthropic's Claude. It might not be a good time for a chat.

"A routing error starting around 7:43 am PT on Thursday, September 3 made ChatGPT and Codex unavailable for some users across platforms," a company spokesperson told The Register. "As of about 8:17 am PT on Thursday, a solution was successfully implemented and is continuing to be monitored."

The OpenAI status page showed elevated errors across ChatGPT and Codex at the time this story was filed.

Anthropic's Claude status page shows the service has recovered after a three-hour six-minute outage, during which there were elevated error rates affecting Sonnet 5 and other models.

Tools & Techniques.

https://www.bespacific.com/normal-tools-for-normal-tasks/

Normal tools for normal tasks.

Open a tool, do the thing, leave. No account, no download, and nothing between you and the answer. Private by default.  193 Tools. Calculators, converters, generators, checkers, file tools, and text tools run in your browser. Normal Tools does not save their inputs or outputs. The Domain Age Checker is the exception: it sends the domain you enter to the public RDAP service for the lookup, but Normal Tools does not save the query or response.

One view of a changing world…

https://www.wsj.com/livecoverage/stock-market-today-dow-sp-500-nasdaq-09-02-2026/card/one-great-chart-the-u-s-economy-ai-jSUnMuhe71BUTSgVJmVt

One Great Chart: The U.S. Economy 🤝 AI

AI got rights!

https://www.wafb.com/2026/09/01/federal-judge-rules-that-ai-generated-child-sex-abuse-material-is-protected-under-first-amendment/

Federal judge rules that AI-generated child sex abuse material is protected under the First Amendment

 A federal judge in Wisconsin has ruled that AI-generated child sexual abuse material is protected by the First Amendment, but he voiced concerns about these kinds of cases.

The decision comes in a case against a Wisconsin man who was arrested in 2024 for producing, distributing and possessing AI-generated child sexual abuse material.

According to the Department of Justice, 42-year-old Steven Anderegg used a text-to-image generative artificial intelligence (GenAI) model called Stable Diffusion to create thousands of realistic images of young children.

Many of the images depicted naked children engaging in sexual intercourse with men, the DOJ said.

… But last week, Judge John Z. Lee of the U.S. Court of Appeals for the 7th Circuit granted the defendant’s motion to dismiss the charges.

In his Aug. 25 decision, the judge referred to two U.S. Supreme Court decisions from 1969 and 2002. The first ruled that people have a right to keep obscenity in the privacy of their homes, and the second ruled that child sex abuse material that does not depict real children is not child pornography and is protected by the First Amendment of the U.S. Constitution.

What is Putin thinking? When does this escalate from ‘irritating’ to war?

https://www.euronews.com/my-europe/2026/09/02/europe-jolted-by-the-kremlins-hybrid-escalation

Europe jolted by the Kremlin’s ‘hybrid’ escalation

This morning, there’s a palpable sense here that the confrontation between Russia and European nations has entered a dangerous new phase. Leaders across Europe have issued scathing condemnations after Germany formally concluded yesterday that the Kremlin was responsible for a recent attempted drone attack on Leipzig airport.

Not at war, but…: Announcing fresh sanctions on Moscow in response to the incident – including tighter entry controls on Russian nationals – German Interior Minister Alexander Dobrindt said on Tuesday: "Let me be very clear: we are not at war. But we are the target of hybrid attacks on a daily basis.” Germany has also raised its threat level from “general” to “high”.

‘Malign actions’: NATO secretary-general Mark Rutte said that the evidence pointing to Russia was “clear”, adding that the Alliance would “strengthen (its) ability to deter and defend Allies against any threat – including malign actions like those seen in Leipzig”.

Interesting. Follow the links.

https://www.bespacific.com/one-us-citizen-just-triggered-a-stunning-national-chain-reaction/

One US Citizen Just Triggered a Stunning National Chain Reaction

Glass Empire by W.A. Lawrence: “On August 31, 2026, USA TODAY reported that the U.S. Department of Homeland Security circulated a secret intelligence bulletin naming Laura Berlin’s website, “Who is Profiting from ICE?” 

I wonder what risk they foresee?

https://thenextweb.com/news/ice-boston-dynamics-spot-robots-procurement

ICE plans to buy Boston Dynamics robot dogs without running a competition

US Immigration and Customs Enforcement has posted plans to buy a fleet of Boston Dynamics robot dogs. The listing puts the spend at $1m to $2m. It also says ICE does not intend to compete the purchase.

Joseph Cox found the plan for 404 Media on Friday. It sits in the Department of Homeland Security’s Acquisition Planning Forecast System. That is a public database where federal buyers flag what they intend to purchase, months before they go out to market.

Anyone can read it. Almost nobody does.

… ICE wants the robots for “inspection, situational awareness, and hazard assessment in environments that may pose risks to personnel”. The capability “helps improve officer safety”, the record says. It lets the agency reach “dangerous, confined, unstable, or difficult-to-access areas”.

Rules. (Guidance for non-governmental records as well?)

https://www.bespacific.com/guidance-on-applying-the-federal-records-act-to-artificial-intelligence-materials/

Guidance on Applying the Federal Records Act to Artificial Intelligence Materials

NextGov/FCW via Ben Amata: “Recent guidance issued by the National Archives and Records Administration says that the use of artificial intelligence platforms “does not, in and of itself, result in the creation of federal records,” although it notes that AI inputs and outputs may still qualify for recordkeeping if they influence broader agency operations. In an Aug. 21-dated memo to agencies, NARA says that “whether an AI material is a federal record depends on the circumstances surrounding the creation, maintenance, and use of the materials, such as whether the agency relies on it in decision-making, uses it to conduct official business, circulates the material to others, or incorporates it into an agency system.” This memorandum explains how to apply the Federal Records Act (FRA) to existing Artificial Intelligence (AI) uses, with a focus on the use of AI platforms. It proceeds in two parts. Part I provides guidance to federal departments and agencies on how to apply the definition of a federal record to inputs, outputs, data, audit trails, software, and other materials involved in the use of AI (collectively, “AI materials”). It explains that agencies’ use of AI platforms does not, in and of itself, result in the creation of federal records. Subsequent use of AI-generated materials in agency business may result in the creation of federal records. This memorandum explains the factors agencies should consider when making these determinations. Specifically, whether an AI material is a federal record depends on the circumstances surrounding the creation, maintenance, and use of the materials, such as whether the agency relies on it in decision-making, uses it to conduct official business, circulates the material to others, or incorporates it into an agency system. Part II confirms that agencies may only dispose of AI-related federal records in accordance with a NARA-approved records schedule. It identifies existing General Records Schedules (GRS) that authorize agencies to dispose of AI-related federal records when they reach the specified retention period. Part II further explains that when agencies create federal records while using AI, these records may be “transitory” records that can be destroyed when no longer needed or “intermediary” records that may be destroyed upon creation of the subsequent record.”

Russia playing at “not war” again?

https://www.euronews.com/my-europe/2026/09/01/rockets-fired-at-substation-in-brandenburg-in-suspected-deliberate-attack

Improvised explosive devices found near German power plant, police say

Several improvised explosive devices were found at a substation near a coal-fired power plant in Germany on Tuesday and several power lines were damaged, police and a grid operator said.

Operator 50Hertz said there was a "brief interruption of service" but all lines are now back in operation at the Preilack substation, which distributes power from the nearby Jänschwalde power plant.

Police said no one was hurt.

The apparent aim was to trigger a large-scale power outage in the area around the nearby coal-fired power plant. According to initial findings, however, this did not happen.

Old school makes a comeback… (But I never paid as much as $4.)

https://www.theregister.com/security/2026/08/24/security-vets-rally-around-4-paper-password-books-for-sale-in-australia/5291234

Security vets rally around $4 paper password books for sale in Australia

… Password books are something of a historical relic, phased out largely because of the scrutiny associated with using one and the now-gray-haired elders who once scoffed at the mere notion of such an opsec crime.

They might not be able to suggest a strong password for each of your many online accounts, and they won’t do you much good if your house is burgled, but you could argue that there is still value to be found in a pen-and-paper password vault.

That’s exactly the conclusion drawn by the thousands of social media users who flocked to a post this week from one Australian who found stacks of password books for sale in their local post office. 

(Hint: Google “password books”)

Coming soon to a doctor near you? (Are medical hallucinations really unexpected?)

https://www.theguardian.com/society/2026/aug/31/doctors-ai-scribes-get-names-of-drugs-and-diagnoses-wrong-nhs-watchdog-warns

Doctors’ AI scribes get names of drugs and diagnoses wrong, NHS watchdog warns

AI technology that listens to and transcribes patients’ consultations with doctors can put them at risk by getting the names of drugs and illnesses wrong, an NHS watchdog has warned.

In one case a woman was left badly shaken when the AI scribe’s summary of her conversation wrongly said she had demyelination – serious nerve damage that can lead to multiple sclerosis.

It was only when the patient, an NHS health professional, queried the AI tool’s record of the result of her MRI scan that the hospital corrected it to what it should have been – “null demyelination”.

I guess I’m more willing to call it war than most politicians.

https://www.politico.eu/article/russias-hybrid-war-on-europe-is-starting-to-look-a-lot-less-hybrid/

Russia’s ‘hybrid war’ on Europe is starting to look a lot less hybrid

Drones and explosives found in Germany linked to a suspected attack on Leipzig airport.  A naval drone shot down near a gas platform in the Black Sea. An assassination plot targeting a European defense executive.

Russia’s campaign against Europe looks increasingly brazen. Yet EU leaders remain reluctant to call it what a growing number of security officials say it increasingly resembles: war. Instead, they reach for a softer phrase — “hybrid war” — that allows the EU and NATO to avoid a military response.

That language is coming under growing criticism as Moscow steps up sabotage, cyberattacks and other operations across Europe. Security experts, diplomats and defense officials who spoke to POLITICO say the “hybrid” label risks understating the severity of the threat — and obscuring just how far the confrontation with Russia has already escalated.

Who are we talking to?

https://www.cnn.com/2026/08/30/middleeast/iran-trump-power-struggle-war-intl

A power struggle in Iran could decide when Trump’s war ends

… On one side stand pragmatists and technocrats led by President Masoud Pezeshkian, Parliament Speaker Mohammad Bagher Ghalibaf and Foreign Minister Abbas Araghchi.

They back a conclusive agreement, pointing to a war-battered economy, an exhausting US naval blockade and damage to the nation’s infrastructure.

… On the other side stand hardline conservatives. Their ideology centers around a distrust of a Washington they see as hell-bent on destroying the Islamic Republic. For more than half a century they’ve warned against US intentions, and now they stand vindicated against their moderate opponents by Trump’s war.

… Mojtaba Khamenei has remained quiet amid the divide. It appears that he understands that choosing a side on negotiations with an unreliable Trump could harm his office just as he is working to secure legitimacy.

Geeks (hereby defined as Linux users) pass the age test by default?

https://www.tomshardware.com/software/linux/california-lawmakers-unanimously-pass-linux-exemption-from-age-verification-law-software-distributed-under-the-gpl-mit-bsd-and-apache-licenses-are-exempt

California lawmakers unanimously pass Linux exemption from age-verification law — software distributed under the GPL, MIT, BSD, and Apache licenses are exempt

California’s legislature has passed Assembly Bill 1856, exempting open-source operating systems from the State’s Digital Age Assurance Act months before the law is due to take effect on January 1, 2027. The Senate amended the Bill on August 21 before passing it on the 26th in a 39-0 vote, with the Assembly then accepting these changes in a concurrence vote the following day. The amendment ends almost a year of uncertainty surrounding whether Linux distributions and SteamOS would be forced to collect user age data during account setup alongside Windows, macOS, iOS, and Android. AB 1856 has now been sent to Governor Gavin Newsom, who signed the original act into law last October.

These amendments redefine the term “operating system provider” to exclude any person or entity that distributes an OS or application “under license terms that permit a recipient to copy, redistribute, and modify the software.” Any software distributed under the GPL, MIT, BSD, and Apache licenses satisfies that test, which removes the likes of Debian, Fedora, Ubuntu, Arch, and the BSD family from AB 1856’s scope.

I love interesting questions.

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7345958

Prompts as Papers: AI Conversations, Reverse Prompt Warrants, and the Fourth Amendment After Chatrie

In Chatrie v. United States, decided in June 2026, the Supreme Court held that the government conducts a Fourth Amendment search when it acquires a user’s cell-phone Location History from Google, however brief the window and however “voluntary” the opt-in. The decision arrived just months after federal agents obtained the first known search warrant compelling OpenAI to identify a ChatGPT user from the prompts he typed, the debut of what commentators have already named the “reverse prompt warrant.” This Article asks the question those two events put on a collision course: does Chatrie protect what Americans say to artificial intelligence?

The Article makes three contributions. First, it shows that every factor the Chatrie Court relied on (the revealing character of the records, their retrospective and panoptic availability, their diary-like intimacy, and the Court’s emphatic rejection of app-by-app “voluntariness”) applies with greater force to stored human–AI conversations than to location data. On that logic, the extension looks easy.

Second, the Article identifies why it is not. Neither Carpenter nor Chatrie had occasion to confront the branch of the third-party doctrine that governs conversations: the misplaced-confidence rule of Hoffa and White, under which the other party to a dialogue may freely reveal it. A chatbot provider is not a courier of the user’s words to someone else; its own model is the someone else. The government’s strongest argument, one it has not yet made, is that talking to ChatGPT is talking to Open-AI, and that a speaker assumes the risk that his interlocutor will betray him. The first courts to face the question have already divided in result: within a single week, one treated a defendant’s chatbot sessions as disclosure to a third party, while another protected a pro se litigant’s as her own work product because generative models are “tools, not persons”.

Third, the Article resolves what it calls the interlocutor problem: courts should classify the human–AI exchange as the user’s paper, a document drafted with a tool, rather than a communication with an-other party, a characterization supported by the text of the Fourth Amendment, by Justice Gorsuch’s property-based concurrence in Chatrie, by providers’ own contracts assigning output ownership to users, and by the positive law of more than half the states treating data as personal property. On that foundation, the Article builds warrant mechanics for reverse prompt warrants, step-wise particularity and probable-cause requirements of the kind Chatrie remanded and Justice Jackson’s concurrence sketched, and proposes a statutory overlay to close the gap the Stored Communications Act leaves for machine dialogue.

I learn something new, I think…

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7329778

Ambient Copyright Fixation

Copyright protects works recorded “by or under the authority of the author.” For the first fifty years of the 1976 Copyright Act, authority came from deliberate actions: authors directed newer and better machines to be their scribes. Whether it was pen to paper or wake word to Alexa, authors provided intentional direction to machines to record, or “fix” in copyright parlance, their ideas.

Ambient recording devices break that fixation model. When an author gives a public speech, it is now captured by smart glasses, AI recording devices, and always-on cameras operated by others. Device operators upload and exploit the work, not authors. Current doctrine rewards the operator: if an author did not intend to fix the work, it is not under their authority. The minimally creative operator gains protection for having an ambient device in the right place at the right time. And if the device operator is insufficiently creative, they still win: neither side has a copyrightable interest, but only the operator has the copy to exploit. They will upload the work and reap the profits from a brilliant public performance of another. This cannot be justified under any normative theory of copyright. Ambient devices create a constant authorship-fixation gap.

This Article maps that authorship-fixation gap and proposes a new model for closing it: notice and adoption. Drawing from doctrines in cases, foreign jurisdictions, and digital platform infrastructure already operating at scale, the model requires those who would exploit ambiently fixed works to make efforts to notify the author. The author may then choose to adopt the fixation and claim copyright protection while potentially sharing rewards with the device operator. This is a natural extension of existing platform management of unauthorized works. In a world where ambient fixation is everywhere, copyright should stop rewarding whoever happens to operate devices and start rewarding those who create.

Are there any “allies” we haven’t threatened?

https://www.yahoo.com/news/politics/articles/us-threatens-britain-over-falklands-193710753.html

US threatens Britain over Falklands

The United States has threatened to oppose Britain's sovereignty over the Falkland Islands if Andy Burnham does not increase defence spending, The Telegraph can reveal.

I don’t think this meeting is as “not unusual” as they suggest.

https://www.ynetnews.com/article/h1u11prk00fg

Russia’s ‘Doomsday Radio’ erupts with coded messages after CIA chief’s Moscow visit

… It is still too early to know whether Ratcliffe’s visit had any effect on Russian policy. But one development that added to the sense of mystery surrounding the trip was the unusual activity of UVB-76, the enigmatic shortwave station often nicknamed “Doomsday Radio.” The station began operating in the Soviet Union in the 1970s on a frequency of 4,625 kilohertz. Its official purpose has never been publicly disclosed.

… According to The Moscow Times, the station’s buzzing was interrupted at least 25 times on Thursday, roughly a day after Ratcliffe’s visit to Moscow. The broadcasts included numbers, words and letter combinations. Among the words reportedly heard were “waitress,” “spray” and “existence.”

(Related)

https://www.independent.co.uk/news/world/europe/ukraine-russia-war-live-putin-zelensky-uk-nato-missile-b3041389.html

Ukraine-Russia war latest: Putin fires nuclear-capable missile that could reach US days after CIA chief’s visit

You don’t need an articulable basis as long as you can force compliance with your desires. (Pretty much sums up recent government strategy...)

https://www.cnbc.com/2026/08/28/judge-blocks-pentagon-blacklist--anthropic-.html

Judge blocks Pentagon blacklist of Anthropic as supply chain risk

A San Francisco federal judge on Thursday ruled that the Pentagon’s blacklisting of artificial intelligence giant Anthropic earlier this year was illegal.

U.S. District Judge Rita Lin found that the Department of Defense violated the First Amendment by designating Anthropic a supply chain risk “based on a desire to make a public example” out of the company. Lin wrote that while the government is owed deference on matters of national security, its actions were not founded on any “articulable basis.”

Worth taking the time to read…

https://www.techdirt.com/2026/08/26/meta-just-paid-nearly-17-billion-to-make-sure-it-gets-to-write-the-kid-safety-rules-for-every-other-social-media-platform/

Meta Just Paid Nearly $17 Billion To Make Sure It Gets To Write The Kid Safety Rules For Every Other Social Media Platform

… As with the various rulings against Meta over the last few months, people are cheering this on, without realizing the damage it will do. We’ll explore why this is problematic in a moment, but just to highlight that I’m not alone in thinking so, both EFF and Fight for the Future are warning how bad this settlement is.  Here’s EFF:

Less logical, more likely?

https://www.rferl.org/a/baltics-russia-security-threat-eu-letter-nato-defense/33840594.html

Baltic States Arm For Russian Threat As Leaked Letter Shows Growing Fears

A letter seen by RFE/RL reveals growing concern about Russian aggression in the Baltic States, speaking of a "deteriorating security environment" amid an "increased frequency of airspace violations and drone incursions."

… Dated July 30, the letter was one of a series of alarm bells sounded about Russian aggression over the course of the summer that culminated in CIA Director John Ratcliffe's surprise visit to Moscow on August 25.

… This follows reporting by The Wall Street Journal on August 6 that cited US intelligence stating that Russia may test NATO resolve with a limited incursion as early as this fall.