Leftovers

• Vereis ☛ My Take On Free Will

  Free will is one of those topics that gets people really worked up.

  A lot of people seem to dislike this notion, but I don't understand why—I don't believe free will exists.

• Tommy Palmer ☛ Is It Shorts Weather Today?

  Back in May of 2011, I was 22 and had just recently finished University. Not working weekends for the first time since I was 17 was glorious. Summer was a few weeks away, but I was already enjoying all the Sun I could.

  Belfast weather during the summer can be sunny, and might reach a level you could describe as warm or even hot.

  During an evening in May, a friend and I were talking about how we couldn’t decide if we should’ve worn shorts that day. We joked that there should be a single serving site for this purpose.

• Russell Graves ☛ EBR-I Museum, Nuclear Jet Engines, and Automation

  The promises of the [Internet] I grew up with have gone whistling past, in the mad rush to addict users to everything, to fight other tech companies for the most addictive apps, to collect the most data, and now, to feed any and all human generated content to the gaping maw of the AI engines for “training.” I’ve, out of curiosity, run an “AI trap nonsense generator,” with a robots.txt file that simply says, “All automated crawlers, stay out.” So, 1.5 million page scrapes later, I think the point has been made. I know I’m feeding that, and I’m tired of it. With the exception of specifically marked AI text in a post of mine exploring early ChatGPT, everything I’ve written has been written by me, a human, with no AI assistance. So, I shall go pursue those routes in a much more aggressively offline way.

  I may be back. I may not. I really, really don’t know right now, but I’ll certainly update the blog if I decide to write on the [Internet] again.

  So long, and thanks for all the fish!

• ARRL ☛ Local, County, and State Governments Proclaim Value of Amateur Radio

  “While ARRL Field Day is a fun, social, occasion to get together and get on the air, it also serves as an opportunity to test equipment in a way that it would be needed in a time of crisis. The same people who come to visit your site under blue skies are the community members who would be served in an identical manner during and after an emergency,” said ARRL Public Relations and Outreach Manager Sierra Harrop, W5DX.

• Science

  • Matt Wedel ☛ Come hear Matt and Mike at Darren’s DinoCon!

    The first ever DinoCon will be this summer — Saturday 16th & Sunday 17th August at the University of Exeter. It’s the successor to the popular TetZooCon, which ran from 2014 to 2024, but with more of a focus on palaeontology in particular.

  • Cell Press ☛ Manipulating trapped air bubbles in ice for message storage in cold regions: Cell Reports Physical Science

    All experiments are conducted in a closed dark room at 20°C ± 1°C room temperature and 30% ± 3% relative humidity. In experiments, deionized water is injected into the Hele-Shaw cell in advance. By adjusting the input voltage of the semiconductor plate, the plate temperature can be controlled. Here, the plate temperature ranges from −15°C to −35°C with a 5°C internal temperature. The freezing direction or the inclination angle of the Hele-Shaw cell is adjusted from 0° to 180°, with a 45° internal temperature. Since the set plate temperature is well below the freezing point of deionized water, the water film may start to freeze before the plate temperature reaches the set value. However, the cooling stage occupies less than 6% of the total time in each freezing experiment (Figure S17), and thus it is reasonable to ignore the influence of the cooling stage on the experimental results.

  • New Scientist ☛ Asteroid 2024 YR4 is on collision course with the moon and could fire shrapnel at Earth

    Wiegert and his team calculated that 2024 YR4 could create a kilometre-wide crater on the moon – the largest lunar impact for at least the past 5000 years, though relatively small compared with a typical crater. An impact of this size would eject a cloud of debris into space, and by simulating its potential behaviour 10,000 times, the team found that this could cause Earth’s satellites to experience a level of collisions equivalent to what we would expect to see in years or even decades, but occurring in just a few days.

• Career/Education

  • Blake Watson ☛ 40

    The past decade has brought with it the increasingly accelerating passage of time. After transitioning to remote work in 2018, the days, weeks, and months blended together and flew by. The beginning of the pandemic feels like a couple of years ago, not five. I need to break up the monotony and slow things down.

  • Vox ☛ AI cheating: With more students using ChatGPT, what should teachers do?

    In other words, while the purpose of other departments is ultimately to create a product, a humanities education is meant to be different, because the student herself is the product. She is what’s getting created and recreated by the learning process.

  • University of Michigan ☛ Mapping major trends at UMich

    These nationwide trends are reflected in the University. Other than plant biology and pediatric dentistry, the majority of the University’s 10 least popular majors are in the humanities field. This seems to follow international trends of declining humanities popularity. In the U.S., humanities majors have dropped by 25% since 2012. Furthermore, throughout the last decade, Forbes reported an almost quarter decline in English graduates, a 15% decline in religious and philosophical studies graduates and a 5% decline in foreign language graduates — aligning with University trends.

• Hardware

  • Digital Camera World ☛ We obsess over specs, but no one asks how the camera "feels" | Digital Camera World

    That’s why some people still swear by their old Fujis, their beat-up Leicas or even a creaky old DSLR that feels like home. Not because it’s the sharpest or the fastest, but because it gets out of the way and enables them to make photographs.

  • Daniël de Kok ☛ Intel MKL on AMD Zen

    Disclaimer: this post investigates how recent MKL versions behave on Zen CPUs. You should read the MKL license before using MKL. I shall not be held responsible for how you use MKL.

    Intel MKL has been known to use a SSE code paths on AMD CPUs that support newer SIMD instructions such as those that use the Zen microarchitecture. A (by now) well-known trick has been to set the MKL_DEBUG_CPU_TYPE environment variable to the value 5 to force the use of AVX2 kernels on AMD Zen CPUs. Unfortunately, this variable has been removed from Intel MKL 2020 Update 1 and later. This can be confirmed easily by running a program that uses MKL with ltrace -e getenv.

  • CoryDoctorow ☛ Pluralistic: Surveillance pricing lets corporations decide what your dollar is worth

    Economists praise "price discrimination" as "efficient." That's when a company charges different customers different amounts based on inferences about their willingness to pay. But when a company sells you something for $2 that someone else can buy for $1, they're revaluing the dollars in your pocket at half the rate of the other guy's.

• Health/Nutrition/Agriculture

  • Andrea Contino ☛ 42

    Digesting hundreds of posts through my RSS feed is becoming a tiring task rather than a pleasant one, so I’m dedicating less and less time to it. The nice thing, though, is that there’s no rush to consume that kind of content, and I know I can return to it whenever I feel like it or maybe when I have more time to spare.

  • Daniel Pocock ☛ Social Control Media, Technology & Catholicism: Synod on Synodality review and feedback

    There is another working group that looked at the topic of polygamy. The psychology and neuroscience of social control media usurps our relationships in such a way that some people feel Facebook's algorithms are exerting as much influence as a third person in their marriage. I created a separate blog post to look at the evidence concerning the phenomena of polygamy and cyberpolygamy. Given how pervasive these technologies are becoming, it is essential reading alongside the content in this analysis.

• Proprietary

  • The checks Microsoft runs before allowing Windows 11 installation to proceed exclude most PCs manufactured before 2018.

    The checks Microsoft runs before allowing Windows 11 installation to proceed exclude most PCs manufactured before 2018. The operating system has become so bloated (with crapware and data collection spyware you don’t need) that these requirements are far higher than they need to be.

  • The Record ☛ Leak of data belonging to 7.4 million Paraguayans traced back to infostealers | The Record from Recorded Future News

    Redline was one of the most widely used tools by cybercriminals until the takedown last year, allowing hackers to steal usernames, browser information, passwords, credit cards, VPN logins and more from infected devices.

    Infostealers are typically spread through phishing emails, malicious downloads, or compromised websites, and quietly collect login credentials, cookies and other sensitive data from infected devices, which are then sold or exploited on the dark web.

  • Mandaris Moore ☛ MacOS 26 Beta 2 is worse

    It’s going to be a long summer.

  • The Record ☛ Israeli officials say Iran exploiting security cameras to guide missile strikes

    “We know that in the past two or three days, the Iranians have been trying to connect to cameras to understand what happened and where their missiles hit to improve their precision,” Franco said, without providing specific evidence.

  • Artificial Intelligence (AI) / LLM Slop / Plagiarism

    • Cyble Inc ☛ Hackers Poisoning AI Tools' Google Search Results To Deliver Infostealers

      The strategy? Ride the hype wave of AI search traffic to quietly drop malware onto the systems of curious users.

    • IT Wire ☛ iTWire - What Is Computer Vision? A Beginner-Friendly Introduction with Real-World Examples

      One area of AI which is being spoken a lot about at the moment is Computer Vision. Computer vision is a field of artificial intelligence (AI) that enables computers to “see,” understand, and interpret visual information, just like humans do. It involves teaching machines to recognise patterns, objects, and even emotions in images or videos. If you haven’t heard of Digitalsense, take a look at their site for some more information on Computer Vision.

    • Neil Selwyn ☛ AI not great for ‘hard’ tasks … and teaching is full of hard tasks

      In contrast, ‘hard-to-learn tasks’ do not have a clear connection between performing an action and achieving the desired outcome. Often the desired outcome is not clear and working out what to do is highly determined by local contextual factors – many of which will not be immediately apparent. This is work thar require problem-solving, intuition, informed guess-work and other forms of expertise and familiarity with the specific people, places and processes being dealt with. Crucially, then, these are tasks where there is often not enough information for an AI system to learn and/or it is unclear what information might be required to address the problem. To illustrate ‘hard-to-learn’ tasks, Acemoglu gives the example of a doctor diagnosing the likely cause of a persistent cough and proposing a course of treatment. This is a highly complex task. There are many past events that that might be contributing to a lingering cough. While there are common causes for coughs, there are also many rare conditions that should be considered. Crucially, there is no neat dataset of successful diagnoses and cures. At best, it is possible to train AI models based on the behaviour of human doctors performing similar tasks, but this is difficult because there is often no clear metric of success. This also means that any AI can only be developed to perform at similar levels to the average doctor.

    • The New Stack ☛ 6 Key Security Risks in LLMs: A Platform Engineer's Guide

      Platform engineers are on the front lines of securing these AI systems, especially as large language models (LLMs) enter production. While cloud-hosted LLMs pose significant security challenges, self-hosted models, such as Meta’s LLaMa, introduce even greater complexity. Self-hosting provides more control over security, customization and costs, but also expands the attack surface.

    • Rlang ☛ Specialized R packages for spatial machine learning: An introduction to RandomForestsGLS, spatialRF, and meteo

      This document provides an overview of three R packages, RandomForestsGLS, spatialRF, and meteo, that implement spatial machine learning methods, but are outside of standard machine learning frameworks like caret, tidymodels, or mlr3.1

    • [Repeat] Zimbabwe ☛ DeepSeek Accused of Helping Chinese Military Evade US Tech Bans - Techzim US Accuses DeepSeek of Helping China’s Military — Why That Matters for Zimbabwe

      It also shows how much we rely on tools we don’t control, built by companies caught up in global power struggles we know little about.

      The U.S. could restrict DeepSeek soon, especially through unofficial platforms many Zimbabweans use. Though honestly, that might not be a big deal, there are so many options now. Maybe that’s even the goal, flood the market with U.S.-friendly tools and starve out the rest.

    • Reuters ☛ Exclusive: DeepSeek aids China's military and evaded export controls, US official says

      "We understand that DeepSeek has willingly provided and will likely continue to provide support to China's military and intelligence operations," a senior State Department official told Reuters in an interview. "This effort goes above and beyond open-source access to DeepSeek's AI models," the official said, speaking on condition of anonymity in order to speak about U.S. government information.

    • Marcy Wheeler ☛ What We Talk About When We Talk About AI (Part Three)

      The parallels with our AIs are not subtle.

      If the golem was not well managed, it could become a violent horror, ripping up anything in its path mindlessly. The metaphors for technology aside, what makes the golem itself such a useful idea for talking about AI is how human shaped it is. Both literally, and in its design as the ultimate desirable servant. The people of Prague mistook the golem for a man, and we mistake AI for a human-like mind.

      Eventually, the rabbis put the golems away forever, but they had managed to do useful things that made life easier for the community. At least, sometimes. Sometimes, the golems got out of hand.

  • Social Control Media

    • 404 Media ☛ AI Models And Parents Don’t Understand ‘Let Him Cook’

      Young people have always felt misunderstood by their parents, but new research shows that Gen Alpha might also be misunderstood by AI. A research paper, written by Manisha Mehta, a soon-to-be 9th grader, and presented today at the ACM Conference on Fairness, Accountability, and Transparency in Athens, shows that Gen Alpha’s distinct mix of meme- and gaming-influenced language might be challenging automated moderation used by popular large language models.

    • Jeff Geerling ☛ How I make thumbnails for YouTube

      Therefore YouTube thumbnails are often a contentious topic. In the early days, thumbnails were just a still frame from the video, meaning it was harder (but not impossible) to 'game the system' and introduce clickbait.

      But many years ago, YouTube (and now most online video platforms) offered custom thumbnails, and once that was available, it was a race to the bottom to see who could tweak and hone their thumbnail artistry to 'get the click'.

    • The Telegraph UK ☛ Ban selfie-takers from museums

      Only one thing will take me from zero to 60 faster than a museum selfie-taker, and that’s a museum selfie-taker who thinks it’s beyond hilarious to assume the same pose as the subject of the painting. A goon like the Uffizi visitor who was trying so hard to perfect the same pose as Anton Domenico Gabbiani’s Ferdinando de’ Medici on Saturday, that he lost his balance and leant back against the canvas, tearing a hole in the bottom corner.

  • Windows TCO / Windows Bot Nets

    • The Record ☛ Data of more than 740,000 stolen in ransomware attack on Michigan hospital network

      The attack last year was launched by an “international ransomware group” and impacted the computer networks of both McLaren Health Care and Karmanos Cancer Institute, the documents said. McLaren did not offer further details about the cybercriminals.

• Security

  • Confidentiality

    • Inside Towers ☛ Court Okays AT&T's $177 Million Data Breach Settlement

      One of the incidents resulted in the illegal downloading of about 190 million wireless customer accounts. AT&T disclosed that its call logs were copied from its workspace on a Snowflake cloud platform covering about six months of customer call and text data from 2022, from nearly all its customers, Inside Towers reported.

    • CCC ☛ Onion Services: Design, Protocol and Implementation - media.ccc.de

      Onion Services are a crucial part of the Tor ecosystem and provide a clever way for anonymously hosting location hidden network services. Almost all of us know about them but how do they work in detail? This talk explains the technical details from .onion addresses up to the transfer of actual TCP stream data.

      This talk aims to give an introduction into the design, protocol and implementation of onion services.

• Defence/Aggression

  • Scoop News Group ☛ Meta confused over WhatsApp ban issued to House staffers

    The chamber’s chief administrative officer issued a memo Monday that the messaging app is not approved for official use.

  • Paul Krugman ☛ Why Big Tech Turned Against Democrats — and Democracy

    This change in attitude didn’t come out of nowhere. There is growing evidence that social media can do real harm, especially to children. There were, as I’ll explain, good reasons to worry that the benefits of technology were no longer flowing to consumers and business as a whole. And the Biden administration was also reflecting a broader change in perceptions: Many Americans have turned sour on the tech sector and tech leaders. In fact, the industry’s fall from grace has been little short of spectacular.

  • Graham Cluley ☛ Twitter refuses to explain what it’s doing about hate speech and misinformation, sues New York State for asking • Graham Cluley

    Elon Musk’s Twitter (surely no-one really calls it X?) is suing New York State.

    Why? Because apparently being asked to explain how your social media platform handles hate speech and misinformation is an unconstitutional burden.

    New York state’s law requires social media companies to report how they moderate hate speech and disinformation. It doesn’t require censorship or even impose fines. It just says “Tell us what your moderation policies are, and how you enforce them.”

  • Axios ☛ Scoop: WhatsApp banned on House staffers' devices

    The U.S. House's chief administrative officer informed congressional staffers Monday that messaging app WhatsApp is banned on their government devices, Axios has learned.

  • The Strategist ☛ Working with Ukraine is how Australia can supercharge defence innovation

    Ukraine aims to produce 4.5 million drones in 2025 alone, three times as many as in 2024. It has tested drones with maximum ranges of 3,000 km. Australia’s 2024 drone innovation program, by contrast, funded just three companies to build 100 prototypes, each with a 5 km range. Australia’s ambition is modest and experimental; Ukraine’s is vast and operational.

  • Wired ☛ Taiwan Is Rushing to Make Its Own Drones Before It's Too Late

    Yet Taiwan, which has set an ambitious target of producing 180,000 drones per year by 2028, is struggling to create this industry from scratch. Last year, it produced fewer than 10,000.

    “Taiwan definitely has the ability to make the best drones in the world,” says Cathy Fang, a policy analyst at the Research Institute for Democracy, Society, and Emerging Technology (DSET).

    So why doesn’t it?

  • [Old] New Yorker ☛ Lawrence Lessig v. Citizens United

    In 2010, in Citizens United v. Federal Election Commission, the Supreme Court, arguing that campaign contributions were a form of “political speech,” struck down limits on the amount of money corporations, unions, and rich individuals could spend on elections. The decision led to a surge of money greater than anyone predicted. Between 2008 and 2012, campaign spending shot up by nearly two billion dollars. Much of that growth came from super PACs, the committees that are allowed to spend whatever they want as long as they don’t work directly with candidates. In 2012, super PACs spent a billion dollars; seventy-three per cent of the money came from a hundred people. The other large source of growth has been in so-called dark money—donations from nonprofits that are allowed to keep their donors anonymous. In late August, before the midterm campaign had reached full speed, dark-money spending had climbed to fifty million dollars, according to the Center for Responsive Politics—seven times the sum that had been spent by that point in 2010.

  • [Old] New Yorker ☛ Why I Dropped Out

    On August 11th, I took the leap and, with a small team, launched my campaign Web site. I promised to be a “referendum candidate,” with the referendum being a commitment to fix our democracy, by passing a package of fundamental reform that would crack the corruption that had captured our government. Once the reform was passed, I would resign. The aim was to build a mandate powerful enough to take on the interests that would defend the status quo. The mandate to do that needed to be extraordinary if it were to have a chance to succeed. Limiting my scope, I believed, would strengthen the mandate. Less could be more.

  • [Old] New Yorker ☛ Why I Ran For President

    Instead, what our democracy needs is a democracy movement: a bipartisan recognition of how our Republic has failed, and a political movement powerful enough to bring about a remedy. And today, the only possible agent of political reform is the President. Congress isn’t going to fix itself. Neither are the state legislatures likely to rise and exercise their power to reform the Constitution (even though that was clearly the Framers’ plan). The only modern political actor capable of putting comprehensive reform before “the People” is the one political actor actually elected by “the People” (or sort of, at least indirectly, through an Electoral College): the President.

• Transparency/Investigative Reporting

  • Futurism ☛ Elon Musk’s Lawyers Claim He Doesn’t Use a Computer, Even Though There's a Vast Amount of Evidence That He Does

    Amid Elon Musk's attempt to kneecap OpenAI, attorneys for the billionaire claim he doesn't use a computer — even though he obviously does, including the fact that he's posted about doing so multiple times in the recent past.

    As flagged by Wired, this bizarre legal gambit came in response to claims from OpenAI — which has counter-sued Musk in the fiery legal battle — that the billionaire's attorneys are trying to "resist discovery."

• Environment

  • Wired ☛ India Is Using AI and Satellites to Map Urban Heat Vulnerability Down to the Building Level

    At present, heat action plans (HAP) are India’s primary approach for managing heat waves and keeping essential services running. Developed annually by state, district, and city governments—the Delhi government released its citywide HAP for 2025 in April—these plans are designed to help cities prepare for, respond to, and recover from extreme heat. And they’re not working.

    This has driven nonprofits and research organizations to show how HAPs can be improved—with one key idea being to use geographic information systems (GIS) that combine satellite imagery with local data to provide cities with a granular, building-by-building views of their heat. Action plans, often plagued by generalization, can then be tailored to better protect those at the highest risk.

  • Omicron Limited ☛ Nearly 40% of grilled eel products in Japanese retail market identified as American eel

    Stocks of eel species, most notably those in temperate Northern Hemisphere regions, have exhibited declines. These include the European eel, currently listed as Critically Endangered (CR) on the IUCN Red List of Threatened Species, as well as the Japanese eel and American eel, which are classified as Endangered (EN).

  • TruthOut ☛ Youth Who Fought the Climate Crisis in Court Are Now Targeting Trump

    “Trump’s fossil fuel orders are a death sentence for my generation,” Eva Lighthiser, the named plaintiff in the case, said in a statement. “I’m not suing because I want to — I’m suing because I have to. My health, my future, and my right to speak the truth are all on the line. He’s waging war on us with fossil fuels as his weapon, and we’re fighting back with the Constitution.”

  • Futurism ☛ Scientists Just Found Something Unbelievably Grim About Pollution Generated by AI

    In a new study in the science journal Frontiers in Communication, German researchers found that large language models (LLM) that provide more accurate answers use exponentially more energy — and hence produce more carbon — than their simpler and lower-performing peers. In other words, the findings are a grim sign of things to come for the environmental impacts of the AI industry: the more accurate a model is, the higher its toll on the climate.

  • Energy/Transportation

    • Wired ☛ This Is Why High-End Electric Cars Are Failing

      And it’s not the only e-flop from a legacy automaker: Just last week Ferrari announced that it is delaying its second EV model (the first will arrive in October) until at least 2028 because of weak demand. Porsche has cut back its plans for EVs amid soft sales of its electric Macan SUV and Taycan models. Indeed, depreciation on the Taycan is so eye-watering that some Porsche dealers have supposedly refused to take their own brand’s performance EV from owners looking to offload or upgrade.

    • Austin Gil ☛ Luger: My favorite Portland bicycle route

      Over the last year, I’ve been unintentionally planning a bike route. Really, a collection of sections I always seem to end up riding when I’ve got a couple of hours of time to kill and want so experience something beautiful. Once I realized that all these sections could make a route, it’s become a bit of an obsession to find the optimal way to connect some of my favorite sections in North East Portland. A sample platter of everything, including: [...]

    • Wired ☛ China’s Electric-Vehicle Factories Have Become Tourist Hot Spots

      As Chinese EV brands expand from competing on low prices to promoting premium features and sleek designs, they are increasingly putting their factories in the spotlight. At least two Chinese EV brands, Xiaomi and Nio, offer regular tours for the general public this year, and three more automakers have announced plans to follow suit.

    • Renewable Energy World ☛ Globeleq to acquire majority stake in Zambia's Lunsemfwa Hydro Power Company

      Globeleq, an independent power company in Africa, has signed a share purchase agreement with Norfund, the Norwegian development institution for the proposed acquisition of a 51% equity stake in the Zambian entity, Lunsemfwa Hydro Power Company (LHPC). The remaining 49% of LHPC is owned by Wanda Gorge Investments, a Zambian based infrastructure investment company.

• AstroTurf/Lobbying/Politics

  • Misinformation/Disinformation/Propaganda

    • DeSmog ☛ A Weaponized AI Chatbot Is Flooding Canadian City Councils with Climate Misinformation

      Thousands of councillors in more than 500 municipalities have received AI emails from a group called KICLEI, a name that mimics the international network ICLEI.

    • Silicon Angle ☛ Snyk acquires Invariant Labs to expand AI agent security capabilities

      Cybersecurity company Snyk Ltd. today announced that it has acquired Invariant Labs AG, a Swiss artificial intelligence security research firm, for an undisclosed price.

    • [Repeat] Evan Hahn ☛ Notes from "Empire of AI: Dreams and Nightmares in Sam Altman's OpenAI"

      Empire of AI covers the history of OpenAI, from before ChatGPT to early 2025. It profiles CEO Sam Altman and several other executives. It also discusses the broader AI industry.

• Censorship/Free Speech

  • The Register UK ☛ Just say no to NO FAKES Act, EFF argues

    The Electronic Frontier Foundation (EFF) says the revised version of the NO FAKES Act, reintroduced in April, would be a disaster for free speech and innovation if signed into law.

  • US Senate ☛ NO FAKES Act [PDF]

  • Jamie Zawinski ☛ Always scrub your travel phone of JD Vance memes.

    Mads from Tromsø reacts strongly to the treatment he received at American border control on his way into the United States: [...]

  • BoingBoing ☛ European tourist deported over JD Vance meme on his phone

    A visitor from Norway was reportedly denied entry to the United States after border officers found an unflattering meme of Vice President JD Vance on it. Mads Mikkelsen, 21, told his hometown newspaper that be was harassed by officers before being deported.

  • Ireland ☛ Man 'refused entry into US' as border control catch him with bald JD Vance meme

    After handing over his password, Mads was told he would not be allowed to go through with his planned vacation after two images were not to the officers' liking. One image was of a meme showcasing JD Vance with a bald, egg-shaped head. Variations of the image were shared endlessly in March on social media, with the Vice President himself posting his own version.

    The other picture showed Mads with a wooden pipe which he had made years prior. "Both pictures had been automatically saved to my camera roll from a chat app, but I really didn't think that these innocent pictures would put a stop to my entry into the country," the 21-year-old admitted.

    Mads told Nordlys he tried to explain the images as being harmless and meant as jokes but the immigration authorities ignored his pleas. He claims he was then strip-searched, forced to give blood samples, a facial scan and fingerprints.

• Freedom of Information / Freedom of the Press

  • The Dissenter ☛ 'Giving Information To The Enemy': Israel's Ban On Al Jazeera Extends To Foreign Broadcasters

  • New York Times ☛ Media Matters Sues F.T.C. Over Advertising Investigation

    Media Matters, a liberal advocacy organization, sued the Federal Trade Commission on Monday, claiming that the agency was waging a “campaign of retribution” against the group on behalf of the Trump administration and Elon Musk.

    The F.T.C. started investigating Media Matters last month over whether the organization had illegally colluded with other advertising advocacy groups to pinch off revenue from X, Mr. Musk’s social media company, and other right-leaning sites. Media Matters reported in 2023 that ads on X appeared alongside antisemitic content.

  • 404 Media ☛ Here's the Video for Our Sixth FOIA Forum: Massive Blue

    Here is the video archive for our FOIA Forum where we explained how we got records about Massive Blue, a company selling AI personas to cops.

  • Nebraska Examiner ☛ Accurate information is our responsibility

    Bacon said the House Speaker assured him that money lost in the massive bill would be restored later. His optimism is noteworthy (or perhaps naive) in a place where the only sure thing these days is that nothing is a sure thing. We’re teeter tottering on tariffs, immigration and war, with policies seemingly changing based on the latest news cycle. Or worse, whims.

• Civil Rights/Policing

  • The Verge ☛ The US is stripping its forests of decades-old protections | The Verge

    The Trump administration wants to open up tens of millions of acres of national forest to development. The US Department of Agriculture (USDA) announced yesterday that it’s rescinding a landmark rule that prevents road construction and timber harvesting in the last unfragmented stretches of national forest.

  • 404 Media ☛ ‘FuckLAPD.com’ Lets Anyone Use Facial Recognition to Instantly Identify Cops

    “We deserve to know who is shooting us in the face even when they have their badge covered up,” McDonald told me when I asked if the site was made in response to police violence during the LA protests against ICE that started earlier this month. “fucklapd.com is a response to the violence of the LAPD during the recent protests against the horrific ICE raids. And more broadly—the failure of the LAPD to accomplish anything useful with over $2B in funding each year.”

  • Jamie Zawinski ☛ FuckLAPD dot com:

    Lets Anyone Use Facial Recognition to Instantly Identify Cops: [...]

  • Rolling Stone ☛ Book Excerpt: Copaganda and Me

    In an excerpt from her book, I Want to Burn This Place Down, Maris Kreizman writes about the way law enforcement officers are mythologized in popular culture — and how she became 'radicalized' by reality

  • CoryDoctorow ☛ Pluralistic: The case for a Canadian wealth tax

    A major problem with letting billionaires decide how your country is run is that they will back whichever psycho promises the lowest taxes and least regulation, no matter how completely batshit and unfit that person is: [...]

• Internet Policy/Net Neutrality

  • APNIC ☛ The scourge of excessive AS-SETs

    In the last edition of Beyond Their Intended Scope, our series analysing BGP leaks, we investigated a recent path leak by a Border Gateway Protocol (BGP)-based Distributed Denial-of-Service (DDoS) mitigation vendor. In it, we touched on the possibility that the size and scope of this leak may have been a product of a rarely discussed problem in the world of automated route filtering: Excessively broad AS-SETs.

    In this post, we’ll discuss what AS-SETs are, their role in automated route filtering, and the challenges that excessive AS-SETs cause. Finally, with the help of bgp.tools founder Ben Cartwright-Cox, we’ll examine some of the worst (most excessive) AS-SETs on the Internet today.

• Digital Restrictions (DRM)

  • Björn Wärmedal ☛ Subscription Costs for My Entertainment

    I wish I could do more, but I need to be picky. The sad truth is that supporting three creators on Patreon is roughly the same cost as a streaming service with hundreds of TV shows and thousands of movies.

  • Digital Music News ☛ Spotify Appears to Be Working on a Music Import Tool

    The feature is still in development, with no confirmed release date yet. APK teardowns typically only provide hints about future releases—references to Spotify’s still missing HiFi tier were discovered years ago. Spotify is once again playing catch-up with Apple Music, which launched a similar music import tool that allows users to migrate away from Spotify or YouTube Music to Apple Music.

Monopolies/Monopsonies

• Bruce Lawson ☛ British Browser Boiz in Brussels! Fun at Microsoft’s DMA compliance meeting

  On Friday I attended the Microsoft’s second DMA enforcement workshop at the European Commission in Brussels, to hear their update on first year of DMA compliance. I was there on behalf of Vivaldi, as part of the Browser Choice Alliance (BCA), but this write-up is personal opinion only, and doesn’t represent the BCA or any other members.

  [...]

  Then it was time for a debrief, so we sampled what are reputedly the best frites in Brussels, from Maison Antoine (with samauri sauce, of course) and a glass of cold Kriek beer. Microsoft has said it will stop some of the most egregious dark patterns – opening some links in Edge, regardless of the default browser, and resetting it as the default browser after a Windows Update. Of course, these are not yet released, so I can’t judge whether Microsoft will comply with the spirit of the DMA.

  However, Windows will continue ignore the default browser in two heavily-used Microsoft applications, Outlook and Edge. They will still require users to dig around in Windows Settings to set a browser as the default instead of asking a user during installation, and on choosing it to become default, it should automatically replace Microsoft Edge in the taskbar.

• Trademarks

  • [Repeat] Digital Music News ☛ OpenAI Scrubs Jony Ive Partnership Following Trademark Dispute

    OpenAI recently removed all mentions of its partnership with Jony Ive’s new hardware startup io from its website and social media channels. The move comes after a federal court order related to a trademark dispute with IYO, a startup specialized in AI-powered hearing devices.

• Copyrights

  • 404 Media ☛ Judge Rules Training AI on Authors' Books Is Legal But Pirating Them Is Not

    A judge rules that Anthropic's training on copyrighted works without authors' permission was a legal fair use, but that stealing the books in the first place is illegal.

  • Simon Willison ☛ Anthropic wins a major fair use victory for AI — but it’s still in trouble for stealing books

    The judgement itself is a very readable 32 page PDF, and contains all sorts of interesting behind-the-scenes details about how Anthropic trained their models.

  • Manton Reece ☛ Training LLMs on books judged as fair use

    In a nutshell, the judge says that legally purchased books can be used to train AI, as long as the models do not reproduce verbatim the original copyrighted works. Pirated books, of course, are a separate issue. They are unlawfully acquired! We can’t steal a book from a store, regardless of what we planned to do with it.

  • Pivot to AI ☛ Anthropic AI wins broad fair use for training! — but not on pirated books

    Anthropic bought a huge amount of print books and scanned them into a private library to train from. But it also downloaded 7 million pirate copies of books from LibGen and Books.

  • Jérôme Marin ☛ The trial that could reshape AI

    Getty was among the first to recognize the threat AI posed to its business. In February 2023, it became the first company to file a lawsuit against Stability AI, then the leading player in AI-generated imagery. The first case — still ongoing — was filed in the US, followed by this second one in the UK. These cases are among the most symbolic in the escalating battle between AI model developers and content rights holders whose works have been scraped and used without permission or compensation.

    Getty’s main piece of evidence is compelling: some images generated by Stable Diffusion include visible watermarks (the branded logos from Getty's publicly viewable photos). Stability does not even try to deny it; it acknowledges that it pulled images from Getty’s library.

  • BoingBoing ☛ Mr. Beast launches AI-powered YouTube plagiarism machine

    Viewstats, a YouTube metrics tracker helmed by Mr. Beast, has announced an AI-powered thumbnail generation tool that allows users to directly feed YouTubers' thumbnails into it without their consent to create something ripping off- sorry, inspired by any creator they want. Mr. Beast plagiarizes himself in the initial demo to avoid any snags, but with the ability to rip off entire channels in one fell swoop or even specific videos, the writing is pretty clearly on the wall. It's as easy as entering a URL and grabbing any thumbnail you like to bastardize for yourself.