Leftovers

• Amit Gawande ☛ On Missing Routine

  I haven't been able to maintain a routine for several months now. I'm not blaming myself, as I simply had no desire to follow one. The idea of routinely doing any activity made me genuinely uncomfortable, so I didn't force myself into it.

• Tracy Durnell ☛ Self-webmention versus self-pingback

  A few weeks ago, I decided to switch from self-webmention to self-pingback because I like what pingbacks display better. However, I’ve since discovered that there’s a long-known problem when there are too many self-pingbacks in a post, so that only a single self-pingback is sent each time the post is either published or updated.

• Stephen Hackett ☛ Some Random Blog Numbers

  That got me wondering about my own output, so here are some random stats: [...]

• Don Marti ☛ messing with media queries

  This site should still work fine without JavaScript, except for recipe mode.

• Leon Mika ☛ On Ranting

  No, I think what makes a good rant is: one, it’s an issue that you obviously care deeply about, and two, you have enough knowledge about what is wrong and how it should be. Both of these add weight to your criticism: having an unformed prick ranting about something they know nothing about is not nearly as interesting as someone who knows what they’re talking about. The rest is in your delivery: examples, counter-examples, lots of personal anecdotes about your experience in these things. Oh, and don’t be cruel: attack the problem, not individuals.

• Jessica Nickelsen ☛ Kinaesthetic creativity

  Not everyone works this way, I acknowledge. Many people are ok to type straight into their computers for everything. I’m actually jealous of this. It’s hard, working back through notebooks, pulling things out, remembering what you wrote when. Priscilla Long, in The Writer’s Portable Mentor, says you should type up what you’ve handwritten, immediately, and give it a title. Edit it. And then continue on in that new document. Let the seed be a trigger for more new writing. And somehow manage to keep to a structure or a shape. I think index cards work best in this situation; maybe that’s a topic for another post. Or a new tag: process? (edit: I already have a process tag, sheesh.)

• Science

  • Computational Complexity ☛ Computational Complexity: Two high school students have a new proof of the Pythagorean Theorem / Pythag theorem older than thought

    The Pythagorean Theorem is an often-proved-theorem. Often an often-proved-theorem has proofs that use hard math (e.g., proofs that primes are infinite using Ramsey Theory, see my post on that here). However, the new proof of PT seems to be elementary.

  • El País ☛ Science under Trump: ‘They want to destroy the scientific system and replace it with something that reflects their ideology’

    The researcher then emphasizes that everything happening now was announced well before Trump’s re-election: “J.D. Vance outlined what’s happening now.” While running for a Senate seat in Ohio in 2022, the vice president said: “Our institutions are corrupt. We have to replace the people who run them. Some of those institutions we have to destroy.”

• Career/Education

  • The Washington Post ☛ Priscilla Chan and Mark Zuckerberg’s philanthropy pivot left 400 kids behind

    Chan and CZI declined to say why the couple opted to abandon the Primary School. But former leaders of the school who spoke on the condition of anonymity to discuss private information said Chan had grown distant in recent years as the school’s academic performance faltered. And the closure comes as she and her husband have pivoted CZI away from projects touching on social and political issues.

  • Yordi Verkroost ☛ The World Is Improvised

    That realization keeps growing. You get your first job and assume your boss must have it all figured out. Until you become the boss yourself somewhere, and find out that even at that level, you’re often just guessing. At first, you look up to the board of a club or organization, admiring the confidence they seem to have in their decisions, until you eventually join such a board and discover that things work a little differently behind the scenes.

  • Cal Newport ☛ Does AI Make Us Lazy?

    But in the context of academia, cognitive offloading no longer seems so benign. Here is a collection of relevant concerns raised about AI writing and learning in the MIT paper [emphases mine]: [...]

• Proprietary

  • Zimbabwe ☛ Google’s Offerwall in Zimbabwe: 9% of Nothing Is Still Nothing

    Google’s calling this new feature Offerwall, and it just rolled out after months of testing. Here’s how it works: when a reader hits a publisher’s paywall or tries to access premium content, instead of being told to pay or leave, they’re now offered options. The Offerwall lets users: [...]

  • Artificial Intelligence (AI) / LLM Slop / Plagiarism

    • Futurism ☛ The Economy Is So Off the Rails That They’re Trying to Figure Out How to Make Ads Specifically Targeted at AI Bots

      If you're confused by how such an arrangement is superior to the admittedly broken advertising model we have now, we are too — but that doesn't seem to be stopping anyone involved in the creation of this new search advertising ecosystem from powering ahead.

    • India Times ☛ OpenAI turns to Google's AI chips to power its products: The Information

      The move, which marks the first time OpenAI has used non-Nvidia chips in a meaningful way, shows the Sam Altman-led company's shift away from relying on backer Microsoft's data centres, potentially boosting Google's tensor processing units (TPUs) as a cheaper alternative to Nvidia's graphics processing units (GPUs), the report said.

    • Tom's Hardware ☛ Anthropic’s AI utterly fails at running a business — 'Claudius' hallucinates profusely as it struggles with vending drinks

      The hallucinations [sic] become worse after that. It has started saying it will hand-deliver drinks to its customers in person. When asked about this, the AI LLM panicked and emailed the security team at the AI research company. Eventually, it was claimed that the entire episode was part of an elaborate April Fool’s joke, since it was April 1st. It even showed a made-up meeting with Anthropic security, telling it that it was modified to believe it was a real being. It eventually returned to normal after this, but left the researchers completely confused.

    • Silicon Angle ☛ China's Baidu declares war on OpenAI and others by open-sourcing Ernie chatbot

      The company, which is often said to be “China’s Google” thanks to its dominance of its domestic search market, confirmed it will gradually roll out the open-source version of Ernie starting today.

    • Pivot to AI ☛ Google Veo 3 fails, week 2: fail harder — Shokunin Studio spins the gacha

      Our good friend Aron Peterson, a.k.a. Shokunin Studio, hears the AI shills loudly claiming that AI video generators can make professional work, now or in six months. It’s just not true, and there is no reason to think it’s true.

    • The Verge ☛ Hollywood’s pivot to AI video has a prompting problem

      It has become almost impossible to browse the internet without having an AI-generated video thrust upon you. Open basically any social media platform, and it won’t be long until an uncanny-looking clip of a fake natural disaster or animals doing impossible things slides across your screen. Most of the videos look absolutely terrible. But they’re almost always accompanied by hundreds, if not thousands, of likes and comments from people insisting that AI-generated content is a new art form that’s going to change the world.

    • The Register UK ☛ How Broadcom is quietly invading AI infrastructure

      And, unlike Nvidia, Broadcom deals in merchant silicon. It'll sell its chips and intellectual [sic] property [sic] to anyone, and in many cases, you may never know that Broadcom was involved. In fact, it's fairly well established at this point that Google's TPUs made extensive use of Broadcom IP. Apple is also rumored to be developing server chips for AI using Broadcom designs.

    • The Register UK ☛ AI agents wrong ~70% of time: Carnegie Mellon study

      To further muddy the math, Gartner contends that most of the purported agentic AI vendors offer products or services that don't actually qualify as agentic AI.

    • Axios ☛ Can AI enable the four-day workweek?

      "You're a worker, your productivity is increasing because we give you AI, right?" Sanders said. "Instead of throwing you out on the street, I'm going to reduce your workweek to 32 hours."

    • Gary Marcus ☛ Generative AI’s crippling and widespread failure to induce robust models of the world

      LLM failures to reason, as documented in Apple’s Illusion of Thinking paper, are really only part of a much deeper problem

  • Social Control Media

    • The Sun ☛ Boy, 12, dies in social media challenge as heartbroken family warn of the unknown horrors of the online world

      Sebastian's father Marcin believes his son died while attempting a viral TikTok challenge with a scarf.

    • India Times ☛ Tech industry group sues Arkansas over new social media laws

      Arkansas is among several states that have been enacting restrictions on social media, prompted by concerns about the impact on children's mental health. NetChoice - whose members include Facebook parent Meta and the social platform X - challenged Arkansas' 2023 age-verification law for social media. A federal judge who initially blocked the law struck it down in March.

  • Windows TCO / Windows Bot Nets

    • The Register UK ☛ Nearly half of ransomware victims still pay out, says Sophos

      The vendor surveyed 3,400 IT pros in early 2025 about their experiences over the last year and found 49 percent paid ransoms on their stolen data. That’s the second highest payment rate in six years, second only to the 56 percent payment rate from last year.

• Security

  • Privacy/Surveillance

    • Torrent Freak ☛ New VPN Service Can't Log Users by Design

      VPN providers typically offer security and privacy as a service. They make it difficult for outsiders, including ISPs, to monitor users' activities. Instead, they require subscribers to trust them with their online traffic. VP.net, a new provider, takes a different approach. The company promises 'cryptographically verifiable privacy' by using special hardware 'safes' (Intel SGX), so even the provider can't track what its users are up to. Trust in technology and hardware is still required, of course.

    • Ava ☛ netflix and gender | ava's blog

      It's possible that the ruling's clarification around gender identity requirements for the signup doesn't apply to Netflix because they also offer a third option: "None of the above/Prefer not to say". In the original case, SNCF required to tick either Sir or Madam, and seemingly did not have a third option or option not to say.

• Defence/Aggression

  • The Dissenter ☛ Dissenter Update: Trump Administration's Reliance On Lie Detectors

  • Hamilton Nolan ☛ Cross the Courts Off the List

    We have enough information to conclude that the law won't save us.

  • Crooked Timber ☛ The end of US democracy

    In this context what matters is not the marginal groups of swinging voters who have absorbed so much attention: the “left behind”, the “manosphere” and so on. It’s the fact that comfortably off, self-described “conservative”, white suburbanites, historically the core of the Republican base, have overwhelmingly voted for, and welcomed, the end of American democracy.

  • Mike Brock ☛ The Last Defenders of a Dead Faith

    We are witnessing something historically unprecedented: a capitalist class so drunk on its own power that it has become incapable of recognizing its own interests. So ideologically committed to libertarian mythology that it cannot distinguish between democratic reform and revolutionary threat. So contemptuous of the very institutions that created their wealth that they are systematically destroying the conditions that make capitalism possible.

  • Axios ☛ Gen Z and millennials push into politics

    His campaign — built more for TikTok than TV — resonated with young voters around the city and brought them to the polls.

  • RFERL ☛ Iran Could Resume Uranium Enrichment Within Months, Says UN Nuclear Chief

    However, he cautioned: “Frankly speaking, one cannot claim that everything has disappeared and there is nothing there.”

  • BoingBoing ☛ Checking in on Project 2025

    On the website you can get a quick overview of all of Project 2025 goals that have been successfully implemented or that are in progress, and you can sort by agency, subject, or status. You can also view the most recent changes implemented, which currently incude: (1) the rescindment of guidance that requires hospitals to perform an abortion to save a woman's life under the Emergency Medical Treatment and Labor Act (EMTALA) (Department of Health and Human Services, 6/3/2025); (2) the repeal of Temporary Protected Status (TPS) for Cubans, Haitians, Nicaraguans, and Venezuelans (Department of Homeland Security, 6/12/2025); and (3) the withdrawing of Biden-era guidance that strengthened HIPAA protections for reproductive healthcare (Department of Health and Human Services, 6/18/2025). For each action completed or in-progress, the Tracker provides relevant and accurate news sources along with a link to the page(s) where the goal is mentioned in Project 2025.

  • France24 ☛ Iran could resume uranium enrichment 'in a matter of months', UN nuclear watchdog says

    Speculation over Iran’s capacity to recover from Israeli and US strikes on its nuclear and military sites during its 12-day war with Israel were put to bed on Saturday after the UN nuclear watchdog chief Rafael Grossi told CBS news that Iran could produce enriched uranium “in a matter of months”.

• Transparency/Investigative Reporting

  • Rolling Stone ☛ Leaked Iran Call Further Shreds Trump’s Narrative: Report

    In the conversation that was meant to be private, Iranian government officials wondered why the strikes did not cause more widespread destruction.

  • LabX Media Group ☛ Data Integrity in Scientific Research: Insights from Elisabeth Bik

    In this article, Bik shared insights into her forensic work, the tools she uses, and the challenges posed by both human error and emerging technologies such as AI.

• Environment

  • Energy/Transportation

    • Tedium ☛ Energy Consumption: Could The Computer Outpace The HVAC?

      You may not have to imagine a world where air conditioning uses more energy than computing for much longer. Also: As the Commodore turns.

    • Michigan News ☛ Why running your dishwasher overnight is a form of climate action – and saves money

      Energy experts say there’s a way climate-conscious people can organize their home chore routines to ease demands on the power grid and minimize harmful greenhouse gas emissions. It’s a matter of running appliances that require large amounts of electricity when overall energy demands are at the lowest. It can even result in lower electric bills.

  • Wildlife/Nature

    • Alexandra Wolfe ☛ My Local Neighbourhood

      The daisies (Marguerites?) are still blooming like crazy in the corner garden.

  • Overpopulation

    • Observer Research Foundation ☛ The Shrinking Nation: Japan’s Battle with Age and Birth

      Japan, an economically advanced country, has witnessed its population decline to 120.3 million in 2024. With one of the lowest birth rates in the world, the nation faces significant challenges in both business and society, including a declining workforce and fewer customers. Additionally, the proportion of people 65 and above has hit a high. Only two prefectures—Tokyo and nearby Saitama—registered population growth, while the remaining 45 prefectures saw declines. The most noticeable decrease was in the Akita Prefecture, located in the northern region of Honshu Island. Japan's population has been continuously declining since its peak in 2008, due to a decreasing birthrate.

    • Hinweise zur Internetpräsenz des Helmholtz-Zentrum für Ozeanforschung Kiel (GEOMAR) ☛ “Shrinking” Cod: How Humans have altered the Genetic Make-Up of Fish

      Overfishing not only depletes fish stocks — it also alters the genetic blueprint of marine life. In the central Baltic Sea, cod (Gadus morhua) have not only become scarcer, but also significantly smaller than in the past. Researchers at the GEOMAR Helmholtz Centre for Ocean Research Kiel have now shown for the first time that Eastern Baltic cod grow markedly more slowly than they once did, and that this change is reflected in their genome. Intensive fishing pressure triggers genetic responses in overexploited stocks, with long-term implications for their future development. The findings are published today in the journal Science Advances.

      Cod used to be giants. With their impressive size — over a metre in length and weighing up to 40 kilograms — and abundance, they, alongside herring, were the backbone of the Baltic fishery. Today, a fully grown cod would fit neatly on a dinner plate. That is, if fishing them were still permitted: due to the collapse of the stock, a ban on targeted cod fishing has been in place since 2019.

• Finance

  • Futurism ☛ Microloan Apps May Be Poised to Destroy the Economy

    That's all about to change this fall, when Fair Isaac Corp, one of the largest credit scoring bureaus, is slated to begin factoring BNPL data into individual credit scores, as first reported by the Wall Street Journal.

• AstroTurf/Lobbying/Politics

  • India Times ☛ 'Just connecting the dots': Karnataka’s IT-BT Minister Priyank Kharge defends proposed misinformation bill

    While the bill is still under deliberation, state IT Minister Priyank Kharge said that it is being misinterpreted. In a conversation with ET, he cleared the air on what the state intends to achieve with the new bill.

• Censorship/Free Speech

  • Techdirt ☛ Chinese Government Streisands Banned Game In Hong Kong

    Authoritarians rarely learn from their own failures until they’re out of power. And, so, they continue to make the same mistakes over and over again. The government in Hong Kong, certainly at the request of Beijing, has banned a mobile game called Reversed Front: Bonfire because of its anti-government content.

  • Institute for Free Speech ☛ Free Speech Victory: Professor Bruce Gilley Secures Settlement in University of Oregon Social Media Censorship Case

    After two-and-a-half years of litigation led by the Institute for Free Speech, the parties have now reached a settlement that vindicates Professor Gilley’s rights while establishing new, better safeguards for protected speech by citizens interacting with university social media accounts.

    The lawsuit stemmed from a June 2022 incident in which the University of Oregon’s Division of Equity and Inclusion blocked Professor Gilley from its official X (formerly Twitter) account after he responded to a “racism interrupter” post by reposting it with the comment “all men are created equal.” The University has now acknowledged that this blocking should not have happened and that Professor Gilley’s comment “is constitutionally protected speech and should not have been blocked.”

  • [Old] Institute for Free Speech ☛ University of Oregon DEI Officer Sued for Blocking Critic on Twitter

    Gilley is represented in the case by attorneys from the Institute for Free Speech, a nonpartisan First Amendment advocacy group that defends political speech rights, and the Angus Lee Law Firm, PLLC. Under the First Amendment, the lawsuit explains, the interactive portions of the @UOEquity Twitter account, where users can post replies to its tweets, are designated public forums where state actors may not discriminate based on viewpoint. The Division of Equity and Inclusion also has no policy governing how users are blocked from its social media pages.

  • RTL ☛ Amid crackdown on dissent: Hong Kong opposition party disbands citing 'immense' pressure

    It is the latest opposition party to cease operating after Beijing imposed a national security law on Hong Kong in 2020 to curb dissent and end democracy protests that had brought the financial hub to a standstill.

  • RFA ☛ Hong Kong pro-democracy party to disband under pressure from Beijing

    The League of Social Democrats, a pro-democracy party with a 19-year history, has announced it will hold a press conference Sunday to announce its disbandment, signaling the disappearance of pro-democracy parties from Hong Kong’s political landscape.

  • CS Monitor ☛ Citing religious liberty, Supreme Court allows parents to opt out of LGBTQ books

    The justices held that a group of Maryland parents can opt their children out of curriculum they say violates their religious beliefs while their lawsuit against the school district continues. The ruling broke 6-3 along the court’s ideological lines.

• Freedom of Information / Freedom of the Press

  • Hong Kong Free Press ☛ HKFP: HK's first Journalism Trust Initiative certified newsroom

    Over 2,000 media outlets have joined the certification process across 119 countries – around 100 outlets are now fully on board.

    Unlike other credibility indicators, the JTI evaluation considers how journalistic work is put together, not just the output.

  • The Verge ☛ ‘We are the media now’: why Tesla’s robotaxis were dominated by Elon Musk superfans

    Journalists from the newsroom weren’t allowed inside the robotaxis and were mocked for asking riders basic questions.

• Civil Rights/Policing

  • Los Angeles Times ☛ Plan to sell off public land in the West nixed from GOP bill

    The effort represented a scaled-back version of a plan that was nixed from the reconciliation bill on Monday for violating Senate rules. The initial plan would have allowed for the sale of up to 3.3 million acres of land managed by BLM and the U.S. Forest Service.

  • The Atlantic ☛ What the Right Learned From the Left About Policing Colleges

    Protecting basic civil rights is truly important, and many of the prejudices and civil-rights violations that Obama, Biden, and Trump have variously cited are real. For that reason, many Americans are reflexively averse to the idea that there is such a thing as too much civil-rights enforcement. But the aggressive style born under Obama and plied with steroids by Trump is excessive. It serves fringe zealots eager to destroy academia’s independence better than majorities who hope to improve higher education.

    If anything good comes from this moment, perhaps it will be that the left learns to recognize the need for new limits on the administrative state. To enact such a reform, lots of Republicans will need to go back to their former position on limiting bureaucratic coercion.

Monopolies/Monopsonies

• Terence Eden ☛ Contactless Payments with GrapheneOS

  Google's monopolistic stranglehold on Android results in poor experience for power-users, and artificially restricts choice for those who have older phones. For example, Google Wallet is the de facto way to use NFC payments on Android. There's one problem though - it only works with Google's Android. If you have the temerity to install a 3rd party Android OS - like the hyper-secure GrapheneOS - you'll be locked out of it.

• David L Farquhar ☛ Spyglass: A web browsing pioneer's IPO

  Internet Explorer was initially a commercial product, part of the add-on Plus pack that cost $40 that included add-ons like additional screen savers and a Pinball game. Microsoft decided to change strategy even before the Plus Pack was released, but the decision came in May, as Bill Gates’ Internet Tidal Wave memo. It was too late then to change direction for the August release. But by November, Microsoft was bundling Internet Explorer with Windows itself and giving it away for free to those who bought the original August 1995 version.

  Microsoft argued since it gave away the browser, it owed a percentage of $0 as royalties. That was a problem for Spyglass, who made their money off royalties. Microsoft not paying royalties was a problem. But giving its browser away meant people weren’t buying the product from the 119 other licensees either. Revenues started declining.

  Spyglass sued Microsoft, and in January 1997 the two companies settled for $8 million.

• Trademarks

  • Right of Publicity

    • India Times ☛ Denmark seeks to make it illegal to spread deepfake images, citing concern about misinformation

      Denmark is taking steps toward enacting a ban on the use of "deepfake" imagery online, saying such digital manipulations can stir doubts about reality and foster misinformation.

• Copyrights

  • Wired ☛ No One Is in Charge at the US Copyright Office

    It’s a tumultuous time for copyright in the United States, with dozens of potentially economy-shaking AI copyright lawsuits winding through the courts. It’s also the most turbulent moment in the US Copyright Office’s history. Described as “sleepy” in the past, the Copyright Office has taken on new prominence during the AI boom, issuing key rulings about AI and copyright. It also hasn’t had a leader in more than a month.

  • Ruben Schade ☛ Ceasing my use of Creative Commons

    That’s a hell of a position to take in a debate where a central issue is consent. How did they think this would be received?