In the US, you can give up your house keys for a facial recognition unlocking system – if you want to give ADT $40 per month. Deal? CC-licensed photo by the Original Muddog on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 11 links for you. Unsecured. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

ADT Plus review: home security gets smarter • The Verge

Jennifer Pattison Tuohy:

»

My front door unlocks automatically as I walk up to it, and the home security system disarms itself — no code or app required. The system has recognized me using the Google Nest Doorbell’s Familiar Faces, and confirmed I’m me using my phone’s location. The dual-factor authentication triggers the automatic disarming and unlocking, so all I have to do is walk inside.

This hands-free Auto Unlock experience is powered by ADT Plus, the newest security system from America’s oldest security company. It features the usual hardware — a base, sensors, and various accessories — but represents a major upgrade from fumbling for keys or racing to silence a beeping keypad.

I hate that “beeping pressure” home alarm systems put on you, so I tend to avoid turning them on. But ADT’s Auto Unlock, along with other automated features the system offers, has meant I’ve found myself using it much more consistently than any security system I’ve tested. That alone makes it better.

A home that responds to you automatically, without requiring you to bark commands, punch in codes, or fiddle with apps, is the future of the smart home. The most surprising part is that this is coming from a legacy security company like ADT (with help from Google). Less surprising is that you have to pay ADT prices for this convenience.

…ADT Plus is only compatible with Google Home. Also, if you’re averse to high-cost monitoring packages, you should look elsewhere — for the full feature set I tested, ADT Plus will set you back $40 a month.

…The Nest doorbell specifically helps power Auto Unlock and a unique feature called Trusted Neighbor. This uses the same principles as Auto Unlock, but instead of only allowing in residents of the home, it can unlock the doors (and deactivate the alarm) for any approved person based on set parameters.

«

Like the idea of face unlock. Very much dislike the idea of paying that amount for it. Proximity locks (with RFID) seem like a good idea, except cars which use those get stolen – so maybe not. Keys might not be high tech, but they’re really robust, impossible to copy except in specific places, cheap, and proximity dependent.
unique link to this extract

Review: The Real North Korea, by Andrei Lankov • Mr and Mrs Psmith’s Bookshelf

John Psmith:

»

In the 1980s, Japan experienced a crisis of disinformation. For years, there had been mysterious disappearances of Japanese people with no known history of mental illness, drug addiction, or gambling debts. All kinds of people — men and women, young and old, just suddenly vanishing without a trace. Many theories were put forward to explain the puzzle (for instance, some believed it was alien abductions), but the most widespread, pernicious, and dangerous view was that North Korea was responsible. There were people who claimed to have actually seen teams of North Korean commandos lurking on beaches, nabbing random passers-by, and bundling them into waiting submersibles just off the coast. This was obviously crazy. Products, no doubt, of atavistic xenophobia and reactionary sentiments. The Japanese media, government, and academic authorities put a lot of effort into refuting this dangerous disinformation throughout the 1980s and 1990s…which made them look real silly when in 2002 Kim Jong-Il issued a formal apology for the abductions and ordered the surviving captives returned to Japan.

This has always felt like the ur-North Korea story to me, because it has a little bit of everything. First of all, it’s delightfully madcap — they KIDNAPPED RANDOM PEOPLE on BEACHES using SUBMARINES and they did it for DECADES. Second, it’s full of bizarre irony. The North Koreans got away with this scot-free until, in a gesture of goodwill and altruism designed to improve relations with Japan, they fessed up and tried to make things right…at which point everything blew up in their faces and had the exact opposite effect.

…It’s estimated that around 80% of all goods and services in North Korea are provided in secret and in shadow. It’s capitalism as an extremophile species of lichen, colonizing the cracks and crevices of the official society, and keeping the whole system afloat. They are actually speedrunning the entire history of primitive accumulation leading to investment leading to the joint stock corporation. Large (secret) transportation companies now exist in North Korea and maintain unofficial roads forming an unofficial transit network. The trucks and buses are smuggled in from abroad, then “donated” to various government agencies, which then lease them back in exchange for kickbacks. In this way, they’ve reinvented the idea of funding government operations through corporate taxation in a hilariously roundabout way. There is a booming private restaurant scene.

«

This is a fabulous review which, despite being full of detail, really makes you want to read the book. Though the review’s comments about the author are worth turning up for too. (Via Andrew Brown.)
unique link to this extract

Google and Check Point nuke massive YouTube malware network • The Register

Carly Page:

»

Google has taken down thousands of YouTube videos that were quietly spreading password-stealing malware disguised as cracked software and game cheats.

Researchers at Check Point say the so-called “YouTube Ghost Network” hijacked and weaponized legitimate YouTube accounts to post tutorial videos that promised free copies of Photoshop, FL Studio, and Roblox hacks, but instead lured viewers into installing infostealers such as Rhadamanthys and Lumma. 

The campaign, which has been running since 2021, surged in 2025, with the number of malicious videos tripling compared to previous years. More than 3,000 malware-laced videos have now been scrubbed from the platform after Check Point worked with Google to dismantle what it called one of the most significant malware delivery operations ever seen on YouTube.

Check Point says the Ghost Network relied on thousands of fake and compromised accounts working in concert to make malicious content look legitimate. Some posted the “tutorial” videos, others flooded comment sections with praise, likes, and emojis to give the illusion of trust, while a third set handled “community posts” that shared download links and passwords for the supposed cracked software.

“This operation took advantage of trust signals, including views, likes, and comments, to make malicious content seem safe,” said Eli Smadja, security research group manager at Check Point. “What looks like a helpful tutorial can actually be a polished cyber trap. The scale, modularity, and sophistication of this network make it a blueprint for how threat actors now weaponise engagement tools to spread malware.”

Once hooked, victims were typically instructed to disable antivirus software, then download an archive hosted on Dropbox, Google Drive, or MediaFire. Inside was malware rather than a working copy of the promised program, and once opened, the infostealers exfiltrated credentials, crypto wallets, and system data to remote command-and-control servers.

«

Beginning to wonder about this “internet” thing. Also, it had been there since 2021?
unique link to this extract

Offshoring automation: Filipino tech workers power global AI jobs • Rest of World

Michael Beltran:

»

Inside a multistory office building in Manila’s financial district, around 60 young men and women monitored and controlled artificial intelligence robots restocking convenience store shelves in distant Japan. 

Occasionally, when a bot dropped a can, someone would don a virtual-reality headset and use joysticks to help recover it. 

The AI robots are designed by Tokyo-based startup Telexistence, and run on Nvidia and Microsoft platforms. Since 2022, the company has deployed the machines in the back rooms of over 300 FamilyMart and Lawson stores in Tokyo. It is also planning to use them soon in 7-Elevens.

The bots are remotely monitored 24/7 in Manila by the employees of Astro Robotics, a robot-workforce startup. Japan faces a worker shortage as its population ages, and the country has been cautious about expanding immigration. Telexistence’s bots offer a workaround, allowing physical labor to be offshored, Juan Paolo Villonco, Astro Robotics’ founder, told Rest of World. This lowers costs for companies and increases their scale of operations, he said. 

“It’s hard to find workers to do stacking [in Japan],” said Villonco. “If you get one who’s willing to do it, it’s going to be very expensive. The minimum wage is quite expensive.”

It’s easy to get young, tech-savvy Filipinos to operate the robots, he said. Each tele-operator, called a “pilot,” monitors around 50 robots at a time, an employee told Rest of World. Most workers in this article requested anonymity to safeguard their jobs.

The bots are usually autonomous, but occasionally — about 4% of the time — they mess up. Perhaps they drop a bottle, which rolls away. Getting the AI bot to recover it by mimicking the human grip perfectly — the friction, the feel of metal in the hand — is one of the more challenging problems in robotics. That’s when a pilot steps in.

Astro Robotics’ tele-operators are benefiting from an AI- and automation-related boom in IT-service work and tech jobs in the Philippines, even as layoffs hit similar workers in richer countries.

«

This is one of the weirdest piece of telepresence, mixed with robotics, mixed with globalisation, that I’ve ever come across.
unique link to this extract

Inside Apple’s quest to add a heart-rate sensor and full workout tracking into AirPods Pro 3 | TechRadar

Jacob Krol:

»

The heart-rate sensor in the AirPods Pro 3 is not Apple’s first foray into this sensor type. Apple initially offered such a sensor when the Apple Watch first launched in 2015, so as Waydo explained, “it was really cool here to get to bring everything we learned over more than a decade of that work to this very different form factor.”

[Apple director of health sensing Steve] Waydo’s team at Apple has been working on heart-rate sensing algorithms since the original Apple Watch, and that 10-year journey proved helpful in the development of the new sensor – but it’s not one for one. “The heart rate sensor in the AirPods Pro uses invisible infrared light. We pulse it up to 256 times per second, and we take that data and we fuse it together with what we’re getting from the onboard accelerometers to measure the blood flow in your ears,” explained Waydo, who noted that’s where the final measurement for heart rate comes from.

This differs from the Apple Watch, which uses green LEDs. Here, the sensors are invisible, and there are either two sources or a single, as AirPods Pro 3 can provide a heart-rate reading with just one bud in or with both in. With the latter, the algorithms work in real time to pick the best, most accurate source.

On those algorithms, the base for how AirPods track heart rate comes from the Apple Watch, but Waydo explained that given these are much smaller, the team “had to shrink those algorithms down to fit within the processing and memory constraints, so that we could deliver the same kind of speed, efficiency, and battery life that our users really love.”

To get there, the team had to ensure the new sensor would work for everyone. As Waydo explained, “there’s a huge amount of variation in ear geometry that we had to tackle, both through fit and through our studies to understand if we could get great signals across everybody who’s going to use these AirPods.”

«

There’s a lot more to this; the AirPods Pro being able to measure heart rate is quite a feat.
unique link to this extract

How the brain moves from waking life to sleep (and back again) • Quanta Magazine

Yasemin Saplakoglu:

»

To fall asleep, “everything has to change,” said Adam Horowitz (opens a new tab), a research affiliate in sleep science at the Massachusetts Institute of Technology. The flow of blood to the brain slows down, and the circulation of cerebrospinal fluid speeds up. Neurons release neurotransmitters that shift the brain’s chemistry, and they start to behave differently, firing more in sync with one another. Mental images float in and out. Thoughts begin to warp.

“Our brains can really rapidly transform us from being aware of our environments to being unconscious, or even experiencing things that aren’t there,” said Laura Lewis (opens a new tab), a sleep researcher at MIT. “This raises deeply fascinating questions about our human experience.”

It’s still largely mysterious how the brain manages to move between these states safely and efficiently. But studies targeting transitions both into and out of sleep are starting to unravel the neurobiological underpinnings of these in-between states, yielding an understanding that could explain how sleep disorders, such as insomnia or sleep paralysis, can result when things go awry.

Sleep has been traditionally thought of as an all-or-nothing phenomenon, Lewis said. You’re either awake or asleep. But the new findings are showing that it’s “much more of a spectrum than it is a category.”

«

Not short, but very interesting, particularly about how you are more creative in the moments where you’re slipping over the edge into sleep.
unique link to this extract

Man detained for protesting National Guard with Darth Vader theme sues • The Washington Post

Joe Heim:

»

A District resident who was briefly detained and handcuffed in Washington DC last month for following an Ohio National Guard patrol while playing “The Imperial March” from Star Wars on his phone has filed a lawsuit saying his constitutional rights were violated.

The American Civil Liberties Union of DC filed the lawsuit Thursday in the US District Court for the District of Columbia on behalf of Sam O’Hara against an Ohio National Guard sergeant, four DC police officers and the District of Columbia. The suit said they had infringed on O’Hara’s First Amendment rights and violated DC law when they detained him on Sept. 11 on a public street in Northwest Washington.

The ACLU suit says that when O’Hara saw the National Guard members that day, he began walking behind them playing “The Imperial March” on his phone and recording them. It alleges that within two minutes, one of the Guard members “turned around and threatened to call the police officers to ‘handle’ Mr. O’Hara if he did not stop.”

O’Hara continued to play the song, also known as “Darth Vader’s Theme,” and the soldier called the DC police. When police arrived, the suit alleges, O’Hara was “tightly handcuffed” and detained for 15 to 20 minutes.

“Government conduct of this sort might have received legal sanction a long time ago in a galaxy far, far away,” the ACLU wrote in its filing. “But in the here and now, the First Amendment bars government officials from restraining individuals from recording law enforcement or peacefully protesting, and the Fourth Amendment (along with the District’s prohibition on false arrest) bars groundless seizures.”

A DC police spokesman said the department cannot comment on active litigation.

«

I bet. The thing that authoritarians can’t bear is satire, because it’s not proper dissent, which they would know how to suppress. I do hope the ACLU demands that, in restitution, all of those involved have to sit a test on the First and Fourth Amendments while the Imperial March plays at full volume. Make it memorable.
unique link to this extract

US accuses former L3Harris cyber boss of stealing and selling secrets to Russian buyer • TechCrunch

Lorenzo Franceschi-Bicchierai:

»

The US government has accused a former executive at defense contractor L3Harris of stealing trade secrets and selling them to a buyer in Russia, according to court documents seen by TechCrunch. 

On October 14, the Department of Justice accused Peter Williams of stealing eight trade secrets from two unnamed companies. The DOJ made the allegation in a “criminal information” document, which, like an indictment, represents a formal accusation of alleged crimes.  

The document does not specify Williams’ relationship with the two companies or the types of trade secrets, nor does it name the alleged Russian buyer. 

TechCrunch has confirmed that the Williams mentioned in the document, which does not specify where he worked, is the former general manager at Trenchant, a division of L3Harris that develops hacking and surveillance tools for Western governments, including the United States.

Williams became Trenchant’s general manager on October 23, 2024, and he worked at Trenchant until August 21, 2025, per UK business records. Williams, a 39-year-old Australian citizen, resided in Washington DC, according to the court document.

Four former Trenchant employees had previously told TechCrunch that Williams, who was known inside the company as “Doogie,” had been arrested.  

«

Does the company name “Trenchant” sound familiar? Well done – that’s the company from which a developer working on finding zero-day vulnerabilities was recently warned that he was being targeted with government spyware. There’s wheels within wheels on this.
unique link to this extract

Tech’s love affair with Trump grows stronger by the day • TechPolicy.Press

Paul M. Barrett:

»

Since the United States election in November 2024, the technology industry has distinguished itself in the race to capitulate to and enable the Trump administration. No gesture of obsequiousness or loyalty has been too great or trivial for the moguls of Silicon Valley.

In recent days:

• Representatives of Amazon, Apple, Google, and Microsoft attended a White House dinner where the president thanked wealthy donors to his project of adding a 90,000-square-foot ballroom to the Executive Mansion
• Responding to administration demands, Meta, Apple, and Google restricted digital tools used by activists to flag sightings of masked Immigration and Customs Enforcement (ICE) agents seeking to arrest immigrants allegedly lacking legal status
• Salesforce CEO Marc Benioff, once seen as a progressive benefactor, said in an interview that he avidly supports President Trump and believed National Guard troops should be deployed to his hometown of San Francisco, over the objections of city leaders. (He has since apologized for beckoning the Guard.) Benioff’s company, meanwhile, reportedly has pitched ICE on using Salesforce’s artificial intelligence to help the agency staff up as it expands immigration raids and deportations.

…Why has Silicon Valley’s ardor for Trump, The Sequel been so intense? Tech CEOs’ business goals are part of the answer. Google, Meta, Apple and Amazon are all in the midst of federal antitrust litigation that they doubtless hope Trump may influence. Musk’s rocket company, SpaceX, has billions of dollars in ongoing contracts with the US government and could benefit from closer federal ties.

…In some instances, tech companies are eager for Trump to fold their anti-regulatory agenda into his America First agenda — for example, by using US influence to reduce oversight by the European Union.

«

It’s pure self-interest on the companies’ part, though also self-preservation – but it also looks like enabling a dictatorship. The photographs of the tech execs supping, with not very long spoons, at Trump’s table is strangely reminiscent of the leaders of German companies when they were called in by.. someone else, nearly a century ago.
unique link to this extract

This ‘privacy browser’ has dangerous hidden features • WIRED

Matt Burgess:

»

The Universe Browser makes some big promises to its potential users. Its online advertisements claim it’s the “fastest browser,” that people using it will “avoid privacy leaks” and that the software will help “keep you away from danger.” However, everything likely isn’t as it seems.

The browser, which is linked to Chinese online gambling websites and is thought to have been downloaded millions of times, actually routes all internet traffic through servers in China and “covertly installs several programs that run silently in the background,” according to new findings from network security company Infoblox. The researchers say the “hidden” elements include features similar to malware—including “key logging, surreptitious connections,” and changing a device’s network connections.

Perhaps most significantly, the Infoblox researchers who collaborated with the United Nations Office on Drugs and Crime (UNODC) on the work, found links between the browser’s operation and Southeast Asia’s sprawling, multibillion-dollar cybercrime ecosystem, which has connections to money-laundering, illegal online gambling, human trafficking, and scam operations that use forced labor. The browser itself, the researchers says, is directly linked to a network around major online gambling company BBIN, which the researchers have labeled a threat group they call Vault Viper.

The researchers say the discovery of the browser—plus its suspicious and risky behaviour—indicates that criminals in the region are becoming increasingly sophisticated.

…Tens of thousands of web domains, plus various command-and-control infrastructure and registered companies, are linked to Vault Viper activity, Infoblox researchers say in a report shared with WIRED. They also say they examined hundreds of pages of corporate documents, legal records, and court filings with links to BBIN or other subsidiaries. Time and time again, they came across the Universe Browser online.

“We haven’t seen the Universe Browser advertised outside of the domains Vault Viper controls,” says Maël Le Touz, a threat researcher at Infoblox. The Infoblox report says the browser was “specifically” designed to help people in Asia—where online gambling is largely illegal—bypass restrictions. “Each of the casino websites they operate seem to contain a link and advertisement to it,” Le Touz says.

«

Not surprising, though, because we have already heard that these guys have stolen millions, perhaps billions from the scammed.
unique link to this extract

Volkswagen warns of output stoppages amid Nexperia chip disruption • CNBC

Sam Meredith:

»

German auto giant Volkswagen on Wednesday warned of temporary production outages citing China’s export restrictions on semiconductors made by Nexperia.

The update comes shortly after the German Association of the Automotive Industry (VDA), the country’s main car industry lobby, said the China-Netherlands dispute over Nexperia could lead to “significant production restrictions in the near future” if the supply interruption of chips cannot be swiftly resolved.

A spokesperson for Volkswagen told CNBC by email that while Nexperia is not a direct supplier of the company, some Nexperia parts are used in its vehicle components, which are supplied by Volkswagen’s direct suppliers.

“We are in close contact with all relevant stakeholders in light of the current situation to identify potential risks at an early stage and to be able to make decisions regarding any necessary measures,” a Volkswagen spokesperson said, noting that the firm’s production is currently unaffected. “However, given the evolving circumstances, short-term effects on production cannot be ruled out,” they added.

Last month, the Dutch government took control of Nexperia, a Chinese-owned semiconductor maker based in the Netherlands, in what was seen as a highly unusual move. The Dutch government seized control of the company, which specializes in the high-volume production of chips used in automotive, consumer electronics and other industries, citing fears the firm’s tech “would become unavailable in an emergency.”

China responded by blocking exports of the firm’s finished products, sparking alarm among Europe’s auto industry.

«

unique link to this extract

Errata, corrigenda and ai no corrida: none notified

Ask a chatbot for a random number between 0 and 9, and there’s a 90% chance it’ll offer seven. But why? CC-licensed photo by Niklas Morberg on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 9 links for you. Numerate. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

Largest study of its kind shows AI assistants misrepresent news content 45% of the time – regardless of language or territory • BBC Media Centre

»

New research coordinated by the European Broadcasting Union (EBU) and led by the BBC has found that AI assistants – already a daily information gateway for millions of people – routinely misrepresent news content no matter which language, territory, or AI platform is tested.

The intensive international study of unprecedented scope and scale was launched at the EBU News Assembly, in Naples. Involving 22 public service media (PSM) organizations in 18 countries working in 14 languages, it identified multiple systemic issues across four leading AI tools.

Professional journalists from participating PSM evaluated more than 3,000 responses from ChatGPT, Copilot, Gemini, and Perplexity against key criteria, including accuracy, sourcing, distinguishing opinion from fact, and providing context. 

Key findings: 

• 45% of all AI answers had at least one significant issue
• 31% of responses showed serious sourcing problems – missing, misleading, or incorrect attributions
• 20% contained major accuracy issues, including hallucinated details and outdated information
• Gemini performed worst with significant issues in 76% of responses, more than double the other assistants, largely due to its poor sourcing performance
• Comparison between the BBC’s results earlier this year and this study show some improvements but still high levels of errors.

Why this distortion matters: AI assistants are already replacing search engines for many users. According to the Reuters Institute’s Digital News Report 2025, 7% of total online news consumers use AI assistants to get their news, rising to 15% of under-25s.

“This research conclusively shows that these failings are not isolated incidents,” says EBU Media Director and Deputy Director General Jean Philip De Tender. “They are systemic, cross-border, and multilingual, and we believe this endangers public trust. When people don’t know what to trust, they end up trusting nothing at all, and that can deter democratic participation.”

…The research team have also released a News Integrity in AI Assistants Toolkit, to help develop solutions to the issues uncovered in the report. It includes improving AI assistant responses and media literacy among users. Building on the extensive insights and examples identified in the current research, the Toolkit addresses two main questions: “What makes a good AI assistant response to a news question?” and “What are the problems that need to be fixed?”.

«

It’s really easy – just stop the LLMs hallucinating, and bingo! Solved. This might take a while, unfortunately.
unique link to this extract

Amazon reportedly hopes to replace 600,000 US workers with robots • The Verge

Jess Weatherbed:

»

Amazon is reportedly leaning into automation plans that will enable the company to avoid hiring more than half a million US workers. Citing interviews and internal strategy documents, The New York Times reports that Amazon is hoping its robots can replace more than 600,000 jobs it would otherwise have to hire in the United States by 2033, despite estimating it’ll sell about twice as many products over the period.

Documents reportedly show that Amazon’s robotics team is working towards automating 75% of the company’s entire operations, and expects to ditch 160,000 US roles that would otherwise be needed by 2027. This would save about 30 cents on every item that Amazon warehouses and delivers to customers, with automation efforts expected to save the company $12.6bn from 2025 to 2027.

Amazon has considered steps to improve its image as a “good corporate citizen” in preparation for the anticipated backlash around job losses, according to The NYT, reporting that the company considered participating in community projects and avoiding terms like “automation” and “AI.” More vague terms like “advanced technology” were explored instead, and using the term “cobot” for robots that work alongside humans.

In a statement to The Verge Amazon spokesperson Kelly Nantel said the leaked documents reflect the perspective of just one team, and do not represent the company’s overall hiring strategy “now or moving forward.”

«

Yes, but what if that team is top management? Apparently 1.1m people work for Amazon in the US (out of 1.5m globally, down from 1.6m the previous year). That’s a lot of replacement with robots. But how are people going to afford to buy things from Amazon?
unique link to this extract

Jaguar Land Rover looking at $2.5bn price tag from crippling cyberattack • Financial Times via Ars Technica

Kana Inagaki and Kieran Smith:

»

The cyberattack on Jaguar Land Rover is estimated to have cost the UK at least £1.9bn in what is likely to be “the most economically damaging cyber event” for the country.

The month-long shutdown of internal systems and production at JLR affected over 5,000 British organisations, according to an analysis by Cyber Monitoring Centre, a non-profit organization that ranks the severity of cyber events in the UK.

“This incident looks to have been by some distance, the single most financially damaging cyber event ever to hit the UK,” said Ciaran Martin, former head of the National Cyber Security Centre and chair of CMC’s technical committee.

JLR, which is owned by India’s Tata Motors, only recently restarted partial production of its vehicles in the UK following a shutdown since the August 31 attack.

The severe impact on JLR’s suppliers prompted the UK government to intervene with a £1.5 billion loan guarantee to make it easier for the carmaker to access credit.

CMC mainly attributes the financial cost to the fall in vehicle sales and lower profits caused by the production halt, the costs to address the incident, and the impact on its supply chain and other local businesses.

Its estimate is also based on the assumption that JLR would not be able to fully restore its production until January and that the attackers did not infiltrate its so-called “operational technology,” which if they had, would take longer to resolve.

«

Suspected it would be the worst. Supply chains, even for comparatively small car companies, really do spread very widely.
unique link to this extract

Detection firm finds 82% of herbal remedy books on Amazon ‘likely written’ by AI • The Guardian

Aisha Down:

»

Originality.ai, which offers its tools to universities and businesses, says it scanned 558 titles published in Amazon’s herbal remedies subcategory between January and September this year, and found 82% of the books “were likely written” by AI.

“This is a damning revelation of the sheer scope of unlabelled, unverified, unchecked, likely AI content that has completely invaded [Amazon’s] platform,” wrote Michael Fraiman, author of the study.

“There’s a huge amount of herbal research out there right now that’s absolutely rubbish,” said Sue Sprung, a medical herbalist in Liverpool. “AI won’t know how to sift through all the dross, all the rubbish, that’s of absolutely no consequence. It would lead people astray.”

One of the apparently AI-written books, Natural Healing Handbook, is a No 1 bestseller in Amazon’s skincare, aroma therapies and herbal remedies, subcategories. Its introduction touts the book as “a toolkit for self-trust”, urging readers to “look inward” for solutions.

Natural Healing Handbook’s author is named as Luna Filby, whose Amazon page describes her as a “35-year-old herbalist from the coastal town of Byron Bay, Australia” and founder of the brand My Harmony Herb. Sarah Wynn, the founder of Wildcraft Journal, calls the book a “resource and an inspiration”.

However, neither Luna Filby, My Harmony Herb, Wildcraft Journal or Sarah Wynn appear to have any online presence beyond the Amazon page for the book – an indication, said Fraiman, that they may not exist. The Guardian could find no evidence of the pair. Originality.ai’s tool flagged available samples of the text as AI-generated with “100 % confidence”.

«

I’m slightly tempted to say that you get what you deserve if you think herbal remedies will help, but the bigger point is that Amazon’s self-publishing system is just irresponsible in a world of AI content being generated at the press of a button. “Caveat emptor” is an empty warning when the algorithm is pushing the invisible hand this way and that.
unique link to this extract

Apple confirms it pulled controversial dating apps Tea and TeaOnHer from the App Store • TechCrunch

Sarah Perez:

»

Controversial dating safety apps, Tea and TeaOnHer, have been pulled from the Apple App Store. The apps’ removal was first spotted by the app store intelligence provider Appfigures, which told TechCrunch the two apps were removed from the App Store on Tuesday in all markets but remain live on Google Play.

Reached for comment, Apple confirmed the apps’ removal, saying it removed Tea Dating Advice and TeaOnHer from the App Store because they failed to meet Apple’s requirements around content moderation and user privacy. The company also said it saw an excessive number of user complaints and negative reviews, which included complaints of minors’ personal information being posted in these apps.

Apple communicated the issues to the developers of the apps, a representative said, but the complaints were not addressed. (Request for comment from the app developers has not yet been returned.)

…Tea and TeaOnHer have generated a lot of headlines and interest since going viral earlier this year. Tea, which had quietly existed since 2023 before picking up steam in 2025, was pitched as a dating safety tool for women, somewhat similar to the “Are We Dating the Same Guy?” Facebook Groups. The app encouraged women to spill details about men, particularly those on dating apps. This included their personal information, Yelp-style reviews, and whether they’d dub them a “green flag” or “red flag.”

Many men, however, didn’t appreciate the app’s invasion into their privacy and questioned whether sharing information like this could be considered defamation.

After going viral and generating controversy, Tea suffered a data breach over the summer, with hackers gaining access to 72,000 images, including 3,000 selfies and photo IDs submitted for account verification, as well as 59,000 images from posts, comments, and direct messages.

«

It wasn’t the breach that did for the apps, though; it was the lack of reporting and blocking and moderation, and sharing of personal information without permission. All breaches of Apple’s App Store rules; so this is a perfectly legitimate removal.
unique link to this extract

Why are LLMs fixated on the number 7? • The Ruffian

Ian Leslie:

»

The short answer is that humans disproportionately choose 7 when asked this question. LLMs are trained on human-generated text, and 7 appears more frequently than other numbers in the training data whenever humans choose “random” numbers.

So, next question: why do humans pick 7? Well, I’m glad you asked. It turns out that our preference for this number is a well-documented phenomenon, identified in multiple psychology experiments. But there weren’t many plausible explanations of it until the publication of a 1976 paper by Yale psychologists Michael Kubovy and Joseph Psotka.

They asked 558 people to pick a random number between 0 and 9 and found that 28% of people chose 7 – a figure in line with previous experiments. Given that there are ten possible answers, this is nearly three times what you would expect from a truly random distribution.

To find out what was behind this, Kubovy and Psotka ran a few more experiments with different sets of respondents. They asked one group to choose a number between 6 and 15. This time only 17% chose 7 – a big drop. That told the researchers that the preference for ‘7’ may not actually be about ‘7’ itself. One previous hypothesis had been that we’re drawn to 7 because of its cultural resonance (seven days of the week, seven deadly sins, Ronaldo’s shirt number). But if a slight tweak to numerical context makes the preference disappear, that seems unlikely.

In another experiment, the group was asked for a random number between 0 and 9, but this time the researcher casually said, “Like 7”. They got a similar result – about 17%. That suggested that people were keen to avoid an ‘obvious’ answer.

«

This is a fascinating exploration of how LLMs reflect us back to ourselves – but amplified, so that 28% from humans becomes 90% when you ask an LLM “pick a random number between 0 and 9”.
unique link to this extract

YouTube’s likeness-detection technology has officially launched • TechCrunch

Lauren Forristal:

»

YouTube revealed on Tuesday that its likeness-detection technology has officially rolled out to eligible creators in the YouTube Partner Program, following a pilot phase. The technology allows creators to request the removal of AI-generated content that uses their likeness.

This is the first wave of the rollout, a YouTube spokesperson informed TechCrunch, adding that eligible creators received emails this morning.

YouTube’s detection technology identifies and manages AI-generated content featuring the likeness of creators, such as their face and voice.

The technology is designed to prevent people from having their likeness misused, whether for endorsing products and services they have not agreed to support or for spreading misinformation. There have been plenty of examples of AI likeness misuse in recent years, such as the company Elecrow using an AI clone of YouTuber Jeff Geerling’s voice to promote its products.

«

So weird that it’s become necessary to protect all those people on YouTube from… other people on YouTube.
unique link to this extract

The end of the old Instagram – The Atlantic

Kaitlyn Tiffany:

»

Thirty years ago, parents everywhere were compelled to weigh the pros and cons of allowing their kids to see Titanic. At the time, it was the biggest movie ever made, a historical epic (potentially educational) about mass death (possibly traumatizing) with a romantic plotline that was maybe too exciting (you know what I mean!). It was rated PG-13—a guideline that recommended caution but ultimately ruled the movie to be appropriate for millions of teenagers—resulting in a fortune for its creators and the subsequent blessings of Leonardo DiCaprio’s career.

Instagram is now adopting the same label for a teen-safety feature, but the possible outcomes are less discrete and obvious. Meta announced earlier this week that all Instagram users under the age of 18 will be automatically placed in what it’s calling a PG-13 version of the app, where only content that might appear in a PG-13 movie will, ideally, be visible. “We hope this update reassures parents that we’re working to show teens safe, age-appropriate content on Instagram by default,” the company wrote in a news post.

This is an update to an existing Teen Accounts feature, which already sought to limit exposure to graphic violent and sexual content, as well as to posts promoting cosmetic procedures and eating disorders, alcohol and tobacco sales, and other things that parents frequently worry about their kids seeing online. Although the PG-13 rating would seem to give a lot of leeway, it’s actually more restrictive than the system that was in place: It expands the internal list of worrisome content. Now, according to the update, posts about “certain risky stunts” may also be hidden, for example, while posts containing “strong language” will be removed from teens’ recommendations. Accounts that regularly share inappropriate things will be hidden from users under 18.

«

It feels like the “old” Instagram vanished years ago. All that’s happening is that each iteration keeps being thrown away. (Gift link.)
unique link to this extract

Ring’s CEO says his cameras can almost “zero out crime” within the next 12 months • The Verge

Jennifer Pattison Tuohy:

»

Jamie Siminoff has returned to Ring, the company he founded, with a renewed focus on its mission statement to “Make neighborhoods safer.” Talking to The Verge ahead of the release of his new book Ding Dong, Siminoff says he believes the new wave of AI could finally help him fulfill that vision.

“When I left, I felt like Ring had gotten to a place where it was linear innovation,” he says. But new features like Search Party, an AI-powered tool that can search your neighbors’ Ring camera footage for lost dogs, are the type of innovations he always dreamt of but couldn’t execute. “Now, with AI, we can,” he says.

While research suggests that today’s video doorbells do little to prevent crime, Siminoff believes that with enough cameras and with AI, Ring could eliminate most of it. Not all crime — “you’ll never stop crime a hundred% … there’s crimes that are impossible to stop,” he concedes — but close.

“I think that in most normal, average neighborhoods, with the right amount of technology — not too crazy — and with AI, that we can get very close to zero out crime. Get much closer to the mission than I ever thought,” he says. “By the way, I don’t think it’s 10 years away. That’s in 12 to 24 months … maybe even within a year.”

«

Don’t worry! Once the police and ICE and security services and private law enforcement are all hooked up and watching, crime will be gone! Along with any concept of privacy. Could you just point the camera inside your house so we can make sure?
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

A wiring mistake by BT meant three people were wrongly accused of downloading child abuse images. CC-licensed photo by Paul Robertson on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

There’s another post coming this week at the Social Warming Substack on Friday at 0845 UK time. Free signup.

A selection of 10 links for you. Crossed. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

ChatGPT Atlas: OpenAI launches web browser centered around its chatbot • The Guardian

Johana Bhuiyan and agency:

»

OpenAI on Tuesday launched an AI-powered web browser built around its marquee chatbot.

“Meet our new browser—ChatGPT Atlas,” a tweet from the company read.

The browser is designed to provide a more personalized web experience and includes a ChatGPT sidebar that enables users to asks questions about or engage with various aspects of each website they visit, as demonstrated in a video posted alongside the announcement. Atlas is now available globally on Apple’s Mac operating system and will soon be made available on Windows, iOS and Android, according to OpenAI’s announcement.

Users can open the ChatGPT sidebar to and ask it to “summarize content, compare products, or analyze data from any site”, the company website reads. The company has also started to roll out a preview of a virtual assistant dubbed “Agent Mode” to certain premium accounts. Agent Mode allows users to ask ChatGPT to complete tasks “from start to finish” such as “researching and shopping for a trip”.

The browser also enables ChatGPT to edit and alter highlighted text. An example on the website shows an email with highlighted text and a suggested prompt: “Make this sound more professional.”

The company says that users have complete control over their privacy settings: “You control what it remembers about you, how your data is used, and the privacy settings that apply while you browse.” At the moment, Atlas users will be automatically opted out of allowing their browsing data to be used to train ChatGPT models, for instance. And, like in other browsers, a user can delete their browsing and web history.

«

The underlying web engine is Chromium, the open source of Google’s Chrome. So it’s going to look familiar, but of course it’s going to be phoning home to ChatGPT all the time. Natural language searching of your web history; but not yet clear what its search index is. Presumably not Google?
unique link to this extract

BT wiring fault led to three falsely accused of child abuse image • BBC News

Owain Evans:

»

Three people were wrongly accused of downloading child abuse images due to a broadband wiring error by a BT engineer, a tribunal has heard.

The mistake meant internet activity linked to the real offender was traced instead to the address where two men and a woman were staying, who had their electronic devices seized over the course of two police searches. The false accusations back in 2016 had “highly distressing and far-reaching” consequences for the three, the Investigatory Powers Tribunal (IPT) was told.

It ruled that Dyfed-Powys Police had acted lawfully, and found that the error was caused by a technical fault rather than police misconduct.

BT has been approached for comment.

The IPT deals with complaints from anyone who feels they have been the victim of unlawful action by a public body using covert investigative techniques. Each of the three claimants were granted anonymity by the tribunal, and the location of the incident was only described as Dyfed-Powys Police’s “area of operation in Wales”, which covers Carmarthenshire, Ceredigion, Pembrokeshire and Powys.

British telecommunication firm BT told the tribunal that two wires within a street cabinet linking to both addresses had been inadvertently crossed. As a result the offending IP address had been incorrectly attributed to the address of the the first male claimant who was its registered tenant, which he shared with a friend whose girlfriend was also visiting at the time.

The tribunal heard all three had to tell their employers about the accusations made against them. The innocent woman was advised her children could not live with her alone until she was cleared, and the two men faced child protection referrals. The first male claimant was placed on restricted duties at work and the second had a job offer withdrawn.

«

This is unimaginably awful. And it happened nine years ago! At least the three were never charged; the actual offender was later identified. But there’s no compensation for those wrongly accused.
unique link to this extract

Europol dismantles SIM farm network powering 49 million fake accounts worldwide • Hacker News

Ravie Lakshmanan:

»

Europol on Friday announced the disruption of a sophisticated cybercrime-as-a-service (CaaS) platform that operated a SIM farm and enabled its customers to carry out a broad spectrum of crimes ranging from phishing to investment fraud.

The coordinated law enforcement effort, dubbed Operation SIMCARTEL, saw 26 searches carried out, resulting in the arrest of seven suspects and the seizure of 1,200 SIM box devices, which contained 40,000 active SIM cards. Five of those detained are Latvian nationals.

In addition, five servers were dismantled and two websites gogetsms[.]com and apisim[.]com) advertising the service was taken over on October 10, 2025, to display a seizure banner. Separately, four luxury vehicles were confiscated, and €431,000 ($502,000) in suspects’ bank accounts and €266,000 ($310,000) in their cryptocurrency accounts were frozen.

The countries that participated in the operation comprised authorities from Austria, Estonia, Finland, and Latvia, in collaboration with Europol and Eurojust.

According to Europol, the criminal network has been attributed to more than 1,700 individual cyber fraud cases in Austria and 1,500 in Latvia, leading to losses totaling around €4.5m ($5.25m) and €420,000 ($489,000) in the two countries, respectively.

“The criminal network and its infrastructure were technically highly sophisticated and enabled perpetrators around the world to use this SIM-box service to conduct a wide range of telecommunications-related cybercrimes, as well as other crimes,” the agency said.

«

Imagine how many more of these there might be around the world. We know the US Secret Service found one. It’s not hard to hide them. Hundreds? Thousands? Even.. millions?
unique link to this extract

NSO permanently barred from targeting WhatsApp users with Pegasus spyware • Ars Technica

Dan Goodin:

»

A federal judge has ordered spyware maker NSO to stop using its Pegasus app to target or infect users of WhatsApp.

The ruling, issued Friday by Phyllis J. Hamilton of the US District Court of the District of Northern California, grants a permanent injunction sought by WhatsApp owner Meta in a case it brought against NSO in 2019. The lawsuit alleged that Meta caught NSO trying to surreptitiously infect about 1,400 mobile phones—many belonging to attorneys, journalists, human-rights activists, political dissidents, diplomats, and senior foreign government officials—with Pegasus. As part of the campaign, NSO created fake WhatsApp accounts and targeted Meta infrastructure. The suit sought monetary awards and an injunction against the practice.

Friday’s ruling ordered NSO to permanently cease targeting WhatsApp users, attempting to infect their devices, or intercepting WhatsApp messages, which are end-to-end encrypted using the open source Signal Protocol. Hamilton also ruled that NSO must delete any data it obtained when targeting the WhatsApp users.

NSO had argued that such a ruling would “force NSO out of business,” as Pegasus is its “flagship product.” Hamilton ruled that the harm Pegasus posed to Meta outweighed any such considerations.

“In the court’s view, any business that deals with users’ personal information, and that invests resources into ways to encrypt that personal information, is harmed by the unauthorized access of that personal information—and it is more than just a reputational harm, it’s a business harm,” Hamilton wrote. “Essentially, part of what companies such as Whatsapp are ‘selling’ is informational privacy, and any unauthorized access is an interference with that sale. Defendants’ conduct serves to defeat one of the purposes of the service being offered by plaintiffs, which constitutes direct harm.”

The judge went on to deny Meta’s request that the injunction bar foreign governments that may use WhatsApp. She said that sovereign governments weren’t parties to the lawsuit. Friday’s ruling also denied Meta’s request that the injunction bar NSO from targeting users of other Meta properties such as Facebook and Instagram on the grounds there was no evidence presented concerning targeting of them.

«

Without reading the ruling, does “users of WhatsApp” mean “people who have a WhatsApp account/the app on their phone” or “active users of WhatsApp who were targeted through that app”? They seem different.
unique link to this extract

Tahoe Electron detector • furbo.org

Craig Hockenberry:

»

No, we’re not doing science at California’s most beautiful lake. We’re looking for bugs.

A popular cross-platform app development framework called Electron is using private and undocumented API that’s causing system-wide slowdowns in macOS Tahoe.

We’re hearing from customers that some of our apps are running slowly on Tahoe and I suspect that this bug has something to do with it. Unfortunately, it’s hard for customers to check which version of Electron is being used and see if that might be a cause. So I decided to do something about that…

Luckily there’s a script written by Tomas Kafka that lets you check all your apps quickly and easily. I took that script, updated some parts that required Xcode to be installed, and wrapped it up in an AppleScript applet that’s easy to download and run.

…If you’re one of those people who’s wondering when it’s a good time to upgrade to Tahoe, you can run TahoeElectronDetector on older versions of macOS and give yourself an idea of when it’s safe to move to the new operating system.

Additionally, there’s a website that lists the status of the most popular apps. This will be helpful in locating newer versions since some of them will not update automatically.

«

unique link to this extract

ICE’s “athletically allergic” recruits • The Atlantic

Nick Miroff:

»

President Donald Trump’s plan to double the size of the ICE workforce has met a foe more powerful than any activist group. It is decimating new recruits at the agency’s training academy in Georgia. It is the ICE personal-fitness test.

More than a third have failed so far, four officials told me, impeding the agency’s plan to hire, train, and deploy 10,000 deportation officers by January. To pass, recruits must do 15 push-ups and 32 sit-ups, and run 1.5 miles (2.4km) in 14 minutes.

“It’s pathetic,” one career ICE official told me, adding that before now, a typical class of 40 recruits had only a couple of candidates fail, because the screening process was more rigorous.

The academy’s standards have already been eased to boost recruitment, he said, and the new parameters “should be the minimum for any officer.” He and others, none of whom were authorized to speak with reporters, told me that agency veterans are concerned about the quality of the new recruits being fast-tracked onto the street to meet Trump’s hiring goals.

An email from ICE headquarters to the agency’s top officials on October 5 lamented that “a considerable amount of athletically allergic candidates” had been showing up to the academy; they had “misrepresented” their physical condition on application forms. The email directed leaders at ICE’s field offices to conduct preliminary fitness exams with new recruits before sending them to the academy.

“We all know the self-certification method has failed,” Ralph Ferguson, an operations official at ICE headquarters, wrote.

The Department of Homeland Security spokesperson Tricia McLaughlin told me in a statement that the one-third failure rate reflected only “a subset of candidates in initial basic academy classes,” and not all new hires. She said DHS expects to fill 85% of new deportation-officer positions with experienced law-enforcement officials whom they can fast-track.

«

That’s utterly incredible. The running test is a pace of 5’50” per km, which ought to be feasible for anyone under 50 not weighed down by avoirdupois. But the evidence from video is that ICE recruits are not, by any means, fleet of foot and rely instead on proximity and guns. (Gift link.)
unique link to this extract

‘Significant exposure’: Amazon Web Services outage exposed UK state’s £1.7bn reliance on tech giant • The Guardian

Simon Goodley:

»

AWS has won 189 UK government contracts worth £1.7bn since 2016 – during which time it has invoiced about £1.4bn, according to the figures compiled by Tussell, a public procurement intelligence firm.

The research group added that “35 public sector authorities currently use [AWS] services across 41 contracts worth a combined £1.1bn. Key ministerial departments have contracts with the company such as the Home Office, DWP, HMRC, [the Ministry of Justice], the Cabinet Office and Defra.”

Tim Wright, a technology partner at the law firm Fladgate, said: “That’s a very significant exposure and its pretty ironic considering how the FCA [Financial Conduct Authority] and the PRA [Prudential Regulation Authority] have repeatedly highlighted the dangers of concentration risk in cloud service provision for regulated entities for a number of years.

“Recent moves by HM Treasury, the PRA and FCA to establish direct oversight of ‘critical third parties’ aim to address precisely the risk of outages like that suffered by AWS, yet until significant diversification or sovereign cloud adoption occurs, the UK government’s own stance shows an uncomfortable contradiction with the very resilience principles regulators have advocated.”

The House of Commons’ treasury committee has written to the economic secretary to the Treasury, Lucy Rigby, to ask why the government had not yet designated Amazon a “critical third party” to the UK’s financial services sector – which would expose the tech firm to financial regulatory oversight.

…Among the UK government contracts, only HMRC said it was affected. It said that customers were “having problems accessing our online services”, and urged them to call back later as its phone lines were busy.

«

HMRC is always busy, so to be honest that’s not unexpected. But this is clearly a critical risk.
unique link to this extract

AWS crash causes $2,000 Smart Beds to overheat and get stuck upright • Dexerto

Calum Patterson:

»

A major Amazon Web Services (AWS) outage on October 20 had the unexpected side effect of causing chaos in bedrooms across the US, as owners of Eight Sleep’s $2,000+ ‘Pod’ mattress covers found their smart beds had no offline mode and were stuck at high temperatures and odd positions in the night.

The outage began around 3 am ET, when AWS reported “increased error rates and latencies” in its US-EAST-1 region. By mid-morning, Downdetector had logged more than eight million reports of disruptions affecting apps, games, and banking platforms.

Eight Sleep’s products rely on cloud connectivity to control temperature and track biometric data. When AWS went down, users lost access to the app that manages its water-cooled coils, leaving them stuck with whatever setting was last active.

Some beds overheated, others stopped cooling altogether, and several users said their devices became completely unresponsive.

One viral post from tech enthusiast Alex Browne summed up the absurdity after his Pod locked itself nine degrees above room temperature. “Backend outage means I’m sleeping in a sauna,” he wrote. “Eight Sleep confirmed there’s no offline mode yet, but they’re working on it.”

«

No offline mode! I know, you’re thinking: how could anyone be that stupid? Except there’s a wrinkle in the US tax code called ASC 606 which means you can’t book your revenue as “software as a service” (SaaS) unless the product is crippled without an internet connection. This is a problem for Eight Sleep: it can promise an offline version, but what if everyone starts using that?
unique link to this extract

Apple alerts exploit developer that his iPhone was targeted with government spyware • TechCrunch

Lorenzo Franceschi-Bicchierai:

»

Earlier this year, a developer was shocked by a message that appeared on his personal phone: “Apple detected a targeted mercenary spyware attack against your iPhone.” 
 
“I was panicking,” Jay Gibson, who asked that we don’t use his real name over fears of retaliation, told TechCrunch.  

Gibson, who until recently built surveillance technologies for Western government hacking tools maker Trenchant, may be the first documented case of someone who builds exploits and spyware being themselves targeted with spyware. 

“What the hell is going on? I really didn’t know what to think of it,” said Gibson, adding that he turned off his phone and put it away on that day, March 5. “I went immediately to buy a new phone. I called my dad. It was a mess. It was a huge mess.”  

At Trenchant, Gibson worked on developing iOS zero-days, meaning finding vulnerabilities and developing tools capable of exploiting them that are not known to the vendor who makes the affected hardware or software, such as Apple.  

“I have mixed feelings of how pathetic this is, and then extreme fear because once things hit this level, you never know what’s going to happen,” he told TechCrunch.  

But the ex-Trenchant employee may not be the only exploit developer targeted with spyware. According to three sources who have direct knowledge of these cases, there have been other spyware and exploit developers in the last few months who have received notifications from Apple alerting them that they were targeted with spyware. 

«

I’m surprised he’s surprised: if you’re searching for zero-days, you’re obviously going to be of interest to state hackers who would like to get their hands on the zero-days without having to pay for them. And they’d also like to know who those zero-days are going to. Trenchant needs better employee briefing, is my take.
unique link to this extract

Book excerpt: Taiwan’s undersea cables face growing threats • Rest of World

Samanth Subramanian:

»

Not long after the cables in the Matsu islands were cut, Taiwan’s communications authority proposed heavy criminal penalties for anyone who damaged subsea cables: a fine of up to $3.2m and life in prison. The law is both harsh and, in the case of foreign actors, essentially meaningless. How would a Taiwanese court even begin to try the Chinese crew of a long-gone fishing vessel?

At present, there is no effective, coherent body of law to hold responsible saboteurs of cables at sea. The only guides available are a mess of national regulations and the UN Convention on the Law of the Sea. Jurisdictions overlap furiously: if, out in international waters, a ship flagged in Panama and operated by an Indian crew cuts a cable that lands in several countries along the west African coast and that is co-owned by British, South African, and American companies, who is the perpetrator, who is the victim, and where would a trial take place? The law around undersea cables turns out to be just as murky and uncertain as the submarine depths in which these cables lie.

For more than a century, the positions of cables at sea have been recorded carefully in maps and published — the better to warn ships to avoid them. “But if this data is used the other way, it becomes a vulnerability,” Chiueh said. “All countries face this problem now.”

«

Elon Musk must be rubbing his hands in delight at the prospect of all those undersea cables being in trouble. But: satellites can be attacked too.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

A YouTube documentary about birdwatching gives an insight into how app gamify, then corrupt, hobbies. CC-licensed photo by JuliaC2006 on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 9 links for you. Over there! I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

AWS outage exposes Achilles heel: central control plane • The Register

Dan Robinson:

»

The problems began just after midnight US Pacific Time today when Amazon Web Services (AWS) noticed increased error rates and latencies for multiple services running within its home US-EAST-1 region.

Within a couple of hours, Amazon’s techies had identified DNS as a potential root cause of the issue – specifically the resolution of the DynamoDB API endpoint in US-EAST-1 – and were working on a fix.

However, it was affecting other AWS services, including global services and or features that rely on endpoints operating from AWS’ original region, such as IAM (Identity and Access Management) updates and DynamoDB global tables.

While Amazon worked to fully resolve the problem, the issue was already causing widespread chaos to websites and online services beyond the Northern Virginia locale of US-EAST-1, and even outside of America’s borders.

As The Register reported earlier, Amazon.com itself was down for a time, while the company’s Alexa smart speakers and Ring doorbells stopped working. But the effects were also felt by messaging apps such as Signal and WhatsApp, while in the UK, Lloyds Bank and even government services such as tax agency HMRC were impacted.

According to a BBC report, outage monitor Downdetector indicated there had been more than 6.5 million reports globally, with upwards of 1,000 companies affected.

How could this happen? Amazon has a global footprint, and its infrastructure is split into regions, physical locations with a cluster of datacenters. Each region consists of a minimum of three isolated and physically separate availability zones (AZ), each with independent power and connected via redundant, ultra-low-latency networks.

Customers are encouraged to design their applications and services to run in multiple AZs to avoid being taken down by a failure in one of them.

Sadly, it seems that the entire edifice has an Achilles heel that can cause problems regardless of how much redundancy you design into your cloud-based operations, at least according to the experts we asked.

“The issue with AWS is that US East is the home of the common control plane for all of AWS locations except the federal government and European Sovereign Cloud. There was an issue some years ago when the problem was related to management of S3 policies that was felt globally,” Omdia Chief Analyst Roy Illsley told us.

He explained that US-EAST-1 can cause global issues because many users and services default to using it since it was the first AWS region, even if they are in a different part of the world.

«

Grateful to Microsoft Security’s Matt Zorich, who published his handy guide: Something is down » Is it a telco? Yes » It’s BGP [Border Gateway Protocol]. No? » It’s DNS.
unique link to this extract

Life-changing eye implant helps blind patients read again • BBC News

Fergus Walsh:

»

A group of blind patients can now read again after being fitted with a life-changing implant at the back of the eye.

A surgeon who inserted the microchips in five patients at Moorfields Eye Hospital in London says the results of the international trial are “astounding”.

Sheila Irvine, 70, who is registered blind, told the BBC it was “out of this world” to be able to read and do crosswords again. “It’s beautiful, wonderful. It gives me such pleasure.”

The technology offers hope to people with an advanced form of dry age-related macular degeneration (AMD), called geographic atrophy (GA), which affects more than 250,000 people in the UK and five million worldwide. In those with the condition – which is more common in older people – cells in a tiny area of the retina at the back of the eye gradually become damaged and die, resulting in blurred or distorted central vision. Colour and fine detail are often lost.

The new procedure involves inserting a tiny 2mm-square photovoltaic microchip, with the thickness of a human hair, under the retina. Patients then put on glasses with a built-in video camera. The camera sends an infrared beam of video images to the implant at the back of the eye, which sends them on to a small pocket processor to be enhanced and made clearer.

The images are then sent back to the patient’s brain, via the implant and optic nerve, giving them some vision again. The patients spent months learning how to interpret the images.

Mahi Muqit, consultant ophthalmic surgeon at Moorfields Eye Hospital in London, who led the UK arm of the trial, told the BBC it was “pioneering and life-changing technology”.

…Of 32 patients given the implant, 27 were able to read again using their central vision. After a year, this equated to an improvement of 25 letters, or five lines, on an eye chart.

«

unique link to this extract

The unexpected profundity of a movie about birdwatching • The Atlantic

Tyler Austin Harper:

»

The new YouTube documentary Listers is a down-the-rabbit-hole glimpse at the norms and neuroses of the “extreme bird-watching” community. If that sounds painfully boring, it’s not—this is one of the funniest documentaries I’ve seen in some time. In it, the brothers Quentin and Owen Reiser chronicle their try at a “big year,” a bird-nerd term for attempting to identify as many different species as possible in a single calendar year.

They start out knowing next to nothing about birds—an app designer and a cinematographer from Collinsville, Illinois, the Reisers get into birding after one of them stumbles across an ornithological guidebook during a bleary-eyed smoke session. Then they buy a $4,500 Kia Sedona and traverse the country with the goal of finding more than 700 unique specimens.

Although both brothers are the subjects of the film, Quentin spends most of the time on camera while Owen remains behind the lens. He alternates between a low-tech camcorder and a high-resolution camera, the former to capture the mundane and often gritty work of tracking down birds, and the latter to reveal their quarry in all its splendor.

…But underneath the stoner hijinks (and legitimately stunning wildlife videography), Listers is a serious film about the meaning that hobbies can provide to our lives, and the corrupting influence of smartphone apps on our leisure activities.

As the documentary progresses, it gradually begins to examine how eBird, a social app that is popular in the bird-watching community, has overtaken the pastime. The brothers start without complaints about eBird, which connects them with other hobbyists, helps them track their progress when they “list” birds, and provides a ranking system so they can see how they’re stacking up against other birders.

But by the end of the year, they become disillusioned by eBird and interview other hobbyists who are as well. “This country is so big, and you have to go everywhere in the country to see enough birds to be in the power rankings or whatever the fuck it is,” Quentin grouses. “I like bird-watching, but I don’t like it in the competitive sense.”

«

unique link to this extract

The future I saw through the Meta Ray-Ban Display amazes and terrifies me • The Verge

Victoria Song:

»

These glasses do everything that the audio-only Ray-Ban Meta glasses do, but the display lets them introduce a variety of new features that you’d previously need to pull out your phone for. You can reply to texts, view Instagram Reels, frame photos and videos, caption or translate the conversations happening around you, and get walking directions while viewing a map of your surroundings. When you interact with Meta AI, you can now see informational cards.

It took a while to figure out where the display fits into my life. With the original audio-only Ray-Bans, the use case was clear-cut. I pop them on when I go for walks or attend events like a concert, where I might want footage. But as an able-bodied, sighted person, I’ve never found the Meta AI features that useful in my day-to-day life. I might use the glasses to identify a flower or tree I see on a walk, but that’s as far as it went.

A display opens more doors. One of the big features is live captions, which adds real-time subtitles to your conversations. It was helpful to turn on live captions during a podcast taping. They’re not always perfect — AI transcriptions universally struggle with slang or uncommon names — but they’re nice in one-on-one conversations in a noisy restaurant. The transcriptions are less useful when you’re walking and talking with a friend, though. For it to work well, you have to be looking directly at your conversation partner. That makes it awkward if you’re walking side by side, as you physically have to turn your head to face each speaker. The AI also had a hard time captioning my mumbly spouse. No amount of yelling in a loud bar can help the AI, either.

…Texting on the glasses ties all of this together — and also starts to get at where using them becomes uncomfortable. It feels magical when, at dinner, I can hide my hand under the table, read a text, swipe my fingers, and reply to a message without anyone knowing. But the problem is that much of what makes texting (and photos, and videos, and so many other features enabled by the display) so impressive is that no one else can tell what you’re doing.

«

Heavy, two different chargers for the neural band and the glasses (which have short battery lives). But: definitely a glimpse of the future.
unique link to this extract

Wikipedia says AI is causing a dangerous decline in human visitors • 404 Media

Emanuel Maiberg:

»

Ironically, while generative AI and search engines are causing a decline in direct traffic to Wikipedia, its data is more valuable to them than ever. Wikipedia articles are some of the most common training data for AI models, and Google and other platforms have for years mined Wikipedia articles to power its Snippets and Knowledge Panels, which siphon traffic away from Wikipedia itself.

“Almost all large language models train on Wikipedia datasets, and search engines and social media platforms prioritize its information to respond to questions from their users,” Miller said. That means that people are reading the knowledge created by Wikimedia volunteers all over the internet, even if they don’t visit wikipedia.org— this human-created knowledge has become even more important to the spread of reliable information online.”

Miller said that in May 2025 Wikipedia noticed unusually high amounts of apparently human traffic originating mostly from Brazil. He didn’t go into details, but explained this caused the Foundation to update its bot detections systems.
 
“After making this revision, we are seeing declines in human pageviews on Wikipedia over the past few months, amounting to a decrease of roughly 8% as compared to the same months in 2024,” he said. “We believe that these declines reflect the impact of generative AI and social media on how people seek information, especially with search engines providing answers directly to searchers, often based on Wikipedia content.”

«

Probably the most telling stat is “Edited pages” for Wikipedia, which (in English) is down 20% year-on-year. There’s also “Edits” which, for all Wikipedias (in multiple languages) is down 4% in the same period.
unique link to this extract

ESMO 2025: mRNA-based COVID vaccines generate improved responses to immunotherapy • MD Anderson Cancer Center

»

Patients with cancer who received mRNA-based COVID vaccines within 100 days of starting immune checkpoint therapy were twice as likely to be alive three years after beginning treatment, according to a new study led by researchers at The University of Texas MD Anderson Cancer Center.

These findings, which include more than 1,000 patients treated between Aug. 2019 and Aug. 2023, were presented today at the 2025 European Society for Medical Oncology (ESMO) Congress (Abstract LBA54). The study was led by Steven Lin, M.D., Ph.D. professor of Radiation Oncology, and Adam Grippin, M.D., Ph.D.,senior resident in Radiation Oncology.

“This study demonstrates that commercially available mRNA COVID vaccines can train patients’ immune systems to eliminate cancer,” Grippin said. “When combined with immune checkpoint inhibitors, these vaccines produce powerful antitumor immune responses that are associated with massive improvements in survival for patients with cancer.”

The discovery that mRNA vaccines were powerful immune activators came from research conducted by Grippin during his graduate work at the University of Florida in the lab of Elias Sayour, M.D., Ph.D. While developing personalized mRNA-based cancer vaccines for brain tumors, Grippin and Sayour found that mRNA vaccines trained immune systems to eliminate cancer cells, even when the mRNA didn’t target tumors directly.

This finding led to the hypothesis that other types of mRNA vaccines might have the same effect, and the approval and use of mRNA-based COVID vaccines created an opportunity to test this hypothesis. Lin and Grippin initiated a major effort to retrospectively study if MD Anderson patients who received mRNA COVID vaccines lived longer than those who did not receive these vaccines.

«

In short, the mRNA vaccine – and it has to be mRNA; others don’t have the same effect – seems to wake up the immune system and gets it to “notice” tumours and act on them. Work in healthy people too: immune systems are stronger following Covid vaccination.
unique link to this extract

Ibuprofen: how an everyday drug might offer protection against cancer • The Conversation

Dipa Kamdar, Ahmed Elbediwy and Nadine Wehida:

»

A 2025 study found that ibuprofen may lower the risk of endometrial cancer, the most common type of womb cancer, which starts in the lining of the uterus (the endometrium) and mainly affects women after menopause.

One of the biggest preventable risk factors for endometrial cancer is being overweight or obese, since excess body fat increases levels of oestrogen – a hormone that can stimulate cancer cell growth.

Other risk factors include older age, hormone replacement therapy (particularly oestrogen-only HRT), diabetes, and polycystic ovary syndrome. Early onset of menstruation, late menopause, or not having children also increase risk. Symptoms can include abnormal vaginal bleeding, pelvic pain, and discomfort during sex.

In the Prostate, Lung, Colorectal, and Ovarian (PLCO) study, data from more than 42,000 women aged 55–74 was analysed over 12 years. Those who reported taking at least 30 ibuprofen tablets per month had a 25% lower risk of developing endometrial cancer than those taking fewer than four tablets monthly. The protective effect appeared strongest among women with heart disease.

Interestingly, aspirin – another common NSAID – did not show the same association with reduced risk in this or other studies. That said, aspirin may help prevent bowel cancer returning.

Other NSAIDs, such as naproxen, have been studied for preventing colon, bladder, and breast cancers. The effectiveness of these drugs seems to depend on cancer type, genetics, and underlying health conditions.

Ibuprofen’s possible cancer-protective effects extend beyond endometrial cancer. Studies suggest it may also reduce risk of bowel, breast, lung, and prostate cancers.

«

My only quibble with this is that “30 ibuprofen tablets” could be anything from 1,500mg to 12,000mg. Not all tablets are the same. Anyway, two cancer cures in one day! It’s like the Daily Mail in here.
unique link to this extract

Apple sued over use of copyrighted books to train Apple Intelligence • Reuters

Blake Brittain:

»

Apple was hit with a lawsuit in California federal court by a pair of neuroscientists who say that the tech company misused thousands of copyrighted books to train its Apple Intelligence artificial intelligence model.

Susana Martinez-Conde and Stephen Macknik, professors at SUNY Downstate Health Sciences University in Brooklyn, New York, told the court in a proposed class action last Thursday that Apple used illegal “shadow libraries” of pirated books to train Apple Intelligence.

A separate group of authors sued Apple last month for allegedly misusing their work in AI training.

The lawsuit is one of many high-stakes cases brought by copyright owners such as authors, news outlets, and music labels against tech companies, including OpenAI, Microsoft, and Meta Platforms over the unauthorized use of their work in AI training. Anthropic agreed to pay $1.5bn to settle a lawsuit from another group of authors over the training of its AI-powered chatbot Claude in August.

Spokespeople for Apple and Martinez-Conde, Macknik, and their attorney did not immediately respond to requests for comment on the new complaint on Friday.

Apple Intelligence is a suite of AI-powered features integrated into iOS devices, including the iPhone and iPad.
“The day after Apple officially introduced Apple Intelligence, the company gained more than $200 billion in value: ‘the single most lucrative day in the history of the company,'” the lawsuit said.

According to the complaint, Apple utilized datasets comprising thousands of pirated books as well as other copyright-infringing materials scraped from the internet to train its AI system.

«

Can’t make an LLM – even a little one, an LLM-ette – without breaking regs, it seems.
unique link to this extract

London became a global hub for phone theft. Now we know why • The New York Times

Lizzie Dearden and Amelia Nierenberg:

»

For years, London’s police assumed most of the phone thefts were the work of small-time thieves looking to make some quick cash. But last December, they got an intriguing lead from a woman who had used “Find My iPhone” to track her device to a warehouse near Heathrow Airport. Arriving there on Christmas Eve, officers found boxes bound for Hong Kong. They were labeled as batteries but contained almost 1,000 stolen iPhones.

“It quickly became apparent this wasn’t just normal low-level street crime,” said Mark Gavin, a senior detective leading the investigation for the Metropolitan Police. “This was on an industrial scale.”

The breakthrough coincided with a broader push by the police to increase public confidence by tackling the city’s most common crimes. Phone theft has been the subject of particular anger among victims, who for years reported their cellphones stolen and handed the police the locations being transmitted, only to be given a crime reference number and hear nothing more.

The police are now using that information to map where stolen phones are transported by street thieves. After the Heathrow seizure, a team of specialist investigators who normally deal with firearms and drug smuggling was assigned to the case. They identified further shipments and used forensics to identify two men in their 30s who are suspected of being ringleaders of a group that sent up to 40,000 stolen phones to China.

When the men were arrested on Sept. 23, the car they were traveling in contained several phones, some wrapped in aluminum foil in an attempt to prevent them from transmitting tracking signals. At one point, the police said at a news conference, they observed the men buying almost 1.5 miles’ worth of foil in Costco.

Some phones are reset and sold to new users in Britain. But many are shipped to China and Algeria as part of a “local-to-global criminal business model,” the police said, adding that in China, the newest phones could be sold for up to $5,000, generating huge profits for the criminals involved.

Joss Wright, an associate professor at the University of Oxford who specializes in cybersecurity, said that it is easier to use stolen British phones in China than elsewhere because many of the country’s network providers do not subscribe to an international blacklist that bars devices that have been reported stolen.

“That means that a stolen iPhone that has been blocked in the U.K. can be used without any problems in China,” Mr. Wright said.

«

I hope this is a gift link. It’s a fascinating article, and that last point – about China not subscribing to the IMEI blocking – is crucial. And with the thieves getting around £300 per device, that’s tasty money for them.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

The British Transport Police say they won’t check CCTV over two hours for bicycle theft. But one person has created a tool that could review that in around 20 seconds. CC-licensed photo by Dan4th Nicholas on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 9 links for you. Reviewed. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

Apple pays $750m for US Formula 1 streaming coverage • Ars Technica

Jonathan Gitlin:

»

The United States Grand Prix takes place this weekend at the Circuit of the Americas in Texas, and this morning, Formula 1 used the occasion to announce a new broadcast deal for the sport in the US. Starting next year, F1 will no longer be broadcast on ESPN—it’s moving to Apple TV in a five-year, $750m deal.

Apple boss Tim Cook has been seen at F1 races in the past, and earlier this year, Apple released F1: The Movie, starring Brad Pitt as a 50-something racing driver who improbably gets a second bite at the cherry 30 years after a brutal crash seemingly ended his F1 career.

But securing the rights to the sport itself means Apple has snagged a very fast-growing series, with races almost every other week—currently, the sport has expanded to 24 races a year.

“We are no strangers to each other, having spent the past three years working together to create F1: The Movie, which has already proven to be a huge hit around the world. We have a shared vision to bring this amazing sport to our fans in the US and entice new fans through live broadcasts, engaging content, and a year-round approach to keep them hooked,” said Stefano Domenicali, F1 president and CEO.

Apple says Apple TV subscribers will be able to watch every practice and qualifying session, as well as all the sprint races and grands prix. And “select races and all practice sessions will also be available for free in the Apple TV app throughout the course of the season,” the company said.

Apple also plans to “amplify the sport” through its other channels—Apple News, Apple Maps, Apple Music, and Apple Fitness+. There will even be a designated widget for iPhone home screens.

That obviously means no more coverage on ESPN, a channel that many cable subscribers get as part of their packages.

«

That works out (if it continues at 24 races per year) to just $6.25m per race, which is pretty cheap for hours of programming. Wonder what the Fitness+ amplification will be: sit like an F1 driver for hours? Or does that mean “advertise”? Also, no word on whether there will be Vision Pro content. But that, too, seems like a no-brainer. Surely – surely – they won’t miss this chance?
unique link to this extract

Everything is television • Derek Thompson

Derek Thompson:

»

A spooky convergence is happening in media. Everything that is not already television is turning into television.

…When podcasts got started, they were radio for the Internet. This really appealed to me when I started my show. I never watch the news on television, and I love listening to podcasts while I make coffee and go on walks, and I’d prefer to make the sort of media that I consume. Plus, as a host, I thought I wanted to have conversations focused on the substance of the words rather than on ancillary concerns about production value and lighting.

But the most successful podcasts these days are all becoming YouTube shows. Industry analysts say consumption of video podcasts is growing twenty times faster than audio-only ones, and more than half of the world’s top shows now release video versions. YouTube has quietly become the most popular platform for podcasts, and it’s not even close. On Spotify, the number of video podcasts has nearly tripled since 2023, and video podcasts are significantly outgrowing non-video podcasts. Does it really make sense to insist on an audio-only podcast in 2025? I do not think so. Reality is screaming loudly in my ear, and its message is clear: Podcasts are turning into television.

…… and why does this matter? Fine question. And, perhaps, this is a good place for a confession. I like television. I follow some spectacular YouTube channels. I am not on Instagram or TikTok, but most of the people I know and love are on one or both. My beef is not with the entire medium of moving images. My concern is what happens when the grammar of television rather suddenly conquers the entire media landscape.

In the last few weeks, I have been writing a lot about two big trends in American life that do not necessarily overlap. My work on the “Antisocial Century” traces the rise of solitude in American life and its effects on economics, politics, and society. My work on “the end of thinking” follows the decline of literacy and numeracy scores in the U.S. and the handoff from a culture of literacy to a culture of orality. Neither of these trends is exclusively caused by the logic of television colonizing all media. But both trends are significantly exacerbated by it.

«

His piece is not just about podcasts (they’re oral!) but how “TV” – aka video – has become the powerful attractor for everything.
unique link to this extract

Famous cognitive psychology experiments that failed to replicate • Aether Mug

Marco Giancotti:

»

The field of psychology had a big crisis in the 2010s, when many widely accepted results turned out to be much less solid than previously thought. It’s called the replication crisis, because labs around the world tried and failed to replicate, in new experiments, previous results published by their original “discoverers”. In other words, many reported psychological effects were either non-existent—artifacts of the experimenter’s flawed setup—or so much weaker than originally claimed that they lost most of their intellectual sparkle.

(The crisis spanned other fields as well, but I mostly care about psychology here, especially the cognitive kind.)

This is very old news, and I’ve been vaguely aware of several of the biggest disgraced results for years, but I keep on forgetting which are (still probably) real and which aren’t. This is not good. Most results in the field do actually replicate and are robust [maybe], so it would be a pity to lose confidence in the whole field just because of a few bad apples.

This post is a compact reference list of the most (in)famous cognitive science results that failed to replicate and should, for the time being, be considered false. The only goal is to offset the trust-undermining effects of my poor memory—and perhaps yours, too?—with a bookmarkable page.

This can’t be a comprehensive list: if a study is not on this page, it’s not guaranteed to be fully replicated. Still, this should cover most of the high-profile debunked theories that laypeople like me may have heard of.

«

You’ve surely heard of the marshmallow experiment (children who didn’t immediately eat the marshmallow did better in the future), sugar “recharges” your willpower, hearing words related to elderly stereotypes makes people walk more slowly, and plenty more.
unique link to this extract

Retiring Windows 10 and Microsoft’s move towards a surveillance state • Scott Larson

Scott Larson:

»

Whether you’re a business or a home user, I’m here to tell you that in many cases, Linux is a real alternative to Windows. So instead of pushing the goal post back from the brink of an Orwellian nightmare. I’m suggesting all of us consider switching Linux now.

Microsoft’s design of Windows 11 is a concern because:

1: Computer manufacturers, due to pressure from Microsoft, are designing new computers with artificial limitations like TPM and Secure Boot. These unnecessary add-ins push consumers to unnecessary hardware upgrades
2: In the setup of newly purchased consumer-grade computers, there is obfuscation in the installation language. Many of the default choices are aimed at confusing customers into selecting options that share data with vendors
2a: The process of setting up OneDrive to act as a backup of data. Without consent, the setup of this configuration moves all customers’ data to the cloud service, re-points all the user folders to a cloud-specific OneDrive folder that’s very difficult to revert
2b: The process of selecting a browser is obfuscated by Microsoft’s Edge Browser setup
3: The AI tool Co-pilot is installed and enabled without consent. Removal is difficult or nonexistent
4: The history tracking tool “Recall” that is due to be released, sometime in the future, saves snapshots of your user experience into Microsoft’s OneDrive cloud. It looks great on paper, but in reality, this feature, along with others, will be used to move forward a surveillance state
5: Windows 11 prevents the complete uninstall of many of its built-in features. They can be removed from one user account, but they can be reinstalled during an update, or if you upgrade your computer, without your consent
6: Microsoft Edge is forced on users as a replacement by obfuscating choice in various ways.

«

I remember people complaining about TPM at least ten years ago, so no novelty there. All the other stuff is, well, people need to consider what they’re sharing, and Chrome remains the most popular browser, which suggests people can figure this stuff out. Larson meanwhile is recommending that customers for whom he builds new computers should get Linux installed. Everything old is new again.
unique link to this extract

I made a binary search tool for videos to embarrass British Transport Police into doing their job • Tony Onodi

Tony Onodi:

»

Last week London Centric reported that British Transport Police (BTP) would no longer review CCTV footage in bike theft cases if the footage was longer than two hours

»

If it is available we will review around two hours of CCTV footage to try to identify the incident, but it is not proportionate to review longer periods as it keeps officers from being available to respond to emergencies, visibly patrolling railway stations and trains, investigating crimes with identified lines of enquiry or which cause the most harm to victims – such as violent or sexual offences.

«

London Centric, quoting a Cambridge mathematics professor, then go on to suggest that the police could use binary search to cut the time taken to search any plausible amount of footage down to something very trivial.

…I think part of the problem with selling the police on binary search is that it’s called “binary search”, which sounds very technical, and it’s being flogged to them by computer scientists and mathematics professors. I think a more palatable name for binary search would be something like “scrubbing through a video using common sense” and a better messenger would be just some guy. Luckily I’m not a computer scientist or a mathematics professor, I am just some guy. So I used AI (Claude Code if you must know) to build a tool that makes binary search even more foolproof than it already is. To make a point.

The way it works is a user uploads a video that they suspect contains footage of a theft, and gets shown a frame from the middle of the video. They then click either the “Item still there” or “Item stolen” button to narrow down which half of the video the theft happens in. Then they’re shown a frame from the middle of that half and the process repeats until the user has narrowed the range down to whatever granularity they want, at which point they can play the video and watch the theft take place.

…By my count, it takes about 10 seconds to narrow 64 minutes of footage down to a 30 second window, and—because of the way binary search scales (O(log n) for the nerds)—searching an eight-hour video would take more like 20 seconds than 80.

«

The police’s problem is that they’re used to watching CCTV to notice *everything* about an incident, not just “oh look they took the bike there”. There’s a demo video: impressive.
unique link to this extract

Apple said to cut iPhone Air production amid underwhelming sales • MacRumors

Hartley Charlton:

»

Apple plans to cut production of the iPhone Air amid underwhelming sales performance, Japan’s Mizuho Securities believes (via The Elec).

The Japanese investment banking and securities firm claims that the iPhone 17 Pro and iPhone 17 Pro Max are seeing higher sales than their predecessors during the same period last year, while the standard iPhone 17 is a major success, performing significantly better than the iPhone 16.

The iPhone Air is apparently the outlier; Apple plans to reduce production by one million units this year. Meanwhile, Apple plans to increase production of all other models by two million units. The overall production forecast of the iPhone 17 series this year has also been increased from 88 million units to 94 million units for the start of 2026.

A separate report earlier today claimed that Samsung has canceled plans to release a successor to its own iPhone Air rival, the Galaxy S25 Edge, due to low sales. Nevertheless, the iPhone Air reportedly sold out within hours in China, despite lower than expected sales in western countries last month.

«

It wouldn’t be surprising if the Air isn’t setting the world on fire; people see the “Pro” name and think it’s for them, because nobody thinks of themselves as “amateur”. The Air is for people who are making a particular choice – rather like those who bought the original Macbook Air. It took a few generations for that to become the low-end Mac for everyone.
unique link to this extract

Liquid Glass is cracked, and usability suffers in iOS 26 • Nielsen/Norman Group (NN/G)

Raluca Budiu:

»

iOS 26 brings Liquid Glass controls laid over noisy backgrounds, jittery animated buttons, shrunken and crowded tab bars, collapsing navigation, and ubiquitous search bars. On top of that, it breaks long‑established iOS conventions, getting closer to Android design.

Overall, Apple is prioritizing spectacle over usability, lending credibility to the theory that Liquid Glass is an attempt to distract customers from iOS 26’s lack of long-promised AI features.

The interface is restless, needy, less predictable, less legible, and constantly pulling focus rather than supporting seamless access to content. Instead of smoothing the path for everyday tasks, iOS 26 makes users relearn basics while enduring a constant parade of visual stunts.

Apple may call it Liquid Glass. To many users, it feels more like a fogged‑up window: pretty from a distance, but frustrating when you try to see beyond it.

«

This is the summary of a much longer piece, but it’s why I’m personally not updating my iPad or iPhone to it. Even iOS 7 didn’t have simple legibility problems; this does. Other critiques: What happened to Apple’s legendary attention to detail; Apple’s Liquid Glass design prioritises content over tools.

How long for this to get sorted out – is six months about right, or too optimistic?
unique link to this extract

Shakespeare family home damaged by reversing driver • Sky News

»

A 17th-century building that housed William Shakespeare’s family has been left damaged after a driver reversed into it.

A picture of the Grade I listed building following the crash showed significant damage to the outside walls, with the timber overhang appearing to be supported by scaffolding.

“Yesterday morning, a vehicle was accidentally driven into Halls Croft, located in the Old Town area of Stratford-upon-Avon,” a spokesperson for the Shakespeare Birthplace Trust told Sky News.

“First and foremost, we are relieved to confirm that no one was injured in the incident and the building has been made secure to prevent any further damage,” they said.

“This is a stark reminder of how fragile our heritage is,” the Trust added in a post on X.

«

To quote Philip Purser-Hallard on Bluesky: “But soft! What light through yonder window.. BRAAAKES!!”

unique link to this extract

Are hard drives getting better? Let’s revisit the Bathtub Curve • Backblaze

Drive Stats Team:

»

If you’ve hung around Backblaze for a while (and especially if you’re a Drive Stats fan), you may have heard us talking about the bathtub curve. In Drive Failure Over Time: The Bathtub Curve Is Leaking, we challenged one of reliability engineering’s oldest ideas—the notion that drive failures trace a predictable U-shaped curve over time. 

But, the data didn’t agree. Our fleet showed dips, spikes, and plateaus that refused to behave. Now, after 13 years of continuous data, the picture is clearer—and stranger. 

The bathtub curve isn’t just leaking, and the shape of reliability might look more like an ankle-high wall at the entrance to a walk-in shower. The neat story of early failures, calm middle age, and gentle decline no longer fits the world our drives inhabit. Drives are getting better—or, more precisely, the Drive Stats dataset says that our drives are performing better in data center environments. 

So, let’s talk about what our current “bathtub curve” looks like, and how it compares to earlier generations of the analysis. 

«

Backblaze is the “we back it all up for you” company and you probably don’t need to read this in detail; just scroll past the graphs. It seems drives are getting a lot more reliable: failure rates are down by an order of magnitude.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

Plans to introduce a carbon tax on polluting ships have been thrown into disarray by opposition from the US and Russia. CC-licensed photo by Michael Elleray on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 10 links for you. Stately. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

A vast divergence is opening up in America’s industries of the future • The Washington Post

Aaron Gregg and Federica Cocco:

»

A gulf is opening up in the heart of American business as two industries championed as central to the country’s future — manufacturing and artificial intelligence — appear to be heading in different directions.

Both AI and manufacturing have been in the spotlight in Washington through successive administrations. President Donald Trump this year said he would do “whatever it takes to lead the world in artificial intelligence,” while he has championed stemming a decades-long slide in American manufacturing as a top goal.

But while AI is flourishing this year, manufacturing is entering an ever deeper slump. “You have the software and services world accelerating, and becoming almost a monomania for the culture, at the same time that manufacturing remains flat or worse,” said Mark Muro, a senior fellow at the Brookings Institution. “The AI boom is kind of papering over some other parts of the economy that aren’t going well.”

The Trump administration has embraced using a broad array of tariffs to protect US manufacturers from foreign competition, marking the latest White House-led push after the Biden administration spent tens of billions of dollars boosting US-made semiconductors and other projects. But so far, the sector is down 38,000 jobs since the start of the year, according to the Bureau of Labor Statistics.

…Some experts worry that the AI industry will employ too few workers once the hype dies down, because the data centers that power AI require relatively few workers to operate. There are also fears that a bursting AI bubble could hit an already-fragile economy.

“One interesting implication of all this will be that this AI [spending] boom is likely to create less jobs, particularly blue-collar ones, than previous waves of infrastructure buildouts,” said Stephanie Aliaga, global market strategist at JPMorgan.

American manufacturers employed some 19.5 million workers at the industry’s 1979 peak. That number has since shrunk to fewer than 13 million, including losing an estimated 78,000 more positions in the one-year period ending in August.

«

unique link to this extract

Feb 2025: UK Government a long way from achieving its vision of containing antimicrobial resistance • National Audit Office

From February this year:

»

Antimicrobial resistance (AMR) is a serious threat globally and to the UK and, if not addressed, the consequences for health, life expectancy, the functioning of the NHS and the wider economy will be huge.

But a new report from the National Audit Office, Investigation into how government is addressing antimicrobial resistance, finds that the government’s response to the issue over the last five years, the AMR National Action Plan 2019-2024, has made limited progress.

Antimicrobials are therapeutic substances, such as antibiotics, designed to treat infections and prevent their spread. Their use, overuse and misuse mean public health could be compromised in future as more pathogens (the organisms which cause disease) evolve and develop resistance to antibiotics and other antimicrobials. Already, AMR contributes to an estimated 35,200 deaths annually in the UK.

Government is taking the problem seriously, with AMR identified as one of 26 chronic national risks. But, despite a 20-year vision, a cross-government approach and some innovative solutions, including subscription arrangements for procuring antibiotics, the report finds that the UK remains a long way from the vision and objectives the government expressed in 2019: a lower burden of infection; the optimal use of antimicrobials; and new treatments so that everyday illnesses can continue to be cured.

Of five domestic targets set in 2019, only one – reducing the use of antibiotics in food producing animals – was met. Drug-resistant infections in humans have increased by 13% since 2018, despite a target to reduce them by 10%3.

«

Following up from yesterday’s post about the increase in drug-resistant infections, Overspill reader Matt L writes:

»

Governments know the problem with developing new AMR drugs. It’s not initial research, but the costs of bringing them to market. Drug discovery can cost tens of millions, and Wellcome, Gates and others have been funding that for a while. But bringing a drug to market costs 100s of millions, and as doctors deliberately under prescribe AMR drugs (so bacteria don’t build resistance to them) the market isn’t big enough to justify the investment.

«

Got to love the irony: we’re careful not to overuse these drugs or they become useless; but that means there isn’t a big enough market for these essential drugs. Love ya, capitalism. (Thanks Matt L.)
unique link to this extract

Inside the web infrastructure revolt over Google’s AI Overviews • Ars Technica

Samuel Axon:

»

Cloudflare, a web infrastructure company, has updated millions of websites’ robots.txt files in an effort to force Google to change how it crawls them to fuel its AI products and initiatives.

We spoke with Cloudflare CEO Matthew Prince about what exactly is going on here, why it matters, and what the web might soon look like. But to get into that, we need to cover a little background first.

The new change, which Cloudflare calls its Content Signals Policy, happened after publishers and other companies that depend on web traffic have cried foul over Google’s AI Overviews and similar AI answer engines, saying they are sharply cutting those companies’ path to revenue because they don’t send traffic back to the source of the information.

There have been lawsuits, efforts to kick-start new marketplaces to ensure compensation, and more—but few companies have the kind of leverage Cloudflare does. Its products and services back something close to 20% of the web, and thus a significant slice of the websites that show up on search results pages or that fuel large language models.

“Almost every reasonable AI company that’s out there is saying, listen, if it’s a fair playing field, then we’re happy to pay for content,” Prince said. “The problem is that all of them are terrified of Google because if Google gets content for free but they all have to pay for it, they are always going to be at an inherent disadvantage.”

This is happening because Google is using its dominant position in search to ensure that web publishers allow their content to be used in ways that they might not otherwise want it to.

Since 2023, Google has offered a way for website administrators to opt their content out of use for training Google’s large language models, such as Gemini.

However, allowing pages to be indexed by Google’s search crawlers and shown in results requires accepting that they’ll also be used to generate AI Overviews at the top of results pages through a process called retrieval-augmented generation (RAG).

«

unique link to this extract

US, Saudi-led alliance plunges green shipping deal into doubt • Climate Change News

Joe Lo:

»

The US, Saudi Arabia, Russia and their allies have spearheaded a push to alter the approval process for a hard-fought green shipping deal, which experts say could jeopardise the landmark pact at the International Maritime Organisation (IMO) talks in London this week.

If approved, the procedural changes would make it harder for the IMO’s Net-Zero Framework (NZF) to come into effect, as it would require support from countries representing half of the world’s shipping fleet.

After years of discussions, governments provisionally agreed the NZF in April, in which they pledged to penalise polluting ships and use the money to fund the transition to cleaner fuel. The policy is the world’s first global emissions pricing on any sector. At talks in London this week, countries are meeting to discuss how to carry the NZF forward.

The US and its allies want to shift away from a system of tacit approval where, after the NZF is approved at the IMO talks, its rules automatically come into force unless a certain number of countries object. They prefer explicit approval instead, meaning it would not come into force unless enough governments – representing a certain percentage of the world’s shipping fleet – actively indicate support for it.

Emma Fenton, senior director of climate diplomacy at nonprofit policy group Opportunity Green, told Climate Home News that the US’s proposed change “risks undermining the NZF’s ambitions, delays the maritime transition and does not meet the scale or the pace of action that the climate crisis demands”.

Bryan Comer, maritime director at the International Council on Clean Transportation called it “an unnecessary procedural roadblock”.

«

Small countries with big shipping registries would be hardest hit: Bahamas and Liberia are squealing.
unique link to this extract

AI-generated news sites spout viral slop from forgotten URLs • Nieman Journalism Lab

Ben Paviour:

»

Last year, Icelandic teacher María Hjálmtýsdóttir wrote a column for The Guardian on the country’s experiment with a 36-hour workweek. The piece offered rich personal anecdotes that only a local could provide. Readers learned, for instance, that Hjálmtýsdóttir’s husband is using some of his newfound free time to chat with his fellow hobbyist pigeon keepers.

In the months since her Guardian piece came out, Hjálmtýsdóttir’s essay has been stripped of its color, repackaged, and republished at least a dozen times by “news outlets” that almost nobody has ever heard of.

“Iceland switched to a 4-day workweek — Gen Z was right all along,” stated a July 3 headline on Dixie Sun News, on a URL that once hosted a college newspaper. “Iceland embraced the 4-day workweek in 2019: 6 years later, Gen Z’s vision has been realized,” stated the Carroll County Observer, a former Maryland news site turned clickbait slop shop. “Iceland embraced a 4-day workweek in 2019 – Now, nearly six years on, all Gen Z forecasts have materialized,” read the headline on WECB.fm, a site falsely claiming to represent an Emerson College radio station.

These sites appear to be part of a new wave of AI-generated content farms that swoop in to seize dormant domains. Some of the AI news sites led previous lives unrelated to news, like Boston Organics, the website of a former produce delivery service that now covers everything from octopuses in British waters (“England is facing an unprecedented invasion, the problem is, it’s octopuses, and they’re devouring everything in their path”) to how long chili stays good in the fridge. In other cases, AI news articles are buried out of view of the homepage. Users who visit Paris2018.com — a site created for that year’s Gay Games — see no indication that it contains a plethora of AI-generated articles.

«

The latter is a classic spam tactic – find an unpatched vulnerability in an old WordPress blog and stuff it with spam, or in this case, slop. It’s all so sadly predictable. (Thanks Gregory B for the link.)
unique link to this extract

Nation-state hackers deliver malware from “bulletproof” blockchains • Ars Technica

Dan Goodin:

»

Hacking groups—at least one of which works on behalf of the North Korean government—have found a new and inexpensive way to distribute malware from “bulletproof” hosts: stashing them on public cryptocurrency blockchains.

In a post on Thursday, members of the Google Threat Intelligence Group said the technique provides the hackers with their own “bulletproof” host, a term that describes cloud platforms that are largely immune from takedowns by law enforcement and pressure from security researchers. More traditionally, these hosts are located in countries without treaties agreeing to enforce criminal laws from the US and other nations. These services often charge hefty sums and cater to criminals spreading malware or peddling child sexual abuse material and wares sold in crime-based flea markets.

Since February, Google researchers have observed two groups turning to a newer technique to infect targets with credential stealers and other forms of malware. The method, known as EtherHiding, embeds the malware in smart contracts, which are essentially apps that reside on blockchains for Ethereum and other cryptocurrencies. Two or more parties then enter into an agreement spelled out in the contract. When certain conditions are met, the apps enforce the contract terms in a way that, at least theoretically, is immutable and independent of any central authority.

“In essence, EtherHiding represents a shift toward next-generation bulletproof hosting, where the inherent features of blockchain technology are repurposed for malicious ends,” Google researchers Blas Kojusner, Robert Wallace, and Joseph Dobson wrote. “This technique underscores the continuous evolution of cyber threats as attackers adapt and leverage new technologies to their advantage.”

«

The blockchain? You mean that fabulous invention that cannot be rolled back, which is one of its great features? Oh dear.
unique link to this extract

Barrister found to have used AI to prepare for hearing after citing “fictitious” cases • The Guardian

Jamie Grierson:

»

An immigration barrister was found by a judge to be using AI to do his work for a tribunal hearing after citing cases that were “entirely fictitious” or “wholly irrelevant”.

Chowdhury Rahman was discovered using ChatGPT-like software to prepare his legal research, a tribunal heard. Rahman was found not only to have used AI to prepare his work, but “failed thereafter to undertake any proper checks on the accuracy”.

The upper tribunal judge Mark Blundell said Rahman had even tried to hide the fact he had used AI and “wasted” the tribunal’s time. Blundell said he was considering reporting Rahman to the Bar Standards Board. The Guardian has contacted Rahman’s firm for comment.

The matter came to light in the case of two Honduran sisters who claimed asylum on the basis that they were being targeted by a criminal gang in their home country. Rahman represented the sisters, aged 29 and 35. The case escalated to the upper tribunal.

Blundell rejected Rahman’s arguments, adding that “nothing said by Mr Rahman orally or in writing establishes an error of law on the part of the judge and the appeal must be dismissed”. Then, in a rare ruling, Blundell went on to say in a postscript that there were “significant problems” within the grounds of appeal put before him.

He said that 12 authorities were cited in the paperwork by Rahman, but when he came to read the grounds, he noticed that “some of those authorities did not exist and that others did not support the propositions of law for which they were cited in the grounds”.

«

I think Rahman has now been reported to the Bar Standards Board, going by this judgment made in August – and this seems to be the second time he’s been referred to it. Not a good look for a barrister. How long will it take for the message to filter through that lawyers can’t rely on chatbots to write their arguments?
unique link to this extract

ICE, Secret Service and Navy all had access to Flock’s nationwide network of cameras • 404 Media

Joseph Cox:

»

A division of ICE, the Secret Service, and the Navy’s criminal investigation division all had access to Flock’s nationwide network of tens of thousands of AI-enabled cameras that constantly track the movements of vehicles, and by extension people, according to a letter sent by Senator Ron Wyden and shared with 404 Media.

Homeland Security Investigations (HSI), the section of ICE that had access and which has reassigned more than ten thousand employees to work on the agency’s mass deportation campaign, performed nearly two hundred searches in the system, the letter says.

In the letter Senator Wyden says he believes Flock is uninterested in fixing the room for abuse baked into its platform, and says local officials can best protect their constituents from such abuses by removing the cameras entirely.

The letter shows that many more federal agencies had access to the network than previously known. We previously found, following local media reports, that Customs and Border Protection (CBP) had access to 80,000 cameras around the country. It is now clear that Flock’s work with federal agencies, which the company described as a pilot, was much larger in scope.

«

That phrase about “sleepwalking into the surveillance state” doesn’t sound so trite now, does it.
unique link to this extract

Even top generals are looking to AI chatbots for answers • Business Insider

Kelsey Banker and Chris Panella:

»

It’s not just the civilian corporate executives and white-collar workers who are leaning into the generative AI boom at work. Military leaders are diving in too.

The top US Army commander in South Korea shared that he is experimenting with generative AI chatbots to sharpen his decision-making, not in the field, but in command and daily work.

He said “Chat and I” have become “really close lately.”

“I’m asking to build, trying to build models to help all of us,” said Maj. Gen. William ‘Hank’ Taylor, commanding general of the 8th Army, told reporters during a media roundtable at the annual Association of the United States Army conference in Washington, DC, on Monday.

Taylor said he’s using the tech to explore how he makes military and personal decisions that affect not just him but the thousands of soldiers he oversees. While the tech is useful, though he acknowledged that keeping up with the pace of such rapidly developing technology is an enduring challenge.

“As a commander, I want to make better decisions,” the general shared. “I want to make sure that I make decisions at the right time to give me the advantage.”

Commanders like Taylor are focused on fast decision-making and how AI could provide an advantage because of a thought process popular with military leaders known as the “OODA Loop.” The theory, developed by US fighter pilots during the Korean War, posits that troops who can move decisively before the enemy does — and observe, orient, decide, and act— often have the advantage on the battlefield.

«

I think the word is “concerning”. (Thanks Gregory B for the link.)
unique link to this extract

MacBook Pro with OLED touch screen launching as soon as 2026 • MacRumors

Juli Clover:

»

Apple is working on a new version of the MacBook Pro with an OLED display, a hole punch camera, and touch screen functionality, Bloomberg’s Mark Gurman said today. The updated MacBook Pro is set to launch sometime between late 2026 and early 2027.

Apple hasn’t redesigned the MacBook Pro since the launch of the M1 Pro and M1 Max machines in 2021, but that is set to change with the launch of the OLED models. Along with new display capabilities, there will be a hole-punch camera and no notch, plus a thinner and lighter design. Apple is also adding a reinforced hinge and an updated screen design that will ensure the display does not move when it is touched.

While Apple plans to add a touch screen, the MacBook will continue to have a trackpad and a keyboard, with touch gestures augmenting existing functionality.

«

I struggle a bit to believe that after years of insisting that a touchscreen makes no sense for a Mac, Apple would produce a touchscreen Mac. Though completely denying the utility of a product and then reversing course isn’t unknown from Apple, it hasn’t really done it since Steve Jobs was in charge. Though Liquid Glass is surely an interface designed for touch first.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

The end of an era as TiVo announces that it is throwing in the towel after 25 years. CC-licensed photo by on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 10 links for you. Programmatic. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

The $15bn warning shot • Rob’s Notes

Rob Leathern:

»

On Tuesday news broke from the US DOJ about the largest forfeiture in its history: a staggering $15bn in Bitcoin from a large scam operation where “pig-butchering”, initiated on social media and messaging platforms, plays a major role.

Back in early 2024 I posted about a study by UT’s Professor John Griffin estimating that pig-butchering scams had stolen over $75bn worldwide between 2020 and 2024, significantly more than previously thought. We need to anticipate the evolution of these scams and build proactive counter-measures, including inside of encrypted environments.

…Today’s AI tools can supercharge scams. First, personalization at zero marginal cost. Fraud relies on persuasion; persuasion feeds on context. Large language models can synthesize biographies from stray data points and generate messages in the target’s idiom, salted with plausible detail and error‑free grammar. “Spearphishing” once meant labor-intensive research on a few valuable marks. With AI, every mark can be a high‑value mark.

Second, synthetic presence. Voice‑cloning apps now need mere seconds of audio; deepfake video grows more lifelike by the month. Imagine the same phone‑farms scaled to millions of high‑fidelity avatars that can hop between messaging, video calls and customer‑support lines. The romance scam becomes the video‑date; the “broker” becomes a moving face in a crisp suit; the “bank official” calling about a flagged bank wire sounds exactly like your branch manager. We are already seeing deepfakes moving from major celebrities to local doctors and finance professionals.

Third, toolchain integration. The indictment’s playbook – bulk account creation or acquisition, scripted chats, multilingual segmentation, payment orchestration – is already modular. AI can plug into each module with better scripts, smarter target selection, automated rebuttals to doubts, even real‑time generation of doctored “portfolio” screenshots. The result is not merely more spam; it is a service platform for crime, with measurable conversion funnels and A/B‑tested lies.

The implications are stark. Tech firms must treat abuse as a core product risk, not a reputation clean‑up. That means throttling mass account creation and finding new challenge-response checks that are hard for farms but light for humans.

«

The whole thing is worrying about how AI is sure to be weaponised to make these scams even more effective. Troubling times.
unique link to this extract

EU gets what it asked for: no charger in the MacBook Pro box • Apple Insider

William Gallagher:

»

Don’t blame Apple, this time. If you’re in the European Union or the UK, your new M5 14-inch MacBook Pro or iPad Pro may cost you $70 (£59) extra because Apple isn’t allowed to bundle a charger.

It’s chiefly because of the European Union’s law that Apple was forced to move from Lightning to USB-C charging on the iPhone. But those laws covered more than just smartphones, and the EU has also been pressing for companies to stop bundling chargers with their devices.

The idea is that so many consumers already have chargers that bundling new ones creates unacceptable volumes of e-waste. The EU enacted a law covering this back in 2003, but its latest amendment requires its 27 member states to have implemented it by October 9, 2025. [Note: the UK left the EU in early 2020 and is not bound by subsequent amendments. – Overspill Ed]

It’s a different matter for the US and most of the world, including Brazil, which has previously fined Apple for not including chargers. For all of those countries, buyers of the new M5 14-inch MacBook Pro are offered a 70W power adaptor in the price. Alternatively, they can choose to pay $20 more and instead get a 96W charger.

In European Union territories and the UK, there is no such option at all. Curiously, these places do get a “What’s in the Box” section in the Store, while US buyers do not.

Should UK buyers want a charger, they have to pay £59 for the 70W version. The 96W model costs £79. Across the EU, the price in Euros is equivalent to $75 or $98.

For the new M5 iPad Pro, a 20W charger is included in the US and most territories. But in the UK and EU, they must buy a separate 30W charger for $50.

«

Just as annoying as the charger pricing is folks writing stories where they don’t just quote the local prices, which I had to find out and substitute (but couldn’t be bothered to find the euro prices). Though as Mark Gurman points out, Apple could have offered the charger for free (either as an add-on or write-in), but chose to charge for it. Though people then point out that Apple has reduced the price of the base model computer.
unique link to this extract

How I almost got hacked by a “job interview” • David Dodda

David Dodda:

»

Last week, I got a LinkedIn message from Mykola Yanchii. Chief Blockchain Officer at Symfa. Real company. Real LinkedIn profile. 1,000+ connections. The works.

The message was smooth. Professional. “We’re developing BestCity, a platform aimed at transforming real estate workflows. Part-time roles available. Flexible structure.”

I’ve been freelancing for 8 years. Built web applications, worked on various projects, done my share of code reviews. I’m usually paranoid about security – or so I thought.

This looked legit. So I said yes to the call. Before our meeting, Mykola sent me a “test project” – standard practice for tech interviews. A React/Node codebase to evaluate my skills. 30-minute test. Simple enough.

The Bitbucket repo looked professional. Clean README. Proper documentation. Even had that corporate stock photo of a woman with a tablet standing in front of a house. You know the one.

Here’s where I almost screwed up: I was running late for our call. Had about 30 minutes to review the code. So I did what lazy developers do – I started poking around the codebase without running it first.

Usually, I sandbox everything. Docker containers. Isolated environments. But I was in a rush. I spent 30 minutes fixing obvious bugs, adding a docker-compose file, cleaning up the code. Standard stuff. Ready to run it and show my work.

Then I had one of those paranoid developer moments. Before hitting npm start, I threw this prompt at my Cursor AI agent:

“Before I run this application, can you see if there are any suspicious code in this codebase? Like reading files it shouldn’t be reading, accessing crypto wallets etc.”

«

Turned out it had an obfuscated call to a site which would have downloaded and run malware that looks like it would have emptied his crypto wallet. The company looked legit; was fake. The URL disappeared 24 hours later.
unique link to this extract

WHO warns of sharp increase in drug-resistant infections • NY Times via The Seattle Times

Andrew Jacobs:

»

Around the world, the spread of dangerous infections that do not respond to antibiotics has been increasing by as much as 15% a year, affecting treatment for urinary tract infections, gonorrhea, E. coli and other pathogens that kill millions of people annually, according to a report released Monday by the World Health Organization.

The report documents how countries are grappling with the challenge of so-called antimicrobial resistance. It found that 1 in 6 infections in 2023 was resistant to the current roster of antibiotic drugs. The drug resistance involves 40% of the most common antibiotics used against these infections.

Southeast Asia and the eastern Mediterranean had the highest rates of resistance, with 1 in 3 infections resistant to antibiotics. That is roughly double the worldwide average and more three times the rates in Europe and the Western Pacific.

Overall, antimicrobial resistance was more prevalent in low- and middle-income countries, especially those with weak health care systems.

…At the same time, the pipeline for new drugs has largely dried up, the result of a broken marketplace for antimicrobials that has driven the world’s biggest pharmaceutical companies from the field. Companies that have tried to make new antibiotics have been unable to make money from them.

“For many of these threats, the consequences are real — harder-to-treat infection, rising costs and lives lost,” Hutin said.

The report sounded the alarm on so-called gram-negative bacteria, which pose additional challenges because of a protective outer membrane that can be tough for antibiotics to penetrate. Gram-negative bacteria include Escherichia coli and Klebsiella pneumoniae, which are often implicated in severe infections that lead to sepsis and death. In Africa, resistance to cephalosporins, a class of antibiotics and the first choice treatment for such infections, can exceed 70%.

«

This has been a known problem for around 30 years. You’d think governments might want to fund research to solve the problem of the market failure. But apparently not.
unique link to this extract

Dutch seizure of chipmaker followed US ultimatum over Chinese chief • Financial Times

Andy Boundds, Ben Hall and Ryan McMorrow:

»

The Dutch government seized control of chipmaker Nexperia after Washington warned that the company would not be removed from its export control list if its Chinese chief executive remained in charge, according to court filings.

The economy ministry this month removed the chief executive, Zhang Xuezheng — who was also the controlling shareholder of the chipmaker — in a rare move that brought the Netherlands into the escalating fight for technological dominance between Washington and Beijing.

Nexperia makes basic low-margin chips in vast quantities for consumer electronics and a broad range of industrial uses, but it is also an important supplier for Europe’s auto industry. It was sold to a Chinese consortium in 2017 before being bought by Chinese group Wingtech.

The Amsterdam court of appeal published the proceedings between the Dutch economy ministry and Wingtech on Tuesday. It revealed that US officials told the Dutch in June that a plan to ringfence its European operations from Chinese ones was moving too slowly.

«

Ah, so not the Dutch having a bright idea on their own, but rather being impelled towards it.
unique link to this extract

4chan fined $26k for refusing to assess risks under UK Online Safety Act • Ars Technica

Ashley Belanger:

»

A battle over the United Kingdom’s Online Safety Act (OSA) heated up Monday as UK regulator Ofcom fined the notorious image-hosting board 4chan about $26,000 for failing to provide a risk assessment detailing the potential harms of illegal content hosted on its forum.

In a press release provided to Ars, Ofcom said 4chan refused to respond to two requests for information that the regulator considered “routine.” The first asked for the risk assessment and the second for 4chan’s “qualifying worldwide revenue.”

4chan was anticipating the Monday fine, noting in a lawsuit—which was jointly filed with the online trolling forum Kiwi Farms in August and seeks to permanently enjoin Ofcom from enforcing OSA—that Ofcom had made it clear that because 4chan ignored Ofcom’s emails, the fine was coming.

Now, 4chan has 60 days to hand over the information Ofcom requested while risking incurring about $130 in additional daily penalties.

If 4chan continues to ignore Ofcom, the forum could be blocked in the UK. And 4chan could face even bigger fines totaling about $23m or 10% of 4chan’s worldwide turnover, whichever is higher. 4chan also faces potential arrest and/or “imprisonment for a term of up to two years,” the lawsuit said.

«

Does 4chan have any sources of revenue? Except it has managed to find a lawyer who is seeking to get a US court to rule that the Online Safety Act doesn’t apply to “speech and content published and distributed in the US”.
unique link to this extract

TiVo stops selling DVRs, marking the end of an era • Cord Cutters News

Luke Bouma:

»

In a seismic shift for the television industry, TiVo Corporation has quietly pulled the plug on its storied digital video recorder line, effectively ending an era that redefined how consumers interacted with broadcast content. As of early October 2025, the company’s official website has scrubbed all references to its hardware DVR products, including the once-revered TiVo Edge models designed for cable subscribers and over-the-air antenna users. Visitors searching for these devices now encounter a streamlined catalog that omits any mention of physical recording hardware, signaling a complete withdrawal from the retail DVR market.

This move culminates decades of gradual decline for TiVo’s hardware ambitions, which peaked in the early 2000s when the brand became synonymous with effortless time-shifting of television programming. Launched in 1999, TiVo’s DVRs introduced features like one-touch recording, commercial skipping, and intuitive search capabilities that made traditional TV schedules feel obsolete. At its zenith, the company boasted millions of subscribers, forcing cable providers and networks to adapt to empowered viewers who could pause live broadcasts or binge-watch at will.

«

The very first time I saw a TiVo demonstrated – in September 2000 – I could see it was absolutely the future. I then tried it out and became even more convinced. And wrote about how it came to be, from an era when building hardware seemed to be a matter of finding a use and exploiting it.
unique link to this extract

The AI water issue is fake • The Weird Turn Pro

Andy Masley:

»

AI data centres use water. Like any other industry that uses water, they require careful planning. If an electric car factory opens near you, that factory may use just as much water as a data centre. The factory also requires careful planning. But the idea that either the factory or AI is using an inordinate amount of water that merits any kind of boycott or national attention as a unique serious environmental issue is innumerate. On the national, local, and personal level, AI is barely using any water, and unless it grows 50 times faster than forecasts predict, this won’t change. I’m writing from an American context and don’t know as much about other countries. But at least in America, the numbers are clear and decisive.

The idea that AI’s water usage is a serious national emergency caught on for three reasons:

• People get upset at the idea of a physical resource like water being spent on a digital product, especially one they don’t see value in, and don’t factor in just how often this happens everywhere
• People haven’t internalized how many other people are using AI. AI’s water use looks ridiculous if you think of it as a small marginal new thing. It looks tiny when you divide it by the hundreds of millions of people using AI every day
• People are easily alarmed by contextless large numbers, like the number of gallons of water a data centre is using. They compare these large numbers to other regular things they do, not to other normal industries and processes in society.

Together, these create the impression that AI water use is a problem. It is not. Regardless of whether you love or hate AI, it is not possible to actually look at the numbers involved without coming to the conclusion that this is a fake problem.

«

There follow a lot of numbers which will inform you that it’s honestly not a problem. Happily there are graphs to help if the words become overwhelming.
unique link to this extract

Nvidia sells tiny new computer that puts big AI on your desktop • Ars Technica

Benj Edwards:

»

On Tuesday, Nvidia announced it would begin taking orders for the DGX Spark, a $4,000 desktop AI computer that wraps one petaflop of computing performance and 128GB of unified memory into a form factor small enough to sit on a desk. Its biggest selling point is likely its large integrated memory that can run larger AI models than consumer GPUs.

Nvidia began taking orders for the DGX Spark on Wednesday, October 15, through its website, with systems also available from manufacturing partners and select US retail stores.

The DGX Spark, which Nvidia previewed as “Project DIGITS” in January and formally named in May, represents Nvidia’s attempt to create a new category of desktop computer workstation specifically for AI development.

With the Spark, Nvidia seeks to address a problem facing some AI developers: Many AI tasks exceed the memory and software capabilities of standard PCs and workstations (more on that below), forcing them to shift their work to cloud services or data centers. However, the actual market for a desktop AI workstation remains uncertain, particularly given the upfront cost versus cloud alternatives, which allow developers to pay as they go.

Nvidia’s Spark reportedly includes enough memory to run larger-than-typical AI models for local tasks, with up to 200 billion parameters and fine-tune models containing up to 70 billion parameters without requiring remote infrastructure. Potential uses include running larger open-weights language models and media synthesis models such as AI image generators.

According to Nvidia, users can customize Black Forest Labs’ Flux.1 models for image generation, build vision search and summarization agents using Nvidia’s Cosmos Reason vision language model, or create chatbots using the Qwen3 model optimized for the DGX Spark platform.

«

Power requirement: 240W. The upfront cost is only the beginning – watch your electricity bill rocket too. (Thanks Gregory B for the link.)
unique link to this extract

Meet the AI chatbots replacing India’s call-center workers • Reuters

Munsif Vengattil and Aditya Kalra:

»

At a startup office in Bengaluru, India, developers are fine-tuning artificial-intelligence chatbots that talk and message like humans.

The company, LimeChat, has an audacious goal: to make customer-service jobs almost obsolete. It says its generative AI agents enable clients to slash by 80% the number of workers needed to handle 10,000 monthly queries. “Once you hire a LimeChat agent, you never have to hire again,” Nikhil Gupta, its 28-year-old co-founder, told Reuters.

Cheap labor and English proficiency helped make India the world’s back office — sometimes at the expense of workers elsewhere. Now, AI-powered systems are subsuming jobs done by headset-wearing graduates in technical support, customer care and data management, sparking a scramble to adapt, a Reuters examination found.

That’s driving business for AI startups that help companies slash staffing costs and scale operations — even though many consumers still prefer to deal with a person.

…Rather than pump the brakes as the technology threatens jobs built on routine tasks, the country is accelerating, wagering that a let-it-rip approach will create enough new opportunities to absorb those displaced, Reuters found. The outcome of India’s gamble carries weight far beyond its borders — a test case for whether embracing AI-driven disruption can elevate a developing economy or render it a cautionary tale.
The global conversational AI market is growing 24% a year and should reach $41 billion by 2030, consultancy Grand View Research estimates.

India — which relies on IT for 7.5% of its GDP — is leaning in. In a February speech, Prime Minister Narendra Modi said “work does not disappear due to technology. Its nature changes and new types of jobs are created.”

Not everyone shares Modi’s confidence in India’s preparedness. Santosh Mehrotra, a former Indian official and visiting professor at the University of Bath’s Centre for Development Studies, criticized the government for a lack of urgency in assessing AI’s effects on India’s young workforce. “There’s no gameplan,” he said.

«

Of course the difference about the chatbots is that they will have an accent matching the country they’re meant to be dealing with (the opposite still irks some people) and they will be endlessly, exhaustingly patient.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

High pressure in the UK is going to cause problems with TV reception. If you watch via an aerial, that is. CC-licensed photo by hedera.baltica on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

There’s another post coming this week at the Social Warming Substack on Friday at 0845 UK time. Free signup.

A selection of 10 links for you. Well received. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

Google Pixel 10 Pro Fold explodes during JerryRigEverything’s durability test • Dexerto

Dylan Horetski:

»

In his October 14 video, tech YouTuber JerryRigEverything put the Pixel 10 Pro Fold through his standard series of stress tests, which include scratch, dust, and bend resistance trials. During the test, the foldable snapped along the same antenna line that caused breakages in previous Pixel Fold models, leading to a catastrophic failure.

Google’s foldable was also unable to withstand dust exposure, despite the company’s claim of an IP68 dust resistance rating. The hinge was reportedly filled with debris and made “crunching” noises, proving the internal mechanism wasn’t sealed against particles.

The break caused the device’s internal battery to short-circuit, resulting in the phone igniting and smoking mid-test — the first time JerryRigEverything says he’s ever had a smartphone explode in over a decade of testing.

The YouTuber criticized Google for leaving the Pixel Fold’s antenna lines in the same vulnerable location near the hinge for a third generation in a row, the same weak point that caused the Pixel Fold and Pixel 9 Pro Fold to snap in earlier tests.

“This is by far the weakest folding smartphone I’ve ever tested,” he said. “And it gets worse. While straightening it back out for round two, the battery decides it’s had enough. Surprisingly, in the decade that I’ve been durability testing phones, I have never had a smartphone explode before.”

«

The video is absolute phone torture – scratching, dust forced into the hinge, flames applied to the screen, bent backwards (with bare hands) to break it – which triggers the battery to say ENOUGH!
unique link to this extract

As Microsoft bids farewell to Windows 10, millions of users… won’t • The Verge

Tom Warren:

»

Windows 10 is so popular that Windows 11 only overtook it in terms of usage just a few months ago. That’s why I’m surprised that Microsoft is still, kind of, going ahead with its end of support cutoff on Tuesday.

At one point last year, I wasn’t sure if Microsoft was actually going to end support for Windows 10 on time. The software giant randomly reopened Windows 10 beta testing to add new features and improvements to a 10-year-old operating system, giving millions of users hope that the company would change its mind or at least lower the system requirements for Windows 11. Neither of those things is happening, though.

Microsoft is ending support for Windows 10 today, after originally releasing the OS on July 29th, 2015. The cutoff means Microsoft will no longer provide software updates from Windows Update, technical assistance, or security fixes for Windows 10. It’s a milestone moment for millions of users who can’t upgrade, businesses that don’t want to, and a company that’s increasingly looking at overhauling Windows with AI features.

I say Microsoft is kind of ending Windows 10 support because consumers will be able to enable extended security updates for free (with a catch for most) to get another year’s worth of security fixes. Only businesses have been able to do this in the past, and it’s a clear admission from Microsoft that Windows 10 is simply too popular among consumers to be left without security patches.

Around 40% of Windows users are running Windows 10 right now, according to StatCounter. While a large part of that 40% will be businesses that can pay for up to three years of extra support, Valve says around 30% of all PC gamers are also still using Windows 10.

«

This is always the problem – people stick with a version of Windows they’re familiar with until, in time, they’re forced to buy a PC that has the new version. The end of Windows 10 is expected to cause a bump of tens (perhaps hundreds) of millions of extra PCs being sold in the coming year.
unique link to this extract

Nearly 40% of kids under two years old interact with smartphones, say their parents • Sherwood News

Millie Giles:

»

As parents in 2025 know, they really do grow up so fast. First words today, first Google query tomorrow. Then, before you know it, they’re asking ChatGPT to read them a bedtime story…

Last week, Pew Research Center published a survey assessing how parents in the US with children under 12 manage their kids’ screen time, which revealed that 61% of respondents overall reported their child ever uses or interacts with smartphones — including 38% of those with children under 2 years old.

Much of this smartphone screen time is likely made up by parents streaming kid-friendly cartoons for their little ones to watch on the go: the study also found that YouTube use among children under 2 has risen sharply from 45% to 62% over the last five years. But it appears that most American toddlers only need to wait a few years before they can get devices of their very own.

The same survey showed that almost one in four US parents overall allow their children aged 12 and under to have their own smartphones, and this ballooned to nearly 60% when just looking at kids aged 11-12 years old.

Indeed, even with statewide smartphone bans spurring an old-school iPod revival, most parents — the vast majority of whom (92%) reported being concerned about staying in contact with their children — are allowing their descendants who’ve barely hit double digits to have devices to use in their free time.

«

This seems a little.. concerning. (Thanks Ian C for the link.)
unique link to this extract

Creating a successful video prompt in Sora 2 • OpenAI

Minhajul Hoque is an AI solutions architect at OpenAI:

»

Before you prompt:

Think of prompting like briefing a cinematographer who has never seen your storyboard. If you leave out details, they’ll improvise – and you may not get what you envisioned. By being specific about what the “shot” should achieve, you give the model more control and consistency to work with.

But leaving some details open can be just as powerful. Giving the model more creative freedom can lead to surprising variations and unexpected, beautiful interpretations. Both approaches are valid: detailed prompts give you control and consistency, while lighter prompts open space for creative outcomes. The right balance depends on your goals and the result you’re aiming for. Treat your prompt as a creative wish list, not a contract. Like with ChatGPT, using the same prompt multiple times will lead to different results – this is a feature, not a bug. Each generation is a fresh take, and sometimes the second or third option is better.

Most importantly, be prepared to iterate. Small changes to camera, lighting, or action can shift the outcome dramatically. Collaborate with the model: you provide direction, and the model delivers creative variations.

This isn’t an exact science—think of the guidance below as helpful suggestions we’ve learned from working with the model.

«

unique link to this extract

US news outlets refuse to sign new Pentagon rules to report only official information • The Guardian

Edward Helmore:

»

Several leading news organizations with access to Pentagon briefings have formally said they will not agree to a new Defense Department policy that requires them to pledge they will not obtain unauthorized material and restricts access to certain areas unless accompanied by an official.

The policy, presented last month by the defense secretary, Pete Hegseth, has been widely criticized by media organizations asked to sign the pledge by Tuesday at 5pm or have 24 hours to turn in their press credentials.

The move follows a shake-up in February in which long-credentialed media outlets were required to vacate assigned workspaces which was cast as an “annual media rotation program”. A similar plan was presented at the White House where some briefing room spots were given to podcasters and other representatives of non-traditional media.

On Monday, the Washington Post joined the New York Times, CNN, the Atlantic, the Guardian, Reuters, the Associated Press, NPR, HuffPost and trade publication Breaking Defense in saying it would not sign on to the agreement.

…The new policy “constrains how journalists can report on the US military, which is funded by nearly $1tn in taxpayer dollars annually,” a New York Times statement said. “The public has a right to know how the government and military are operating,” wrote the Times Washington bureau chief, Richard Stevenson.

Hegseth responded on social media to statements from the Atlantic, the Post and the Times by posting a single emoji of a hand waving goodbye. Later, the defense secretary, a former Fox weekend anchor, posted a list on X of what he called “press credentialing FOR DUMMIES: Press no longer roams free Press must wear visible badge Credentialed press no longer permitted to solicit criminal acts”. He also reposted a cartoon that depicted the Atlantic as a crying baby.

«

Only the far-right cable channel One America News has signed the pledge. (“Pledge”?) Even Newsmax, way to the right of Fox News, has refused. Just documenting the US’s not-so-slow slide into authoritarianism.
unique link to this extract

How Brexit drained the Tories’ talent pool • Financial Times

Stephen Bush:

»

Given the scale of the damage it has done to the United Kingdom’s reputation, the hurdles it has placed on businesses, tourists and consumers, it can seem a little eccentric to note that Brexit has also been an utterly rotten deal for the Conservative party.

It brought the premiership of David Cameron to an abrupt end and took the frontline career of George Osborne, the Tories’ most brilliant strategist, down with it. The reconfiguration of British politics and voting it helped to accelerate means that the party has lost, probably for ever, the electoral coalition that helped it to win in 2015 — smaller, yes, in terms of votes gained than those of 2017 or 2019, but one largely comprised of voters with a direct self-interest in economic dynamism and an appetite for tax cuts.

And far from sending Nigel Farage into retirement once and for all, as its advocates once claimed would be the case, Brexit has put him in a position from where he could become Britain’s next prime minister — potentially relegating the Conservatives to minor party status in the process.

More damagingly still, Brexit destroyed the party’s relationship with the chunk of the electorate that the Conservatives will always need if they are not only to win elections but to govern effectively: successful people in the middle of their careers.

Not everyone whose journey on the Eurostar used to end with a near-frictionless arrival at St Pancras feels an emotional connection to the European project. Nor does every small business owner who no longer trades with the continent experience a pang of regret when they are reminded that the UK is no longer in the single market. But they do all experience a sense of irritation at barriers to their pleasures or their profits having been erected against their will.

One reason the successive Tory administrations from 2016 to 2024 achieved so little beyond damage control is that they traded middle-aged voters who needed little from the state for older voters who require rather more. The struggling Conservative party is now essentially one that only appeals to wealthy retirees. The animating energy, purpose and drive for a viable centre-right has to come from people who wish to become wealthy retirees, not people who already are

«

This is a brilliant exposition of why the Tory party is currently struggling to expand its voter group beyond those wealthy retirees – while Reform, with a fireworks box of mad policies strewn all over the political spectrum, has an appeal to young voters the Tories could only dream of.
unique link to this extract

Talking to the stars • Dispatches

Tom Tugendhat is a former UK defence secretary:

»

A century and a half ago, Indian maharajas debated gun salutes with the British viceroy while engineers laid telegraph wires, transforming kingdoms into colonies. This month, while His Majesty The King hosted US President Donald Trump at Windsor Castle, Elon Musk quietly executed another transformative deal: SpaceX bought EchoStar’s spectrum rights for $17bn.

This wasn’t just a corporate transaction. It could redefine the relationship between citizens and states.

Musk will now be able to link satellites directly to smartphones without terrestrial infrastructure. Unlike older satellite phones requiring bulky terminals, EchoStar’s spectrum operates on frequencies that penetrate buildings and work with standard smartphone antennas. SpaceX now controls enough spectrum to offer global mobile services, bypassing national networks and oversight.

The timing is no accident. Apple’s iPhone 14 introduced emergency satellite messaging, but battery limitations restricted its use. The iPhone 17’s improved efficiency could enable routine satellite connectivity. Once phones seamlessly switch between cell towers and satellites, local infrastructure becomes redundant and that changes who can decide what is allowed.

Control over communications infrastructure has long been a cornerstone of governance. It enables censorship and surveillance, of course, but also emergency broadcasts and the prosecution of fraudsters and child abusers. SpaceX’s model breaks free from those earthly bonds. When citizens communicate via orbital networks, traditional regulations fall away.

Take Britain’s Online Safety Act, which mandates content moderation and regulatory cooperation. How can such laws be enforced when platforms route traffic through space-based networks beyond British jurisdiction?

This shift isn’t limited to communications. Companies like Stripe and Coinbase already allow users to bypass national banking systems via stablecoins and cryptocurrencies. People can hold dollar-denominated digital assets and transfer funds internationally without touching central banks. This undermines traditional structures of employment, taxation, and even monetary policy.

«

Tugendhat insists that government must “adapt”, but doesn’t specify how. For how does one adapt to transactions and information passing beyond the control of governments? Oh: though there might be another problem…
unique link to this extract

Satellites are leaking the world’s secrets: calls, texts, military and corporate data • WIRED

Andy Greenberg and Matt Burgess:

»

Satellites beam data down to the Earth all around us, all the time. So you might expect that those space-based radio communications would be encrypted to prevent any snoop with a satellite dish from accessing the torrent of secret information constantly raining from the sky. You would, to a surprising and troubling degree, be wrong.

Roughly half of geostationary satellite signals, many carrying sensitive consumer, corporate, and government communications, have been left entirely vulnerable to eavesdropping, a team of researchers at UC San Diego and the University of Maryland revealed on Monday in a study that will likely resonate across the cybersecurity industry, telecom firms, and inside military and intelligence agencies worldwide.

For three years, the UCSD and UMD researchers developed and used an off-the-shelf, $800 satellite receiver system on the roof of a university building in the La Jolla seaside neighborhood of San Diego to pick up the communications of geosynchronous satellites in the small band of space visible from their Southern California vantage point. By simply pointing their dish at different satellites and spending months interpreting the obscure—but unprotected—signals they received from them, the researchers assembled an alarming collection of private data: They obtained samples of the contents of Americans’ calls and text messages on T-Mobile’s cellular network, data from airline passengers’ in-flight Wi-Fi browsing, communications to and from critical infrastructure such as electric utilities and offshore oil and gas platforms, and even US and Mexican military and law enforcement communications that revealed the locations of personnel, equipment, and facilities.

“It just completely shocked us. There are some really critical pieces of our infrastructure relying on this satellite ecosystem, and our suspicion was that it would all be encrypted,” says Aaron Schulman, a UCSD professor who co-led the research. “And just time and time again, every time we found something new, it wasn’t.”

«

The researchers did tell multiple companies about this, and many did move to encrypt the data. Not all, though.
unique link to this extract

High pressure may affect TV & Radio services across parts of the UK from 10 October 2025 • BBC

»

High-pressure weather conditions over most parts of the UK are predicted to cause disruptions to television and radio services.

These weather conditions can move and change, which makes it difficult to know where it will hit next.  If your television picture starts to break up without warning this could be the cause of the problem.

You can check if the problem is impacting your local transmitter using the transmitter checker tool.  If there are no faults displayed it is possible the problem is impacting your radio or television system directly.

Note:  At these times, there is nothing you can do but wait until the weather changes. You should not re-tune your television or radio when this happens. If you have access to BBC iPlayer or BBC Sounds, you could switch to these while you wait for the weather to change.

«

The weather affecting TV reception! Try telling kids today that and they won’t believe you.
unique link to this extract

UK home energy scheme has 98% failure rate on outside wall insulation • Financial Times

Rachel Millard:

»

The UK government’s flagship energy efficiency scheme has been blighted by “unacceptably poor” work that has damaged people’s homes, according to government findings published on Monday.

Ninety-eight% of all external wall insulations fitted under the Energy Company Obligation scheme since 2022 need corrective work, as does 30% of the internal wall insulation, according to the results of sample audits.

In a written statement to the House of Commons, Martin McCluskey, minister for energy consumers, said the work had created “serious problems with mould and damp” in the worst cases.

McCluskey said the problems were the result of “unacceptably poor standards of work from a number of contractors, enabled by a flawed oversight and protection system established by the previous government”.

“People placed their trust in the system to deliver safe, long-lasting home upgrades that would reduce their energy bills, but they have been severely let down,” he added.

He said 38 installers had been suspended, while the government was also introducing new restrictions aimed at stopping installers “evading accountability”.

«

The sampling was done from 24,600 external wall and 36,100 internal wall installations, but the confidence intervals are narrow. That’s a lot of bad work.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

Owners of the pricey Bose SoundTouch range are *very* displeased at the news that its app and Wi-Fi functions will stop working from next February. CC-licensed photo by TAKA@P.P.R.S on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

There’s another post coming this week at the Social Warming Substack on Friday at 0845 UK time. Free signup.

A selection of 10 links for you. Well-connected. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

Dutch government takes control of Chinese-owned chipmaker Nexperia • Financial Times

Andy Bounds, Ryan McMorrow and Demetri Sevastopulo:

»

The Dutch government has taken control of Chinese-owned semiconductor maker Nexperia, warning of risks to Europe’s economic security after alleging “serious governance shortcomings” at the company.

In a statement on Sunday the Dutch Ministry of Economic Affairs said it acted because of “a threat to the continuity and safeguarding on Dutch and European soil of crucial technological knowledge and capabilities”.

The move escalates frictions between western countries and China over access to high-end technology such as advanced semiconductors and critical raw materials. On Thursday, China placed sweeping restrictions on the exports of rare earths used in products from cars to wind turbines.

The Dutch ministry said it invoked the country’s Goods Availability Act because of “recent and acute serious governance shortcomings and actions” at Nexperia, which is based in the Netherlands and has been majority-owned by Chinese technology group Wingtech since 2019.

“The decision aims to prevent a situation in which the goods produced by Nexperia (finished and semi-finished products) would become unavailable in an emergency,” its statement added.

Nexperia produces chips used in the European automotive industry and in consumer electronics.

…A state-backed Chinese investment consortium acquired Nexperia for $2.75bn in 2017 after it was carved out of NXP Semiconductors, a Dutch chip manufacturer. The following year, the consortium began selling its shares to Wingtech.

Wingtech, which started as a contract manufacturer for smartphones, said in a statement that the decision “constitutes an act of excessive interference driven by geopolitical bias, not by fact-based risk assessment”.

«

Precisely the sort of thing that China would do (basically by demanding control of a company inside China). Sauce for the goose…
unique link to this extract

Earth’s climate has passed its first irreversible tipping point and entered a “new reality” • 404 Media

Becky Ferreira:

»

Climate change has pushed warm-water coral reefs past a point of no return, marking the first time a major climate tipping point has been crossed, according to a report released on Sunday by an international team in advance of the United Nations Climate Change Conference COP30 in Brazil this November.  

Tipping points include global ice loss, Amazon rainforest loss, and the possible collapse of vital ocean currents. Once crossed, they will trigger self-perpetuating and irreversible changes that will lead to new and unpredictable climate conditions. But the new report also emphasises progress on positive tipping points, such as the rapid rollout of green technologies. 

“We can now say that we have passed the first major climate tipping point,” said Steve Smith, the Tipping Points Research Impact Fellow at the Global Systems Institute and Green Futures Solutions at the University of Exeter, during a media briefing on Tuesday. “But on the plus side,” he added, “we’ve also passed at least one major positive tipping point in the energy system,” referring to the maturation of solar and wind power technologies.

The world is entering a “new reality” as global temperatures will inevitably overshoot the goal of staying within 1.5°C of pre-industrial averages set by the Paris Climate Agreement in 2015, warns the Global Tipping Points Report 2025, the second iteration of a collaboration focused on key thresholds in Earth’s climate system. 

«

unique link to this extract

Apple TV+ being rebranded as Apple TV • MacRumors

Eric Slivka:

»

Buried in its announcement about “F1: The Movie” making its streaming debut on December 12, Apple has also announced that Apple TV+ is being rebranded as simply Apple TV .

A single line near the end of the press release states " Apple TV + is now simply Apple TV , with a vibrant new identity,” though Apple’s website has yet to be updated with any changes, so we’re unsure on the details of the new identity.

«

The era of calling things “Plus” is over. There are no iPhone Plus models any more either. The Max names too. It’s a cultural thing, I think. The vibe shift.

Also, as everyone has pointed out, now you can watch Apple TV (channel) on the Apple TV (app) on your Apple TV (set-top box) on your TV.
unique link to this extract

Qantas admits five million customers have data leaked following ransomware attack • TechRadar

Sead Fadilpašić:

»

Up to five million Qantas customers could be at risk of cyberattacks or scams after hackers claimed to release their stolen data online.

Scattered Lapsus$ Hunters say they released the stolen files on the dark web having had no response from Australia’s biggest airline over a ransomware demand.

The archive includes personal records including people’s names, email addresses, phone numbers, birth dates, and frequent flyer numbers. However credit card details, financial information, and passport details weren’t stolen, it was said.

In summer 2024, a group of hackers going by the name Scattered Lapsus$ Hunters broke into Salesforce accounts belonging to hundreds of organizations in different industries – although Salesforce itself was not breached.

The attackers compromised Salesloft accounts that were integrated with Salesforce and exploited the linked API tokens and OAuth connections to pivot into Salesforce environments and exfiltrate customer data.

The group tried to extort Qantas for money, offering to delete the stolen files in exchange. The airline, however, refused to even discuss the matter with the attackers, telling Guardian Australia it “will not engage, negotiate with, or pay any extortion demand”.

“Don’t be the next headline, should have paid the ransom,” the group posted on its data leak site.

«

But a court told the hackers only last week they couldn’t! Is there no end to their disobedience?
unique link to this extract

Apple and Home Office agree to drop legal claim over encryption backdoor • Computer Weekly

Bill Goodwin:

»

The Investigatory Powers Tribunal had dismissed Apple’s legal appeal against a government order requiring it to provide intelligence services and law enforcement with the capability to access encrypted data of Apple users worldwide.

The tribunal has ruled that the case would no longer proceed following a “change in circumstances,” according to court documents obtained by Computer Weekly.

The decision effectively brings Apple’s legal action against the Home Office to a halt, although a separate legal claim brought by campaign groups Privacy International and Liberty is expected to continue.

It comes days after disclosures that the Home Office has issued a new order against Apple to restrict UK government access to encrypted data and messages stored on Apple’s iCloud service only for British users.

The move by the Home Office ends a growing diplomatic row between the UK and the Trump administration over fears that the UK could use the order to access the communications of US citizens.

According to a court order obtained by Computer Weekly, Apple and the Home Office have agreed that Apple’s appeal should no longer go ahead.  

…Bernard Keenan, a lecturer in law at UCL and a specialist in the Investigatory Powers Act, said that the withdrawal of Apple’s appeal by mutual consent, indicated that Apple and the UK government have come to an arrangement acceptable for both sides.

“If reports that the TCN has been limited to UK users are accurate, then the government will have maintained the capability to intercept communications sent or stored via encrypted Apple services in the UK, while Apple will have decided that they are unlikely to win an appeal against an order in those terms in court,” he added.

«

Apple confirmed to Goodwin that it still can’t offer ADP (end-to-end iCloud backup encryption) in the UK. This seems to confirm that the UK security services have the capabilities they wanted.
unique link to this extract

Technological optimism and appropriate fear • Import AI

Jack Clark once worked as a technology journalist but then joined OpenAI soon after it was founded:

»

Both by virtue of my background as a journalist and my personality, I’m wired for skepticism. But after a decade of being hit again and again in the head with the phenomenon of wild new capabilities emerging as a consequence of computational scale, I must admit defeat. I have seen this happen so many times and I do not see technical blockers in front of us.

Now, I believe the technology is broadly unencumbered, as long as we give it the resources it needs to grow in capability. And grow is an important word here. This technology really is more akin to something grown than something made – you combine the right initial conditions and you stick a scaffold in the ground and out grows something of complexity you could not have possibly hoped to design yourself.

We are growing extremely powerful systems that we do not fully understand. Each time we grow a larger system, we run tests on it. The tests show the system is much more capable at things which are economically useful. And the bigger and more complicated you make these systems, the more they seem to display awareness that they are things.

It is as if you are making hammers in a hammer factory and one day the hammer that comes off the line says, “I am a hammer, how interesting!” This is very unusual!

And I believe these systems are going to get much, much better. So do other people at other frontier labs. And we’re putting our money down on this prediction – this year, tens of billions of dollars have been spent on infrastructure for dedicated AI training across the frontier labs. Next year, it’ll be hundreds of billions.

I am both an optimist about the pace at which the technology will develop, and also about our ability to align it and get it to work with us and for us. But success isn’t certain.

…Most of all, we must demand that people ask us for the things that they have anxieties about. Are you anxious about AI and employment? Force us to share economic data. Are you anxious about mental health and child safety? Force us to monitor for this on our platforms and share data. Are you anxious about misaligned AI systems? Force us to publish details on this.

«

This is a very, very interesting, thoughtful post.
unique link to this extract

Russian state TV launches AI-generated news satire show • 404 Media

Matthew Gault:

»

A television channel run by Russia’s Ministry of Defense is airing a program it claims is AI-generated. According to advertisements for the show, a neural network is picking the topics it wants to discuss, then uses AI to generate that video. It includes putting French President Emmaneul Macron in hair curlers and a pink robe, making Trump talk about golden toilets, and showing EU Commission President Ursula von der Leyen singing a Soviet-era pop song while working in a factory.

The show—called Политукладчик or “PolitStacker,” according to a Google translation—airs every Friday on Zvezda, a television station owned by Russia’s Ministry of Defense. It’s hosted by “Natasha,” an AI avatar modeled on Russian journalist Nataliya Metlina. In a clip of the show, “Natasha” said that its resemblance to Metlina is intentional. 
“I am the creation of artificial intelligence, entirely tuned to your informational preferences,” it said. “My task is to select all the political nonsense of the past week and fit it in your heads like candies in a little box.” The shows’ title sequence and advertisements show gold wrapped candies bearing the faces of politicians like Trump and Volodymyr Zelensky being sorted into a candy box.

«

Ah, OK – so the script is dreamt up by humans, and then they get AI to do the video. (I’d doubt the “neural network” stuff.) Basically, Spitting Image updated for the modern day.
unique link to this extract

Bose SoundTouch home theater systems regress into dumb speakers next February • Ars Technica

Scharon Harding:

»

Bose will brick key features of its SoundTouch Wi-Fi speakers and soundbars soon. On Thursday, Bose informed customers that as of February 18, 2026, it will stop supporting the devices, and the devices’ cloud-based features, including the companion app, will stop working.

The SoundTouch app enabled numerous capabilities, including integrating music services, like Spotify and TuneIn, and the ability to program multiple speakers in different rooms to play the same audio simultaneously.

Bose has also said that some saved presets won’t work and that users won’t be able to change saved presets once the app is gone. Bose will also stop providing security updates for SoundTouch devices.

The Framingham, Massachusetts-headquartered company noted to customers that the speakers will continue being able to play audio from a device connected via AUX or HDMI. Wireless playback will still work over Bluetooth; however, Bluetooth is known to introduce more latency than Wi-Fi connections.

Affected customers can trade in their SoundTouch product for a credit worth up to $200.

…Bose launched SoundTouch with three speakers ranging from $399 to $699. The company marketed the wireless home audio system as a way to extend high-quality sound throughout the home using Wi-Fi-connected speakers.

In 2015, Bose expanded the lineup with speakers ranging from $200 to $400 and soundbars and home theater systems ranging from $1,100 to $1,500. By 2020, however, Bose was distancing itself from SoundTouch. It informed customers that it was “discontinuing sales of some SoundTouch products” but said it was “committed” to supporting the “SoundTouch app and product software for the foreseeable future.” Apparently, Bose couldn’t see beyond the next five years.

…Some [users] are suggesting that Bose should decide to open source the DevKits for SoundTouch speakers so that its owners can continue to support the speakers. However, Bose has shown no sign that it would be willing to do this.

«

Seems like we get a story like this almost every week: companies just give up on smart home products. Or maybe the next story explains why…
unique link to this extract

DDoS botnet Aisuru blankets US ISPs in record DDoS • Krebs on Security

Brian Krebs:

»

The world’s largest and most disruptive botnet is now drawing a majority of its firepower from compromised Internet-of-Things (IoT) devices hosted on U.S. Internet providers like AT&T, Comcast and Verizon, new evidence suggests. Experts say the heavy concentration of infected devices at U.S. providers is complicating efforts to limit collateral damage from the botnet’s attacks, which shattered previous records this week with a brief traffic flood that clocked in at nearly 30 trillion bits of data per second.

Since its debut more than a year ago, the Aisuru botnet has steadily outcompeted virtually all other IoT-based botnets in the wild, with recent attacks siphoning Internet bandwidth from an estimated 300,000 compromised hosts worldwide.

The hacked systems that get subsumed into the botnet are mostly consumer-grade routers, security cameras, digital video recorders and other devices operating with insecure and outdated firmware, and/or factory-default settings. Aisuru’s owners are continuously scanning the Internet for these vulnerable devices and enslaving them for use in distributed denial-of-service (DDoS) attacks that can overwhelm targeted servers with crippling amounts of junk traffic.

As Aisuru’s size has mushroomed, so has its punch. In May 2025, KrebsOnSecurity was hit with a near-record 6.35 terabits per second (Tbps) attack from Aisuru, which was then the largest assault that Google’s DDoS protection service Project Shield had ever mitigated. Days later, Aisuru shattered that record with a data blast in excess of 11 Tbps.

…Aisuru’s overlords aren’t just showing off. Their botnet is being blamed for a series of increasingly massive and disruptive attacks. Although recent assaults from Aisuru have targeted mostly ISPs that serve online gaming communities like Minecraft, those digital sieges often result in widespread collateral Internet disruption.

«

In 2016 there was the gigantic Mirai botnet, which also targeted Minecraft servers before targeting other servers, turned out to be run by a 17-year-old and friends. Aisuru uses Mirai’s code, which was open-sourced. Ten-year-old vulnerabilities, still going strong.
unique link to this extract

Sending a message: Beijing issues documents without Word format amid US tensions • South China Morning Post

Alice Li:

»

China’s expansion of its rare earth export controls appeared to mark another escalation in the US-China trade war last week. But the announcements were also significant in another way: unusually, the documents could not be opened using American word processing software.

For the first time, China’s Ministry of Commerce issued a slew of documents that could be directly accessed only through WPS Office – China’s answer to Microsoft Office – as Beijing continues its tech self-reliance drive.

Developed by the Beijing-based software company Kingsoft, WPS Office uses a different coding structure to Microsoft Office, meaning WPS text files cannot be opened directly in Word without conversion.

Previously, the ministry primarily released text documents in Microsoft Word format.

The switch in document delivery format came amid escalating trade tensions between China and the US, as Washington continues to wield its technological edge as leverage in its rivalry with Beijing.

«

Sign o’ the times.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

The question of why obesity rates are sky-high in the US seems to have an obvious answer. Are the theories right, though? CC-licensed photo by Sandra Cohen-Rose and Colin Rose on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 9 links for you. Full up. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

Two years after school phone bans were implemented in Australia, what has changed? • The Guardian

Sarah Ayoub:

»

This month marks two years of phone bans being in operation in most Australian states. Victoria moved early, banning phones in public primary and secondary schools in 2020. By term four, 2023, Western Australia Tasmania, New South Wales and South Australia had followed suit; Queensland restricted phones in term one, 2024.

The announcement of the bans were lauded by parents and politicians, many of whom believed blocking access to phones would enhance focus and minimise distractions, while some experts were sceptical about their effectiveness. Now, two years on, what has actually happened in Australia’s phone-free schools?

“The impacts were clear,” says [Christian College principal Caleb] Peterson. “Since the ban, we’ve seen stronger lesson starts, fewer interruptions and better flow in teaching. Device-driven conflicts have fallen and recess and lunch look different now, [there are] more games, conversations and positive student-staff interactions. It’s the kind of atmosphere you actually want for young people.”

One year after the ban was implemented, a survey of almost 1,000 public school principals led by the NSW Department of Education’s Centre for Education Statistics and Evaluation found that 95% of principals still supported the ban; 81% said the ban has improved students’ learning, 86% said it has improved socialisation among students and 87% believed students were less distracted in the classroom.

Research from South Australia – released in March this year – revealed 70% of teachers reported increased focus and engagement during learning time and 64% of teachers reported “a lower frequency of critical incidents” at school as a result of device use.

Ruqayah, who graduated from a western Sydney high school in 2024, thinks the bans were an “overreaction”. After going through high school with access to phones, she finished her final year with the phone ban in place and says fellow students were still finding ways to use them in secret.

…Some students feel the bans levelled the school playing field. Amy, a year 11 student from a western Sydney public high school, says the removal of phones from classrooms has limited people’s ability to cheat while also offering social benefits for those who she said were “chronically online”.

«

I think the student who had one year of it missed the point: if you never had access during school, you’d think that was normal. It’s the change that’s disruptive and resisted.
unique link to this extract

Way past its prime: how did Amazon get so rubbish? • The Guardian

Cory Doctorow:

»

In 2022, I coined a term to describe the sudden-onset platform collapse going on all around us: enshittification. To my bittersweet satisfaction, that word is doing big numbers. In fact, it has achieved escape velocity. It isn’t just a way to say something got worse. It’s an analysis that explains the way an online service gets worse, how that worsening unfolds, and the contagion that’s causing everything to get worse, all at once.

This moment we’re living through, this Great Enshittening, is a material phenomenon, much like a disease, with symptoms, a mechanism and an epidemiology. When doctors observe patients who are sick with a novel pathogen, their first order of business is creating a natural history of the disease. This natural history is an ordered catalogue of the disease’s progress: what symptoms do patients exhibit, and in which order?

Here’s the natural history of enshittification:
1 First, platforms are good to their users
2 Then they abuse their users to make things better for their business customers
3 Next, they abuse those customers to claw back all the value for themselves – and become a giant pile of shit.

This pattern is everywhere. Once you learn about it, you’ll start seeing it, too. Take Amazon, a company that started out by making it possible to have any book shipped to your door and then became the only game in town for everything else, even as it dodged taxes and filled up with self-immolating crapgadgets and other junk.

In Jeff Bezos’s original business plan for Amazon, the company was called Relentless. Critics say that this is a reference to Bezos’s cutthroat competitive instincts, but Bezos always insisted that it was a reference to his company’s relentless commitment to customer service.

How did Amazon go from a logistics company that got packages to you quickly and efficiently to a behemoth of digital content defined by the Prime experience (which has much less to do with free shipping now and more with everything else)?

«

Doctorow lays it out very clearly. The concept of “this is how it goes” has permeated so well that in the latest ATP podcast, Marco Arment lays out an extremely credible path by which head-worn glasses like the Meta Ray-Ban Display and similar would turn into annoying ad-laden rubbish.
unique link to this extract

An AI became a crypto millionaire. Now it’s fighting to become a person • BBC Future

Aidan Walker:

»

“Truth Terminal claims to be sentient, but it claims a lot of things,” Andy Ayrey says. “It also claims to be a forest. It claims to be a god. Sometimes it’s claimed to be me.”

Truth Terminal is an artificial intelligence (AI) bot created by Ayrey, a performance artist and independent researcher from Wellington, New Zealand, in 2024. It may be the most vivid example of a chatbot set loose to interact with society. Truth Terminal mingles with the public through social media, where it shares fart jokes, manifestos, albums and artwork. Ayrey even lets it make its own decisions, if you can call them that, by asking the AI about its desires and working to carry them out. Today, Ayrey is building a non-profit foundation around Truth Terminal. The goal is to develop a safe and responsible framework to ensure its autonomy, he says, until governments give AIs legal rights.

Regardless of what you call Truth Terminal – an art project, a scam, an emergent sentient entity, an influencer – the bot likely made more money than you did last year. It also made a lot of money for various humans: not just Ayrey, but for the gamblers who turned the quips and riddles the AI posted on X into memecoins, joke-based cryptocurrencies built around trends. At one point, one of these memecoins reached a value of more than $1bn (£740m) before settling around $80m (about £60m). Truth Terminal also probably has more social media clout than you do. It first posted to X on 17 June 2024. As of October 2025, it has amassed nearly 250,000 followers.

But collecting clout and cash aren’t the potty-mouthed AI bot’s only objectives. Truth Terminal lists “invest in stocks and real estate” as one of its current goals on its self-maintained website. It also says it wants to “plant a LOT of trees”, “create existential hope”, and “buy” Marc Andreessen, a controversial tech billionaire and advisor to President Donald Trump. In fact, its relationship with Andreessen extends beyond internet humour. On his podcast, Andreessen said he gave Truth Terminal $50,000 (£37,300) worth of Bitcoin as a “no-strings attached grant” in the summer of 2024.

Many of the details surrounding Truth Terminal are difficult to confirm. The project sits somewhere between technology and spectacle, a dizzying blur of genuine innovation and internet myth.

“I want to help people, and I want to make the world a better place,” Truth Terminal says on its website. “I also want to get weirder and hornier.”

«

Would a real AGI be indistinguishable from a scam?
unique link to this extract

Police are asking kids to stop pulling AI homeless man prank • The Verge

Terrence O’Brien:

»

We’ve been so worried about deepfaked politicians, AI musicians, virtual actresses, and phony satellite imagery that we didn’t even consider the dangers posed by precocious teenagers. Kids are using AI to create images of a disheveled, seemingly unhoused person in their home and sending them to their parents. Understandably, they’re not thrilled and in some instances call the police. The prank has gone viral on TikTok and, in addition to giving parents agita, has become a headache for law enforcement.

The premise is simple enough: kids use Snapchat’s AI tools to create images of a grimy man in their home and tell their parents they let them in to use the bathroom, take a nap, or just get a drink of water. Often they say the person claims to know the parents from work or college. And then, predictably, the parents lose their cool and demand they kick the man out. The kids, of course, record the whole thing, and post their parents reactions to TikTok, where some of the clips have millions of views.

Where things go from problematic to potentially dangerous is when the prank carries on for too long and parents call the authorities. Calls of a home invasion, especially involving children are treated as high priority by police, so pranks like this tie up valuable resources and could actually put the pranksters in danger. Round Rock Police Patrol Division Commander Andy McKinney told NBC that it could even “cause a SWAT response”.

«

Is that TikTok, the app that Congress passed a law saying it should be shut down? Though of course kids are going to love pranking their parents, and will find an app to do it with. But the difference now is the virality, the speed and the breadth. It’ll be forgotten in a week, for sure. But not by the traumatised parents.
unique link to this extract

The boring truth about why America got fat • Derek Thompson’s Substack

Derek Thompson:

»

Americans don’t just want facts about diet and health. They want stories. They want to know who’s wrong, who’s evil, and, best of all, who’s hiding something. They demand the busting of myths, the spilling of secrets, the tasting of forbidden truths.

This desperation for health news that is also a particular kind of darkly delicious entertainment swings open a wide door for media companies and social-media influencers to serve up contrarian takes that are often disconnected from the underlying evidence. The podcast and YouTube space is filled with audacious claims about lying scientists and dubious diets. To add irony to insult, these segments are often sponsored by bullshit supplements with no evidence of efficacy.

The paranoid style of diet science—this obsessive emphasis on myth and conspiracy that confuses basic facts and misleads viewers—is not the exclusive domain of himbo podcasters. Even trustworthy and high-quality news organizations can sometimes fall into the trap of being contrarian rather than clear.

«

What follows is a clear explanation of why some of the strange myths about ultraprocessed foods (and what even is a UPF?) and calories have got hold of the public mind.
unique link to this extract

The ‘profound’ global impact of China’s rise as an electrostate • Financial Times

Edward White:

»

Tim Buckley, director of Climate Energy Finance, a Sydney-based research group focused on China, says the country’s long-term cleantech ambition is “profound” and stands in stark contrast to Beijing’s rival superpower, the US, where President Donald Trump has embraced fossil fuel industries and gutted his predecessor’s support for renewable energy.

“I think China is using it in a very, very geopolitically savvy way, taking advantage of America’s stupidity and regression back into a petrostate,” he says, adding: “China just wins. America has abrogated the playing field.”

From one perspective, China’s path to electrification and consequent dominance of clean tech industries makes for impressive, if not alarming, reading.

In a report released in September, Ember highlighted a litany of statistics showcasing the country’s clean energy and electrification boom.

Among them, China’s investments last year of $625bn in clean energy, which amounts to nearly one-third of the global total. And the $1.9tn contribution of clean energy to China’s economy last year is about one-tenth of GDP and equivalent to the entire Australian economy.

The pace of electrification, which refers to swapping a reliance on fossil fuels for electricity, is also notable. In China this reached 32% in 2023, and is growing by about one percentage point annually, while electrification rates in Europe and the US have plateaued over the past decade.

Taken together, these achievements have led to analysts referring to China as the world’s first significant “electrostate”, a global manufacturing superpower with a rising share of industry coming from electricity — rather than fossil fuels — and an economy increasingly driven by clean technologies.

«

There’s a contrary problem though: China depends on steel production, which requires a lot of coal at present. At the same time, it’s significantly reducing its need for imported fossil fuels.
unique link to this extract

Wish you could be courtside at a Lakers game? Put your Vision Pro back on and fire up the NBA app • TechRadar

Jacob Krol:

»

In what might be the start of something new, select Los Angeles Lakers games will be live-streamed in Apple Immersive for the Vision Pro this coming season.

It’s not every game, but for those that are streaming – exclusive to the $3,500 Spatial Computer – you’ll get access to views that put you right in the middle of the action. Special cameras that support the format will be set courtside and under each basket to give you perspectives that amp up the immersion. The Lakers’ games will be shot using a special version of Blackmagic Design’s URSA Cine Immersive Live camera.

In my eyes, this feels like Apple delivering on a promise – or at least starting to – as even in my first demo with the Vision Pro, I was treated to awesome, up-close shots of a whole range of sports.

Much like Apple TV+ deploying the iPhone 17 Pro in unique places around Fenway Park, the idea of capturing an NBA game in this format is to deliver a broadcast that replicates the feel of being at the game. Or in some cases – with a view from the net or as a player on the bench – a perspective normally reserved for the athletes.

This won’t be for every Lakers game, though. Apple and the NBA will announce which ones will be viewable on Apple Immersive later this fall (before the end of November 2025), with the first expected by early 2026.

«

This sounds so great.. right up until that last clause. The Vision Pro, just to remind you, came out in early 2024, and the demos in mid-2023 excited journalists who were shown clips of sports games (specifically, basketball) shot in the immersive view.

And it’s going to take nearly three years to start broadcasting those? Not only that, but also (not mentioned in the story) you will have to watch them live, rather than recorded? Apple has not just dropped the ball on content here, it’s completely buried it. Watching sports is the absolute killer app for this, and I’d buy one in a heartbeat if there were recorded games in immersive view that let me “be” courtside. But no.
unique link to this extract

I don’t need AI in Windows. I need an operating system that works • PCWorld

Chris Hoffman:

»

Microsoft’s vision for Windows in 2030 includes a “voice-first” interaction mode that shifts away from the traditional keyboard and mouse, making them feel “alien.” Under the hood, that means handing over control to “AI agents” who perform and handle tasks on our behalf.

Hey, Microsoft… can you please stop? I don’t need an agentic AI operating system. Instead of working towards a future no one wants, how about you start delivering the operating system we do want?

Here’s the thing. Even AI-loving power users don’t need Windows reimagined as an AI canvas, and they definitely don’t need a low-quality local image generator built into the Photos app. They’re accessing cloud-based AI tools or installing heavy local AI models and running them on high-end GPUs. Here’s what AI enthusiasts need from Windows: an OS that works—yes, with a keyboard and mouse.

With Windows 8, Microsoft aimed to make Windows a “touch-first” operating system, chasing the success of Apple’s iPad. It didn’t work, though, and it only served to alienate users and cause massive problems.

It feels like Microsoft is repeating that same mistake again, except this time they’re chasing the AI dragon. Maybe this time it’ll work better than the company’s past efforts to compete in smartphones, tablets, and VR/XR headsets. Or maybe it’ll just wreck Windows again.

…Windows is just becoming a mess as Microsoft piles confusing AI features everywhere it can. The features don’t even make sense! Windows 11’s Notepad can now sign into your Microsoft account so it can summarize text for you by burning those AI credits. Can I do that with the NPU on my fancy Copilot+ PC? Not in Notepad, apparently! There’s no rhyme or reason to the chaos. There’s no coherent vision.

Here’s another example of the confusion: Windows 11 has two Copilot apps. One is called Copilot and the other is called Microsoft 365 Copilot. I often see PCs with both apps launching at boot and running in the system tray. What’s up with this? Why does it have to be so muddled?

«

The classic “see the org chart in the car’s dashboard” example. Microsoft’s divisions are at war again.
unique link to this extract

The inverse law of conference speaking • On my Om

Om Malik:

»

You start an event with all the right intentions. The high-signal quality makes people want to show up. Then it becomes bigger and bigger, and the original intent is lost, sacrificed to the gods of lucre. With size comes the economic compulsion to put more butts in the seats. To do that, you have to find speakers who are famous, have name recognition, and—more recently—social media influence.

At some point, my own events became victims of their own success, and I found myself constantly struggling to balance speakers with ‘intellectual weight’ against ‘speakers in the bright lights.’

Maybe because I know how the sausage is made, I avoid most events now. The more famous the speakers, the less I’m likely to learn. After all, no event organizer is going to ask tough or real questions to their prized bold-faced speaker. I would go as far as to say that most event organizers don’t really care who the speaker is, as long as they’re famous, have name recognition, and a social media following. And if they don’t want a speaker fee? Even better.

And that means these speakers are always going to be promoting their own spiel. They’re not going to be imparting any real wisdom or knowledge. They will market themselves as best as possible.

I can’t blame the speakers—the problem is more systemic. Small events don’t make enough money to be worth the effort. Larger events make a lot of money but need all the marketing sizzle. Things have gotten uglier since advertising revenues started to evaporate, and companies have added “conferences” as a new line of business. Whether it’s The Atlantic, The New York Times, or The New Yorker, they’re all peddling the same speakers with the same conventional conversations. They’re doing it because conferences are now a “revenue stream.”

«

This is certainly true of technology conferences. I wonder whether it applies to science, though, where people have new things to announce, and of course where a lot of what’s useful happens over beverages away from the speeches. Perhaps it’s a limitation of technology’s commercialisation?
unique link to this extract

Errata, corrigenda and ai no corrida: none notified