Can you smile like Mark Zuckerberg? (What do you mean, he doesn’t smile?) A new webcam game will test you. CC-licensed photo by Enrique Dans 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. We are not amused. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

DOGE plans to rebuild SSA code base in months, risking benefits and system collapse • WIRED

Makena Kelly:

»

The so-called Department of Government Efficiency (DOGE) is starting to put together a team to migrate the Social Security Administration’s (SSA) computer systems entirely off one of its oldest programming languages in a matter of months, potentially putting the integrity of the system—and the benefits on which tens of millions of Americans rely—at risk.

The project is being organized by Elon Musk lieutenant Steve Davis, multiple sources who were not given permission to talk to the media tell WIRED, and aims to migrate all SSA systems off COBOL, one of the first common business-oriented programming languages, and onto a more modern replacement like Java within a scheduled tight timeframe of a few months.

Under any circumstances, a migration of this size and scale would be a massive undertaking, experts tell WIRED, but the expedited deadline runs the risk of obstructing payments to the more than 65 million people in the US currently receiving Social Security benefits.

“Of course, one of the big risks is not underpayment or overpayment per se; [it’s also] not paying someone at all and not knowing about it. The invisible errors and omissions,” an SSA technologist tells WIRED.

The Social Security Administration did not immediately reply to WIRED’s request for comment.

SSA has been under increasing scrutiny from president Donald Trump’s administration. In February, Musk took aim at SSA, falsely claiming that the agency was rife with fraud. Specifically, Musk pointed to data he allegedly pulled from the system that showed 150-year-olds in the US were receiving benefits, something that isn’t actually happening. Over the last few weeks, following significant cuts to the agency by DOGE, SSA has suffered frequent website crashes and long wait times over the phone, The Washington Post reported this week.

…Sources within SSA expect the project to begin in earnest once DOGE identifies and marks remaining beneficiaries as deceased and connecting disparate agency databases. In a Thursday morning court filing, an affidavit from SSA acting administrator Leland Dudek said that at least two DOGE operatives are currently working on a project formally called the “Are You Alive Project,” targeting what these operatives believe to be improper payments and fraud within the agency’s system by calling individual beneficiaries.

«

This is going to go wrong. One can predict that with certainty. The only question is how wrong. Badly wrong? Quickly wrong? Enormously wrong? All of those wrong? And the other question is: why? What’s to be gained? Nobody seems able or willing to explain that crucial question.
unique link to this extract

iOS 19.4 rumoured to revamp Health app with new coaching feature • MacRumors

Joe Rossignol:

»

In his Power On newsletter today, Gurman said Apple plans to offer a new AI-powered health coaching feature that offers personalized health recommendations.

The information provided by the coaching feature would be accompanied by videos from health experts that inform users about various health conditions and ways to make lifestyle improvements. For example, Gurman said if the Apple Watch tracks poor heart-rate trends, a video could explain the risks of heart disease.

It is possible that the feature could eventually be part of an Apple Health+ service.

Food tracking will be another big part of the revamped Health app, which could compete with the MyFitnessPal app, according to Gurman.

Apple is also aiming for the AI-powered coaching feature to provide users with fitness-related tips, such as how to improve their technique during workouts. This feature could eventually be built into the Apple Fitness+ service.

«

I can’t remember who observed that Apple increasingly seems to be thinking of new features which are aimed at semi-geriatric users, but they’re right. I’ll reserve judgement on the AI-powered coaching, but I don’t have high hopes. If Apple hasn’t got its LLM-based Siri/AI out of the door this year, is it really going to blow the doors off with AI-powered advice on how you’re lifting wrong?
unique link to this extract

Nuclear startup Terrestrial Energy goes public via SPAC, netting $280m in merger • TechCrunch

Tim De Chant:

»

Terrestrial Energy, a small nuclear startup, merged with an acquisition company on Wednesday. 

The North Carolina-based company is developing small modular reactors and expects to net $280m from the deal. Before the SPAC merger, Terrestrial Energy had raised $94m, according to PitchBook. The combined entity expects to list on Nasdaq under the symbol IMSR.

The ticker is a reference to Terrestrial Energy’s flavor of small modular reactor (SMR), which it calls an integral molten salt reactor. In such a device, uranium fuel is mixed with various salts, such as lithium fluoride or sodium fluoride, that serve to suspend the nuclear fuel and act as the reactor’s main coolant.

Terrestrial Energy’s reactor core is designed to be entirely replaced every seven years, in part to head off some of the problems earlier molten salt reactors experienced like corrosion. The reactor core includes not only the fuel and graphite modulators that regulate the speed of the fission reactions, but also the heat exchangers and pumps that keep the salt cool and flowing.

The startup is targeting a range of markets, including electric power, data centers, and industrial applications that require heat.

There are many extant proposals to build commercial-scale molten salt reactors, but to date, none have been built. The basic technology was invented in the 1950s, but two experiments from that era were plagued with problems.

«

The link there in the final sentence is extremely gloomy about the potential for molten salt reactors, so we’ll have to see whether Terrestrial can live up to its promises.
unique link to this extract

The Myanmar quake is the first major disaster to suffer the brunt of Donald Trump’s devastating cuts • Sky News

Dominic Waghorn:

»

Trump’s decision to shut down the US Agency for International Development was already reported to have decimated US aid operations in Myanmar. Its global impact is hard to overstate. American aid had provided 40% of developmental aid worldwide.

Yesterday, Trump promised Myanmar aid for the earthquake. In reality, his administration has fired most of the people most experienced at organising that help and shut down the means to provide it.

The last of its staff were ironically only let go yesterday even as the president was making lofty promises to help.

The US State Department says it has maintained a team of experts in the country. But former USAID officials say the system is now ‘in shambles’ without the wherewithal to conduct search and rescue or transfer aid.

As they count the cost of this massive earthquake, the people of Myanmar will be hoping, though, for a silver lining, that the disaster may hasten the fall of their despised dictator.

The catastrophe comes at a very bad time for General Min Aung Hlaing, who seized power in a coup four years ago. The Myanmar junta is losing a civil war against an array of opposition forces, ceding territory and now largely kettled into the country’s big cities. Some of the earthquake’s worst damage has been done in the junta’s urban strongholds.

«

More than 1,600 people have died in the earthquake, which is a terrible human price to pay for both Trump (and Musk) and the junta.
unique link to this extract

How not to build an AI Institute • Chalmermagne

Alex Chalmers looks in detail at how the Alan Turing Institute (ATI) got to its current position – ie effectively useless:

»

The story of the ATI is, in many ways, the story of the UK’s approach to technology.

Firstly, drift. The UK has chopped and changed its approach to technology repeatedly, choosing seven different sets of priority technologies between 2012 and 2023. Government has variously championed the tech industry as a source of jobs, a vehicle for exports, a means of fixing public services, and a way of expanding the UK’s soft power. These are all legitimate goals, but half-heartedly attempting all of them over a decade is a surefire means of accomplishing relatively little.

Connected to this is our second challenge: everythingism. My friend Joe Hill described this aptly as “the belief that every government policy can be about every other government policy, and that there are no real costs to doing that”. This results in policymakers loading costs onto existing projects, at the expense of efficiency and prioritisation.

As the ATI’s original goals were so vague, it was a prime target. Even before the ATI was up and running, the government announced that it would also be a body responsible for allocating funding for fintech projects. It then had a data ethics group bolted onto it as a result of a 2016 select committee report. As one ex-insider put it, “there was never a superordinate goal”.

Finally, the perils of the UK government’s dependence on the country’s universities for research. The UK has historically channelled 80% of its non-business R&D through universities, versus 40-60% for many peer nations.

«

Back in December 2024 there was this story about “Staff at Britain’s AI institute in open revolt“. Everyone, including the staff, seems to think it hasn’t kept pace with the changes in AI in the past three years.
unique link to this extract

AI models miss disease in Black and female patients • Science

Rodrigo Pérez Ortega:

»

From programs designed to detect irregular heartbeats in electrocardiograms to software that tracks eye movements to diagnose autism in children, artificial intelligence (AI) is helping physicians fine-tune the care they provide patients. But for all the technology’s potential for automating tasks, a growing body of evidence also shows that AI can be prone to bias that disadvantages already vulnerable patients. A new study, published in Science Advances, adds to this work by testing one of the most cited AI models used to scan chest x-rays for diseases—and finding the model doesn’t accurately detect potentially life-threatening diseases in marginalized groups, including women and Black people.

These results “are interesting and timely,” says Kimberly Badal, a computational biologist at the University of California (UC), San Francisco, who was not involved in the new study. “We are at the point in history where we’re moving to deploy a lot of AI models into clinical care,” she says, but “we don’t really know” how they affect different groups of people.

The model used in the new study, called CheXzero, was developed in 2022 by a team at Stanford University using a data set of almost 400,000 chest x-rays of people from Boston with conditions such as pulmonary edema, an accumulation of fluids in the lungs. Researchers fed their model the x-ray images without any of the associated radiologist reports, which contained information about diagnoses. And yet, CheXzero was just as good as the radiologists in reading the diseases associated with each x-ray.

Given AI models’ tendencies for bias, Yuzhe Yang, a computer scientist at UC Los Angeles wanted to assess the Stanford team’s model for such biases. His team selected a subset of 666 x-ray images from the same data set that was used to train the model: the data set’s only images that also came with radiologists’ diagnoses and information about each patient’s age, sex, and race. The team then fed these images to CheXzero and compared the results against the radiologists’ diagnoses.

Compared with the patients’ doctors, the AI model more often failed to detect the presence of disease in Black patients or women, as well in those 40 years or younger. When the researchers looked at race and sex combined, Black women fell to the bottom, with the AI not detecting disease in half of them for conditions such as cardiomegaly, or enlargement of the heart.

«

Are we surprised? No we are not. The problem of using insufficient training data is a persistent one in AI.
unique link to this extract

smile like zuck

»

The secret to a perfect Zuck smile is his mouth and eyes. Focus on those.

Remember to keep the right amount of life in your eyes!

This site requires webcam access, but your face never leaves your computer.

«

There’s more explanation at this site. You can also do a “no smile” to “total Zuck smile” at this site, so you know what your target is. Remember, a Zuck smile isn’t like a human smile. It’s not meant to express happiness, empathy or any of that nonsense.
unique link to this extract

Agencies say Google reps are pressuring clients on AI ad tools • Digiday

Marty Swant:

»

Ad agencies have long fielded relentless pitches from Google Ads reps pushing new products in search of more money. But now, agencies of all sizes say the pressure is intensifying, with reps pushing harder to drive adoption of automated tools like Performance Max and generative AI features.

Google Ads sales reps are increasingly contacting agencies’ clients with advice that at times contradicts agency strategies — and in some cases mismanages campaigns — according to range of media agencies in the U.S. and U.K. Sources. They also say the tactics feel more aggressive and more inappropriate than in the past.

Many agencies say the efforts seem designed to sow confusion, discredit agencies and ultimately cut them out of the picture. For example, agencies claim that when they reject Google reps’ misaligned advice, the reps go around them — directly to clients — discrediting the agency by implying they don’t understand how Google Ads work. PPC professionals also shared similar concerns in a recent Reddit thread, where one user likened the issue to “being told your house needs paint by the guy who sells the paint and does the job.”

Agencies have been left agencies scrambling, said Ian Harris, founder of Agency Hackers, a U.K.-based community of indie agencies. To address the problem, Agency Hackers has created a new “Don’t Be Evil” campaign for agencies to share their experiences and frustrations about Google’s ad-sales tactics.

«

unique link to this extract

Reflecting on TikTok’s role in society as new ban deadline approaches • The New York Times

Brian Chen:

»

TikTok’s effectiveness at keeping people scrolling has been a topic of widespread concern among parents and academic researchers who wonder whether people could be considered addicted to the app, similar to video game addiction.

Studies on the topic are continuing and remain inconclusive. One, published last year and led by Christian Montag, a professor of cognitive and brain sciences at the University of Macau in China, examined TikTok overuse. Very few people in the study, which involved 378 participants of various ages, reported feeling addicted to TikTok.
Yet broadly speaking, the consensus from multiple studies on TikTok and other social media apps is that younger people are more likely to report feeling addicted, Dr. Montag said in an interview.

“I think children should not at all be on these platforms,” he said about TikTok and similar apps. People’s brains can take at least 20 years to mature and self regulate, he added.

A TikTok spokeswoman said the app included tools for people to manage their screen time, including a new setting for parents to block TikTok from working on their children’s phones during certain hours of the day.

…TikTok was banned in the first place because American government officials worry that ByteDance could share the data it has collected on its American users with the Chinese government for espionage purposes.

Those concerns culminated in a Supreme Court hearing in January, where the Biden administration made its case for banning the app, citing concerns that TikTok could create a new pathway for Chinese intelligence services to infiltrate American infrastructure. But officials did not present evidence that TikTok was connected to such threats.

TikTok has, however, been linked to smaller data scandals in the United States. TikTok confirmed in 2022 that four of its employees had been fired for using the app to snoop on several journalists in an effort to track down their sources.

«

At this stage it would be astonishing if Trump banned TikTok. Even the proposal of a sale of its US assets or control to an American company looks remote.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

Earlier this century Volvo had an anti-theft device that relied on your heartbeat. Why do you think it died? CC-licensed photo by Lars P. 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.

It’s Friday, so there’s another post due at the Social Warming Substack at about 0845 UK time.

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

OpenAI halts Studio Ghibli-style images trend • EWeek

Fiona Jackson:

»

If you’ve been wondering why your social media feeds have been awash with Studio Ghibli-style images this week, OpenAI’s new image generator is the answer. On Tuesday, the company embedded the multimodal tool into GPT-4o, and users have been transforming their photos into vibrant, whimsical scenes reminiscent of the Japanese animation studio behind “Spirited Away” and “My Neighbor Totoro.”

OpenAI’s new image generator allows you to upload an existing image to ChatGPT and simply ask it to change the art style in the accompanying text prompt. Even OpenAI’s CEO Sam Altman Ghibli-fied his X profile picture and playfully lamented about how his decade of work in AI has boiled down to people editing their photos into these images. 

However, the fun didn’t last long. The system card for GPT-4o’s native image generator states as of Thursday that OpenAI “added a refusal which triggers when a user attempts to generate an image in the style of a living artist.” OpenAI acknowledged that the fact its tool can emulate named artists’ styles “has raised important questions and concerns within the creative community.”

The same was written in DALL-E 3’s system card from 2023, so OpenAI is not making a wild U-turn on its position when it comes to AI art ownership and copyright. Nevertheless, there is an ongoing debate about whether AI companies violate copyright law by training their models on publicly available content. 

Ironically, Studio Ghibli co-founder Hayao Miyazaki is vocally anti-AI. While not specifically addressing the copyright debate, he referred to AI-generated artwork as “creepy stuff” and that he “would never wish to incorporate this technology into (his) work at all” in 2016, as shown in NHK’s “Never-Ending Man” documentary.

«

Earlier on Thursday there had been some chatter that the Ghibli-style stuff was OK (given it would have had to have come from vacuuming up entire films) because Japan has different copyright laws which do allow AI training. No confirmation of that has emerged, though, and the “living artist” block has come down.
unique link to this extract

Facebook launches Friends tab: a new feed that ditches algorithmic junk • The Verge

Emma Roth:

»

Facebook is trying to go back to its roots with a new “Friends” tab that filters out the algorithmic recommendations that have taken over its main feed. The new tab is rolling out in the US and Canada, and will show your friends’ stories, reels, posts, and birthdays.

In a blog post announcing the feature, the company says, “Over the years, Facebook evolved to meet changing needs and created best-in-class experiences across Groups, Video, Marketplace and more, but the magic of friends has fallen away.”

Facebook previously housed friend requests and friend suggestions within the Friends tab, but now it’s been replaced by a dedicated friends-only feed. I’m not sure I’ll find many recent posts there, but it might be a reprieve from the sponsored content and random group posts that have drowned out all other content on my home feed.

«

This is interesting: essentially it’s accepting that all the AI slop which has taken over the algorithmic feed is a big turnoff for a lot of people, and so this redesign is trying to make that go away, or at least offer an option that makes that vanish. The first sign that slop is properly detrimental to user experience and, more importantly, to time spent on the app, which is what Meta really notices.
unique link to this extract

You need to use Signal’s nickname feature • 404 Media

Joseph Cox:

»

According to screenshots of the chats and the group chat’s members published by The Atlantic on Wednesday, the outlet’s editor Jeffrey Goldberg used the display name “JG” on Signal. He also said in the original article that he displayed as JG. Presumably National Security Adviser Michael Waltz, who accidentally added Goldberg, added the wrong JG. This is a big, big mistake obviously.

But there is a somewhat overlooked setting inside Signal that can ensure you don’t make the same mistake. It’s the nickname feature.

…You can add your own nickname to a Signal contact by clicking on the person’s profile picture in a chat with them then clicking “Nickname.” Signal says “Nicknames & notes are stored with Signal and end-to-end encrypted. They are only visible to you.” So, you can add a nickname to a Jason saying “co-founder,” or maybe “national security adviser,” and no one else is going to see it. Just you. When you’re trying to make a group chat, perhaps.

«

Simple advice. One could probably charge a very high consultancy fee for doing this for, say, a government struggling with catastrophic information leakage.
unique link to this extract

Volvo’s heartbeat-sensing anti-carjacking technology was Y2K safety at its weirdest • Hagerty Media

Benjamin Hunting:

»

Every once in a while, an automaker debuts a feature that is truly unique, standing apart from everything else on the market. In some cases, these novel technologies serve as the thin end of a wedge that opens up a whole new world of vehicle development, with rivals quickly co-opting and branding their own version of the innovation.

In others, you end up standing in a parking lot in the year 2006, watching “Ray the Burglar” be apprehended by a security system that’s been specifically designed to detect his heartbeat as he hides in the backseat, like some digital combination of Sherlock Holmes and Robocop.

Such was the media introduction given to the most intriguing capability of Volvo’s Personal Car Communicator, a thin slab of plastic that initially served as a super-fob for the flagship S80 sedan. Billed as a bulwark against the threat of kidnappers, carjackers, and bloodthirsty murderers who were apparently lurking around every corner of a Volvo owner’s action-packed existence, the heartbeat sensor was an entirely singular high-tech achievement that no other car company ever came close to implementing in any of their vehicles.

What prompted Volvo, a brand known for sensibility and safety above all else, to infer that its customers were living a lifestyle that required this level of electronic overwatch to keep them from becoming just another statistic on a police blotter? And why didn’t the heartbeat sensor catch on long-term in the industry—or even with Volvo’s modern-day lineup? It turns out that the answers to those questions are tied directly to the times that produced this utterly unusual, and unforgettable, feature.

«

It’s a wild ride of a story. Odd how this technology hasn’t continued, eh. (Might there be people out there still using it?)
unique link to this extract

Six unsettling thoughts Google’s former CEO has about artificial intelligence • NPR

Steve Inskeep:

»

Eric Schmidt, the former Google CEO, is thinking about artificial intelligence – how it interacts with humans, and how it may reshape democracy. Or replace it.

Schmidt coauthored a new book, Genesis, with former Microsoft executive Craig Mundie and the late Henry Kissinger, who died in 2023 about a year before the book’s publication.

Kissinger, Schmidt says, had been thinking about the nature of reality “since before we were born,” and used some of his final years exploring how technology might warp our understanding of reality.

Genesis includes a story from history: the Spanish conquistadors who invaded present-day Mexico in 1519. The ruling Aztecs seem to have mistaken the newcomers for gods. Their emperor first met them, took their advice, and then became their hostage before the conquistadors simply took over.

That’s the unsettling start to a chapter that asks if AI might conquer us.

«

Those six thoughts:
• AI will be available to almost anyone
• AI can be a tool for demagogues
• People are interacting with tech they don’t fully grasp
• Tech leaders may not grasp the implications either
• People might allow themselves to be governed by AI
• The recent presidential inauguration showed a concentration of [tech] power.

Plenty of expansion in the article.
unique link to this extract

Turkish female coders leave tech industry over sexism • Rest of World

Kaya Genç:

»

Turkish women who become software engineers generally expect a secure career that can vault them to a better position in society. Instead, they’ve found they are a mistreated and undervalued minority. In Turkey’s tech industry, less than a quarter of workers are women, according to various surveys — lower than the global average of 28%. Female IT engineers have complained about male-dominated workplaces, with 70% saying it negatively affected their careers, according to a 2018 report by consulting firm Deloitte Turkey and the Information Technology Industrialists Association. 

At  the same time, the government is pushing for women to prioritize their families over work. Turkish President Recep Tayyip Erdoğan has said that women should have at least three children. In 2021, Turkey withdrew from the Istanbul Convention, which prevents and fights violence against women. 

Rather than face the turbulent social conditions, many young IT professionals, particularly women, are choosing to leave the country, according to Füsun Sarp Nebil, a tech entrepreneur and CEO of the technology journalism website Turk Internet.

“The increasing pressure in recent years, the obstructions on social life, like the cancellation of concerts, the lack of freedom of expression, the environment of violence … disturbs young people and especially women,” Nebil told Rest of World.

“Salaries are at the bottom of the list of reasons for leaving Turkey,” she said.

«

Turkey withdrawing from the Istanbul Convention is quite the irony.
unique link to this extract

California now has 48% more EV chargers than gasoline nozzles in the state • Governor of California

California Governor’s office:

»

Governor Gavin Newsom has announced that California has reached 178,549 public and shared private electric vehicle (EV) chargers installed statewide. 

In a significant milestone for the state, California now has 48% more public and shared private EV chargers than the number of gasoline nozzles. The California Energy Commission (CEC) estimates there are about 120,000 gas nozzles in the state, compared to 178,000 public and shared private chargers.

Governor Gavin Newsom said: “As the federal government works to make it harder for you to charge your electric car, California is doing the opposite. We now have nearly 50% more chargers than gas nozzles in the state, meaning you have more options than ever to charge your vehicle.

“We’re embracing our clean car future and providing consumers more choices – no matter what ‘big government’ mandates come out of Washington.”

Out of the more than 178,000 EV chargers installed in the state, over 162,000 are Level 2 chargers, and nearly 17,000 are fast chargers. In addition to the public network, the CEC estimates that more than 700,000 Level 2 chargers are installed statewide in single-family homes.

“The California EV driver experience is getting better by the day,” said CEC Chair David Hochschild. “The state continues to invest in EV infrastructure, with particular emphasis in hard-to-reach areas, making these vehicles an easy choice for new car buyers.”

California dominates in zero-emission vehicle (ZEV) infrastructure efforts, dedicating billions to support clean transportation goals. With more EVs on the road every day, consumers are responding to the state’s efforts to build a bigger, better, and more reliable charging network. 

«

California has about 13.2m vehicles, of which at December 2023 – the last data update – a total of 1.8m are electric. Probably moved on a fair bit since then.
unique link to this extract

Carbon credit auditors suspended in sham rice-farming offsets • Climate Change News

Matteo Civillini:

»

Carbon credit registry Verra has suspended activities by four auditors related to carbon credit projects they vetted in China which claimed bogus emission reductions.

In an unprecedented move, TÜV Nord, China Classification Society Certification Company, China Quality Certification Center and CTI Certification will be prevented from auditing agriculture and forestry offsetting schemes on Verra’s registry. For German certification giant TÜV Nord, the measures will only apply to its operations in China. It is the first time Verra has taken such measures.

The auditors certified the activities of 37 programmes that aimed to slash planet-heating methane gas releases from rice fields across China, resulting in the generation of millions of carbon offsets. But Verra revoked the projects in August 2024 after a 17-month review found a string of integrity failures that the auditors had failed to identify.

Before this week’s suspension, Climate Home previously reported on ten of these projects closely linked to energy company Shell and revealed evidence raising serious doubts over whether any emission-cutting activities had been carried out on the ground at all.

Nearly two million worthless carbon credits produced by the projects – and partly used to offset emissions from Shell’s gas business – still need to be compensated.

«

So not only are the carbon credit schemes often a scam, the auditors do a poor job.
unique link to this extract

Paralysed man stands again after receiving “reprogrammed” stem cells • Nature

Smriti Mallapaty:

»

A paralysed man can stand on his own after receiving an injection of neural stem cells to treat his spinal cord injury. The Japanese man was one of four individuals in a first-of-its-kind trial that used reprogrammed stem cells to treat people who are paralysed.

Another man can now move his arms and legs following the treatment, but the two others did not show substantial improvements. The trial was run by Hideyuki Okano, a stem-cell scientist at Keio University in Tokyo, and his colleagues.

The results, which were announced at a press conference on 21 March and have not yet been peer reviewed, suggest that the treatment is safe, say researchers.

“That’s a great positive outcome. It’s very exciting for the field,” says James St John, a translational neuroscientist at Griffith University in the Gold Coast, Australia.

Previous trials using other types of stem cell have also demonstrated that the therapy is safe, but have so far shown mixed results. “Nothing’s really worked so far,” says St John.

Larger trials will be needed to establish whether the improvements observed in the two individuals in the current study were a result of the treatment. It’s possible the patients experienced a natural recovery, says St John.

In 2019, roughly 0.9 million people globally experienced a spinal cord injury, and some 20 million people were living with the condition1.

«

Wait for the peer review! But being able to use arms and legs is quite a peer review in itself. Perhaps there’s a breakthrough just coming here.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

Image generation from OpenAI took another step forward on Wednesday, and people discovered it could create Studio Ghibli-style images. CC-licensed photo by Choo Yut Shing 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. Cartoonish. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

OpenAI upgrades image generation and rolls it out in ChatGPT and Sora • The Verge

Kylie Robison:

»

Users can now use GPT-4o to generate images within ChatGPT itself.

This initial release focuses solely on image creation and will be available across ChatGPT Plus, Pro, Team, and Free subscription tiers. The free tier’s usage limit is the same as DALL-E, spokesperson Taya Christianson told The Verge, but added that they “didn’t have a specific number to share” and ”these may change over time based on demand.“ Per the ChatGPT FAQ, free users were previously able to generate “three images per day with DALL·E 3.” As for the fate of DALL-E, Christianson said “fans” will “still have access via a custom GPT.”

“This model is a step change above previous models,” research lead is Gabriel Goh told The Verge, adding that the team used the GPT-4o “omnimodal” — or a model that can generate any kind of data like text, image, audio, and video — foundation for this feature.

Some of the improvements Goh noted include “binding,” which refers to how well AI image generators maintain correct relationships between attributes and objects; a model with poor binding, for instance, might get a prompt for a blue star plus a red triangle and create a red star and no triangle. Most image models struggle with this, Goh said, often mixing up colors and shapes when asked to render multiple items — typically around 5 to 8. He says this new image generation tool can correctly bind attributes for 15 to 20 objects without confusion, representing a significant improvement in accuracy and reliability.

«

And with all that said, what did people discover? That you can give it an image and it will create a version that looks like it was drawn by Studio Ghibli. Consequently social media has been flooded with versions of memes redrawn in that manner. (This is a wonderful thread. Pick your favourite, and note too which faces it makes angry and which sad.)

The question of quite how OpenAI absorbed a ton of Studio Ghibli content remains unanswered.
unique link to this extract

A breakthrough moment: researchers discover new class of antibiotics • Phys.org

»

The last time a new class of antibiotics reached the market was nearly three decades ago—but that could soon change, thanks to a discovery by researchers at McMaster University.

A team led by researcher Gerry Wright has identified a strong candidate to challenge even some of the most drug-resistant bacteria on the planet: a new molecule called lariocidin. The findings were published in the journal Nature on March 26, 2025.

The discovery of the all-new class of antibiotics responds to a critical need for new antimicrobial medicines, as bacteria and other microorganisms evolve new ways to withstand existing drugs. This phenomenon is called antimicrobial resistance—or AMR—and it’s one of the top global public health threats, according to the World Health Organization.

“Our old drugs are becoming less and less effective as bacteria become more and more resistant to them,” explains Wright, a professor in McMaster’s Department of Biochemistry and Biomedical Sciences and a researcher at the university’s Michael G. DeGroote Institute for Infectious Disease Research. “About 4.5 million people die every year due to antibiotic-resistant infections, and it’s only getting worse.”

Wright and his team found that the new molecule, a lasso peptide, holds great promise as an early drug lead because it attacks bacteria in a way that’s different from other antibiotics. Lariocidin binds directly to a bacterium’s protein synthesis machinery in a completely new way, inhibiting its ability to grow and survive.

“This is a new molecule with a new mode of action,” Wright says. “It’s a big leap forward for us.”

«

If this works, it’s a huge discovery. And where did it come from? A soil sample in a backyard.
unique link to this extract

Solar panel windows that could turn whole buildings into power plants smash electricity record • Euronews

Pascale Davies:

»

Researchers in Denmark have set a new world record in efficiency for converting sunlight into electricity by using new windows that allow light to pass through while simultaneously generating power. 

The transparent solar cell technology could provide a breakthrough for renewable energy by transforming skyscrapers and offices into power plants, using their windows to become solar panels. 

The innovation from the CitySolar project could also help Europe meet its ambitions to make all new buildings nearly zero energy and fully decarbonise the European building sector by 2050.

The researchers from the University of Southern Denmark combined organic solar cells with the material perovskite, which saw an efficiency of 12.3%, which is on par with commercial solar cells. 

The international team say the panels also have a transparency of 30%.
 
Until now, transparent solar windows have not been able to absorb enough energy to be able to generate the amount of electricity needed for a building and the panels have previously not been transparent enough for use.  The CitySolar project says it has now overcome these issues.  

“Transparent solar cells could be the next big step in building integrated energy solutions,” said Morten Madsen, a professor from the University of Southern Denmark who was one of the key researchers behind the breakthrough. “The large glass facades found in modern office buildings can now be used for energy production without requiring additional space or special structural changes… This represents a massive market opportunity”.

Furthermore, Madsen said that the two materials used in the cells are highly affordable and could be scaled for commercial deployment.

When added to the organic solar cell, the perovskite layer absorbs near-ultraviolet light and the cell absorbs near-infrared light. 

«

No word on price or how much energy is generated, but those are decent conversion figures. They’d only be truly useful on south-facing (north, in the southern hemisphere) areas though?
unique link to this extract

Google makes Android development private, but will continue open source releases • Ars Technica

Ryan Whitwam:

»

Google is planning a major change to the way it develops new versions of the Android operating system. Since the beginning, large swaths of the software have been developed in public-facing channels, but that will no longer be the case. This does not mean Android is shedding its open source roots, but the process won’t be as transparent.

Google has confirmed to Android Authority that all Android development work going forward will take place in Google’s internal branch. This is a shift from the way Google has worked on Android in the past, which featured frequent updates to the public AOSP branch. Anyone can access AOSP, but the internal branches are only available to Google and companies with a Google Mobile Services (GMS) license, like Samsung, Motorola, and others.

According to the company, it is making this change to simplify things, building on a recent change to trunk-based development. As Google works on both public and private branches of Android, the two fall out of sync with respect to features and API support. This forces Google to tediously merge the branches for every release. By focusing on the internal branch, Google claims it can streamline releases and make life easier for everyone.

When new versions of Android are done, Google says it will continue to publish the source code in AOSP as always. Supposedly, this will allow developers to focus on supporting their apps without keeping track of pending changes to the platform in AOSP. Licensed OEMs, meanwhile, can just focus on the lively internal branch as they work on devices that can take a year or more to launch.

…This change to private development doesn’t come out of the blue—Android feels less open today than it did in the early days. For example, Google has been moving Android features from AOSP into closed source packages for years. This gives Google greater control over the platform while also making it easier to update core components without a full OS update.

Currently, most Android development takes place in the internal branch, but a handful of components, like Bluetooth and the kernel, are developed in the open. They’ll be moving to internal under the new system.

«

The most important part anyway has always been Google Play, which gets updated much more regularly, and closed source.
unique link to this extract

The new Substack universe • NY Mag

Charlotte Klein:

»

Substack today has all of the functionalities of a social platform, allowing proprietors to engage with both subscribers (via the Chat feature) or the broader Substack universe in the Twitter-esque Notes feed. Writers I spoke to mentioned that for all of their reluctance to engage with the Notes feature, they see growth when they do. More than 50% of all subscriptions and 30% of paid subscriptions on the platform come directly from the Substack network. There’s been a broader shift toward multimedia content: Over half of the 250 highest-revenue creators were using audio and video in April 2024, a number that had surged to 82% by February 2025.

Substack initially rolled out livestreaming in September as a feature available only to publishers with a lot of subscribers; in January, it made live video available to everyone. There’s a collaboration feature that allows you to co-host streams with up to three people, “essentially what is one step up from a public FaceTime call,” says Substack co-founder Hamish McKenzie. “It lowers the barrier to entry to saying something important in public; it’s also kind of fun.” Acosta has kept his videos free for now, but others are offering the feature only to paid subscribers, as Tina Brown has for recent conversations with Maureen Dowd and Janice Min.

“Email is something that we will absolutely never abandon,” said McKenzie, noting it’s “important as a guarantor of that direct owned relationship that the writers and creators can have with their audience.” But by limiting yourself to just email, “you’re tying one of your hands behind your back.” When someone like Acosta goes live on Substack, they announce that they’re doing so to their entire email list, who can then jump on and watch. At the end of the stream, they are presented with the full video, which they can then blast out to their Substack followers as well as post on other social platforms.

But even those who’ve resisted new features and kept their newsletters largely text based continue to see an uptick, like journalist Max Read, who writes the twice-weekly Read Max newsletter. “To the extent there’s been a Trump bump in the media, it’s all going to Substackers. I cover tech, not really politics. But I saw incredibly fast growth — it seems to finally be plateauing — between late October/the election last year and two or three weeks ago,” he said. “I was growing twice as fast as I’d been in the months before that.”

Sports journalism has basically migrated to the platform, and food media seems not far behind — if not already there.

«

The inevitable next step is going to be someone or something which aggregates various Substacks into a sort of “daily Substack”. They could charge for it. You could even print it out and bind it and read it over breakfast.
unique link to this extract

Exclusive: DOGE staffer ‘Big Balls’ provided tech support to cybercrime ring, records show • Reuters

Raphael Satter:

»

The best-known member of Elon Musk’s U.S. DOGE Service team of technologists once provided support to a cybercrime gang that bragged about trafficking in stolen data and cyberstalking an FBI agent, according to digital records reviewed by Reuters.

Edward Coristine is among the most visible members of the DOGE effort that has been given sweeping access to official networks as it attempts to radically downsize the U.S. government.

Past reporting, opens new tab had focused on his youth – he is 19 – and his chosen nickname of “bigballs,” which became a pop culture punchline, opens new tab. Musk has championed the teen on his social media site X, telling his followers, opens new tab last month that “Big Balls is awesome.”

Beginning around 2022, while still in high school, Coristine ran a company called DiamondCDN, opens new tab that provided network services, according to corporate and digital records reviewed by Reuters and interviews with half a dozen former associates. Among its users was a website run by a ring of cybercriminals operating under the name “EGodly,” according to digital records preserved by the internet intelligence firm DomainTools and the online cybersecurity tool Any.Run.

The details of Coristine’s connection to EGodly have not been previously reported.

On Feb. 15, 2023, EGodly thanked Coristine’s company for its assistance in a post on the Telegram messaging app.
“We extend our gratitude to our valued partners DiamondCDN for generously providing us with their amazing DDoS protection and caching systems, which allow us to securely host and safeguard our website,” the message said.

The digital records reviewed by Reuters showed the EGodly website, dataleak.fun, was tied to internet protocol addresses registered to DiamondCDN and other Coristine-owned entities between October 2022 and June 2023, and that some users attempting to access the site around that time would hit a DiamondCDN “Security check.”

Coristine did not return messages seeking comment.

«

I think the administration is a bit backed up with firing people, can this one wait a day or two? But also: what an incredible lack of background checking. Plus anyone who was hit by a DDOS caused by this ring who lost money could sue him as an accessory to crime.
unique link to this extract

Oops, DOGE did it again • The Bulwark

William Kristol:

»

Late last month, Elon Musk set off a firestorm when he announced that government employees would be required to email five bullet points describing their accomplishments from the prior week—and that, if they didn’t, it would be considered an act of resignation. Musk insisted this was management 101 and that people were being entitled or, worse, lazy if they couldn’t handle it. Perhaps they were dead.

Four weeks later, the “five bullets” program appears to be a bust, in part because Musk’s DOGE team didn’t set up an email system that could handle the incoming. Four government employees from different agencies tell The Bulwark that they or their colleagues recently attempted to send in five-bullet emails only to receive a bounce-back response that the OPM mailbox was full.

“If you sent your message to OPM and received this response, then please save your email to OPM for your records, but otherwise there is nothing more you need to do today,” Securities and Exchange Commission COO Ken Johnson told employees in an email this week.

“We are aware that emails to HR@opm.gov are being returned as undeliverable. Please send your weekly accomplishments to HR10@opm.gov and cc your supervisor,” read an email from a leadership official at Health and Human Services.

«

It’s just Pelion upon Ossa. Incompetence that can’t think two steps ahead.
unique link to this extract

Exclusive: DOD has deployed Signal on government devices, overriding their own policy • On Democracy

FPWellman:

»

A high level information security source inside the Department of Defense has informed me that a month ago they were ordered by political appointees to ignore information security regulations and install Signal on government phones for senior leaders.

This story likely confirms that Signal is a primary means of communications for Trump Administration senior leaders in direct violation of the Presidential Records Act, the Espionage Act, and numerous national security regulations. It appears that much of our national security communications are vulnerable to foreign intelligence agencies to access at the highest levels of our government.

With the explosive news that senior Trump Administration officials had used a Signal chat group that included Atlantic editor-in-chief Jeffrey Goldberg to plan attacks on Yemen questions have been raised about how prevalent the use of the off the shelf civilian app is being used. I can confirm that it has become a primary means of communication.

An anonymous senior level source in the Pentagon’s information technology field reached out to tell me that upon the arrival of newly installed senior military officials in February they were asked to install Signal on their government phones. There had already been instances where newly appointed Trump officials attempted to bring their personal phones in secure areas.

They relented on leaving their phones outside the classified spaces but demanded that CIO install Signal on their government devices.

«

Tulsi Gabbard, the alleged Director of National Intelligence – has someone so stupid ever held that role – told the US House Intelligence Committee on Wednesday that Signal comes “preinstalled” on government devices. Sure, if you tell them to, it does. (Thanks Gregory B for the link.)
unique link to this extract

Here are the attack plans that Trump’s advisers shared on Signal • The Atlantic

Jeffrey Goldberg and Shane Harris:

»

So, about that Signal chat.

On Monday, shortly after we published a story about a massive Trump-administration security breach, a reporter asked the secretary of defense, Pete Hegseth, why he had shared plans about a forthcoming attack on Yemen on the Signal messaging app. He answered, “Nobody was texting war plans. And that’s all I have to say about that.”

At a Senate hearing yesterday, the director of national intelligence, Tulsi Gabbard, and the director of the Central Intelligence Agency, John Ratcliffe, were both asked about the Signal chat, to which Jeffrey Goldberg, the editor in chief of The Atlantic, was inadvertently invited by National Security Adviser Michael Waltz. “There was no classified material that was shared in that Signal group,” Gabbard told members of the Senate Intelligence Committee.

Ratcliffe said much the same: “My communications, to be clear, in the Signal message group were entirely permissible and lawful and did not include classified information.”

President Donald Trump, asked yesterday afternoon about the same matter, said, “It wasn’t classified information.”

These statements presented us with a dilemma.

«

Although Mike Waltz (who, the screenshots clearly show, added Goldberg to the group) might have set messages to disappear after one week, he forgot that screenshots, like diamonds, are forever. And there are some real pearls in here. I wonder how long it will take from this publication to the adviser saying “well, if we put secret information into an unclassified app, then it’s not classified any more, so no foul.”
unique link to this extract

Hegseth, Waltz, Gabbard: private data and passwords of senior U.S. security officials found online • DER SPIEGEL

Fidelius Schmid, Friederike Röhreke, Roman Lehberger, Roman Höfner, Jörg Diehl and Patrick Beuth:

»

Private contact details of the most important security advisers to U.S. President Donald Trump can be found on the internet. DER SPIEGEL reporters were able to find mobile phone numbers, email addresses and even some passwords belonging to the top officials.

To do so, the reporters used commercial people search engines along with hacked customer data that has been published on the web. Those affected by the leaks include National Security Adviser Mike Waltz, Director of National Intelligence Tulsi Gabbard and Secretary of Defense Pete Hegseth.

Most of these numbers and email addresses are apparently still in use, with some of them linked to profiles on social media platforms like Instagram and LinkedIn. They were used to create Dropbox accounts and profiles in apps that track running data. There are also WhatsApp profiles for the respective phone numbers and even Signal accounts in some cases.

As such, the reporting has revealed an additional grave, previously unknown security breach at the highest levels in Washington. Hostile intelligence services could use this publicly available data to hack the communications of those affected by installing spyware on their devices. It is thus conceivable that foreign agents were privy to the Signal chat group in which Gabbard, Waltz and Hegseth discussed a military strike.

It remains unclear, however, whether this extremely problematic chat was conducted using Signal accounts linked to the private telephone numbers of the officials involved. Tulsi Gabbard has declined to comment. DER SPIEGEL reporting has demonstrated, though, that privately used and publicly accessible telephone numbers belonging to her and Waltz are, in fact, linked to Signal accounts.

«

Words fail.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

Excited news stories say a network of tunnels has been found beneath the pyramids. In fact: nope. CC-licensed photo by Vincent Brown 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 11 links for you. Plentiful. 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.I. saved his life by discovering new uses for old drugs • The New York Times

Kate Morgan:

»

A little over a year ago, Joseph Coates was told there was only one thing left to decide. Did he want to die at home, or in the hospital?

Coates, then 37 and living in Renton, Wash., was barely conscious. For months, he had been battling a rare blood disorder called POEMS syndrome, which had left him with numb hands and feet, an enlarged heart and failing kidneys. Every few days, doctors needed to drain liters of fluid from his abdomen. He became too sick to receive a stem cell transplant — one of the only treatments that could have put him into remission.

“I gave up,” he said. “I just thought the end was inevitable.”

But Coates’s girlfriend, Tara Theobald, wasn’t ready to quit. So she sent an email begging for help to a doctor in Philadelphia named David Fajgenbaum, whom the couple met a year earlier at a rare disease summit.

By the next morning, Dr. Fajgenbaum had replied, suggesting an unconventional combination of chemotherapy, immunotherapy and steroids previously untested as a treatment for Coates’s disorder.
Within a week, Coates was responding to treatment. In four months, he was healthy enough for a stem cell transplant. Today, he’s in remission.

The lifesaving drug regimen wasn’t thought up by the doctor, or any person. It had been spit out by an artificial intelligence model.

In labs around the world, scientists are using A.I. to search among existing medicines for treatments that work for rare diseases. Drug repurposing, as it’s called, is not new, but the use of machine learning is speeding up the process — and could expand the treatment possibilities for people with rare diseases and few options.

Thanks to versions of the technology developed by Dr. Fajgenbaum’s team at the University of Pennsylvania and elsewhere, drugs are being quickly repurposed for conditions including rare and aggressive cancers, fatal inflammatory disorders and complex neurological conditions. And often, they’re working.

The handful of success stories so far have led researchers to ask: how many other cures are hiding in plain sight?

«

OK, but then again, there’s only a limited number of rare diseases. Sometimes what’s needed is known, but drugs companies can’t be persuaded to make the treatment for a reasonable price.
unique link to this extract

Research: streaming video services struggling for identity • Advanced Television

Hub Entertainment Research:

»

Many viewers simply can’t identify where they can watch “signature” programmes. A clutter of original shows that could easily play across different services has made it hard for viewers to find specific shows.

While over half (58%) of consumers know that Stranger Things is on Netflix, less than half of consumers could correctly place where to watch signature shows like Game of Thrones (on Max), The Bear (on Hulu/Disney+) and Ted Lasso (on Apple TV+), among others.

In the blurry landscape of scripted content, live sports events have stood out as a key driver for new sign-ups and strengthening subscriber retention.

Netflix’s strong push into live sports with Christmas NFL games paid off well with big subscriber gains – nearly half (49%) of people agree that it increases their interest in both signing up for and keeping the service.

“Services that lean into broad-appeal scripted programmes may not be enough for viewers who struggle to identify what makes services distinct from one another,” commented Jason Platt Zolov, Senior Consultant for Hub. “Emphasising more brand-defining features and value drivers beyond just exclusive originals could have more upside for streamers looking to improve viewer loyalty.”

«

So the streaming services are turning into channels which have scripted entertainment and distinguish themselves through live sports? Does this seem familiar?
unique link to this extract

A sneaky phish just grabbed my Mailchimp mailing list • Troy Hunt

Troy Hunt:

»

You know when you’re really jet lagged and really tired and the cogs in your head are just moving that little bit too slow? That’s me right now, and the penny has just dropped that a Mailchimp phish has grabbed my credentials, logged into my account and exported the mailing list for this blog. I’m deliberately keeping this post very succinct to ensure the message goes out to my impacted subscribers ASAP, then I’ll update the post with more details. But as a quick summary, I woke up in London this morning to the following message [a screenshot suggesting it’s from Mailchimp and that a “spam complaint” means “sending privileges” are restricted so you have to log in.. to the phishing site.]

I went to the link which is on mailchimp-sso.com and entered my credentials which – crucially – did not auto-complete from 1Password. I then entered the OTP [one-time password] and the page hung. Moments later, the penny dropped, and I logged onto the official website, which Mailchimp confirmed via a notification email which showed my London IP address:

I immediately changed my password, but not before I got an alert about my mailing list being exported from an IP address in New York.

And, moments after that, the login alert from the same IP.

This was obviously highly automated and designed to immediately export the list before the victim could take preventative measures.

…I’m enormously frustrated with myself for having fallen for this, and I apologise to anyone on that list. Obviously, watch out for spam or further phishes and check back here or via the social channels in the nav bar above for more. Ironically, I’m in London visiting government partners, and I spent a couple of hours with the National Cyber Security Centre yesterday talking about how we can better promote passkeys, in part due to their phishing-resistant nature. 🤦‍♂️

«

Happens to the best of us; and here’s the proof.
unique link to this extract

Apple can’t intervene in battle over Google search, court confirms • Mediapost

Wendy Davis:

»

The ruling, issued Friday by the D.C. Circuit Court of Appeals, upholds a decision by U.S. District Court Judge Amit Mehta in Washington, D.C., who said in January that Apple could submit written testimony and file friend-of-the-court briefs, but couldn’t present live testimony or cross-examine witnesses at the hearing, slated to begin in April.

The appellate ruling comes in an antitrust battle dating to 2020, when the U.S. Department of Justice and a coalition of states alleged that Google monopolized search.

In August, after conducting a trial, U.S. District Court Judge Amit Mehta in Washington, D.C. ruled that Google violated antitrust law by arranging to serve as the default search engine on browsers operated by Apple and Mozilla, as well as on Android devices.

The federal government in October proposed numerous remedies to Google’s monopoly, including that the company end a long-running revenue-sharing partnership with Apple. That deal involves Google serving as the default search engine for Safari, and paying Apple around 36% of search revenue for queries originating on Safari.

In late December, Apple sought to intervene in the hearing over remedies, arguing that it’s entitled to fully participate because it has a stake in the outcome of the case.

Mehta rejected that request as untimely. 

Apple appealed, arguing that if it couldn’t fully participate in the hearing, it would “be a mere spectator” in a fight that could affect its multi-billion dollar partnership with Google. That deal resulted in an estimated $20bn payment from Google to Apple in 2022 alone.

«

Pretty brutal to tell lawyers they’ve been too slow when they thought they were being methodical. But Tim Cook will be gazing at the spreadsheets with a big $20bn hole in them, wondering how to fill that gap.
unique link to this extract

Signal founder: don’t be fooled by WhatsApp’s marketing fluff • Cybernews

Anton Mous:

»

Last month, [Signal founder Meredith] Whittaker lashed out at WhatsApp’s data collection practices, saying that the messaging app collects too much sensitive metadata.

“It tells you exactly who you’re communicating with, at what time, how often, and where you are. You can derive so much from that. WhatsApp can link that information to Facebook, to Instagram and to payment data that they could buy into. Signal simply doesn’t have all that data,” she said.

Her comment didn’t go unnoticed. Will Cathcart, WhatsApp’s head, spoke to a handful of Dutch journalists last week and told them false rumors are circulating regarding WhatsApp’s security and privacy.

“We strongly believe in private communication,” he said, adding that WhatsApp uses the same security protocol as Signal. In addition, WhatsApp doesn’t keep track of whom and when people communicate, and location data and information about a user’s contact aren’t shared with other companies.

In a statement published on Monday, Whittaker says WhatsApp is making a mockery of things.

“In some ways, we’re amused to see WhatsApp stretching the limits of reality to claim that they are just like Signal – we take it as a bit of a compliment when a massive big tech platform strives to meet the bar we set and be cool like us. But the stakes of such marketing fluff are high, so we need to set the record straight,” she wrote.

Whittaker acknowledges that WhatsApp licenses Signal’s end-to-end encryption technology. Nevertheless, a lot of personal and intimate information isn’t protected. According to Signal’s president, this involves users’ location data, contact lists, when they send someone a message, when they stop, what users are in their group chats, their profile picture, and much more.

“These differences may be marketing gloss to Meta, but to us, they’re fundamental life or death issues that the public deserves to understand so they can make an informed choice,” Whittaker concludes.

«

unique link to this extract

Days after the Signal leak, the Pentagon warned the app was the target of hackers • NPR

Quil Lawrence and Tom Bowman:

»

Several days after top national security officials accidentally included [on March 11] a reporter in a Signal chat about bombing Houthi sites in Yemen, a Pentagon-wide advisory warned against using the messaging app, even for unclassified information.

“A vulnerability has been identified in the Signal Messenger Application,” begins the department-wide email, dated March 18 and obtained by NPR.

The memo continues, “Russian professional hacking groups are employing the ‘linked devices’ features to spy on encrypted conversations.” It notes that Google has identified Russian hacking groups that are “targeting Signal Messenger to spy on persons of interest.”

Moreover, a memo in 2023, obtained by NPR, warned of using Signal for any nonpublic official information.

A Signal spokesman said the Pentagon memo is not about the messaging app’s level of security, but rather that users of the service should be aware of what are known as “phishing attacks.” That’s when hackers try to gain access to sensitive information through impersonation or other deceptive tricks.

“Once we learned that Signal users were being targeted and how they were being targeted, we introduced additional safeguards and in-app warnings to help protect people from falling victim to phishing attacks. This work was completed months ago,” said Signal spokesman Jun Harada.

The March 18, 2025, Pentagon memo adds, “Please note: third party messaging apps (e.g. Signal) are permitted by policy for unclassified accountability/recall exercises but are NOT approved to process or store nonpublic unclassified information.”

«

The whole Signal debacle is top to bottom illegal by all those in the administration, but the fact they were using an app which had a weakness on their personal phones (official phones can’t even download Signal) is just totally predictable.
unique link to this extract

Notification summary miscues • One Foot Tsunami

Paul Kafasis:

»

Since they were first enabled last year, I have frequently found Apple Intelligence’s notification summaries for emails to be something less than helpful. Here are some I spotted in just the past few days.

«

These are all pretty egregious. Understandable – you can see how the machine processed the data to make the mistake – but the problem is, can you ever get rid of them? As I explained to some friends this evening, the trouble with LLMs is that they’re not deterministic, like computers as we usually think of them, but probabilistic: their output is inherently unpredictable. And that is a problem.
unique link to this extract

Dubai’s drone show boom is creating jobs, despite instability • Rest of World

Amar Diwakar:

»

On a chilly night in January, Ajay Sreekumar stood outside Dubai’s Museum of the Future and craned his head up to the sky. He watched a swarm of 600 drones whirl some 400 feet above in the dark, humming as they formed an LED-illuminated portrait of the city’s ruler, Sheikh Mohammed bin Rashid Al Maktoum. Over the next nine minutes, the drones transitioned seamlessly between multicolored and animated 3D images, including a spinning globe and a scientist inspecting the DNA double helix.

Sreekumar was not only a spectator — he was also the spectacle’s lead designer. The drone show was part of the closing ceremony for the World Health Expo Dubai, and the company Sreekumar worked for, Skyvertise, had organized it. When the show concluded and the last drone touched down, Sreekumar felt a glow of satisfaction. His team had spent the previous two days designing, programming, and rehearsing the show, and it had gone off without a hitch, he told Rest of World. 

Part of the thrill was seeing his work as a 3D artist come to life immediately. “You don’t have to wait months to see the product,” as with his former job in film and TV, he said. “Your portfolio can be seen in the sky in real time.”

Two years ago, Sreekumar was working at an Abu Dhabi media studio when he came across a job listing for Skyvertise. He had no experience with drones, but his 3D design skills were transferable. He soon found himself doing everything from operating the drones to changing the batteries to designing the show animations.

…In 2023, the global drone-show market was valued at $339m, with the Middle East accounting for about $41m. While North America is the largest market, the shows in the Middle East are more spectacular and use more drones, according to a report by SPH Engineering, a drone tech company. An average show in the region costs about $112,250 and has 401 drones. That’s far more expensive than a traditional fireworks show, which can range from $13,600 to $41,000. Drones, however, are reusable and more eco-friendly.

«

Rest of World continues to capture the most fascinating stories. This one in particular might be the future of fireworks (which will make a lot of wild animals and domestic pets happy).
unique link to this extract

Yahoo is still here—and it has big plans for AI • WIRED

Steven Levy:

»

In September 2021, Jim Lanzone took over a company whose name once embodied the go-go spirit of the internet but had, over the years, become a joke: Yahoo. He accepted the CEO post from the new private-equity owner Apollo Global Management, which had bought the property from Verizon, the most recent and possibly most clueless caretaker (high bar alert) in a long series of management shifts. Visiting him at the company’s offices in New York City, I ask him why he took the job. “I love turnarounds,” he says.

Lanzone’s résumé confirms that. In 2001 he took over a sagging search property called AskJeeves—its share price was less than a dollar, down from a high of $196—and built it back to the point where Barry Diller’s IAC Corp bought it for $1.85bn. At CBS Interactive and then CBS’s chief digital office during the 2010s, he yanked the stuffy Tiffany network into the streaming age. Yahoo, celebrating its 30th anniversary this month, might be his biggest challenge yet.

…One of Lanzone’s canniest AI moves was acquiring Artifact, the AI-powered news aggregator created by Instagram cofounders Kevin Systrom and Mike Krieger. When the pair decided it would not become a viable business, they announced its closure and Lanzone was among multiple suitors vying for the underlying technology. It became the centerpiece of the homepage that Yahoo relaunched earlier this year. “Instead of incorporating their technology into our product, we did it the other way,” Lanzone says. “Essentially Yahoo News is now Artifact.” Systrom approves. “We partnered with Yahoo because they made a strong offer, but also because they planned on deploying our hard work to many millions of people,” he says.

…When I suggest that Yahoo is less than the sum of its parts, Lanzone pushes back, saying that a Yahoo Finance user will get drawn into the Yahoo-sphere and use other services. Bolstering the effort is a nod to 2025 behavior: Yahoo has made deals with over 100 influencers to help establish it as a home for viral content. In a sense, he says, the company is returning to its early mission of delivering the bounty of the internet to a mass audience. In a symbolic reunion, he recently hosted cofounder Jerry Yang at an all-hands, implying a restored legacy.

«

Each new CEO of Yahoo is another go-around of the Arrested Development meme: “it never worked for those people. But it might work for us.”
unique link to this extract

Napster to become a music-marketing metaverse firm after $207m acquisition • Ars Technica

Scharon Harding:

»

Infinite Reality, a media, ecommerce, and marketing company focused on 3D and AI-powered experiences, has entered an agreement to acquired Napster. That means that the brand originally launched in 1999 as a peer-to-peer (P2P) music file-sharing service is set to be reborn again. This time, new owners are reshaping the brand into one focused on marketing musicians in the metaverse.

Infinite announced today a definitive agreement to buy Napster for $207m. The Norwalk, Connecticut-based company plans to turn Napster into a “social music platform that prioritizes active fan engagement over passive listening, allowing artists to connect with, own, and monetize the relationship with their fans.” Jon Vlassopulos, who became Napster CEO in 2022, will continue with his role at the brand.

Since 2016, Napster has been operating as a (legal) streaming service. It claims to have over 110 million high-fidelity tracks, with some supporting lossless audio. Napster subscribers can also listen offline and watch music videos. The service currently starts at $11 per month.

Since 2022, Napster has been owned by Web3 and blockchain firms Hivemind and Algorand. Infinite also develops Web3 tech, and CEO John Acunto told CNBC that Algorand’s blockchain background was appealing, as was Napster’s licenses for streaming millions of songs.

«

It’s so strange to see the brand names that were so enormous in shaping the web we have today turned into these weird contraptions for random conjugations of latter-day enthusiasms. How much competition is there to “market musicians in the metaverse”? I’m guessing zero.
unique link to this extract

No credible evidence supports claims of vast underground structures found beneath Egyptian pyramids • Snopes.com

Joey Esposito:

»

Claims that researchers discovered previously unknown structures beneath the Pyramid of Khafre — the pyramid situated in the center of the Great Pyramids of Giza — using radar technology circulated online in March 2025. 

The purported discovery was that of “five identical structures near the Khafre Pyramid’s base, linked by pathways, and eight deep vertical wells descending 648 meters underground.” 

Users took to social media to express their excitement over the alleged findings, posting on social media platforms like X (archived), Instagram (archived) and TikTok (archived). Some referred to the discovery as “a vast underground city.” One YouTube video sharing the claim stood at over 35,000 views as of this writing. 

Despite the popularity of the claim, there is no evidence to support it. In addition, no credible news outlets or scientific publications have reported on this rumor.

Rather, this appears to be a spin on already questionable research conducted in 2022 that was subsequently embellished by a variety of right-wing content creators like conspiracy website Infowars contributor Greg Reese, who publishes The Reese Report, listed as a source for many of the claims pertaining to this topic. Infowars founder and well-known conspiracy theorist Alex Jones shared a version of the same claim on X (archived).

In short, the Reese Report theorized the alleged discovery supported the idea that the pyramids were not built as tombs but as a sort of ancient power plant

«

Just in case you’d seen this claim repeated somewhere. The Jerusalem Post wrote a breathless writeup which also says, at the end, “The article was written with the assistance of a news analysis system.” Not increasing my trust in AI systems here, people.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

Would you buy a dishwasher if you knew you had to connect to the company’s cloud account to run certain cleaning cycles? CC-licensed photo by Kevin Dooley 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 9 links for you. Sparkling clean. 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 Trump administration accidentally texted me its war plans • The Atlantic

Jeffrey Goldberg:

»

On Tuesday, March 11, I received a connection request on Signal from a user identified as Michael Waltz. Signal is an open-source encrypted messaging service popular with journalists and others who seek more privacy than other text-messaging services are capable of delivering. I assumed that the Michael Waltz in question was President Donald Trump’s national security adviser.

I did not assume, however, that the request was from the actual Michael Waltz. I have met him in the past, and though I didn’t find it particularly strange that he might be reaching out to me, I did think it somewhat unusual, given the Trump administration’s contentious relationship with journalists—and Trump’s periodic fixation on me specifically. It immediately crossed my mind that someone could be masquerading as Waltz in order to somehow entrap me. It is not at all uncommon these days for nefarious actors to try to induce journalists to share information that could be used against them.

I accepted the connection request, hoping that this was the actual national security adviser, and that he wanted to chat about Ukraine, or Iran, or some other important matter.

Two days later—Thursday—at 4:28 p.m., I received a notice that I was to be included in a Signal chat group. It was called the “Houthi PC small group.”

A message to the group, from “Michael Waltz,” read as follows: “Team – establishing a principles [sic] group for coordination on Houthis, particularly for over the next 72 hours. My deputy Alex Wong is pulling together a tiger team at deputies/agency Chief of Staff level following up from the meeting in the Sit Room this morning for action items and will be sending that out later this evening.”

The message continued, “Pls provide the best staff POC from your team for us to coordinate with over the next couple days and over the weekend. Thx.”

The term principals committee generally refers to a group of the senior-most national-security officials, including the secretaries of defense, state, and the treasury, as well as the director of the CIA. It should go without saying—but I’ll say it anyway—that I have never been invited to a White House principals-committee meeting, and that, in my many years of reporting on national-security matters, I had never heard of one being convened over a commercial messaging app.

«

The creator of the group broke the law in multiple ways; the participants too. Why or how Goldberg was added – autocomplete? – may emerge in time, when someone is fired, as has to happen. But it also tells us how government, like private life, has shifted to messaging apps. (Using messaging apps break US government records rules. Maybe they should be updated, especially if the messages are set to disappear. But who can police that?)
unique link to this extract

I won’t connect my dishwasher to your stupid cloud • Jeff Geerling

Geerling’s old General Electric dishwasher died, so he bought a Bosch one:

»

So I turned it on, and immediately hated the new touch sensor stuff on it.

The old GE had buttons: you press them in, they click and you know that you pressed a button.

The touch sensor, you kind of touch it and the firmware—like this new dishwasher actually takes time to boot up! I had to reset it like three times and my wife meanwhile was like laughing at me like look at this guy who does tech stuff and he can’t even figure out how to change the cycle on it. That took about five minutes, sadly.

But eventually I pulled out the manual book because I was like… “this is actually confusing.” It should be like: I touch the button and it changes to that mode! But that was not how it was working.

I wanted to run just a rinse cycle to make sure the water would go in, the water would pump out through the sump, and everything worked post-install. But I couldn’t find a way to do a rinse cycle on the control panel.

So I looked in the manual and found a note: it says options with an asterisk—including Rinse, Machine Care (self-cleaning), HalfLoad, Eco, and Delay start, are “available through Home Connect app only and depending on your model.”

The 500 series model I bought isn’t premium enough to feature a seven-segment display like the $400-more-expensive 800 series, so these fancy modes are hidden behind an app and cloud service.

I was like, “Okay, I’ll look up this app and see if I can use it over Bluetooth or locally or whatever.”

Nope! To use the app, you have to connect your dishwasher to your Wi-Fi, which lets the dishwasher reach out on the internet to this Home Connect service. You have to set up an account on Home Connect, set up the Home Connect app on your phone, and then you can control your dishwasher through the Internet to run a rinse cycle.

That doesn’t make any sense to me.

An app? I mean, I can understand maybe adding some neat convenience features for those who want them. Like on my new fridge—which I chose not to connect to WiFI—it has an app that would allow me to monitor the inside temperature or look up service codes more easily. If I wanted those add-on features, which my old fridge didn’t have, I could get them.

But requiring an app to access features that used to be controllable via buttons on the dishwasher itself—or are still if you pay $400 more for the fancy “800” model? That’s no bueno.

«

I wouldn’t mind if my dishwasher could notify me in some way other than an annoying loud repetitive beeping that it had finished its cycle. But joining a cloud service? Sorry, that won’t wash.
unique link to this extract

UK on alert after H5N1 bird flu spills over to sheep in world-first • Ars Technica

Beth Mole:

»

The H5N1 bird flu has spilled over to a sheep for the first time, infecting a domesticated ruminant in the United Kingdom much like it has in US dairy cows, according to UK officials.

The single sheep—a ewe—in Yorkshire, England, was confirmed infected after captive birds on the same property had tested positive for the virus, according to an announcement Monday. The ewe’s milk was found to be positive for the virus through a PCR test, which detected genetic signatures of the virus. The ewe also had H5 antibodies in its blood. At the time of the confirmation, the ewe had symptoms of the infection in the way of mastitis, inflammation of the mammary glands.

This mirrors what US dairy farmers have been seeing in cows. An outbreak of H5N1 in dairy cows erupted a year ago, on March 25, 2024. Since then, at least 989 herds across 17 states have been infected with bird flu. In previous reports, farmers and researchers have noted that the virus appears to attack the animal’s mammary glands and their milk is teeming with the virus.

In the US, at least 70 people have been infected with the virus, 41 of whom were dairy workers. In some cases, workers reported having milk splashed on their faces before developing an infection. While nearly all of the cases have been relatively mild so far—some only with eye inflammation (conjunctivitis)—one person in the US has died from the infection after being exposed via wild or backyard birds.

In the UK, officials said further testing of the rest of the sheep’s flock has found no other infections. The one infected ewe has been humanely culled to mitigate further risk and to “enable extensive testing.”

“Strict biosecurity measures have been implemented to prevent the further spread of disease,” UK Chief Veterinary Officer Christine Middlemiss said in a statement. “While the risk to livestock remains low, I urge all animal owners to ensure scrupulous cleanliness is in place and to report any signs of infection to the Animal Plant Health Agency immediately.”

While UK officials believe that the spillover has been contained and there’s no onward transmission among sheep, the latest spillover to a new mammalian species is a reminder of the virus’s looming threat.

«

It’s not so much that it can infect sheep – we already know humans can catch it – but that if it goes to other species than birds, the virus could recombine with something else and become much more dangerous. (Watching brief, but slightly concerning for all our ovine readers.)
unique link to this extract

Understanding live facial recognition statistics • Big Brother Watch

»

The vast majority of matches recorded by the Metropolitan Police from its deployments of live facial recognition (LFR) over the years have been false.

• 85%, or six out of every seven matches, have been false
• 15%, one in seven alerts, were a correct match.
 
Not all of the matches the Met claims to be true have been confirmed as definite true matches, meaning the false match figure may be even higher. In Big Brother Watch’s observations of LFR deployments in London, we have seen a number of people trigger an alert who were not then stopped by officers, yet these matches have sometimes been recorded as true without additional verification.

The 84.7% figure is the number of false matches, as a percentage of the total number of facial recognition matches obtained by the Met Police since its first deployment in 2016. There have been 175 matches in total, of which 150 have been false and 25 have been recorded as true.

False positive rate = 100 * number of false matches/ total number of matches.

Professor Peter Fussey, from the University of Essex, used similar methodology to calculate the accuracy rate of the Live Facial Recognition deployments he assessed in the Independent Report On The London Metropolitan Police Service’s Trial Of Live Facial Recognition.

The study, commissioned by the Met Police, found that in the limited number of deployments it observed, 63.64% of matches leading to a stop were inaccurate (14 of 22 total matches), and just 36.36% (8 of 22) were accurate. Similarly South Wales Police has returned false matches for 2,825 of its 3,140 LFR flags, giving it a false match rate of 89.9%.

The Metropolitan Police chooses to use different metrics which present Live Facial Recognition as much more accurate than it is.

The False Positive Identification Rate (FPIR) used by the Met Police is measured as the number of false matches against the total number of faces seen, with the figure quoted by the Met Police being 1 in 6,000 or 0.017%. This figure is reached independently of the number of true matches, allowing the Metropolitan Police to overstate the algorithm’s accuracy.

«

Tricky: what is the false negative rate? How many criminals walk past the cameras and don’t get spotted? Because if that’s zero, then we don’t really mind false positives as much, do we? But of course the false negative is impossible to know.
unique link to this extract

The LibGen data set: what authors can do • The Society of Authors

»

The Atlantic says that court documents show that staff at Meta discussed licensing books and research papers lawfully but instead chose to use stolen work because it was faster and cheaper. Given that Meta Platforms, Inc, the parent company of Facebook, Instagram and WhatsApp, has a market capitalisation of £1.147 trillion, this is appalling behaviour.

According to The Atlantic, Meta argued that it could then use the US’s ‘fair use exception’ defence if it was challenged legally.

It is not yet clear whether scraping from copyright works without permission is unlawful under the US fair use exception to copyright, but if that scraping is for commercial purposes (which what Meta is doing surely is) it cannot be fair use. Under the UK fair dealing exception to copyright, there is no question that scraping is unlawful without permission.

We wrote to Meta in August 2024 to assert our members’ rights around uses of their works by generative AI. As a matter of urgency, Meta needs to compensate the rightsholders of all the works it has been exploiting.

This is yet more evidence of the catastrophic impact generative AI is having on our creative industries worldwide. From development through to output, creators’ rights are being ignored, and governments need to intervene to protects authors’ rights.

In the UK, and globally, we need to see strong legislation from governments to uphold and strengthen copyright law, ensure transparency and fair payment, and to penalise big tech companies who ride roughshod over the law.

• We are continuing to explore all options available for collective action on behalf of our members
• We are continuing to raise this matter with Government through letters and briefings to MPs
• We are continuing our campaign work on AI and Copyright; working with policy makers on the issue of the unremunerated use of copyrighted works in large language model (LLM) training.

«

The Atlantic making waves again.
unique link to this extract

23andMe files for bankruptcy protection – BBC News

Lily Jamali:

»

Popular DNA testing firm 23andMe has filed for bankruptcy protection, and announced that its co-founder and CEO, Anne Wojcicki, has resigned with immediate effect.

The company will now attempt to sell itself under the supervision of a court.

23andMe said in a press release that it plans to continue operating throughout the sale process and that there “are no changes to the way the company stores, manages, or protects customer data.”

The Information Commissioner’s Office (ICO), the UK’s data protection watchdog, said on Monday it had notified the company of its intent to hand down a £4.59m fine over a 2023 data breach.

The ICO, which launched a joint investigation with Canada’s data watchdog into the genetic testing company last June, said the findings were provisional. And deputy commissioner Stephen Bonner said the regulator was aware of the company’s bankruptcy filing in the US and “monitoring the situation closely”.

“As a matter of UK law, the protections and restrictions of the UK GDPR continue to apply and 23andMe remains under an obligation to protect the personal information of its customers,” he said.

The Attorney General in 23andMe’s home state of California issued a consumer alert on Friday advising customers to delete their data from the site given the company’s “reported financial distress.”

23andMe’s saliva-based test kits were once celebrated among customers and investors, who helped to push the company’s value as high as $6bn (£4.6bn). But it has been struggling for survival. Founded in 2006, the company went public in 2021 but has never turned a profit.

«

The value crashed in 2013 when the US FDA told the company it couldn’t use its tests for analysing “health conditions and traits” without authority and passing tests for accuracy. Even so, by February 2019 more than 26 million people had taken an “at-home ancestry test”, and 23andMe was one of the main suppliers.

It’s not clear that any company is going to meet the FDA’s requirements. But maybe it’s not a bad thing that we don’t know every prediction about our future? Would you want to know the day on which you’re going to die?
unique link to this extract

China unveils a powerful deep-sea cable cutter that could reset the world order • South China Morning Post

Stephen Chen:

»

A compact, deep-sea, cable-cutting device, capable of severing the world’s most fortified underwater communication or power lines, has been unveiled by China – and it could shake up global maritime power dynamics.

The revelation marks the first time any country has officially disclosed that it has such an asset, capable of disrupting critical undersea networks.

The tool, which is able to cut lines at depths of up to 4,000 metres (13,123 feet) – twice the maximum operational range of existing subsea communication infrastructure – has been designed specifically for integration with China’s advanced crewed and uncrewed submersibles like the Fendouzhe, or Striver, and the Haidou series.

Developed by the China Ship Scientific Research Centre (CSSRC) and its affiliated State Key Laboratory of Deep-sea Manned Vehicles, the device targets armoured cables – layered with steel, rubber and polymer sheaths – that underpin 95% of global data transmission.

While it was created as a tool for civilian salvage and seabed mining, the dual-use potential of the tool could send alarm bells ringing for other nations.

For example, cutting cables near strategic chokepoints such as Guam, which is a linchpin of the US military’s second island chain, a defence strategy used to contain China, the tool could essentially destabilise global communications during a geopolitical crisis.

«

Of course China says it’s for “repairs”, but the publicising of this is a statement in its own right – as overt as the USSR having a military parade of tanks rolling past the Kremlin.
unique link to this extract

CEO of AI ad-tech firm pledging “world free of fraud” sentenced for fraud • Ars Technica

Kevin Purdy:

»

In May 2024, the website of ad-tech firm Kubient touted that the company was “a perfect blend” of ad veterans and developers, “committed to solving the growing problem of fraud” in digital ads. Like many corporate sites, it also linked old blog posts from its home page, including a May 2022 post on “How to create a world free of fraud: Kubient’s secret sauce.”

These days, Kubient’s website cannot be reached, the team is no more, and CEO Paul Roberts is due to serve one year and one day in prison, having pled guilty Thursday to creating his own small world of fraud. Roberts, according to federal prosecutors, schemed to create $1.3m in fraudulent revenue statements to bolster Kubient’s initial public offering (IPO) and significantly oversold “KAI,” Kubient’s artificial intelligence tool.

The core of the case is an I-pay-you, you-pay-me gambit that Roberts initiated with an unnamed “Company-1,” according to prosecutors. Kubient and this firm would each bill the other for nearly identical amounts, with Kubient purportedly deploying KAI to find instances of ad fraud in the other company’s ad spend.

Roberts, prosecutors said, “directed Kubient employees to generate fake KAI reports based on made-up metrics and no underlying data at all.” These fake reports helped sell the story to independent auditors and book the synthetic revenue in financial statements, according to Roberts’ indictment.

Before Kubient’s IPO in August 2020, Kubient issued a prospectus noting research figures that suggested $42bn lost to ad fraud in 2019. Kubient’s technology was touted as fast enough to work in the 300-millisecond real-time ad auction window. It leveraged “machine learning powered [sic] pre-bid ad fraud prevention technology” and a “self-learning neural network always getting smarter.”

«

I suppose you could say that he has made the world a little bit more free of fraud by getting caught?
unique link to this extract

Plants can take up CWD-causing prions from soil in the lab. What happens if they are eaten? • CIDRAP

Mary Van Beusekom:

»

When Christopher Johnson, PhD, set out to study whether lab mice fed prion-contaminated plants developed neurodegenerative disease, he expected the plants to take up only small prion clusters, but they absorbed large clusters characteristic of prion diseases in deer and other animals.

Then again, “prions are constantly surprising,” Johnson, a study coauthor and deputy director of the Office of Science Quality and Integrity at the US Geological Survey in Reston, Virginia, told CIDRAP News. “But perhaps we shouldn’t ever be allowed to be surprised by them, because they are so resistant to degradation, and they are so resilient that finding prions in unusual settings is maybe something that we should all begin to just expect.”

Prions are infectious misfolded proteins that cause fatal neurodegenerative diseases such as chronic wasting disease (CWD) in cervids like deer and elk, scrapie in sheep and goats, bovine spongiform encephalopathy (BSE, or “mad cow” disease) in cattle, and Creutzfeldt-Jakob disease in humans.

In the case of CWD, once an animal is infected, it can spread the disease through direct contact, saliva, antler velvet, urine, feces, and carcasses, and the prions can persist in the environment for years. Once an animal is exposed, the incubation period in a host—the time before symptoms appear—is thought to be up to 2 years.

But given the rapid spread of CWD throughout North America and parts of Europe and Asia, scientists question whether it is also being transmitted through a different route, such as the ingestion of contaminated plants.

While researchers have been experimenting with protein uptake into plants since the 1970s, Johnson and colleagues’ laboratory study, published in iScience in December, takes those investigations a step further. They demonstrated that alfalfa, barley, and Arabidopsis thaliana, a small plant from the mustard family called thale cress and other names, all accumulated sufficient prions from contaminated soil in their above-ground tissues to cause mice that ingested the plant tissues to develop prion disease.

«

So it doesn’t have to be an intermediate host, though this doesn’t quite explain how CWD would spread across the US. Might it be in feed where farmed venison has been kept? Thanks to Natalie Bennett (ex-Guardian) for the link.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

In the US, government webpages referencing the Enola Gay bomber were wiped – then restored. Guess why? CC-licensed photo by chris favero 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. Flying high. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

Meta settles UK ‘right to object to ad-tracking’ lawsuit by agreeing not to track plaintiff • TechCrunch

Natasha Lomas:

»

A human rights campaigner, Tanya O’Carroll, has succeeded in forcing social media giant Meta not to use her data for targeted advertising. The agreement is contained in a settlement to an individual challenge she lodged against Meta’s tracking and profiling back in 2022.

O’Carroll had argued that a legal right to object to the use of personal data for direct marketing that’s contained in U.K. (and E.U.) data protection law, along with an unqualified right that personal data shall no longer be processed for such a purpose if the user objects, meant Meta must respect her objection and stop tracking and profiling her to serve its microtargeted ads.

Meta refuted rebutted [tch – Overspill Ed.] this — claiming its “personalized ads” are not direct marketing. The case had been due to be heard in the English High Court on Monday, but the settlement ends the legal action.

For O’Carroll it’s an individual win: Meta must stop using her data for ad targeting when she uses its services. She also thinks the settlement sets a precedent that should allow others to confidently exercise the same right to object to direct marketing in order to force the tech giant to respect their privacy.

Speaking to TechCrunch about the outcome, O’Carroll explained she essentially had little choice to agree to the settlement once Meta agreed to what her legal action had been asking for (i.e. not to process her data for targeted ads). Had she proceeded and the litigation failed, she could have faced substantial costs, she told us.

«

O’Carroll is actually a very experienced lawyer, and her case was handled by a company which does a lot of similar cases. Whether this sets a useful precedent is a different question, though. The BBC’s writeup suggests.. only if you go through the Information Commissioner’s office.
unique link to this extract

A big drop in programmers may be the first sign of job loss to AI • The Washington Post

Andrew Van Dam:

»

More than a quarter of all computer programming jobs have vanished in the past two years, the worst downturn that industry has ever seen. Things are sufficiently abysmal that computer programming ranks among 10 hardest-hit occupations of 420-plus jobs for which we have data from the Bureau of Labor Statistics.

Learning to code was supposed to save millions of would-have-been liberal arts majors. But today there are fewer programmers in the United States than at any point since 1980. That’s a 45-year period in which America’s total workforce has grown by about 75%! It’s so long ago that millennials hadn’t been invented, the oldest Gen Xers were barely in high school, and even many boomers were too young for their first real coding jobs.

…Upon perusing the fine print, we saw that while programmers do in fact program, they “work from specifications drawn up by software and web developers or other individuals.” That seems like a clue.

In the real world, “developer” and “programmer” can seem almost interchangeable. But in the world of government statistics, where we have legal permanent residency, there’s a clear distinction.

In the government’s schema, programmers do the grunt work while the much more numerous — and much faster-growing — software developers enjoy a broader remit. They figure out what clients need, design solutions and work with folks such as programmers and hardware engineers to implement them.

Their pay reflects this gap in responsibilities. The median programmer earned $99,700 in 2023, compared with $132,270 for the median developer. And while 27.5% of programming jobs vanished, jobs for developers have only fallen 0.3%, similar to the broader industry.

So it’s not just industry-wide headwinds holding programming back. What could account for the difference between the coder collapse and everyone else?

«

Yes, it is: the most obvious and best explanation is AI taking over slog programming.
unique link to this extract

Humming along in an old church, the Internet Archive is more relevant than ever • NPR

Emma Bowman:

»

As one of the few large-scale archivists to back up the web, the Internet Archive finds itself in a particularly unique position right now. After President Trump’s inauguration in January, some federal web pages vanished. While some pages were removed entirely, many came back online with changes that the new administration’s officials said were made to conform to Trump’s executive orders to remove “diversity, equity, inclusion, and accessibility policies.” Thousands of datasets were wiped — mostly at agencies focused on science and the environment — in the days following Trump’s return to the White House.

Information about climate change, reproductive health, gender identity and sexual orientation also have been on the chopping block. For example, pages referencing the Enola Gay — the B-29 aircraft that dropped an atomic bomb on Hiroshima and is not particularly related to LGBTQ history — were among a leaked list of posts the Pentagon flagged for removal. Some deleted pages, including those related to the Enola Gay, have resurfaced as agencies figure out how to comply with Trump’s directives.

The Internet Archive is among the few efforts that exist to catch the stuff that falls through the digital cracks, while also making that information accessible to the public. Six weeks into the new administration, Wayback Machine director Graham said, the Internet Archive had cataloged some 73,000 web pages that had existed on U.S. government websites that were expunged after Trump’s inauguration. 

Graham noted that, for example, the Internet Archive is currently the only place the public can find a copy of an interactive timeline detailing the events of Jan. 6. The timeline is a product of the congressional committee that investigated the Capitol attack, and has since been taken down from their website. Graham said it’s in the public’s interest to save such records.

“How much money did our tax dollars pay to make it?” he said, referring to the timeline and committee proceedings. “It was a non-trivial exercise and it’s part of our history — and for that reason alone, worthy of preservation and worthy of exploration, of understanding.”

«

The Trump admin is, let’s say it again, utterly insane in its desire to expunge its own content from the web. Removing the Enola Gay? Can we guess, just take a guess, why?
unique link to this extract

What does a “Personal Net Zero” look like? • Terence Eden’s Blog

Terence Eden:

»

Five years ago today, we installed solar panels on our house in London. Solar panels are the ultimate in “boring technology”. They sit on the roof and generate electricity whenever the sun shines. That’s it.

This morning, I took a reading from our generation meter: 19MWh of electricity stolen from the sun and pumped into our home. That’s an average of 3,800 kWh every year. But what does that actually mean?

The UK’s Department for Energy Security and Net Zero publishes quarterly reports on energy prices. Its most recent report suggests that a typical domestic consumption is “3,600 kWh a year for electricity”. Ofgem, the energy regulator, has a more detailed consumption breakdown which broadly agrees with DESNZ.

On that basis, our solar panels are doing well! A typical home would generate slightly more than it uses.

…We imported 2,300 kWh over 2024. Quick maths! Our total electricity consumption was 4,500 kWh during the year. Very roughly, we imported 2,300 and exported 1,500. That means our “net” import was only 800kWh.

There’s a slight wrinkle with the calculations though. Our battery is aware that we’re on a a dynamic tariff; the price of electricity varies every 30 minutes. If there is surplus electricity (usually overnight) the prices drop and the battery fills up for later use. In 2024, our battery imported about 990 kWh of cheap electricity (it also exported a negligible amount). If our battery hadn’t been slurping up cheap energy, we would be slightly in surplus; exporting 190 kWh more than we consumed.

So, I’m happy to report that our panels take us most of the way to a personal net zero for domestic electricity consumption.

«

Solar panels and/or batteries (for the cheap rate) really are easy wins for reducing consumption. The payback is remarkably fast, especially as electricity prices (decided by gas!) keep going up.
unique link to this extract

GB Energy announces first major rooftop solar project • Energy Voice

Mathew Perry:

»

The Department for Energy Security and Net Zero (DESNZ) said the investment in rooftop solar projects will help the schools and NHS trusts save “hundreds of millions on their energy bills”.

The funding includes £80m to support rooftop solar for around 200 schools in England, alongside a further £100m for nearly 200 NHS sites. Meanwhile, there will be £9.3m funding for devolved governments to use for renewable energy schemes on either public sector buildings or new community projects. This includes £4.85m for Scotland, £2.88m for Wales and £1.62m for Northern Ireland.

Elsewhere, community energy groups will be able to apply for a share of £5m in grant funding for local clean energy projects. There will also be £6.8m in funding included for existing local net zero hubs across England.

A DESNZ spokesperson told Energy Voice that GB Energy will commit £90m from its initial £8.3bn budget to fund the solar partnership. This includes £50m for NHS hospitals and £40m for schools, with the health and education departments providing match funding.

…Chancellor Rachel Reeves had allocated £100m for GB Energy over its first two years as part of the government’s first budget in October.

It comes amid reports that the Labour government is considering cutting the £8.3bn budget for GB Energy amid efforts to increase UK defence spending. However, the DESNZ spokesperson told Energy Voice the government “remains fully committed” to the GB Energy budget.

DESNZ said the NHS is the “single biggest public sector energy user” in the UK, with an estimated annual energy bill of £1.4bn.

…Currently, only about 20% of schools and under 10% of hospitals have solar panels installed, according to the government.

«

Putting solar panels on schools and hospitals is shooting at an open goal: schools in particular tend to need their electricity during the day. Hospitals maybe can get some batteries too to store their surplus energy (if any?).
unique link to this extract

Bicycle renderings based on people’s attempts to draw them from memory • Booooooom

»

In 2009 designer Gianluca Gimini started asking friends and strangers to draw a men’s bicycle from memory. While some got it right, most made technical errors — missing fundaments parts of the frame or chain.

The exercise is similar to psychological tests used to demonstrate how little we know compared to what we think we do. However, for Gimini this isn’t about proving how dumb we are but, rather, how extraordinary our imaginations can be! Having now amassed a collection of 376 drawings from participants ranging from 3 – 88 years of age, Gimini has started building realistic renderings of the bikes based on these sketches, in a 3D program.

«

Do you think you could sketch a bicycle from memory? It’s worth having a try and then comparing it to these efforts. A lot of them would collapse the first time somebody sat on them – which also raises the question of how many designs the prototype product needed.
unique link to this extract

The world will only get weirder • Steve Coast’s Musings

Steve Coast (the brains behind OpenStreetMap) back in 2015:

»

As we find more rules to fix more things we are encountering tail events. We fixed all the main reasons aircraft crash a long time ago. Sometimes a long, long time ago. So, we are left with the less and less probable events.

We invented the checklist. That alone probably fixed 80% of fatalities in aircraft. We’ve been hammering away at the remaining 20% for 50 years or so by creating more and more rules.

We’ve reached the end of the useful life of that strategy and have hit severely diminishing returns. As illustration, we created rules to make sure people can’t get in to cockpits to kill the pilots and fly the plane in to buildings. That looked like a good rule. But, it’s created the downside that pilots can now lock out their colleagues and fly it in to a mountain instead.

It used to be that rules really helped. Checklists on average were extremely helpful and have saved possibly millions of lives. But with aircraft we’ve reached the point where rules may backfire, like locking cockpit doors. We don’t know how many people have been saved without locking doors since we can’t go back in time and run the experiment again. But we do know we’ve lost 150 people with them.

And so we add more rules, like requiring two people in the cockpit from now on. Who knows what the mental capacity is of the flight attendant that’s now allowed in there with one pilot, or what their motives are. At some point, if we wait long enough, a flight attendant is going to take over an airplane having only to incapacitate one, not two, pilots. And so we’ll add more rules about the type of flight attendant allowed in the cockpit and on and on.

…On a personal level we should probably work in areas where there are few rules.

To paraphrase Peter Thiel, new technology is probably so fertile and productive simply because there are so few rules. It’s essentially illegal for you to build anything physical these days from a toothbrush (FDA regulates that) to a skyscraper, but there’s zero restriction on creating a website. Hence, that’s where all the value is today.

If we can measure economic value as a function of transactional volume (the velocity of money for example), which appears reasonable, then fewer rules will mean more volume, which means better economics for everyone.

«

unique link to this extract

A first look at how Apple’s C1 modem performs with early adopters • Ookla

Sue Marek, Mark Giles, Luke Kehoe and Kerry Baker:

»

Speedtest data shows the iPhone16e recorded faster median download speeds than the iPhone 16 on both AT&T and Verizon’s networks, but was markedly slower on T-Mobile’s network. 

iPhone 16e users on T-Mobile’s network experienced median download speeds of 264.71 Mbps, which is at least 47% faster than iPhone 16e users on Verizon’s network that experienced median download speeds of 140.77 Mbps. The download speed performance for iPhone 16e users on AT&T’s network was 226.90 Mbps, closer to that of T-Mobile users. 

However, when comparing median download speeds for T-Mobile users with the iPhone 16e (264.71 Mbps) to T-Mobile users with the iPhone 16 device (357.47 Mbps), the iPhone 16 outperformed the iPhone 16e by at least 24%.

The iPhone 16e’s underperformance in median download speed compared to the iPhone 16 on T-Mobile’s network is most likely due to the fact that T-Mobile is the only US carrier to have a nationwide commercialized 5G standalone network (SA) and one of the few operators globally to deploy significant spectrum depth and advanced features like carrier aggregation (CA) on the new 5G architecture. 

«

Those are all very respectable speeds, though. The real takeaway seems more to be that the C1 performs just fine, which is going to be a relief for Apple now the product is out in the real world.
unique link to this extract

Yahoo is selling TechCrunch • The Verge

Emma Roth:

»

TechCrunch has a new owner, again. Yahoo has sold the tech news site to the private equity firm Regent for an undisclosed sum, according to an announcement on Friday.

Regent is the same company that snapped up Foundry, the firm behind outlets like PCWorld, Macworld, and TechAdvisor on Thursday. Founded in 2005, TechCrunch has experienced many shakeups in ownership after AOL acquired the site in 2010.

When Verizon acquired AOL in 2015 and Yahoo in 2017, the company folded TechCrunch, Engadget, Yahoo Sports, and other sites into a new division called Oath, which later became Verizon Media. In 2021, Verizon sold its media division to Apollo Global Management for $5bn, and it was renamed Yahoo!

“Yahoo decided to sell TechCrunch because, in the end, our DNA is simply different from the rest of its portfolio,” TechCrunch editor-in-chief Connie Loizos writes in the announcement, noting that Yahoo will still have a “small interest” in TechCrunch.

«

In the “announcement” on Techcrunch’s site, editor-in-chief Connie Loizos says

»

“While the financial terms remain undisclosed, one thing is clear: Regent is acquiring an iconic brand. TechCrunch isn’t just a tech news site; it’s the most influential voice chronicling innovation in Silicon Valley and beyond.”

«

Is that true any more, though? Does it chronicle innovation, or just recycle press releases? It’s not the blog that changed the tech world back in 2005. Things have changed a lot in 20 years.
unique link to this extract

USDA announces funding for bird flu research • Iowa Capital Dispatch

Cami Koons:

»

The U.S. Department of Agriculture will fund research projects that explore the highly pathogenic avian influenza in poultry, as well as novel vaccines and therapeutics to treat the bird flu, according to a Thursday announcement. 

The $100m investment is part U.S. Agriculture Secretary Brooke Rollins’ $1bn plan to combat avian influenza and inflated egg prices. 

Rollins, in a conference call with poultry and farm stakeholders, said the agency has made “significant progress” on its five-pronged approach to the bird flu issue, especially with a decrease in the wholesale price of eggs. 

“While we’re noting today that prices are exponentially down, and we’re really, really encouraged by that, there is always a possibility those prices could tick back up,” Rollins said, noting the increased egg demand associated with the upcoming Easter holiday. 

The USDA egg market report from March 14 shows a “sharp downward trajectory” of wholesale prices for loose eggs, with prices dropping from more than $8 per dozen in late February to $4.15 per dozen for white large shell eggs. 

Rollins said the department confirmed agreements with South Korea and the west-Asian nation of Türkiye [aka “Turkey” – Overspill Ed.] to provide temporary increased egg imports to the country, which was also part of her plan.

…Rollins said she has been in communication with U.S. Health Secretary Robert F. Kennedy Jr. and other departments, about preventing the spread to humans and limiting impact on farmers. Kennedy recently suggested farmers allow the virus to spread in a flock, rather than culling it after a detection, to see which birds survive.

Rollins left the call before taking questions, but staff members declined to share if her conversations with Kennedy were about this approach.

«

So is that $100m on vaccine and $900m on buying eggs? It’s not clear, but wouldn’t be surprising. Some relief though that RFK Jr doesn’t hold sway over everything vaccine-related.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

In a new lawsuit, Google claims to have found thousands of scammy locksmith entries on its maps. Which is hardly news to most people. CC-licensed photo by Steve Snodgrass 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.

No Social Warming Substack today. Maybe next week?

A selection of 9 links for you. Achievement locked. 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 unbelievable scale of AI’s pirated-books problem • The Atlantic

Alex Reisner:

»

When employees at meta started developing their flagship AI model, Llama 3, they faced a simple ethical question. The program would need to be trained on a huge amount of high-quality writing to be competitive with products such as ChatGPT, and acquiring all of that text legally could take time. Should they just pirate it instead?

Meta employees spoke with multiple companies about licensing books and research papers, but they weren’t thrilled with their options. This “seems unreasonably expensive,” wrote one research scientist on an internal company chat, in reference to one potential deal, according to court records. A Llama-team senior manager added that this would also be an “incredibly slow” process: “They take like 4+ weeks to deliver data.” In a message found in another legal filing, a director of engineering noted another downside to this approach: “The problem is that people don’t realise that if we license one single book, we won’t be able to lean into fair use strategy,” a reference to a possible legal defence for using copyrighted books to train AI.

Court documents released on Wednesday night show that the senior manager felt it was “really important for [Meta] to get books ASAP,” as “books are actually more important than web data.” Meta employees turned their attention to Library Genesis, or LibGen, one of the largest of the pirated libraries that circulate online. It currently contains more than 7.5 million books and 81 million research papers. Eventually, the team at Meta got permission from “MZ”—an apparent reference to Meta CEO Mark Zuckerberg—to download and use the data set.

This act, along with other information outlined and quoted here, recently became a matter of public record when some of Meta’s internal communications were unsealed as part of a copyright-infringement lawsuit brought against the company by Sarah Silverman, Junot Díaz, and other authors of books in LibGen. Also revealed recently, in another lawsuit brought by a similar group of authors, is that OpenAI has used LibGen in the past.

«

That title of the new book about Facebook, “careless people”, rings ever more true. Spend money? To get approved access to content? Copyright is for the little people.
unique link to this extract

Google finds 10,000 fake listings on Google Maps, sues alleged network of scammers • CBS News

Kara Fellow and Cait Bladt:

»

Google says it uncovered thousands of illegitimate listings, including for fake businesses, on Google Maps and has announced a lawsuit against the alleged scammers behind the fraud.

The lawsuit, announced Wednesday, claims a man working within a wider network, created and sold fake business profiles on Google Maps.

An initial alert came from a Texas business that flagged an unlicensed locksmith impersonating them on Google Maps. That was just the tip of the iceberg.

The claim sparked an investigation that led Google to uncover and eliminate more than 10,000 illegitimate listings, the company said. The scams ranged from outright fake businesses to legitimate accounts that had been hacked or hijacked.

“Once we’re alerted to the actual fraud, we take extreme efforts to identify similar fraudulent listings,” Halimah DeLaine Prado, Google’s general counsel, said on “CBS Mornings Plus” Wednesday.

Google found many of the scams were concentrated in what they call “duress verticals” – services people need in urgent or stressful situations, like locksmiths or towing companies.

“Scammers are becoming increasingly sophisticated,” DeLaine Prado said.

Google’s investigation also uncovered that these alleged scammers aren’t working alone. They collaborate with agents around the world and leverage social media to increase their reach. One example in the lawsuit shows an alleged scammer posted in multiple Facebook groups to advertise “5 star reviews” that can bypass Google’s guidelines.

«

Oh, the scams were for things like locksmiths? Because I linked to a piece in January 2016 about “Fake online locksmiths may be out to pick your pocket, too” (congratulations if you were a subscriber then) which pointed out exactly this sort of problem. Google never seems to learn from anything – such as that “duress verticals” attract scammers and so ought to be monitored closely.

So if Google “found” that’s where these scams are, it can’t have been paying much attention in the past nine years.
unique link to this extract

Claude can now search the web • Anthropic

»

You can now use [the chatbot] Claude to search the internet to provide more up-to-date and relevant responses. With web search, Claude has access to the latest events and information, boosting its accuracy on tasks that benefit from the most recent data.

When Claude incorporates information from the web into its responses, it provides direct citations so you can easily fact check sources. Instead of finding search results yourself, Claude processes and delivers relevant sources in a conversational format. This enhancement expands Claude’s extensive knowledge base with real-time insights, providing answers based on more current information.

«

But of course there’s a catch:

»

Web search is available now in feature preview for all paid Claude users in the United States. Support for users on our free plan and more countries is coming soon.

«

The free plan isn’t that great, to be honest. I asked it the tennis question I asked ChatGPT and Grok last week and it responded “As of my knowledge cutoff in October 2024, the head-to-head record between Jack Draper and Taylor Fritz was 1-1.”

Which is wrong. At that point, it was 2-1 Fritz: he had a win at the Paris Olympics in the middle of 2024, and they were split before that. I remain unimpressed by these chatbots, and by people’s reliance on them. Given that the web is pretty unreliable to start with, this can only make things worse.
unique link to this extract

Apple TV+ is losing $1bn every year, report says • 9to5Mac

Ryan Christoffel:

»

Wayne Ma wrote on Thursday at The Information about the state of Apple TV+ as a business, analyzing where the streamer stands in terms of revenue and subscribers after five years.

In the heavily-paywalled article, Ma reports that Apple’s losses on TV+ amount to over $1 billion per year. While it’s long been known that the streamer was not yet profitable, this is the first time I can recall that we’ve had a solid number to quantify the losses.

The report also claims Apple TV+ had 45 million subscribers last year.

It’s unclear, though, how many Apple has added during the current Severance hot streak. For example, a recent Antenna report said the streamer added two million new subscribers in a single month, which was a sizable boost.

We’ve seen reports in the past of Apple working to rein in spending, especially with its film division. Apple’s original movies have achieved some success, including CODA being the first from a streaming service to win the Best Picture Oscar. But several theatrical disappointments reportedly led to the company cutting budgets and prioritizing direct-to-streaming.

On the TV side though, while there may be some budget cuts happening, generally Apple has seemed content stomaching its losses from TV+ as it builds up the service.

It’s entirely normal for streaming services to experience losses for a while before achieving profitability, and Apple happens to be in a great position to ride that train longer than most. While $1bn is no doubt a big number, it’s a relatively small loss in light of Apple’s overall very profitable business.

For example, last quarter alone Apple reported $124 billion in revenue, with $36 billion of that being profit.

«

TV+ only launched in November 2019, with pretty much no licensed content (unlike Netflix and other services, which have shows coming and going all the time), so it’s not surprising that it has been slow to get anywhere near breakeven. Amazon Prime used to be $700m in the hole a few years back, but is now vaguely profitable. Ditto Disney+, which has plenty of content it hasn’t had to pay for. TV+ is building up a back catalogue of programmes that aren’t linked to any time – Slow Horses and especially Severance – so it might license them in future, as Amazon does.
unique link to this extract

Apple shuffles AI executive ranks in bid to turn around Siri • Bloomberg via Yahoo

Mark Gurman:

»

Chief executive Officer Tim Cook has lost confidence in the ability of AI head John Giannandrea to execute on product development, so he’s moving over another top executive to help: Vision Pro creator Mike Rockwell. In a new role, Rockwell will be in charge of the Siri virtual assistant, according to the people, who asked not to be identified because the moves haven’t been announced.

Rockwell will report to software chief Craig Federighi, removing Siri completely from Giannandrea’s command. Apple is poised to announce the changes to employees this week. The iPhone maker’s senior leaders — a group known as the Top 100 — just met at a secretive, annual offsite gathering to discuss the future of the company. Its AI efforts were a key talking point at the summit, Bloomberg News has reported.

The moves underscore the plight facing Apple: its AI technology is severely lagging industry rivals, and the company has shown little sign of catching up. The Apple Intelligence platform was late to arrive and largely a flop, despite being the main selling point for the iPhone 16. [It was not in fact “the main selling point” for the iPhone 16, because it was released well after the phone, and Gurman knows this – Overspill Ed]

Rockwell is currently the vice president in charge of the Vision Products Group, or VPG, the division that developed Apple’s headset. As part of the changes, he’ll be leaving that team and handing the reins to Paul Meade, an executive who has run hardware engineering for the Vision Pro under Rockwell.

«

John Gruber points to that mention of the Top 100 and infers that two separate people (because that’s how many Bloomberg would, or should, demand for reliable sourcing) told Gurman about this, because it hasn’t been announced by Apple at the time of writing.

That Giannandrea is being moved doesn’t surprise me at all (five days ago I commented “I wonder if John Giannandrea will be shuffled away from being in charge of Siri soon, because if you were giving him a performance review, what could he present in his favour?”). That Siri is being put under the software banner, rather than living in the “machine learning” group, says that Federighi is gaining power. Though of course Siri is currently a poisoned chalice. The software group, though, might be the ones who can actually set it alight.
unique link to this extract

How three alleged Tesla vandals got caught • 404 Media

Jason Koebler:

»

The first case relates to a March 7 arson of a set of Tesla charging stations in South Carolina. Witnesses said that a man used red spray to write “Fuck Trump” and “Long Live Ukraine” in a Tesla charging station parking spot, according to court records. The male then lit beer bottles on fire and threw them at the charging stations, with some setting on fire, the documents say.

The Bureau of Alcohol, Tobacco, Firearms and Explosives (ATF) collected evidence from the scene, including a charred piece of fabric suspected to be a wick and shards of glass bottle, the documents continue. Investigators reviewed surveillance video from a nearby restaurant and saw a white male in a grey sweater, black facemask, black shorts, and black shoes. During the footage, the man was carrying a green item, the documents say.

ATF investigators then reviewed more footage from the North Charleston Police Department (NCPD). In that clip, the man was not holding the green item. Investigators then found it: a cardboard bottle carrier for Holland 1839 beer. More footage showed the man getting into a white van and leaving the area, a Tanger Outlet mall, the court documents say.

Investigators then contacted the outlet mall’s security who said they had access to license plate reader (LPR) technology. LPR cameras are typically set up in a fixed area which continuously monitor which vehicles drive by and record their license plates. These systems are run by both government agencies and private businesses, and some surveillance contractors sell access to such data. The LPR footage identified the vehicle as a white 2006 Chrysler Town and Country van with South Carolina license plate 331ANL, according to the court documents.

Investigators then queried the South Carolina Department of Motor Vehicles to find who the vehicle was registered to. That led to the name Clarke-Pounder. The Tanger Outlet mall security consultants were then also able to pull a photograph of the man without his mask from their surveillance cameras, the documents say.

«

The police procedural that is detailed here – there are plenty more steps after this one – demonstrates how it’s not as easy as the TV shows might make you think. It’s pretty hard to remain anonymous in the modern world.
unique link to this extract

‘Don’t call it zombie deer disease’: scientists warn of ‘global crisis’ as infections spread across the US • The Guardian

Todd Wilkinson:

»

In a scattershot pattern that now extends from coast to coast, continental US states have been announcing new hotspots of chronic wasting disease (CWD).

The contagious and always-fatal neurodegenerative disorder infects the cervid family that includes deer, elk, moose and, in higher latitudes, reindeer. There is no vaccine or treatment.

Described by scientists as a “slow-motion disaster in the making”, the infection’s presence in the wild began quietly, with a few free-ranging deer in Colorado and Wyoming in 1981. However, it has now reached wild and domestic game animal herds in 36 US states as well as parts of Canada, wild and domestic reindeer in Scandinavia and farmed deer and elk in South Korea.

In the media, CWD is often called “zombie deer disease” due to its symptoms, which include drooling, emaciation, disorientation, a vacant “staring” gaze and a lack of fear of people. As concerns about spillover to humans or other species grow, however, the moniker has irritated many scientists.

“It trivialises what we’re facing,” says epidemiologist Michael Osterholm. “It leaves readers with the false impression that this is nothing more than some strange fictional menace you’d find in the plot of a sci-fi film. Animals that get infected with CWD do not come back from the dead. CWD is a deathly serious public and wildlife health issue.”

«

CWD is like BSE (bovine spongiform encepalopathy, in cows) and variant CJD (in humans) – a prion disease. All of them can only be spread by eating infected meat (or brains 😬), which does raise the question of how all these wild deer are catching it; is there an intermediate host?

However the human risk exists:

»

The risk of a CWD spillover event is growing, the panel of experts say, and the risk is higher in states where big game hunting for the table remains a tradition. In a survey of US residents by the Centers for Disease Control and Prevention, 20% said they had hunted deer or elk, and more than 60% said they had eaten venison or elk meat.

Tens of thousands of people are probably eating contaminated game meat either because they do not think they are at risk or they are unaware of the threat. “Hunters sharing their venison with other families is a widespread practice,” Osterholm says. The Centers for Disease Control and Prevention advises people who suspect they have killed an animal infected with CWD not to eat it, and states advise any hunters taking animals from infected regions to get them tested. Many, however, do not.

«

unique link to this extract

U.S. Secretary of Health and Human Services suggests letting bird flu spread naturally through poultry farms • Phys.org

I Edwards:

»

A controversial proposal from U.S. Secretary of Health and Human Services Robert F. Kennedy Jr. to let bird flu naturally spread through poultry farms is raising alarms among scientists, who say the move could be inhumane and dangerous.

Kennedy recently suggested that instead of culling infected birds, farmers should instead allow the virus to run through flocks to identify naturally immune birds. “We can identify the birds and preserve the birds that are immune to it,” Kennedy recently told Fox News.

Though Kennedy has no direct control over farms, Agriculture Secretary Brooke Rollins has also expressed interest in testing the idea. “There are some farmers that are out there that are willing to really try this on a pilot as we build the safe perimeter around them to see if there is a way forward with immunity,” Rollins told CBS News.

But veterinary experts say this could backfire. “That’s a really terrible idea, for any one of a number of reasons,” said Dr. Gail Hansen, a former state veterinarian for Kansas, in a report published by The New York Times.

Since January 2022, bird flu has affected more than 166 million birds across every U.S. state. Experts warn that allowing the virus to spread could increase the risk of it mutating. If the bird flu were to run through a flock of five million birds, “that’s literally five million chances for that virus to replicate or to mutate,” Hansen said. It could also put farm workers and other animals at risk.

«

Truly the stupidest people who are also unwilling to listen to the advice of those who have actually studied the topic. The slightly concerned watching brief. (Thanks Joe S for the link.)
unique link to this extract

Apple and Google in the hot seat as European regulators ignore Trump warnings • Ars Technica

Ryan Whitwam:

»

The European Commission is not backing down from efforts to rein in Big Tech. In a series of press releases today, the European Union’s executive arm has announced actions against both Apple and Google. Regulators have announced that Apple will be required to open up support for non-Apple accessories on the iPhone, but it may be too late for Google to make changes. The commission says the search giant has violated the Digital Markets Act, which could lead to a hefty fine.

Since returning to power, Donald Trump has railed against European regulations that target US tech firms. In spite of rising tensions and tough talk, the European Commission seems unfazed and is continuing to follow its more stringent laws, like the Digital Markets Act (DMA). This landmark piece of EU legislation aims to make the digital economy more fair. Upon coming into force last year, the act labeled certain large tech companies, including Apple and Google, as “gatekeepers” that are subject to additional scrutiny.

Europe’s more aggressive regulation of Big Tech is why iPhone users on the continent can install apps from third-party app markets while the rest of us are stuck with the Apple App Store. As for Google, the European Commission has paid special attention to search, Android, and Chrome, all of which dominate their respective markets.

Apple’s mobile platform plays second fiddle to Android in Europe, but it’s large enough to make the company subject to the DMA. The EU has now decreed that Apple is not doing enough to support interoperability on its platform. As a result, it will be required to make several notable changes. Apple will have to provide other companies and developers with improved access to iOS for devices like smartwatches, headphones, and TVs. This could include integration with notifications, faster data transfers, and streamlined setup.

«

Europe is too big for Apple to ignore, but will it bring these changes to users outside the EU?
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

A proposed law in California might make “emitters” liable for damages such as the LA fires: but would it really stand up in court? CC-licensed photo by Scott 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 9 links for you. Nonuninflammable. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

China delays approval of BYD’s Mexico plant amid fears tech could leak to US • Financial Times

Gloria Li, Cheng Leng, Thomas Graham and Kana Inagaki:

»

Beijing is delaying approval for carmaker BYD to build a plant in Mexico amid concerns that the smart car technology developed by China’s biggest electric-vehicle maker could leak across the border to the US.

BYD first announced plans for a car plant in Mexico in 2023, along with intentions to make cars in Brazil, Hungary and Indonesia. It said the Mexican plant would create 10,000 jobs and produce 150,000 vehicles a year.

But domestic automakers require approval from China’s commerce ministry to manufacture overseas and it has yet to give approval, according to two people familiar with the matter.

Authorities feared Mexico would gain unrestricted access to BYD’s advanced technology and knowhow, they said, even possibly allowing US access to it. “The commerce ministry’s biggest concern is Mexico’s proximity to the US,” said one of the people.

Beijing is also giving preference to projects in countries that are part of China’s Belt and Road infrastructure development programme, according to the people.

Shifting geopolitical dynamics have also contributed to Mexico cooling on the plant. Mexico has sought to maintain relations with US President Donald Trump, who has put tariffs on cross-border trade, threatening exports and jobs.

«

Hard to know whether to laugh or cry at the reversal: it was always the worry of American companies that if they shifted their manufacturing to Chinese factories, their precious intellectual property would leak out. (They always did, and it always did.) BYD, of course, is the company which earlier this week announced a superfast electric vehicle charging system that would be nearly as quick as filling a car’s fuel tank.
unique link to this extract

How much does Big Oil owe Californians for the LA fires? • Breakthrough Journal

Alex Trembath, Lauren Teixeira, and Patrick Brown:

»

Last year, both New York and Vermont passed “climate Superfund” laws that impose financial liabilities on large carbon emitters for the damages caused by extreme weather and other climate impacts. California lawmakers introduced similar legislation last year, and are widely expected to do so again in the wake of Southern California’s recent devastating fires—the most expensive natural disasters in American history. With an estimated $250bn in economic damages and about 30 confirmed deaths caused by the fires, there will likely be significant political support for a law like this.

So how much would large carbon emitters owe Californians if the bill becomes law?

The legislative text allows one year for the California Environmental Protection Agency to conduct a “climate cost study” determining the total amount of climate damages owed by each large emitter, defined as institutions that are “responsible for more than 1,000,000,000 metric tons of covered fossil fuel emissions, as defined, in aggregate, globally during the covered period” of 2000-2020.

According to data collected by the UK nonprofit group InfluenceMap, there are 86 eligible entities, but most of them operate outside the United States. For the purposes of this exercise we confined our analysis to the 19 eligible entities headquartered in America.

If the bill becomes law, these large emitters collectively would be, hypothetically, liable for 0.2% of the damages from the LA fires, or just under $495.27m.

«

It’s a fascinating idea, but I do wonder how enforceable it would really be in the courts. The two biggest “entities” are ExxonMobile and Chevron, both oil companies. But, your honour, are they actually emitters? My clients simply supply the product in its unconsumed form. It is not their fault what people then do with it, such as burning it in their vehicle engines.
unique link to this extract

Apple restricts Pebble from being awesome with iPhones • Eric Migicovsky

The Pebble Is Back guy points out that you still can’t do everything you’d like to with a third-party smartwatch connected to an iPhone:

»

The problem is that 40% of everyone who signed up on rePebble.com still uses an iPhone. So we’re going to make a damn iOS app. I guess we’re gluttons for punishment. Just understand a few things:

• Our watch will always appear to have less developed functionality on iOS than Android. This is Apple’s fault, not ours.

• Some features will appear first on our Android app, and then eventually we’ll add them to the iOS app. This is because the majority of our development team uses Android phones, and generally we’re building things for ourselves, so naturally Android comes first.

I don’t want to see any tweets or blog posts or complaints or whatever later on about this. I’m publishing this now so you can make an informed decision about whether to buy a new watch or not. If you’re worried about this, the easiest solution is to buy an Android phone.

«

John Gruber’s response is, well, roll with it. I’m not sure the European Commission would see it that way.

The Pebble is still going to feel like a retro statement. Not that that won’t sell; vinyl is making a comeback, after all.
unique link to this extract

Superior replication, pathogenicity, and immune evasion of a Texas dairy cattle H5N1 virus compared to a historical avian isolate • Nature Scientific Reports

Cassio Pontes Octavani et al:

»

The current outbreak of highly pathogenic avian influenza (HPAI) viruses of the H5N1 subtype clade 2.3.4.4b in dairy cattle in the United States has affected nearly 900 dairy farms and resulted in at least 39 human infections, putting health authorities and the scientific community on high alert.

Here we characterize the virus growth properties and host-pathogen interactions of an isolate obtained from a sick dairy cow in Texas in vitro and in vivo and compare it to an older HPAI isolate. Despite so far being associated with mild disease in human patients, the cattle H5N1 virus showed superior growth capability and rapid replication kinetics in a panel of human lung cell lines in vitro. In vivo, cattle H5N1 exhibited more intense pathogenicity in mice, with rapid lung pathology and high virus titers in the brain, accompanied by high mortality after challenge via different inoculation routes.

Additionally, the cattle H5N1 demonstrated efficient antagonism of overexpressed RIG-I- and MDA5-mediated innate antiviral signaling pathways. In summary, this study demonstrates the profound pathogenicity and suggests a potential innate immune escape mechanism of the H5N1 virus isolated from a dairy cow in Texas.

«

Octavani and the other authors are all based at the University of Texas. Basically, what they found is: it’s bad! Only a watching brief, of course. 🤞
unique link to this extract

No one knows what the hell an AI agent is • TechCrunch

Maxwell Zeff and Kyle Wiggers:

»

Silicon Valley is bullish on AI agents. OpenAI CEO Sam Altman said agents will “join the workforce” this year. Microsoft CEO Satya Nadella predicted that agents will replace certain knowledge work. Salesforce CEO Marc Benioff said that Salesforce’s goal is to be “the number one provider of digital labour in the world” via the company’s various “agentic” services.

But no one can seem to agree on what an AI agent is, exactly.

In the last few years, the tech industry has boldly proclaimed that AI “agents” — the latest buzzword — are going to change everything. In the same way that AI chatbots like OpenAI’s ChatGPT gave us new ways to surface information, agents will fundamentally change how we approach work, claim CEOs like Altman and Nadella.

That may be true. But it also depends on how one defines “agents,” which is no easy task. Much like other AI-related jargon (e.g. “multimodal,” “AGI,” and “AI” itself), the terms “agent” and “agentic” are becoming diluted to the point of meaninglessness.

That threatens to leave OpenAI, Microsoft, Salesforce, Amazon, Google, and the countless other companies building entire product lineups around agents in an awkward place. An agent from Amazon isn’t the same as an agent from Google or any other vendor, and that’s leading to confusion — and customer frustration.

Ryan Salva, senior director of product at Google and an ex-GitHub Copilot leader, said he’s come to “hate” the word “agents.”

«

I’ve been hearing about software agents since, let me see, the late 1990s when British Telecom did a concept demo (yes, that vague) about how “software agents” would patrol the telephone system discovering bugs and fixing them. I have no idea if this ever happened. Anyway, now they’re back, for about the third time.
unique link to this extract

Global VR market declines 12% YoY in 2024; ‘AR+AI’ smart glasses to take centre stage in 2025 • Counterpoint Research

»

Global virtual reality (VR) headset shipments fell 12% YoY in 2024, the market’s third consecutive year of declines, according to the latest update from Counterpoint’s Global XR (AR/VR) Headset Model Tracker. In Q4 2024, the shipments fell 5% YoY. Hardware limitations, lack of compelling VR content and usage scenarios, and decreased consumer engagement continued to impact the market. However, demand from the enterprise market, though relatively limited in size, remained more resilient, particularly in large-scale immersive Location-Based Entertainment (LBE), education, healthcare and military.

Meta continued to dominate the global VR headset market in 2024 with a share of 77%. In Q4 2024, Meta’s market share rose to 84% primarily due to the launch of the more affordable Quest 3S headset. Sony’s PSVR2 shipment share surged to 9% in Q4 2024, fuelled by aggressive promotions and discounts during the Black Friday and Christmas sales. Apple’s Vision Pro shipments saw a steep 43% QoQ decline in Q4 2024, reflecting a slowdown after the initial market hype. In Q4, Apple expanded the Vision Pro’s availability to new markets, including South Korea, UAE and Taiwan, which helped partially offset the overall decline. The device’s enterprise sales also saw an uptick.

«

Meta dominates the market, with more than two-thirds of sales. Guess why? Because it’s affordable and has lots of content. Apple meanwhile isn’t even in double figures (the Q4 estimate is 2% of what should be the Christmas market). Guess why? Because it’s expensive and has basically no content.

(In passing: all hail Counterpoint, using the British spelling for “centre” in its prediction of the future for smart glasses.)
unique link to this extract

Factcheck: why Conservative leader Kemi Badenoch is wrong about UK’s net-zero goal • Carbon Brief

Simon Evans:

»

In a speech launching a “policy renewal programme” to shape the Conservatives’ approach to key issues, Badenoch disowned the target passed into law by her own party in 2019.

She offered no alternative to the 2050 net-zero target and failed to cite any evidence in support of her assertion that meeting it would be “impossible” without “bankrupting” the country.

…Despite the clear evidence of the need to reach net-zero emissions to stop global warming, Badenoch said in her speech that reaching the target by 2050 was “impossible”. She did not offer any evidence to support this supposedly “unvarnished truth”.

Announcing the adoption of the target in 2019, Conservative then-secretary of state Greg Clark said that it was “necessary and feasible”, pointing to the CCC’s advice as evidence.

Indeed, the 2019 advice set out in detail how it would be “feasible” to cut UK emissions to net-zero by 2050. In its latest advice to the government, the CCC set out a “balanced pathway” to net-zero by 2050 that showed the target was “feasible and deliverable”.

Similarly, in 2024 the National Energy System Operator (NESO) published three “credible” and “affordable” pathways to net-zero by 2050, as part of its annual “future energy scenarios”. It said: “Our net-zero pathways identify three credible, strategic routes to reach net-zero…Decisive action is needed within the next two years to deliver the fundamental change required for a fair, affordable, sustainable and secure net-zero energy system by 2050.”

A peer-reviewed research paper in 2022 identified and compared seven pathways to net-zero by 2050, published by four different organisations.

Directly contradicting Badenoch’s speech, the study concluded that “the breadth of pathways analysed in this paper has shown that there are several possible routes to net-zero”. Moreover, the Conservative government in 2021 published its own strategy for reaching net-zero by 2050, including an entire section titled “why net-zero”.

In a foreword to the 2021 strategy, then-Conservative prime minister Boris Johson wrote that “reaching net-zero is entirely possible”.

«

It is astonishing how the Tories’ latest policies suggest that whoever was in charge between 2010 and 2024 was completely mad and useless. Robert Hutton, sketch writer for The Critic, calls them Hot Dog Tories – after the sketch of the Hot Dog Guy who has crashed his Hot Dog Car into a shop saying “we’re all trying to find the guy who did this” – and it’s completely true.
unique link to this extract

The Environmental Protection Agency (EPA) shifts its mission • The New York Times

Lisa Friedman and Hiroko Tabuchi:

»

The E.P.A. has “no obligation to promote agriculture or commerce; only the critical obligation to protect and enhance the environment,” the first administrator, William D. Ruckelshaus, said as he explained its mission to the country weeks after the E.P.A. was created by President Richard M. Nixon. He said the agency would be focused on research, standards and enforcement in five areas: air pollution, water pollution, waste disposal, radiation and pesticides.

Perhaps the most significant of the agency’s regulatory changes is an effort to revise a 2009 legal opinion known as the E.P.A. “endangerment finding,” which concluded that rising greenhouse gas emissions are a danger to public health.

The finding gives the agency the authority to regulate greenhouse gas emissions. Eliminating it would make it virtually impossible for the E.P.A. to curb climate pollution from automobiles, factories, power plants or oil and gas wells.

Some of the other significant policy changes Zeldin said he planned include:
• Rolling back restrictions on carbon dioxide emissions from power plants. Currently the E.P.A. requires existing coal-burning power plants and new gas plants built in the United States to cut their greenhouse-gas emissions by 90% by 2039.

• Rewriting tailpipe pollution standards that were designed to ensure that the majority of new passenger cars and light trucks sold in the United States are all-electric or hybrids by 2032.

• Easing limits on mercury emissions from power plants, as well as restrictions on soot and haze from burning coal. A Biden-era rule had aimed to slash by 70% emissions from coal-burning power plants of mercury, which has been linked to developmental damage in children.

• Greatly reducing the “social cost” of carbon, an economic estimate of the damage caused by each additional ton of carbon dioxide emissions in the atmosphere. That figure plays a significant role in weighing the costs and benefits of regulating industries.

«

This is completely insane. It’s the Environmental “Eh Who Cares” Agency. It might as well not exist; it isn’t going to do anything to protect the environment, nor the people who rely on it (ie everyone). Pollution levels will rise, people will get ill and/or die, and that cohort will include the rich who are old – air doesn’t discriminate.

Also, excuse me, but: “the EPA shifts its mission“? I know the American papers and the NYT particularly are terrible at headlines, but wouldn’t “abandons its mission” be more truthful? Or “says yes to pollution”? Almost anything would be more accurate.
unique link to this extract

Bryan Burrough on Graydon Carter’s memoir and Vanity Fair’s heyday • The Yale Review

Bryan Burrough is a successful author in his own right, but also wrote for Vanity Fair in its heyday – the 1990s:

»

By 1994, Vanity Fair was “hugely profitable,” and Graydon received unyielding support from the magazine’s parent company, Condé Nast, and its owners, the Newhouse family—especially the enigmatic Si Newhouse, who adored magazines and adored Graydon. Not to mention the river of Newhouse cash that flowed Vanity Fair’s way. Oh, the money. Good Lord, the money.

As i look back today, Graydon’s Vanity Fair does feel like some lost world, a gold-encrusted Atlantis ultimately inundated by economic and technological tsunamis, its glories only now being picked over by media anthropologists. I’ve never talked much about what it was like to write there. Because I have always worried about how I’d come off. I mean, the money alone. I’m probably breaking some unwritten law of publishing, but here it is: For twenty-five years, I was contracted to produce three articles a year, long ones, typically ten thousand words. For this, my peak salary was $498,141. That’s not a misprint—$498,141, or more than $166,000 per story. Then, as now, $166,000 was a good advance for an entire book. Yes, I realized it was obscene. I took it with a grin.

Then there was the Hollywood money. Every third or fourth article I wrote ended up optioned for the movies. Most were in the $15,000 to $25,000 range for a renewable eighteen-month option. A handful crossed into six figures. (You haven’t lived until you’ve sat across from Robert De Niro on a film set as he reads your own words back to you—although, sadly, that adaptation of my piece “The Miranda Obsession” never made it past development.) This was an era when management allowed writers to keep that movie money. These days? One magazine I love takes 90% off the top.

I am aware of peers who did just as well. Nowadays, though, such windfalls are a distant memory. Today, for a rare magazine article, I’m lucky to receive two dollars a word, or $20,000 for that same ten-thousand-word story. (Don’t even ask what they’re paying me for this piece.) People sometimes wonder why I don’t write more. It’s a chore to explain that, at these rates, it is hard to get that excited.

«

Burrough gets TWENTY THOUSAND DOLLARS for a magazine article?! I don’t think he realises what life is like for the many, many freelancers scratching around for £150 per thousand words, and never getting a sniff of being asked to write ten thousand.

It’s like glimpsing another world. The article is fascinating. You can read it for free, so don’t ask me how or what the Yale Review pays for it.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

The UK’s DVLA managed to update many of its processes for the web – but still needs to rely on daily batch jobs. CC-licensed photo by Amy Whitney 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 9 links for you. Licensed. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

GLP-1s are crazy effective — just not cost-effective • Medscape

F Perry Wilson:

»

How much is your life worth?

To an economist, the answer is basically $100,000 per year of perfect health. The ways they arrive at this number are pretty fascinating, but a lot of it is done by looking at what we as a society are willing to pay for. A cost of $100,000 per “quality-adjusted life year” (QALY) is our standard candle here. More than that is not particularly cost-effective. Less than that is. The “quality” in the QALY is important too. Economists know that a year of perfect health is worth more, in dollar terms, than a year of moderate or poor health.

…The authors modeled the cost of the [semaglutide] drugs, the cost of medical care for people taking the drugs, even the cost of lifestyle modification alone (gym memberships aren’t free). Over the lifetime of a given individual, the total cost of a policy of using lifestyle interventions to treat overweight and obesity would be $244,000. A policy of lifestyle interventions plus tirzepatide? $313,000 dollars. Sure, the cost of medical care is lower with the tirzepatide strategy; those averted cases of heart disease and diabetes save money — about $30,000 per person over their lifetime. But the cost of the drug itself adds up. At current prices (about $1000 a month), we’re talking $111,000 dollars per person. 

Now that we have a measure of the effectiveness of the drugs and a measure of the cost, we can do some division and calculate the cost per QALY gained. And here’s what you see.

For tirzepatide, the most effective of the drugs, about $200,000 per quality-adjusted life year. For semaglutide? $470,000 per QALY (since it’s less effective and similarly priced). The older, less used drugs are remarkably more cost-effective: $85,000 for the phentermine drug, and Contrave actually saves money. Society saves $2500 per QALY for using this drug, even though it doesn’t work as well as the pricier stuff.

Does that mean we should abandon these amazingly effective agents? Definitely not. I love a drug that works, and these drugs work. They’re just, quite simply, too expensive. If we want to bring them down below the $100,000 per QALY threshold for cost-effective treatments, the price of tirzepatide needs to decrease by 30%, and the price of semaglutide by 82%. Honestly, I suspect Lilly and Novo would do fine at these price points, but what do I know? I’m not an economist.

«

I remain fascinated by the social effects of these drugs. If they get to the point where NICE in the UK thinks the price v QALYs equation is right, everyone gets on board.
unique link to this extract

Apple’s Passwords app was vulnerable to phishing attacks for nearly three months after launch • 9to5Mac

Arin Waichulis:

»

In iOS 18, Apple spun off its Keychain password management tool—previously only tucked away in Settings—into a standalone app called Passwords. It was the company’s first move at making credential management more convenient for users. It’s now been revealed that a serious HTTP bug left Passwords users vulnerable to phishing attacks for nearly three months, from the initial release of iOS 18 until the patch in iOS 18.2.

Security researchers at Mysk first discovered the flaw after noticing that their iPhone’s App Privacy Report showed Passwords had contacted a staggering 130 different websites over insecure HTTP traffic. This prompted the duo to investigate further, finding that not only was the app fetching account logos and icons over HTTP—it also defaulted to opening password reset pages using the unencrypted protocol. “This left the user vulnerable: an attacker with privileged network access could intercept the HTTP request and redirect the user to a phishing website,” Mysk told 9to5Mac.

“We were surprised that Apple didn’t enforce HTTPS by default for such a sensitive app,” Mysk states. “Additionally, Apple should provide an option for security-conscious users to disable downloading icons completely. I don’t feel comfortable with my password manager constantly pinging each website I maintain a password for, even though the calls Passwords sends don’t contain any ID.”

Most modern websites nowadays allow unencrypted HTTP connections but automatically redirect them to HTTPS using a 301 redirect. It’s important to note that while the Passwords app before iOS 18.2 would make a request over HTTP, it would redirected to the secure HTTPS version. Under normal circumstances, this would be totally fine, as the password changes occur on an encrypted page, ensuring that credentials are not sent in plaintext.

However, it becomes a problem when the attacker is connected to the same network as the user (i.e. Starbucks, airport, or hotel Wi-Fi) and intercepts the initial HTTP request before it redirects.

«

Mysk were “surprised”? It’s a shockingly bad piece of app design that should never have come through testing. How could someone write a piece of code that says “http” and not think “wait, that should probably be https, shouldn’t it”?

Apple has made lots of software mistakes in the past, but I can’t think of such an obviously avoidable one offhand. The discussion about software quality at Apple will intensify; apart from anything, who was asking for passwords to be hived off into a new app with such a wonderful bug?
unique link to this extract

Apple innovation and execution • Benedict Evans

Evans considers the delay to the new Siri in the context of what Apple has done in the past, then looks ahead:

»

a year is a long time given the speed of AI progress right now, especially given the ferocity of competition that Apple faces in China and the waves of new features that the OEMs there are pushing. And ‘Apple Intelligence’ certainly isn’t going to drive a ‘super-cycle’ of iPhone upgrades any time soon.

Indeed, a better iPhone feature by itself was never going to drive fundamentally different growth for Apple, but failures like Humane and Rabbit point to what else Apple (or others) might do with this technology once it works. The rumoured new home smart-screen device is probably a lot less appealing without this, and the AR glasses would need this too, except that those really are years away.

However, it clearly is a problem that the Apple execution machine broke badly enough for Apple to spend an hour at WWDC and a bunch of TV commercials talking about vapourware that it didn’t appear to understand was vapourware. The decision to launch the Vision Pro looks like a related failure. It’s a big problem that this is late, but it’s an equally big problem that Apple thought it was almost ready.

And the failure of Siri 2 is by far the most dramatic instance of a growing trend for Apple to launch stuff late. The software release cycle used to be a metronome: announcement at WWDC in the summer, OS release in September with everything you’d seen. There were plenty of delays and failed projects under the hood, and centres of notorious dysfunction (Apple Music, say), and Apple has always had a tendency to appear to forget about products for years (most Apple Watch faces don’t support the key new feature in the new Apple Watch) but public promise were always kept. Now that seems to be slipping. Is this a symptom of a Vista-like drift into systemically poor execution?

On the other hand, I’m old enough to remember when people said Apple was going to miss Machine Learning, and narratives are always easy to build when something’s gone wrong.

«

Certainly there are lots of people prepared to say “nobody knew what the LLM bit was going to do! Ordinary people weren’t looking forward to it, so it’s no loss if it’s delayed!” But people who watch Apple closely know that that isn’t the point. To announce and then backpedal is not the Apple way.
unique link to this extract

The first new Pebble smartwatches are coming later this year • The Verge

David Pierce:

»

The first watch that Migicovsky and [his company] Core plan to ship is called the Core 2 Duo (not to be confused with the old Intel processor), which Migicovsky says will cost $149 and will ship in July. The name explains the whole idea, he says: “It’s like a Pebble 2, but it’s made by Core devices. And then ‘Duo’ is for do-over.” It has the exact same black-and-white e-paper display as the old Pebble 2 (technically a transflective LCD, if you’re curious), and it even comes in the exact same frame. “We were able to find a supplier that still had the frames for Pebble Time 2 and Pebble 2,” he says. “They were never used. So we’ve been able to just draft on that.”

The Core 2 Duo does get a couple of upgrades, mostly by virtue of overall technological progress — Migicovsky says the new watch will last more than 30 days, instead of the Pebble 2’s seven, largely because Bluetooth chips have become so much more efficient. There’s also a speaker in the device now, which Migicovsky uses for chatting with AI assistants. Overall, though, this is an 8-year-old device simply made new again. (This is part of the idea behind the Pebble reboot: Migicovsky is convinced that Pebble mostly had it right a decade ago and simply wants to get back to that.) He estimates there will be around 10,000 Core 2 Duos available and figures a lot of developers and hardcore fans will be happy to have a new watch to play with as soon as possible.

«

I recall an analyst saying once that you can probably sell 100,000 of any new hardware – it’s the next 7,999,900,000 that are the problem. Can Migicovsky make a profit from that small group?
unique link to this extract

Sobering revenue stats of 70K mobile apps show why devs beg for subscriptions • Ars Technica

Scharon Harding:

»

If you’re frustrated by some of your favorite apps pestering you to sign up for a subscription, some new data may help you empathize with their developers more. According to revenue data from “over 75,000” mobile apps, the vast majority have a hard time making $1,000 per month.

The data is detailed in RevenueCat’s 2025 State of Subscription Apps report. RevenueCat makes a mobile app subscription tool kit and gathered the report’s data from apps using its platform. The report covers “more than $10bn in revenue across more than a billion transactions,” and RevenueCat’s customer base ranges from indie-sized teams to large publishers. Buffer, ChatGPT, FC Barcelona, Goodnotes, and Reuters are among the San Francisco-based firm’s customer base.

Additionally, the report examines apps that rely primarily on in-app subscriptions, as well as those that only generate some revenue from subscriptions. All apps examined, though, actively generate subscription revenue and “meet a minimum threshold of installs or revenue (to ensure statistically meaningful findings,” according to the report.

RevenueCat’s report doesn’t cover every single mobile app available, but it paints a picture of the challenges related to monetizing mobile apps across different types of categories, as well as how uneven the distribution of app revenue is.

RevenueCat’s report concluded that most apps fail to make $1,000 in monthly revenue within their first two years. It says: “Across all categories, nearly 20% reach $1,000 in revenue, while 5% reach the $10,000 mark. Revenue drop-off is steep, with many categories losing ~50% of apps at each milestone, emphasizing the challenge of sustained growth beyond early revenue benchmarks.”

«

$12,000 per year really isn’t living wages. So it’s not surprising that scams proliferate from less honest developers. The internet is big, but it’s also parsimonious.
unique link to this extract

Britain’s car parking is a complete disaster • The Value of Nothing

Martin Robbins:

»

Car parks are a key part of something I like to call the ‘National User Experience’: a category of things that have been largely abandoned by government yet have a major impact on people’s daily lives and their perception of how the country is being run. Car parks, potholes, neglected public spaces, boarded-up shopping centres, persistent antisocial behaviour, the punctuality of trains – things people encounter day after day after day that make life a little bit harder, a little bit more miserable.

When Labour tries to get reelected in 2029, yes national issues will play a big part, but I think Westminster politicians massively underestimate the impact of unglamorous daily drudgery on much of the population: fix potholes and parking and you’re showing visible change to a lot of grateful people. Fail, and whatever you do at national level is overshadowed by the continuing enshittification of Britain.

In a sane world, all public transport including car parking would be wrapped up in a single, easy payment system. You should be able to tap in and out of car parks with an Oyster card or equivalent, with the parking charge rolled into your ticket. If you want to get people out of cars and onto public transport, providing a frictionless way to move between the two is a good start, particularly outside London where bus services will never be sufficient and personal transport is a necessity for many people.

«

I know what you’re thinking: don’t worry, Martin, the government’s funding a single-payer app! But as Robbins points out, the government cancelled it in February, citing fiscal straitjackets.

From the article about the cancellation:

»

There are thought to be at least 30 different parking apps in the UK, and it is not unusual for someone to have a number on their phone. Among the biggest are RingGo, PayByPhone and JustPark.

A survey published in 2024 by Autocar, however, found that “more than four in five motorists dislike using car parking apps”, with 83% saying they preferred to use cash or contactless card payments.

«

This is not good for the National User Experience.
unique link to this extract

Why some DVLA digital services don’t work at night • Dafydd Vaughan

Vaughan worked at the UK’s Driver and Vehicle Licensing Agency, and was there when it tried to make everything webby:

»

As part of the GDS [Government Digital Service] exemplar programme in 2013, DVLA committed to delivering a set of new digital services for managing vehicles and personal registrations. To deliver these services, we had to navigate the complexity of the existing tech in place.

Building a new front-end service would be relatively straightforward. However, updating the vehicle record would be more complex – we’d have to integrate with the legacy systems and deal with IBM/Fujitsu to do it. But the even bigger issue – how would we deal with the fragile overnight batch jobs?

We faced a choice. Step back and spend the next few years redesigning and rebuilding the underlying infrastructure to remove/remediate the overnight batch jobs, or accept the service couldn’t initially operate overnight.

Organisations often fall into this trap – spending years and huge amounts of money fixing the underlying foundations before starting to do new things. It’s difficult for an organisation to keep its focus and attention on a complex upgrade – particularly without getting noticeable benefits along the way. DVLA tried this in the early 2000s when migrating away from the mainframe. They ran out of money, and ended up in an even worse half-state.

I pushed for us to press on and deliver a service that could operate normally during the day, but would be turned off overnight. This would allow us to get some value early – giving people access to a new service quickly, while we looked to fix the issues behind the scenes. Luckily the political pressure of the exemplar programme supported us to do that.

«

So they did create a system that in effect still ran batch jobs – stopping taking new entries at some point in the 24 hours so all the records could be updated without creating conflicts.

Two neat codas in his writeup:

»

It’s now 2024 – 10 years on from the launch of the first service. The legacy infrastructure, which really should have been replaced by now, is probably still the reason why the services are still offline overnight.

Is this acceptable? Not really. Is it understandable? Absolutely.

«

And:

»

Transforming government services isn’t as easy as the tech bros and billionaires make it out to be.

«

unique link to this extract

Former Meta executive Sarah Wynn-Williams on her “Careless People” memoir • Business Insider

Pranav Dixit interviewed Sarah Wynn-Williams, who makes some interesting points about why she wrote her book:

»

PD: You left Facebook in 2017. Why did you decide to release it now, after all these years?

SWW: Because I think we’re on the cusp of this new era of technology. We’re stepping into this AI era, and at a high level, I don’t want the mistakes that were made during the social media era to be applied to the AI era.

One of the things that I’ve worked on since leaving [Meta] is the US-China AI dialogue on AI in weapons. So, I really understand the existential nature of AI. I also understand these people and how decisions are made. That’s why, as we go into this new era, we have to do it better. China is such a big part of the story of AI. It’s this growing strategic rivalry and how technology is so central to that rivalry.

And yet, this company has been doing things in the shadows for so long with the Chinese Communist Party, and their line is, “Oh, you know, we tried to get our services [into China] and we told you in 2019 that didn’t happen.” Have a look at how much of [Meta’s] revenue comes from China — it’s $18 billion.

(Editor’s note: According to Meta’s 2024 annual report, the company made $18.35 billion from China, primarily through resellers serving Chinese advertisers targeting global users.) So it seems that everyone is operating under the false notion that Meta is not operating in China when actually, it is fundamental to its current valuation, it’s fundamental to its future growth. And we don’t talk openly about it at the very time that we’re about to enter this new AI era.

…PD: A Meta spokesperson, Andy Stone, has said that your book wasn’t fact-checked and that nobody reached out to Meta for comment. Did you get the book fact-checked?

SWW: I think Meta’s problem is using this to not answer the questions themselves. What I would love is for us not to fall into the distraction. There’s a real risk that we talk about things that don’t matter. We’ve got these huge issues like China and I notice they’re not providing any detail on that. There are so many smart people who’ve worked at this company and who are covering this company. Like, we have to do better.

«

She gave the interview just before the arbitration ruling she couldn’t publicise her book. So this is worth reading now.

Also, it only just occurred to me that the title comes from The Great Gatsby: “They were careless people, Tom and Daisy – they smashed up things and creatures and then retreated back into their money or their vast carelessness or whatever it was that kept them together, and let other people clean up the mess they had made.”
unique link to this extract

How bribes helped a crime ring steal thousands of iPhones from porches • WSJ via MSN

Esther Fung:

»

[Phone shop] Wyckoff Wireless looked like many other mom-and-pop shops around New York City—except federal agents were surveilling the scene, as they later recounted in a criminal complaint.

They suspected the wireless shop was a fence, or middleman, that authorities say was being used to move thousands of stolen iPhones. Last month, federal authorities arrested 13 people in connection with what they say was an international crime ring that targeted FedEx deliveries nationwide.

Porch thefts aren’t new, but they have become increasingly sophisticated. There was a spree last year—captured on doorbell cameras—where thieves stole iPhones just moments after they were dropped on front steps. They knew when the packages were coming and what was inside.

The Wyckoff Wireless case reveals how authorities say they did it: by harnessing technology and old-fashioned bribery.

The group created software to scrape FedEx tracking numbers and bribed AT&T store employees to get order details and delivery addresses, according to a criminal complaint filed in New Jersey. The group then sent thieves to pick off the packages and bring them back to destinations like the Brooklyn shop.

The software was created by Demetrio Reyes Martinez, who is known online as “CookieNerd,” according to the complaint. The 37-year-old wrote code to get around FedEx limits on delivery-data requests and sold it via Telegram with instructions on how to run the program, prosecutors said.

Reyes Martinez, a citizen and resident in the Dominican Republic, is still in the Caribbean nation, according to the U.S. Attorney’s Office in New Jersey, which declined to provide further information on his status.

…An AT&T store employee in Paterson, New Jersey, Alejandro Castillo, used his employment credentials to track hundreds of shipments that were subsequently reported stolen in transit, prosecutors said. He took photos of customers’ names, addresses and tracking numbers and shared them with the criminal group, according to the complaint.

He also worked with another store employee in Fort Lee, New Jersey, and recruited other employees at the cellular carrier, prosecutors said. Law-enforcement officials believe Castillo was receiving $2,000 to $2,500 if he recuirted other employees.

«

Because Americans capitalise first letters in headlines (I take them down), I saw “From Porches” and thought the phones were stolen from Porsches and thought that was a bit niche to hit thousands. A theft from a porch, though? Sure.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

The tolerances required for the ball bearing of a ballpoint pen eluded Chinese manufacturing for decades, and only in the past few years has it met them. CC-licensed photo by Andrew Magill 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.

Last week’s Social Warming Substack post is available for your delectation.

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

AI slop is a brute force attack on the algorithms that control reality • 404 Media

Jason Koebler:

»

Consider, for a moment, that this AI-generated video of a bizarre creature turning into a spider turning into a nightmare giraffe inside of a busy mall has been viewed 362 million times. That means this short reel has been viewed more times than every single article 404 Media has ever published, combined and multiplied tens of times. 

Any of these Reels could have been and probably was made in a matter of seconds or minutes. Many of the accounts that post them post multiple times per day. There are thousands of these types of accounts posting thousands of these types of Reels and images across every social media platform. Large parts of the SEO industry have pivoted entirely to AI-generated content, as has some of the internet advertising industry. They are using generative AI to brute force the internet, and it is working. 

One of the first types of cyberattacks anyone learns about is the brute force attack. This is a type of hack that relies on rapid trial-and-error to guess a password. If a hacker is trying to guess a four-number PIN, they (or more likely an automated hacking tool) will guess 0000, then 0001, then 0002, and so on until the combination is guessed correctly. 

As you may be able to tell from the name, brute force attacks are not very efficient, but they are effective. An attacker relentlessly hammers the target until a vulnerability is found or a password is guessed. The hacker is then free to exploit that target once the vulnerability is found.

The best way to think of the slop and spam that generative AI enables is as a brute force attack on the algorithms that control the internet and which govern how a large segment of the public interprets the nature of reality. It is not just that people making AI slop are spamming the internet, it’s that the intended “audience” of AI slop is social media and search algorithms, not human beings.

What this means, and what I have already seen on my own timelines, is that human-created content is getting almost entirely drowned out by AI-generated content because of the sheer amount of it. On top of the quantity of AI slop, because AI-generated content can be easily tailored to whatever is performing on a platform at any given moment, there is a near total collapse of the information ecosystem and thus of “reality” online. I no longer see almost anything real on my Instagram Reels anymore, and, as I have often reported, many users seem to have completely lost the ability to tell what is real and what is fake, or simply do not care anymore.

«

It’s that latter point which is relevant. People don’t care. They don’t reject it and so it inveigles its way into their content, and the platforms don’t care because it all gets viewed and keeps people on there.
unique link to this extract

AI search has a citation problem • Columbia Journalism Review

Klaudia Jaźwińska and Aisvarya Chandrasekar:

»

AI search tools are rapidly gaining in popularity, with nearly one in four Americans now saying they have used AI in place of traditional search engines. These tools derive their value from crawling the internet for up-to-date, relevant information—content that is often produced by news publishers. 

Yet a troubling imbalance has emerged: while traditional search engines typically operate as an intermediary, guiding users to news websites and other quality content, generative search tools parse and repackage information themselves, cutting off traffic flow to original sources. These chatbots’ conversational outputs often obfuscate serious underlying issues with information quality. There is an urgent need to evaluate how these systems access, present, and cite news content.
Building on our previous research, the Tow Center for Digital Journalism conducted tests on eight generative search tools with live search features to assess their abilities to accurately retrieve and cite news content, as well as how they behave when they cannot.

We found that…
• Chatbots were generally bad at declining to answer questions they couldn’t answer accurately, offering incorrect or speculative answers instead
• Premium chatbots provided more confidently incorrect answers than their free counterparts
• Multiple chatbots seemed to bypass Robot Exclusion Protocol preferences
• Generative search tools fabricated links and cited syndicated and copied versions of articles
• Content licensing deals with news sources provided no guarantee of accurate citation in chatbot responses.

Our findings were consistent with our previous study, proving that our observations are not just a ChatGPT problem, but rather recur across all the prominent generative search tools that we tested. 

«

One of the sections is headed “Chatbots’ responses to our queries were often confidently wrong”, which certainly sums up a lot of what one sees.
unique link to this extract

AI failed to detect critical health conditions: study • Axios

Maya Goldman:

»

Hospitals increasingly use tools that harness machine learning, a subset of AI that focuses on systems that continuously learn and adjust as they’re given new data.

A separate study recently published in Health Affairs found that about 65% of U.S. hospitals use AI-assisted predictive models, most commonly to figure out inpatient health trajectories.

Researchers looked at several machine learning models commonly cited in medical literature for use in predicting patient deterioration and fed them publicly available sets of data about the health and metrics of patients in ICUs or with cancer.

The researchers then created test cases for the models to predict potential health issues and risk scores if some patient metrics were altered from the initial data set.

The models for in-hospital mortality prediction could only recognize an average of 34% of patient injuries, the study found.

What they’re saying: “We are asking the models to make big decisions, and so we really need to figure out … in what kind of situations they can perform,” said Danfeng (Daphne) Yao, an author of the study and a computer science professor at Virginia Tech.

«

unique link to this extract

Here’s a “dead” person on Social Security in Seattle, with plenty to say • The Seattle Times

Danny Westneat:

»

“DOGE Has 10 Staffers at Social Security in Hunt for Dead People,” the headlines read this past week.

I found a dead person on Social Security. Right here in Seattle, on Capitol Hill.

Of course the circumstances of Ned Johnson’s death were completely the opposite of what Elon Musk and his Department of Government Efficiency had claimed was rampant.

“You wake up one day and discover you’re dead,” Johnson told me. “It’s been truly surreal.”

That’s the biggest difference — my deceased guy turns out to be very much alive. Musk is contending that hordes of dead people are listed as alive in the Social Security databases, and are fraudulently still drawing benefits (which the Social Security director disputes).

Johnson is 82 and still kicking. Yet sometime last month, someone or something led Social Security to both tag him as dead and start clawing back his benefits.

Johnson’s strange trip through the netherworld began in February, when a letter from his bank arrived addressed to his wife, Pam. “We recently received notification of LEONARD A. JOHNSON’s passing,” it began. “We offer our sincerest condolences …”

At first she figured it was a scam — her husband, after all, was sitting right there. But then the bank got to the point. “We know this is a difficult time, and we’re here to help,” the bank wrote. “We received a request from Social Security Administration to return benefits paid to LEONARD A. JOHNSON’s account after their passing.”

“There’s nothing you need to do — we’ve deducted the funds from LEONARD A. JOHNSON’s account.” Uh oh. It itemized how $5,201 had been stricken from their bank account, on the grounds that Ned wasn’t justified to get those benefits — because he was dead. That was for payments he’d received in December and January.

…What followed was a nearly three-week battle to resurrect himself. He called Social Security two or three times a day for two weeks, with each call put on hold and then eventually disconnected. Finally someone answered and gave him an appointment for March 13. Then he got a call delaying that to March 24.

In a huff, he went to the office on the ninth floor of the Henry Jackson Federal Building downtown. It’s one of the buildings proposed to be closed under what the AP called “a frenetic and error-riddled push by Elon Musk’s budget-cutting advisers.”

It was like a Depression-era scene, he said, with a queue 50-deep jockeying for the attentions of two tellers. The employees were kind but beleaguered.

«

This scene is going to be repeated all over the US, and journalism is going to be almost powerless to describe it – individual stories like this don’t give any scale, and stories of thousands of people being denied benefits lack focus.
unique link to this extract

Top broadband official exits Commerce Department with sharp Musk warning • POLITICO

John Hendel:

»

A top Commerce Department official sent a blistering email to his former colleagues on his way out the door Sunday warning that the Trump administration is poised to unduly enrich Elon Musk’s satellite internet company with money for rural broadband.

The technology offered by Starlink, Musk’s company, is inferior, wrote Evan Feinman, who had directed the $42.5bn broadband program for the past three years.

“Stranding all or part of rural America with worse internet so that we can make the world’s richest man even richer is yet another in a long line of betrayals by Washington,” Feinman said.

Feinman’s lengthy email, totaling more than 1,100 words and shared with POLITICO, is a sign of deep discomfort about the changes underway that will likely transform the Broadband Equity, Access and Deployment Program. Commerce Secretary Howard Lutnick recently pledged a vigorous review of BEAD, with an aim to rip out what he sees as extraneous requirements and remove any preference for particular broadband technologies like fiber.

Musk, who runs the Starlink satellite broadband service, stands to reap a greater share of these subsidies under the revised rules.

Musk and Starlink did not respond to requests for comment.

The program, created in the 2021 infrastructure law program, became a source of partisan fighting last year on the campaign trail as Republicans attacked the Biden administration for its slow pace. No internet expansion projects have begun using BEAD money, although some states were close at the beginning of this year.

Feinman’s critique: In his email, Feinman notes Friday was his last day leading BEAD and that he’s “disappointed not to be able to see this project through.”

«

Certainly the problem with fixed internet is.. getting it in. Once that’s done, it’s cheap. Satellite internet is quick to get in, but expensive forever.
unique link to this extract

Alphabet spins off laser-based internet project from “moonshot” hub • Financial Times

Stephen Morris:

»

Alphabet is spinning out laser-based internet company Taara from its “moonshot” incubator, hoping to turbocharge the start-up that provides high-bandwidth services to hard-to-reach areas in competition with Elon Musk’s Starlink network of satellites.

Taara is the latest project to spring from X — Alphabet’s experimental hub that produced artificial intelligence lab Google Brain and Waymo’s self-driving cars — and has its origins in a concept called Loon. That envisaged shooting beams of light between thousands of balloons floating on the edge of space to provide phone and internet services across remote areas.

Loon was wound up in 2021 because of political and regulatory hurdles to flying the balloons and the difficulty of servicing the 20-mile-high equipment. However, its lasers found a second life on Taara’s towers under engineer Mahesh Krishnaswamy.

The technology works by firing a beam of light the width of a pencil from one traffic light-sized terminal to another, using a system of sensors, optics and mirrors to fix it on a 1.5 inch receiver. Alphabet says the system can transmit data at 20 gigabits per second over 20km, extending traditional fibre-optics networks with minimal construction and lower costs.

Based in Sunnyvale near Google’s headquarters in Mountain View, California, Taara has two dozen staff and is hiring aggressively. The start-up has secured backing from Series X Capital and Alphabet will retain a minority stake, but the company refused to disclose any details about its seed funding or financial targets.

«

Feels like a long time since Alphabet spun off something from its moonshot.
unique link to this extract

Regional newspaper ABCs: no daily now has print circulation of 20,000 or more • Press Gazette

Bron Maher:

»

Print circulations at UK regional daily newspapers collectively declined 16% between 2023 and 2024, according to the latest data from ABC.

That marks a slight slowing after circulations dropped 17% between the first halves of 2023 and 2024.

But it nonetheless marks a milestone as circulations at the last two regional dailies to circulate 20,000 print copies or more per day, the Irish News and Aberdeen’s Press & Journal, sank below that benchmark and the Evening Standard [in London] ceased as a daily title.

The Standard’s average daily circulation in its final month reported to ABC, August, was 273,631. The weekly London Standard is now the non-daily regional paper with the largest circulation, averaging 148,021 copies per issue between October and December.

Total circulations per issue at the 417 non-daily UK local newspapers audited by ABC fell at a slightly slower rate of 14% between 2023 and 2024.

Across the 69 dailies, the total number of single copies sold at newsstands circulated per day also dropped 16%, to 279,000. Paid subscriptions fell 15% to 67,000 per day and free copies dropped more than a third to a total average of 1,364 per day.

«

The golden days for local papers were 1985 to 2004, according to this article in the New Statesman (free registration). The slope now is very steep, and downhill.

Losing local coverage means the loss of so much accountability.
unique link to this extract

BYD unveils new super-charging EV tech, to build charging network in China • Reuters

Qiaoyi Li, Zhang Yan and Brenda Goh:

»

BYD on Monday unveiled a new platform for electric vehicles (EVs) that it said could charge EVs as quickly as it takes to pump gas and announced for the first time that it would build a charging network across China.

The so-called “super e-platform” will be capable of peak charging speeds of 1,000 kilowatts (kW), enabling cars that use it to travel 400 km (249 miles) on a five-minute charge, founder Wang Chuanfu said at an event livestreamed from the company’s Shenzhen headquarters.

Charging speeds of 1,000 kW would be twice as fast as Tesla’s superchargers whose latest version offers up to 500 kw charging speeds. Fast-charging technology has been key to increasing EV adoption as it is seen to help assure EV drivers’ concerns over being able to charge their cars quickly.

“In order to completely solve our user’s charging anxiety, we have been pursuing a goal to make the charging time of electric vehicles as short as the refuelling time of petrol vehicles,” Wang said.

“This is the first time in the industry that the unit of megawatt (charge) has been achieved on charging power,” he said.
The new charging architecture will be initially available in two new EVs – Han L sedan and Tang L SUV priced from 270,000 yuan ($37,330) and BYD said it would build over 4,000 ultra-fast charging piles, or units, across China to match the new platform.

The company didn’t specify the time frame or how much it would invest in building such facilities. To date, BYD owners have largely relied on other automakers’ charging facilities or public charging poles run by third-party operators to charge their vehicles.

«

This is certainly the part of the trifecta to make EVs completely acceptable: plentiful locations with rapid charging at low prices. It’s the latter which may still be out of reach.
unique link to this extract

2017: Finally, China manufactures a ballpoint pen all by itself • The Washington Post

Adam Taylor, in 2017:

»

To anyone outside of the ballpoint pen manufacturing world, it might seem hard to understand what, exactly, is so surprising about this development. China already produces 38 billion ballpoint pens a year, according to China Daily, which is about 80% of all ballpoint pens in the world. That’s a lot of pens, but there was a catch: China had long been unable to produce a high-quality version of the most important part of the pen, its tip.

The tip of a ballpoint pen is what makes it a ballpoint pen. At the tip, a freely rotating ball is held in a small socket which connects it to an ink reservoir that allows the pen to write or draw lines. Manufacturing a ballpoint pen tip that can write comfortably for a long period of time requires high-precision machinery and precisely thin steel, but for years China was unable to match those crafted by foreign companies.

While there were over 3,000 companies manufacturing pens in China, none had their own high-end technology for the tip. Instead, about 90% of the pen tips and refills, too, were imported from Japan, Germany and Switzerland, according to Chinese state media. This cost the industry $17.3m a year, according to the China National Light Industry Council.

«

The fact that ballpoints require such very careful manufacturing isn’t that obvious – we take it for granted – but shows how China has speedrun through manufacturing history.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified