Remember how, in 2023, the U.S. Office of the Director of National Intelligence published a report acknowledging mass stockpiling of third-party data it had purchased? It turns out there is so much private information about people it is creating a big headache for the intelligence agencies — not because of any laws or ethical qualms, but simply because of the sheer volume.

Sam Biddle, the Intercept:

The Office of the Director of National Intelligence is working on a system to centralize and “streamline” the use of commercially available information, or CAI, like location data derived from mobile ads, by American spy agencies, according to contract documents reviewed by The Intercept. The data portal will include information deemed by the ODNI as highly sensitive, that which can be “misused to cause substantial harm, embarrassment, and inconvenience to U.S. persons.” The documents state spy agencies will use the web portal not just to search through reams of private data, but also run them through artificial intelligence tools for further analysis.

Apparently, the plan is to feed all this data purchased from brokers and digital advertising companies into artificial intelligence systems. The DNI says it has rules about purchasing and using this data, so there is nothing to worry about.

By the way, the DNI’s Freedom of Information Act page was recently updated to remove links to released records and FOIA logs. They were live on May 5 but, as of May 16, those pages have been removed, and direct links no longer resolve either. Strange.

Update: The ODNI told me its “website is currently under construction”.

Berber Jin, Wall Street Journal:

Altman and Ive offered a few hints at the secret project they have been working on [at a staff meeting]. The product will be capable of being fully aware of a user’s surroundings and life, will be unobtrusive, able to rest in one’s pocket or on one’s desk, and will be a third core device a person would put on a desk after a MacBook Pro and an iPhone.

Ambitious, albeit marginally less hubristic than considering it a replacement for either of those two device categories.

Stephen Hackett:

If OpenAI’s future product is meant to work with the iPhone and Android phones, then the company is opening a whole other set of worms, from the integration itself to the fact that most people will still prefer to simply pull their phone out of their pockets for basically any task.

I am reminded of an April 2024 article by Jason Snell at Six Colors:

The problem is that I’m dismissing the Ai Pin and looking forward to the Apple Watch specifically because of the control Apple has over its platforms. Yes, the company’s entire business model is based on tightly integrating its hardware and software, and it allows devices like the Apple Watch to exist. But that focus on tight integration comes at a cost (to everyone but Apple, anyway): Nobody else can have the access Apple has.

A problem OpenAI could have with this device is the same as was faced by Humane, which is that Apple treats third-party hardware and software as second-class citizens in its post-P.C. ecosystem. OpenAI is laying the groundwork for better individual context. But this is a significant limitation, and it is one I am curious to see how it is overcome.

Whatever this thing is, it is undeniably interesting to me. OpenAI has become a household name on a foundation of an academic-sounding product that has changed the world. Jony Ive has been the name attached to entire eras of design. There is plenty to criticize about both. Yet the combination of these things is surely intriguing, inviting the kind of speculation that used to be commonplace in tech before it all became rote. I have little faith our world will become meaningfully better with another gadget in it. Yet I hope the result is captivating, at least, because we could use some of that.

Jessica Conditt, Engadget:

A group of GeoGuessr map creators have pulled their contributions from the game to protest its participation in the Esports World Cup 2025, calling the tournament “a sportswashing tool used by the government of Saudi Arabia to distract from and conceal its horrific human rights record.” The protestors say the blackout will hold until the game’s publisher, GeoGuessr AB, cancels its planned Last Chance Wildcard tournament at the EWC in Riyadh, Saudi Arabia, from July 21 to 27.

Those participating in this blackout created some of the most popular and notable maps in the game. Good for them.

Update: GeoGuessr says it is withdrawing from the EWC.

Thinking about the energy “footprint” of artificial intelligence products makes it a good time to re-link to Mark Kaufman’s excellent 2020 Mashable article in which he explores the idea of a carbon footprint:

The genius of the “carbon footprint” is that it gives us something to ostensibly do about the climate problem. No ordinary person can slash 1 billion tons of carbon dioxide emissions. But we can toss a plastic bottle into a recycling bin, carpool to work, or eat fewer cheeseburgers. “Psychologically we’re not built for big global transformations,” said John Cook, a cognitive scientist at the Center for Climate Change Communication at George Mason University. “It’s hard to wrap our head around it.”

Ogilvy & Mather, the marketers hired by British Petroleum, wove the overwhelming challenges inherent in transforming the dominant global energy system with manipulative tactics that made something intangible (carbon dioxide and methane — both potent greenhouse gases — are invisible), tangible. A footprint. Your footprint.

The framing of most of the A.I. articles I have seen thankfully shies away from ascribing individual blame; instead, they point to systemic flaws. This is preferable, but it still does little at the scale of electricity generation worldwide.

Casey Crownhart, MIT Technology Review:

Today, new analysis by MIT Technology Review provides an unprecedented and comprehensive look at how much energy the AI industry uses — down to a single query — to trace where its carbon footprint stands now, and where it’s headed, as AI barrels towards billions of daily users.

We spoke to two dozen experts measuring AI’s energy demands, evaluated different AI models and prompts, pored over hundreds of pages of projections and reports, and questioned top AI model makers about their plans. Ultimately, we found that the common understanding of AI’s energy consumption is full of holes.

This robust story comes on the heels of a series of other discussions about how much energy is used by A.I. products and services. Last month, for example, Andy Masley published a comparison of using ChatGPT against other common activities. The Economist ran another, and similar articles have been published before. As far as I can tell, they all come down to the same general conclusion: training A.I. models is energy-intensive, using A.I. products is not, lots of things we do online and offline have a greater impact on the environment, and the current energy use of A.I. is the lowest it will be from now on.

There are lots of good reasons to critique artificial intelligence. I am not sure its environmental impact is a particularly strong one; I think the true energy footprint of tech companies, of which A.I. is one part, is more relevant. Even more pressing, however, is our need to electrify our world as much as we can, and that will require a better and cleaner grid.

Last month, the Information reported OpenAI was considering buying io Products — unfortunate capitalization theirs — for around $500 million. The company, founded by Jony Ive and employing several ex-Apple designers and engineers, was already known to be working with OpenAI, but it was still an external entity. Now, it is not, to the tune of over $6 billion in equity.

OpenAI today published a press release and video — set in LoveFrom’s distinctive proprietary serif face — featuring Ive and Sam Altman in conversation. There is barely a hint of what they are working on but, whether because of honesty or just clever packaging, it comes across as an earnest attempt to think about the new technologies OpenAI has successfully brought to the world as part of our broader cultural fabric. Of course, it will be expressed in something that can be assembled in a factory and sold for money, so let us not get too teary-eyed. We have heard a similar tune before.

The video promises revealing something “next year”.

Albert Burneko, Defector:

Over this past weekend, the Chicago Sun-Times and Philadelphia Inquirer’s weekend editions included identical huge “Best of Summer” inserts; in the Inquirer’s digital edition the insert runs 54 pages, while the entire rest of the paper occupies 36. Before long, readers began noticing something strange about the “Summer reading list for 2025” section of the insert. Namely, that while the list includes some very well-known authors, most of the books listed in it do not exist.

This is the kind of fluffy insert long purchased by publishers to pad newspapers. In this case, it appears to be produced by Hearst Communications, which feels about right for something with Hearst’s name on it. I cannot imagine most publishers read these things very carefully; adding more work or responsibility is not the point of buying a guide like this.

What I found very funny today was watching the real-time reporting of this story in parallel with Google’s I/O presentation, at which it announced one artificial intelligence feature after another. On the one hand, A.I. features can help you buy event tickets or generate emails offering travel advice based on photos from trips you have taken. On the other, it is inventing books, experts, and diet advice.

Katie Razzall, BBC News:

When the increasingly expensive contracts to provide broadcast channels and digital terrestrial services like Freeview come to an end, the UK’s broadcasters are likely to pivot to offering digital-only video on demand. (However this won’t happen without a campaign to ensure older people are protected, as well as rural and low-income households who may not have high quality internet access.)

But if the aerials are turned off in 2035, is this the moment TV as we know it changes forever? If it becomes a battle between online-only British streamers and their better-funded US rivals, can the Brits survive? And, crucially, what will audiences be watching?

One of the things British television has in its favour is its established cultural relevance. The BBC has problems, but it matters to people. If a country values its domestic media — particularly public broadcasting — it should watch the future of British media closely and figure out what is worth emulating to stay relevant. The CBC is worth it, too.

Tim Higgins, Wall Street Journal:

Each company is accused of being overly aggressive in holding back internal documents under special legal standing — known as privilege — that should have, in fact, been turned over to the government or lawyers suing on behalf of Epic. The videogame company has been fighting separate multiyear battles against Apple and Google over its desire to load its app on smartphones outside the tech giants’ 30% commission.

Come for the story, stay for the photos of boxes on trolleys.

Joseph Cox, 404 Media:

Flock’s new product, called Nova, will supplement license plate data with a wealth of personal information sourced from other companies and the wider web, according to the material obtained by 404 Media. “You’re going to be able to access data and jump from LPR to person and understand what that context is, link to other people that are related to that person […] marriage or through gang affiliation, et cetera,” a Flock employee said during an internal company meeting, according to an audio recording. “There’s very powerful linking.” One Slack message said that Nova supports 20 different data sources that agencies can toggle on or off.

According to Cox’s reporting, this includes the usual smattering of creepy data brokers but, also, private data made public due to illegal breaches. One may consider that “open source” in much the same way as Meta benefitting from pirated books, which is to say that it is something for which a normal person would be criminally prosecuted, but is apparently fine for corporations. It may be unwise for the limits of our privacy to be governed by the greed of capitalism.

Update: The good news is that Flock has decided not to use leaked data. The bad news is this is apparently a decision the company can make instead of a law it needs to follow.

Over the weekend, Mark Gurman and Drake Bennett, of Bloomberg, published a lengthy examination of Apple’s fumbling history of artificial intelligence features. It is surprisingly warm to John Giannandrea, who is portrayed as someone who tried hard to build talent and resources at Apple, only to hit walls imposed by other senior leadership figures.

There is a fair bit of news about Siri and Apple Intelligence, which you would know if you read MacRumors because it built four separate articles from various pieces of it. I guess rewriting the whole thing as a single article would have been unethical. Anyway, I thought the one about changing the default virtual assistant was notable:

Apple is planning to give users in the EU the ability to set a default voice assistant other than Siri, according to Bloomberg’s Mark Gurman and Drake Bennett.

[…]

Apple is working on this change in response to expanding EU regulations, the report said.

The way Joe Rossignol phrased this surprised me because I had assumed the DMA already covered default virtual assistant, but it seems that none were designated gatekeepers. I can imagine how difficult it will be for third-party services to act as a drop-in replacement for Siri, too.

Federico Viticci, of MacStories, writing last month about Perplexity’s Voice Assistant feature, a piece which I am chopping up here to make a point but you should read in full anyhow:

[…] Perplexity’s iOS voice assistant isn’t using any “secret” tricks or hidden APIs: they’re simply integrating with existing frameworks and APIs that any third-party iOS developer can already work with. […]

[…] Then there are all the integrations that are exclusive to Siri, which Perplexity can’t implement because Apple doesn’t offer related developer APIs. Only Siri can run shortcuts, set timers, call App Intents, send messages, create notes, open and change device settings, and more. […]

There is a long way to go from this to a full Siri replacement, but I will be hugely envious of those who will be able to take advantage of changing the default. The state of Siri was embarrassing ten years ago. The condition it is in today is a testament to the power of unchangeable defaults and a lack of competition within the iOS universe.

Make no mistake, however: Apple is barely about to “let” E.U. users switch from Siri to something else, as the MacRumors headline claims. It is doing so with the reasonable anticipation Siri’s in-universe monopoly will fall on the wrong side of regulations already established. Good.

Do you want to block all YouTube ads in Safari on your iPhone, iPad, and Mac?

Then download Magic Lasso Adblock — the ad blocker designed for you.

Magic Lasso Adblock: YouTube Ad Blocker for Safari

As an efficient, high performance, and native Safari ad blocker, Magic Lasso blocks all intrusive ads, trackers, and annoyances — delivering a faster, cleaner, and more secure web browsing experience.

Magic Lasso Adblock is easy to set up, doubles the speed at which Safari loads, and also blocks all YouTube ads; including all:

  • video ads

  • pop up banner ads

  • search ads

  • plus many more

With over 5,000 five star reviews, it’s simply the best ad blocker for your iPhone, iPad, and Mac.

And unlike some other ad blockers, Magic Lasso Adblock respects your privacy, doesn’t accept payment from advertisers and is 100% supported by its community of users.

So, join over 350,000 users and download Magic Lasso Adblock today.

Rebecca Bellan, TechCrunch:

Ride-hail and delivery giant Uber is introducing cheap, fixed-route rides along busy corridors during weekday commute hours in major U.S. cities — one solution to a world that feels, for most people, more expensive every day.

This is a bus. A worse, less efficient bus. And, with a half-priced discount compared to a standard Uber ride, a very expensive bus at that.

I suppose this is an obvious next step for the progenitor company for undermining existing infrastructure to the benefit of rich people. Uber is rolling this out in cities like Chicago, New York, Philadelphia, and San Francisco — and I think they may already have public transit. What would be better for everybody is if we just improved trains and buses for everyone. But that requires money in the public purse, and Uber dodges taxes despite benefitting from taxpayer-funded programs.

Jacob Eiting, of RevenueCat:

Turns out, in-app purchases are good for conversion rates. In fact, at least 30% better. That’s one of the things we found while running the first large-scale, side-by-side test of in-app vs web purchases in history.

[…]

The initial conversion rate for variant B is between 27% and 30%, while the equivalent web flow in variant D is between 17% and 19%. This is a large decrease, a 25% to 45% relative drop between the two. Digging into the funnel, most of that drop occurs from the payment sheet through to purchase. That’s a lot of fall off.

This is limited data, but maybe one reason In-App Purchases convert better is because Apple says other means of purchasing must be kneecapped.

Sarah Perez, TechCrunch:

[…] Last week, the Patreon iOS app (version 125.5.0) added an option that let users pay via the web using a variety of payment methods, including credit cards, Venmo, PayPal, and even Apple Pay.

These options appeared within an in-app browser, providing a more seamless checkout experience for Patreon users. Now, Patreon’s checkout flow opens in an external browser instead — a change Patreon made based on Apple’s feedback.

Again, it is worth emphasizing this approach is just for digital goods purchases and remains U.S.-only. Apple has no problem with people entering their credit card details from within Amazon’s app, but apparently a browser sheet offers an experience too straightforward for users buying Patreon subscriptions.

Austin Carr and Dina Bass, in a largely defensive but, somehow, still pretty positive profile of Microsoft CEO Satya Nadella in Bloomberg:

Working out of offices in London and Mountain View, California, Suleyman’s employees set about creating a version of Copilot designed for life away from the office. As they’d done at Inflection, they taught this Copilot empathy, humor and kindness. Do people want an emotional connection with the company that makes Excel? Suleyman thinks so. Or at least he believes it’ll be harder to switch to a competitor if the user sees Copilot as a friend or therapist.

For those in the camp seeing artificial intelligence technologies as one tool among many to be incorporated into larger products and services — and I count myself among them — the idea of a faux emotional companion as a standalone goal appears utterly ludicrous. I do not think we should view these tools in this way. They do not substitute for real-life friendships. And why would anyone trust something which can be so easily swayed by bad actors? Using an emotional connection as a product differentiator or a matter of lock-in seems entirely unethical.

Also:

Nadella points out that Microsoft was once less relevant in videoconferencing too. “Everybody would say, ‘Hey, Zoom, Zoom, Zoom,’” Nadella says. “We won that in the enterprise.” (A Zoom spokesperson declined to comment.)

It did so by illegally advantaging Teams, and what Nadella seems to be saying here is that it would again be happy to use its disproportionate power to push its version of other new technologies.

Viktor Maric:

first time seeing this. Apple will punish the apps with external payment system

Maric includes a screenshot of an App Store listing for Instacar, which features a exclamation mark-decorated banner at the top reading “This app does not support the App Store’s private and secure payment system. It uses external purchases.”

Via Michael Tsai:

It’s confusing to follow all the changes, but apparently — unlike in the US — external purchases in the EU don’t need to have corresponding IAP versions.

You have probably already seen this, but I am catching up on a week’s worth of news. And the thing about waiting a week to write about it is that it already generated substantial discussion, including a response from Apple.

Sarah Perez, TechCrunch:

However, the iPhone maker confirmed to TechCrunch that these user-disclosure screens have been live on the EU App Store since the beginning of Apple’s DMA Compliance Plan back in March 2024. They were not newly added, as some had reported.

[…]

In its response to TechCrunch, Apple also noted that it intended to update the message after initial pushback. In August 2024, the company announced a series of changes to its DMA plan that would have included a change to the user disclosure screen. Instead of warning users of the dangers of using external purchases, the new message would have read: “Transactions in this app are supported by the developer and not Apple.” (See below).

Apple wanted to scare users, as we learned during its U.S. trial; its preference would have been the terrifying phrasing. But it softened the language — how much is the result of public criticism and how much is because of regulators’ concerns is a good question.

In any case, does this message show on listings for any applications accepting payments through means other than In-App Purchases? I assume Apple is not warning users about the dangers of paying for a ride through Uber or a hotel room through Kayak. But subscribing to something without using Apple’s own payment mechanism? May as well shout your credit card number in a crowded room.

It is not like Apple is taking an elevated level of responsibility for payments made through In-App Purchases, either. This warning tone carried through in documentation may not be lying to users, but it is bullshitting them and that, in most places, is not a sign of trust or respect. Perhaps things are different in Silicon Valley.

Mark Gurman, Bloomberg:

The company is planning an AI-powered battery management mode for iOS 19, an iPhone software update due in September, according to people with knowledge of the matter. The enhancement will analyze how a person uses their device and make adjustments to conserve energy, said the people, who asked not to be identified because the service hasn’t been announced.

Set aside for now the question of how the notoriously power-hungry group of technologies we call “artificial intelligence” can improve battery life. Gurman says this is “part of the Apple Intelligence platform”, but also says it “will be available for all iPhones that have iOS 19”. This is confusing. Apple has so far marketed Apple Intelligence as being available on only a subset of devices supporting iOS 18. Either Apple’s delineation of “Apple Intelligence” features is about to get even fuzzier, or one of the two statements Gurman made is going to be wrong.

Also:

Besides the AI additions and interface changes, Apple is pushing engineers to ensure that this year’s releases are more functional and less glitchy. Past upgrades were criticized for bugs and features that sometimes didn’t work properly.

I will believe it when I see it. But how is this not the highest priority every year?

Rina Raphael, New York Times:

But what makes this influencer unusual is her age. She’s only 17, and a high school junior.

Ms. [Ava] Noe, a self-described “crunchy teen,” is just one of a number of young influencers who appeal to other health-conscious kids their age. At times, their anti-establishment viewpoints fall in line with those of Robert F. Kennedy Jr. and the “Make America Healthy Again” movement, which has expressed skepticism of the scientific community and large food corporations.

The teens’ videos, while at times factually questionable, highlight a desire among some to avoid the chronic illnesses and other conditions that have plagued their elders.

This article keeps rattling around my brain because, I think, this is a subject potentially deserving media coverage but not with this framing. The most appropriate context for these teenagers’ beliefs would undoubtably be seen as pretty brutal — they are not “skeptical”, they are young people who have no idea what they are talking about. Trying to rebut them would seem unkind and I do not think it would be effective.

This article could instead be hung on the dieticians who are concerned about the number of young people getting their health information in this way. Raphael quotes one doctor and one dietician, and they do provide necessary context, but the story is heavily weighted toward the claims these kids are spreading. The Times hired three photographers for this article, entirely for the influencers. They link to the social media accounts for two of them. Perhaps the reason this story is being told from this perspective is because these teenagers are not actually that influential. That is potentially good news. But the Times has now platformed these accounts and is once again guiding readers to information marketed as edgy and anti-establishment, when it is simply bad.

To the extent there is any good information spread by these influencers, it is the kind of stuff you already know. Choose more fruits and vegetables — that kind of thing. Noe may protest that is the advice she is giving; in a response to this Times story, she says “literally all I do is eat whole foods”. It is “common knowledge that fruit is good for you, for the most part, and Skittles are bad for you”. But that is not descriptive of what Noe is doing. She, according to Raphael’s reporting, “use[s] ChatGPT to ask about the benefits of, say, red light therapy, then read the suggested studies”. Whether Noe can actually understand and weigh the credibility of those studies is a good question. But she happens to have a partnership with some red light therapy company.

It seems to me these influencers have succumbed to the same bizarre information space as a bunch of adults, and are selling that back to their peer group. That is an interesting story to me, and I wish it is the one told by the Times. Instead, we get a too-soft profile of some kids who have found online success repeating the things they hear on podcasts, balanced by a few quotes from two professionals. While I have no food or nutrition expertise, I do know what skepticism looks like. I am not sure the Times does.

Blake Brittain, Reuters:

In the first court hearing on a key question for the AI industry, U.S. District Judge Vince Chhabria grilled lawyers for both sides over Meta’s request for a ruling that it made “fair use” of books by Junot Diaz, comedian Sarah Silverman and others to train its Llama large language model.

Worth noting Thompson Reuters, Reuters’ parent company, just won a similar case.

Kate Knibbs, Wired:

US District Court Judge Vince Chhabria spent several hours grilling lawyers from both sides after they each filed motions for partial summary judgment, meaning they want Chhabria to rule on specific issues of the case rather than leaving each one to be decided at trial. The authors allege that Meta illegally used their work to build its generative AI tools, emphasizing that the company pirated their books through “shadow libraries” like LibGen. The social media giant is not denying that it used the work or that it downloaded books from shadow libraries en masse, but insists that its behavior is shielded by the “fair use” doctrine, an exception in US copyright law that allows for permissionless use of copyrighted work in certain cases, including parody, teaching, and news reporting.

So let me get this straight: if an artist makes a song sharing only a similar vibe with another one, that is copyright infringement, but if a massive corporation downloads illicit copies of a bunch of books to enrich its A.I. training, that is fair use? Perhaps copyright law is entirely the wrong domain for both of these issues.

The Irish Data Protection Commission:

The decision, which was made by the Commissioners for Data Protection, Dr Des Hogan and Mr Dale Sunderland, and has been notified to TikTok, finds that TikTok infringed the GDPR regarding its transfers of EEA User Data to China and its transparency requirements. The decision includes administrative fines totalling €530 million and an order requiring TikTok to bring its processing into compliance within 6 months. The decision also includes an order suspending TikTok’s transfers to China if processing is not brought into compliance within this timeframe.

It turns out data use expectations can be consistent and it is possible to penalize violators when there are privacy laws. This seems preferable over ad hoc responses.