The 2003 Space Shuttle Columbia disaster from a damaged wing was blamed, in part, on PowerPoint. Will chatbots get the same blame for future calamities? CC-licensed photo by NASA on The Commons 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. Sliding. 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 Mozart of the attention economy’: why MrBeast is the world’s biggest YouTube star • The Guardian

Mark O’Connell:

»

A couple of weeks ago, I had a drink with a friend who’d just got back from Alexandria, where he’d gone to get a complicated root canal situation seen to, at presumably a fraction of what it would have cost him at home. I mentioned I was writing about MrBeast, and my friend – whose dental status has since improved – told me he’d heard from a number of Egyptians that the country’s tourist trade was currently experiencing a significant upswing, and that that upswing could be accounted for by the release, a few months back, of I Spent 100 Hours Inside the Pyramids! (194m views thus far), in which Donaldson and his boys were granted unprecedented access to the deep interiors of the pyramids of Giza.

One obvious question to ask about all of this is why is MrBeast so successful? Another, less obvious, question is: what does his success say about the culture that gave rise to it? The first question is, in a sense, a fairly straightforward one to answer. MrBeast is successful because his videos are highly entertaining. I can personally vouch for this. I don’t imagine, being in my mid-40s (among other limiting demographic and cultural factors), that I’m anywhere close to the MrBeast target viewership, and yet I have consumed more than my fair share of his content, and have been steadily and straightforwardly entertained. Troubled by the general tenor and implications of that state of being entertained, yes, but entertained nonetheless.

[27-year-old Jimmy] Donaldson is not by any means one of God’s chosen entertainment-industry stars. He’s not especially handsome, and neither is he particularly funny-looking. At 6ft 5in, and with the sparse reddish beard he nowadays sports, he has the charmingly awkward aspect of a teen who has recently put on a growth spurt and hasn’t quite settled into himself. He’s likable, and is possessed of a goofy and anarchic sense of humour, but more in a guy-you-went-to-school-with sense than, say, the Eric André or the Jack Black sense. You definitely wouldn’t call him cool, either, and he’s certainly not edgy, but neither is he staid or offensively corny. My wife – who has, in passing, taken in a fair measure of MrBeast content over the time I have been working on this essay, consuming them in a manner roughly analogous to passive smoking – described Donaldson and his crew of sidekicks thus: “They just seem like good kids.”

«

I’ve never watched a single second of MrBeast content, but clearly lots and lots of people spend a lot longer on it.
unique link to this extract

---

The Methaphone is a phone (that’s not a phone) to help you stop using your phone • WIRED

Arielle Pardes:

»

Earlier this year, Eric Antonow was in a coffee shop with his family when he felt the familiar, twitchy urge to reach for his phone. He patted his pockets for relief—the cool, thin slab was still there. He joked to his family that, like an addict jonesing for a hit, he would one day need a medical-grade solution to detox from his phone. Opioid addicts had methadone. iPhone addicts would need … methaphones.

“It was a joke, but I got two laughs from my two teenagers, which is gold,” Antonow says. “I was like, ‘I’m going to commit to the bit.’”

Antonow, a former marketing executive at Google and Facebook, has been committing to bits for half a decade, making what he calls “mindless toys.” His online shop features projects like a “listening switch” to indicate when one is paying attention, and a vinyl for silent meditation, with 20 minutes of recorded silence on each side (record player not required).

So within days of his latest joke, he had enlisted ChatGPT to mock up an image of a gadget in the shape of a phone, without all of the contents: a translucent rectangle that one could look at, or through. From that original generative sketch emerged a more realized design: a six-inch slab of clear acrylic with rounded corners, like the iPhone, and green edges that resembled glass. Antonow placed an order for samples, and started an Indiegogo campaign for the Methaphone: to “leave your phone without the cravings or withdrawal.”

«

Couldn’t he just.. not charge his phone?
unique link to this extract

---

The dumbest phone is parenting genius • The Atlantic

Rheana Murray:

»

When Caron Morse’s nine-year-old daughter asked for a smartphone last year, her reaction, she told me, was unambiguous: “A hard ‘Hell no.’” Morse is a mental-health provider in the Portland, Maine, public-school system, and she was firmly against smartphones, having seen how social media and abundant screen time could shorten students’ attention spans and give them new anxieties. But she wanted her children to have some independence—to be able to call friends, arrange playdates, and reach out to their grandparents on their own. She also needed a break. “I was so sick,” she said, “of being the middle person in any correspondence.”

So when her daughter turned 10, Morse did get her a phone: a landline.

For that gift to provide all the benefits she wanted, Morse had to lay some groundwork. It would be annoying if her daughters—she also has an eight-year-old—were to start calling their friends’ parents’ smartphones all the time, so she told her neighbours about her plan and suggested that they consider getting landlines too. Several bought in immediately, excited for the opportunity to placate their own smartphone-eager kids. And over the next couple of months, Morse kept nudging people. She appealed to their sense of nostalgia by sharing photos of her older daughter sitting on the floor and twirling the landline’s cord around her fingers. She wrote messages: “Guys, this is adorable and working and important.”

The peer pressure paid off. Now about 15 to 20 families in their South Portland neighborhood have installed a landline. They’ve created a retro bubble in which their children can easily call their friends without bugging a parent to borrow their phone—and in which the parents, for now, can live blissfully free of anxieties about the downsides of smartphones.

«

Future pollsters are going to love this. People with landlines!
unique link to this extract

---

Key fair use ruling clarifies when books can be used for AI training • Ars Technica

Ashley Belanger:

»

Artificial intelligence companies don’t need permission from authors to train their large language models (LLMs) on legally acquired books, US District Judge William Alsup ruled Monday.

The first-of-its-kind ruling that condones AI training as fair use will likely be viewed as a big win for AI companies, but it also notably put on notice all the AI companies that expect the same reasoning will apply to training on pirated copies of books—a question that remains unsettled.

In the specific case that Alsup is weighing—which pits book authors against Anthropic—Alsup found that “the purpose and character of using copyrighted works to train LLMs to generate new text was quintessentially transformative” and “necessary” to build world-class AI models.

Importantly, this case differs from other lawsuits where authors allege that AI models risk copying and distributing their work. Because authors suing Anthropic did not allege that any of Anthropic’s outputs reproduced their works or expressive style, Alsup found there was no threat that Anthropic’s text generator, Claude, might replace authors in their markets. And that lacking argument did tip the fair use analysis in favor of Anthropic.

“Like any reader aspiring to be a writer, Anthropic’s LLMs trained upon works not to race ahead and replicate or supplant them—but to turn a hard corner and create something different,” Alsup wrote.

Alsup’s ruling surely disappointed authors, who instead argued that Claude’s reliance on their texts could generate competing summaries or alternative versions of their stories. The judge claimed these complaints were akin to arguing “that training schoolchildren to write well would result in an explosion of competing works.”

“This is not the kind of competitive or creative displacement that concerns the Copyright Act,” Alsup wrote. “The Act seeks to advance original works of authorship, not to protect authors against competition.”

Alsup noted that authors would be able to raise new claims if they found evidence of infringing Claude outputs. That could change the fair use calculus, as it might in a case where a judge recently suggested that Meta’s AI products might be “obliterating” authors’ markets for works.

«

Alsup is famous in tech legal cases for having overseen Oracle v Google, where Oracle claimed Android copied Java, and showing great understanding of programming. So this will be similarly strongly argued.
unique link to this extract

---

What a set of knockoff headphones taught me about headphones — and knockoffs • The Verge

David Pierce:

»

I kept hearing that Picun headphones were roughly as good as the AirPods Max for a fraction of the price. A few TikToks I saw argue that you’re not the problem if you buy knockoffs — you’re the problem if you’re spending $500 more just to get a brand name. Some videos purport to perform scientific noise-canceling tests; others just hold up a pair of AirPods Max and then a pair of Picuns, as if the side-by-side proves the point.

All the sales-creators made it clear that I needed to buy these headphones now. Some videos spread a rumor that Apple was suing Picun over the design, so they might be off the market soon. (This is not the case, as far as I know — Apple didn’t respond to a request for comment.) Others continuously claim that the headphones are about to be taken off the TikTok Shop; I’ve been seeing that for weeks, and they’re still for sale.

I don’t believe any of it! And yet, after a few taps I barely even remember, I’d spent $63.58 to get a pair shipped to my door. I also ran to the Apple Store and dropped $581.94 on blue AirPods Max. I had testing to do.

I’ve been using both for the past several weeks, and I’ve come to a conclusion I didn’t expect. The Picun F8 Pros sound a smidge worse than the AirPods Max, but in a few ways, I actually prefer them, and given the price I’d easily pick Picun. The bass in the F8s is a little more pronounced than I like, and can be a little muddy on extra-thumpy songs. They were crisper on the high notes in a song like “Welcome to the Black Parade,” though, and for the most part both brands sound pretty similar.

The limiting factor for headphones, I suspect, is not the headphones themselves but the context. Buy all the great gear you want, but if you’re still streaming Spotify playlists over Bluetooth, there’s only so much fidelity available. Yes, the AirPods Max now support lossless audio over a wired connection, but that’s not how most people listen to music. Most listen on loud subways, in the gym, or while walking the dog; unless you’re in a dedicated listening environment, I’d wager that good-enough sound is usually good enough. Especially for the price.

«

This is very much the point about “high quality” headphones: most real listening environments are high-quality-hostile.
unique link to this extract

---

Delayed Scottish NHS app cut back to just one service • The Times

Henry Anderson:

»

Scotland’s ill-fated NHS app will only be available to dermatology patients after being scaled back by ministers but still faces “significant issues” being delivered on time, an internal review has found.

Known as the digital front door, the health and social care app was first announced in 2021 but has been branded a “national embarrassment” as it is yet to get off the ground.

A stocktake, seen by The Times, said the app had been reduced to a “more limited scope … than originally envisaged” after ministers brought the launch date forward by four months.

At first patients will only be able to receive appointment letters for one specialty, dermatology, at a single health board, NHS Lanarkshire. The system will also let users view personal information and search for health services.

In contrast, England’s NHS App, which has 35 million users, allows patients to message GPs, order repeat prescriptions and receive test results.

Critics say the report confirms the “chaos, delays and mind-boggling expense” caused by the SNP’s mismanagement of the NHS.

Under initial plans the app was due to go live in March 2026, but first minister John Swinney made a commitment to launching it by December in a speech earlier this year. Swinney said the app would “be a much-needed addition to improve patients’ interaction with the NHS”.

«

Got to love government computing projects. No indication here of how much was spent on this one (so far), but you do have to wonder at the “not invented here” syndrome that made them think they could do it better – or should do it separately – of the English NHS. But the Scottish government seems to be very good at screwing things up.
unique link to this extract

---

US House of Representatives bans WhatsApp on government devices • Financial Times

Hannah Murphy, Stephanie Stacey and Alex Rogers:

»

The US House of Representatives has warned staff members not to use Meta’s messaging platform WhatsApp due to privacy concerns.

The warning marks a blow to WhatsApp, whose $1.8tn parent Meta has long battled concerns that it has been lax with user data in its hunt for commercial growth and advertising revenue.

The House’s Chief Administrative Officer told staffers on Monday that WhatsApp had been deemed “a high-risk to users”, according to a copy of the memo seen by the Financial Times.

The email ordered staff not to download or keep the messaging service on any House laptop or mobile device from June 30, adding that anyone who had the application would be asked to remove it.

The decision was taken due to “a lack of transparency in how [WhatsApp] protects user data, absence of stored data encryption, and potential security risks involved with its use”, read the memo, which was first reported by Axios.

A spokesperson for Meta said the company disagreed with the characterisation “in the strongest possible terms”.

The person added that WhatsApp messages were “end-to-end encrypted by default”, meaning that neither the company nor third parties could read them, adding that the platform offered “a higher level of security than most of the apps on the CAO’s approved list”.

«

Bonkers. (Thanks Gregory B for the link.)
unique link to this extract

---

The great PowerPoint panic of 2003 • The Atlantic

Jacob Stern:

»

Sixteen minutes before touchdown on the morning of February 1, 2003, the space shuttle Columbia disintegrated into the cloudless East Texas sky. All seven astronauts aboard were killed. As the broken shuttle hurtled toward Earth in pieces, it looked to its live TV viewers like a swarm of shooting stars.

The immediate cause of the disaster, a report from a NASA Accident Investigation Board determined that August, was a piece of insulating foam that had broken loose and damaged the shuttle’s left wing soon after liftoff. But the report also singled out a less direct, more surprising culprit. Engineers had known about—and inappropriately discounted—the wing damage long before Columbia’s attempted reentry, but the flaws in their analysis were buried in a series of arcane and overstuffed computer-presentation slides that were shown to NASA officials. “It is easy to understand how a senior manager might read this PowerPoint slide and not realize that it addresses a life-threatening situation,” the report stated, later continuing: “The Board views the endemic use of PowerPoint briefing slides instead of technical papers as an illustration of the problematic methods of technical communication at NASA.”

PowerPoint was not then a new technology, but it was newly ubiquitous. In 1987, when the program was first released, it sold 40,000 copies. Ten years later, it sold 4 million. By the early 2000s, PowerPoint had captured 95% of the presentation-software market, and its growing influence on how Americans would talk and think was already giving rise to a critique. A 2001 feature in The New Yorker by Ian Parker argued that the software “helps you make a case, but it also makes its own case: about how to organize information, how much information to organize, how to look at the world.” Vint Cerf, one of the “fathers of the internet,” took to quipping that “power corrupts, and PowerPoint corrupts absolutely.”

«

Nowadays, of course, it would be death caused by hallucinating chatbot which someone didn’t check. And, indeed, perhaps it will come to that.

unique link to this extract

---

Apple updates its on-device and cloud AI models, introduces a new developer API • Deeplearning.ai

»

Apple updated the Apple Foundation Models (AFM) family, including smaller on-device and larger server-hosted versions, to improve their capabilities, speed, and efficiency. It also released the Foundation Models framework, an API that enables developers to call the on-device model on Apple devices that have Apple Intelligence enabled.

• Input/output: Text, images in (up to 65,000 tokens), text out
• Architecture: AFM-on-device: 3 billion-parameter transformer, 300-million parameter vision transformer. AFM-server: custom mixture-of-experts transformer (parameter count undisclosed), 1 billion-parameter vision transformer
• Performance: Strong in non-U.S. English, image understanding
• Availability: AFM-on-device for developers to use via Foundations Models framework, AFM-server not available for public use
• Features: Tool use, 15 languages, vision
• Undisclosed: Output token limit, AFM-server parameter count, details of training datasets, vision adapter architecture, evaluation protocol

…Apple may be behind in AI, but its control over iOS is a huge advantage. If the operating system ships with a certain model and loads it into the limited memory by default, developers have a far greater incentive to use that model than an alternative. Limited memory on phones and the large size of good models make it impractical for many app developers to bundle models with their software, so if a model is favored by Apple (or Android), it’s likely to gain significant adoption for on-device uses.

«

Developers are, while annoyed generally with Apple, pleased at the onboard LLMs that are going to ship in its next OS and be available to their apps. It’s going to make quite a difference. Does Google have the same, or is the RAM capability of Android phones too variable? I’d have thought they would have lots of RAM – they always used to – and that means the capability to run LLMs.
unique link to this extract

---

Elon Musk’s lawyers claim he “does not use a computer” • WIRED

Caroline Haskins:

»

Elon Musk’s lawyers claimed that he “does not use a computer” in a Sunday court filing related to his lawsuit against Sam Altman and OpenAI. However, Musk has posted pictures or referred to his laptop on X several times in recent months, and public evidence suggests that he owns and appears to use at least one computer.

Musk and his artificial intelligence startup xAI sued OpenAI in February 2024, alleging the company committed breach of contract by abandoning its founding agreement to develop AI “for the benefit of humanity,” choosing instead “to maximize profits for Microsoft.”

The Sunday court filing was submitted in opposition to a Friday filing from OpenAI, which accused Musk and xAI of failing to fully comply with the discovery process. OpenAI alleges that Musk’s counsel does not plan to collect any documents from him. In this weekend’s filing, Musk’s lawyers claim that they told OpenAI on June 14 that they were “conducting searches of Mr. Musk’s mobile phone, having searched his emails, and that Mr. Musk does not use a computer.”

Musk and xAI Corp’s lawyers did not immediately respond to requests for comment. In the filing, Musk’s legal team disputed claims that it was resisting discovery efforts.

Multiple employees at X tell WIRED that while Musk primarily works from his mobile phone, he has occasionally been seen using a laptop.

«

It does seem like a very hard claim to substantiate.
unique link to this extract

---

• Why do social networks drive us a little mad? • Why does angry content seem to dominate what we see? • How much of a role do algorithms play in affecting what we see and do online? • What can we do about it? • Did Facebook have any inkling of what was coming in Myanmar in 2016? Read Social Warming, my latest book, and find answers – and more.

---

Errata, corrigenda and ai no corrida: none notified