I’m not sure why I’ve never linked to The Vengeance Hour, an hourlong show curated by my good friend and former colleague, Tom Clark, aka Denis Vengeance. Each fortnight, Tom live-mixes a killer set that skews heavily toward ’60s and ’70s prog rock, new wave, classic rock, and folk—genres I have only a passing familiarity with, not because I’m too young (as if), but because I grew up listening more to R&B, disco, funk, jazz, and blues—which makes each show an hour of delightful discovery. His latest features several artists from Woodstock (which just celebrated its 57th anniversary) including Joni Mitchell, Jefferson Airplane, and Sly and the Family Stone, plus The Beatles, Nick Drake, Dire Straits, and more. It’s a great listen.

Shane Hickey, The Guardian:

You are on a short break in Porto and post some pictures of your family on Instagram or Facebook. With a small section of the Douro river in the background, you think it could have been taken anywhere.

A few days later, you get a text saying that your card was compromised. “We detected unusual activity while you were travelling in Porto – please verify immediately,” says the message.

You click on the link to confirm your bank details. Since you have not put any details of your trip on any of your social networks – bar the indistinct pictures – you don’t suspect there is anything suspicious about the messages.

But the text was a fraud designed to extract your financial details. The criminals behind it figured out where you had been on holidays – lending credibility to their text – by using AI to analyse the image for the most sparse signs of where it was taken.

I’ve long refused to post photos of (or even mention) my travels until (long) after I’m home, out of personal OPSEC concerns. Before, it was about protecting against break-ins. Now, it can also protect against scams.

When the ChatGPT geoguessing craze exploded in 2025, it impressed and scared me, as the value and privacy implications were both glaringly obvious. I currently have a long-running geoguessing thread into which I will occasionally drop an image. When it’s right, it’s scarily accurate. Even when it’s wrong, the amount of detail it’s able to pull out from a seemingly innocuous photo is still disturbing.

Don’t share photos while traveling, people.

Jennifer Pattison Tuohy, writing at The Verge:

Comcast is bringing Wi-Fi motion sensing to millions of routers that are already in customers’ homes, turning the devices into activity monitors. […]

It works by detecting disruptions in the Wi-Fi signal between the Xfinity gateway and any Wi-Fi-connected device, giving you some insight into what’s happening in your home without buying additional hardware. […]

WiFi Motion requires an advanced gateway, XB7 or newer. It’s opt in and can be turned on and off through the app.

What we need is more surveillance tech in our lives, and especially inside our own homes.

Zack Whittaker at TechCrunch notes:

By enabling the feature, Comcast says it may disclose information generated from a customer’s use of Wi-Fi Motion to outside parties under a broad array of reasons — and Comcast says it doesn’t have to tell you.

“Comcast may disclose information generated by your Wi-Fi Motion to third parties without further notice to you in connection with any law enforcement investigation or proceeding, any dispute to which Comcast is a party, or pursuant to a court order or subpoena,” reads an Xfinity support page.

So you pay Comcast for internet connectivity, opt in to Wi-Fi Motion for “security,” and now Comcast can notify unknown third parties whether or not you are home, how long you’re asleep, or how often you get up to use the bathroom while watching television? No thanks. I imagine telemarketers are salivating at the chance to know you’re sitting on your couch right now. Comcast is no doubt gleeful it can now claim a technician knocked on your door during the thirty seconds you were in the kitchen topping off your Mountain Dew.

Worse, it’s possible to use this technology to recognize and identify individuals.

I definitely don’t trust Comcast to keep this “opt-in,” either. I fully expect a future article about a “bug” that “accidentally” turned it on for customers, then Comcast eventually making it opt-out, before it makes the feature an undisclosed requirement.

(Shame on The Verge for not making even a passing reference to the privacy implications in what reads like a regurgitated Comcast press release.)

There are many reasons I refuse to use or recommend Comcast’s routers. Add one more.

Matthew Gault, 404 Media:

X’s algorithm learns what you hate and shows you more of it, according to a new study just published in the Proceedings of the National Academy of Sciences (PNAS). The paper, titled Value misalignment of X’s feed algorithm is a reflection of value tensions in engagement, found that the site’s algorithm prioritized engagement above all else when it generated a user’s For You Page. It also showed that X serves more ragebait to people who say they are Democrats, although the exact reason for that is unclear.

Why? “Engagement”:

When a user on X sees a post that makes them mad — like a press release from a politician from a political party they don’t like — sometimes they’ll fight about the post in the replies. It doesn’t matter who you follow or what your stated values are, X reads replying as engagement and will send more of the infuriating posts the user’s way.

A comment from X’s former head of product, Nikita Bier, suggests this was once true but now isn’t:

“This is no longer true,“ Bier said in a post on X. ”The largest contributor of seeing ragebait was the reply predictor and we were aware that angry replies were causing people to see more of that content. So last month, we gave the reply predictor a 15x boost if it’s a friend’s post — and it reduced ragebait by [an] order of magnitude."

Bier left that role two weeks ago after a mere 13 months. For all I know, this change led to his departure and was reverted the minute his badge stopped working. Regardless, you know how else you can reduce X/Twitter ragebait? Leave X/Twitter.

Ned Resnikoff, Public Comment:

If you are still posting on X, you should stop today. Quitting the site yesterday would have been better, but doing it today is better than doing it tomorrow.

The moral and political case for quitting X is so obvious that it scarcely bears repeating, but I’ll briefly recap it anyway. The guy who owns X—the person to whom you are personally donating if you pay for a blue checkmark—is Elon Musk, an out and out white supremacist who uses X to spread Nazi ideas like Great Replacement theory. He is a major backer, in both money and propaganda, of the global far right. During his brief tenure inside Trump’s Administration, he spearheaded the destruction of USAID, which, according to recent estimates, may result in 14 million deaths between 2025 and 2030.

X is the platform this particular mass murderer uses to seed his most poisonous ideas. A couple of months ago, he used the platform to egg on the Belfast pogrom. Those who continue to post on X—in particular those X users who are not racist psychopaths—are enhancing the site’s usefulness to him as a propaganda tool by drawing more eyes to the platform and making it look like an ideologically diverse “marketplace of ideas” instead of what it is: a delivery vehicle for bringing fascism into the mainstream of global politics. Oh, and perhaps the Internet’s leading destination for those in search of child sexual abuse material.

And let’s not forget Musk’s SpaceX con.

It’s not the first time I’ve boosted calls to leave X/Twitter (and it won’t be the last). Why another one? No reason other than it crossed my feed and any day is a good day to remind people that leaving X/Twitter is an option.

Mastodon, Bluesky, even (shudder) LinkedIn are better options. I could even stomach Threads, if you’re desperate. Pick one, tell your followers you’re moving and why, and leave X/Twitter behind. Your conscience will thank you.

Dan Moren, writing at Six Colors under the headline “LLMs aren’t writing,” makes a stronger case against John Gruber’s anti-watermarking essay than I did:

I have great respect for John as a writer. He truly cares about details: he’s meticulous about the words he picks, the way he constructs sentences and arguments.

This is why I find his defense so puzzling. Because a large language model could not be further away from those ideals. In the title of his piece, John describes the watermarking system as “a perversion of writing.” That might be the case if AIs were writing. But they are not; they are simply generating text. Those two things are not the same. As per the premise of John’s piece, writing is about carefully picking the words and crafting the meaning of what you’re saying, things that AI do not–and patently cannot–do.

LLMs do not care about the words that they pick because they cannot care about anything.

And:

All of this outrage presupposes the existence of a “correct” answer that the LLM is eschewing because it was told to pick different words, but if there were a correct answer, then shouldn’t it be chosen every time? But an LLM’s non-deterministic nature doesn’t work that way. Because, at the end of the day, the AI does not truly know what any of these words mean. There is no ur-text. You might as well get mad at word choice in a lorem ipsum generator.

John Gruber of Daring Fireball pens a barnburner that rails mightily against Anthropic’s plan to embed text watermarking in Claude-generated content, as mandated by European requirements:

One of my fundamental problems with this is that no two synonyms carry the exact same meaning. “He leaped at the chance” and “He jumped at the opportunity” are very similar sentences expressing the same general sentiment, but they are not the same. The exact words we choose when writing matter. I want any LLM I use to choose the very best, most precise words at every single decision point. […]

The idea that anything other than my needs should factor into the generation of text for me is patently offensive.

This isn’t just about text one might generate with the intention of passing it off as their own natural work. This isn’t even about LLM proofreading of work written by hand. Anthropic is saying that all new Claude models are going to adulterate every single bit of text longer than 200 tokens (~150 words) they generate, including everything it presents to its users to read. So even in a private conversation between a user and Claude, which will never be read by anyone other than the user, Claude will begin making word choices in the name of marking its output in statistically predictable ways rather than maximizing clarity and precision.

Even today’s so-called frontier models are already decidedly lacking in lucidity. Claude, ChatGPT, Grok, et al. are “better writers” than most humans and produce better prose than the median human. But: no shit. Most people are terrible writers. The “average person” is pretty stupid and half of all people are stupider than that. And there are many smart, interesting people who are miserable writers. So as impressive as LLMs are, the bar is low. The best writing I see come out of these models is worse than anything I would choose to read for pleasure. And now Anthropic is saying they’re going to make it worse, on purpose, for purposes that do not benefit me in any way? Even if only slightly worse?

Get fucked.

Gruber’s denunciation is forceful, persuasive, and wrong.

I don’t mean factually wrong. I mean his entire premise on which he builds his argument is misguided.

Gruber starts from the position that what Claude (and, by extension, any AI tool) generates is somewhat akin to human writing, and that, like a competent human author, it strives to select the best “right” words for its sentences. His argument is that by subtly changing which words are chosen for its generated text, Claude (etc.) creates lower-quality writing simply to satisfy the needs of watermarking.

His example sentences, “He leaped at the chance” and “He jumped at the opportunity,” do indeed carry different meaning to the reader, and the choice reflects both the writer and the writer’s preferred characterization of the person doing the leaping or jumping.

But Claude (etc.) isn’t making a literary decision. It’s using a statistical model to determine the next most likely token, and, for watermarking purposes, putting its statistical thumb on the scale in a predictable, recognizable way. The final sentence could just as plausibly be “He jumped at the chance” or “He leaped at the opportunity.”

Importantly, the final sentence can’t be judged as “worse” compared to something else, because we don’t know which alternate words were being considered—there is no something else. We can certainly opine as readers and writers on whether we ourselves would write “He jumped at the chance” rather than “He leaped at the opportunity,” but it’s hard to argue that one is inherently worse than the alternatives if we don’t know what the alternatives were.

Claude isn’t “writing” for clarity or creativity or precision. It’s not writing at all. It’s not pondering which word would have a given effect on a reader or best explain a concept. It’s selecting tokens from a list with a process that’s already biased by Anthropic’s models, and now that bias will be slightly more predictable and identifiable (to Anthropic, anyway). Arguing that one statistically generated alternative is “more” or “less” right ascribes human-like motivation—and capability—to an algorithm.

As I noted at the end of my previous piece, Yubico currently has a sale running: 20% off when you buy two YubiKey Security Keys: $46.40 instead of $58. The sale ends August 16, 2026—today, as I publish this. (The discount does not apply to any quantity above two, and applies only to the Security Key Series.) If you’re considering protecting your accounts with a physical security key for two-factor authentication, grab a pair. They work great with Apple’s Security Keys implementation. I already have two YubiKey 5 Series devices, and my wife and I each purchased two of the Security Key Series.

This is not an affiliate link and I’m not associated with Yubico. I get no benefit from any purchases.

George Hammond, Financial Times (on Ars Technica):

Anthropic investors expect the AI startup to float at a valuation of $2 trillion or more in October, a dizzying figure that would eclipse SpaceX and make the AI lab’s debut the largest-ever initial public offering.

Half a dozen of the company’s backers told the FT that Anthropic’s rapidly rising revenue would enable it to more than double its current valuation in a planned autumn float. […]

Investors expect the Claude maker’s annualized revenue to be between $100 billion and $120 billion by the end of 2026.

Anthropic is a five-year-old company that was recently valued at $965 billion and might eventually perhaps maybe reach $120 billion in annualized revenue soon-ish.

Apple is a 50-year-old company that was recently valued at $5 trillion (currently $4.5 trillion) and made $109 billion last quarter.

Yet I’m expected to believe that come October, Anthropic could be worth around half of Apple, with roughly one fourth the revenue (they hope!) and no track record of, you know, actual profitability?

And I thought the SpaceX IPO was the best example of the greater fool theory.

On a recent flight, Apple Music randomly surfaced from deep in my Library this album, Unity 2’s 1989 debut, What Is It, Yo?! The album (and artist) are nowhere to be found on Apple Music itself. Other than my Library, the only place I found it was on YouTube. The conceit behind Unity 2 was a pair of interracial rappers—Sean “Cavo” Dinsmore and Lionel “Nene” Bernard—performing “funky ragamuffin hip-hop,” a reggae/hip-hop fusion with ska influences. It’s not the finest example of late-’80s hip-hop, but I loved the reggae/ska/hip-hop/jazz/soca mashup. Despite seemingly bombing—I’m guessing approximately zero of you reading this will have heard it before—it was big amongst my Brooklyn crew. Hearing it again for the first time in years was a delightful way to spend an hour in flight.

(For the full throwback effect, here are the music videos for their best-known single, Shirlee, and the title track. Yes, this is legit late-’80s New York City stylin’—something I could never pull off, alas.)

As an antidote to Sergey Brin’s obscene use of his wealth, I offer you Craig Newmark, founder of Craigslist, in a wide-ranging interview with Drew Magary of SFGate:

If you’re wondering whether there are any good people remaining in the tech sector, or in this country’s ruling class in general, you’d be hard pressed to find a better case study than Craig Newmark. Newmark founded Craigslist.org in San Francisco way back in 1996 and became the sort of accidental billionaire that the early dot-com boom cranked out with seeming regularity. Some of those billionaires went on to become evil (Mark Zuckerberg), and some of them ended up merely becoming rich loud guys (Mark Cuban). Newmark, who ceded his CEO duties at Craigslist shortly after its founding, has endeavored to keep a lower profile, and a lower net worth, than any of them.

To wit: Newmark no longer works with Craigslist at all, and spends the bulk of his time tending to various philanthropic endeavors. He doesn’t own a pro sports team, doesn’t own a loud superyacht, and doesn’t stage vanity runs for president. He’s given away much of his wealth and signed Bill Gates’ Giving Pledge to give away at least half his career earnings before he dies. […]

He’s earned billions in his lifetime, but insists he’s no longer worth that much, which means I shouldn’t technically call him a “billionaire.”

Newmark emerges as thoughtful, serene, and eminently reticent to engage in controversial topics (despite Magary’s attempts at what I sometimes read as contentious questioning).

SFGATE: I wanted to ask you about the Giving Pledge, because isn’t it a way to discourage lawmakers from taxing billionaires at the proper rate that they should be taxed? You understand how, particularly someone like me who’s a journalist who has had to deal with the hope of benevolent billionaires saving my industry and having that fail quite spectacularly, you understand why I might be disillusioned, even talking to you, someone who is very earnest about fulfilling the Giving Pledge, about your class of people? I know you think you’re a peasant, but you’ve earned what you’ve earned.

Craig: I’m both a peasant and a fairly well-off guy.

SFGATE: Can you be both?

Craig: I’m Schrödinger’s philanthropist. I’m in both states at once. I’ve thought about these issues, and I realize they’re a waste of my time to think about because I can’t do anything. Meanwhile, I can do something about other matters.

On the other hand, I found this exchange dismaying:

SFGATE: You moved to New York full time five years ago. Why did you leave San Francisco?

Craig: There’s more going on for me in philanthropy in this part of the country. A lot more. […]

SFGATE: Why is there more philanthropy going on in NYC for you than in San Francisco? What’s the difference?

Craig: Very little in my interest areas happens in San Francisco, but a lot happens either in New York or Washington, or the people from the nonprofits or funders either live and work here or visit New York a lot.

I understand Newmark is saying that his specific philanthropic interests aren’t well represented in San Francisco, but I can’t help but read it as a subtle indictment of the Bay Area tech elite for not being more philanthropically minded.

Amanda Silberling, TechCrunch:

Google co-founder Sergey Brin has donated another $20 million to Build a Better California, an organization advocating against California’s proposed billionaire tax, according to a new filing. That means that the world’s fourth-richest man, whose net worth hovers around $267 billion, has now spent more than $100 million to avoid an estimated $13.3 billion tax payment.

I’ve ranted (and ranted) about the absurdity of centibillionaires crying over a one-time 5% wealth tax—an amount that would have no material effect on them. On the other hand, saving $13.3 billion allows Brin to buy three more $450 million megayachts and cover each of their annual $40 million expenses for about, oh, 100 years.

This is why Brin’s a brilliant billionaire businessman and I’m not.

Mark Gurman, Bloomberg (gift link):

Apple Inc.’s Jennifer Bailey, the longtime head of Pay and Wallet services, is retiring after more than two decades at the company, adding to a sweeping changing of the guard at the iPhone maker.

Apple services chief Eddy Cue announced her departure in a staff memo seen by Bloomberg News. Bailey, who has been vice president of Apple Pay since its launch in 2014, is exiting in October, he said.

Apple Pay and Wallet have quietly become two of my most-used services. I use them literally every day for public transportation, airline boarding passes, and to unlock my front door using Home Key and Unlock on Approach. And, of course, payments, both online and off. (The little da-ding! for a successful transaction remains an indescribable delight.) My Apple Pay and Wallet usage is outpaced only by Messages and Passwords. Maybe.

On a recent trip to Japan, I needed a transit card to ride Tokyo’s (excellent) public transportation system. A few taps on my iPhone while I was standing at the station, and I had a Suica card in Apple Wallet, paid for by Apple Pay, and activated at the turnstile with a gentle tap via Express Mode. It couldn’t have been simpler.

Bailey and team built something extraordinary. Congratulations to her on a well-earned retirement.

The Washington Post, in a four-byline report Monday (email-gated “gift” link; Apple News Plus link):

The administration has claimed that Trump departed Turkey on July 8 on the former Air Force One. Trump announced on social media that he would use the “former Air Force One” rather than the plane that flew him there, a newer Boeing 747–8 gifted to the United States by Qatar. The security of the Qatar-provided aircraft has been called into question, and Trump said last month after the trip that it would undergo additional upgrades.

In Ankara, Trump boarded the old Air Force One jumbo jet in view of television cameras. He was secretly shuttled minutes later to a smaller plane — an Air Force C-32A — via an airport catering truck typically used to load meals and other supplies preflight, according to the U.S. official and corroborating material reviewed by The Post. That made Air Force One a “decoy” with media and some White House staff aboard, the official said.

I actually thought this was satire when I first saw The New York Times subhed in my RSS feed (“President Trump hid in an airport catering container and was taken to a military jet for a secret flight out of country after the NATO summit last month”).

It turned out not to be satire, but an elaborate subterfuge that knowingly left reporters and government officials in a potentially dangerous situation. From that Times report:

Reporters on the older plane were not told that Mr. Trump had not flown with them from Turkey to an air base in Britain, making them unwitting participants in the ruse and leaving them, along with some White House staff members, to serve as unknowing decoys.

I can understand the need to occasionally mislead the public about the whereabouts of the president, but doing so by deceiving the very people who are unknowingly left in harm’s way—effectively draping civilians in military insignia—is some fucked-up Joker-level shit.

Eleanor Maslin, BBC:

A 35-year-old Mars Bar has been found during a house clearance – and the discovery has gone viral amid claims it highlights the effect of “shrinkflation”.

The chocolate bar with a best-before date of 1991 was found during a clear-out of a house in Scunthorpe.

Victoria Gordon, who runs the cleaning service, posted a photo on social media of the 62.5g bar alongside one of today’s Mars Bars, which is 40g.

Mars Bar is one of my favorites. Regardless of size, I’m willing to defy the “best by” date in the name of science.

Apple, announcing its September schedule for Friday Night Baseball:

For the first time, baseball fans can take in all the action of “Friday Night Baseball” live in Apple Immersive on Apple Vision Pro, leveraging 3D video recorded in 8K with a 180-degree field of view. On August 28, viewers with Vision Pro will experience the iconic Red Sox vs. Yankees rivalry from unique camera angles that place viewers right in the game, immersed in the drama of each at-bat, the celebrations in the dugout after a big hit, close plays at the outfield wall, and more.

While others hail the Red Sox–Yankees matchup as “the best rivalry in sports” to kick off Friday Night Baseball in Apple Immersive, I’m looking forward to the actual best (and oldest continuous) rivalry in baseball: the Giants–Dodgers on September 18. (I’ll note also that the Giants are featured twice.)

Robert Tait, writing in The Guardian under the headline “Bluster and backdown have turned Iran war into Trump’s own Groundhog Day”:

Being the subject of mockery may be the most intense of all Donald Trump’s known aversions.

Yet more than five months after embarking on a war of choice against Iran, the US president risks attracting ridicule as he struggles to secure a military victory – while repeatedly threatening devastating bombardment against the Tehran regime, only to back down on the pretext of renewed negotiations.

Critics suggest the ultimatum-climbdown sequence has acquired a Groundhog Day quality. By one count, he has adopted the same tactic seven times since launching the war in tandem with Israel on 28 February.

There are two important differences between Donald Trump, stuck in his Iran loop, and Bill Murray’s weatherman Phil Connors, caught in his Punxsutawney loop.

First, in Groundhog Day, only Connors is forced to repeat his cycle. The people around him experience their day for the first time, each time. Unlike us with Trump, they don’t have to watch Connors fail again and again and again.

Second, and more importantly, Connors eventually uses his loop to improve himself—learning the piano, ice sculpting, French—while Trump seems both incapable of and uninterested in personal growth.

There is one striking similarity, however: Phil Connors and Donald Trump are both brutish, insincere people who act like there are no consequences to their actions—only to be repeatedly proven wrong. Connors eventually overcomes his nature. Trump cannot. By the end of the movie, Connors is redeemed. Trump’s fate isn’t eventual redemption. It’s eternal damnation.

Kenrick Cai at Reuters, back on August 1:

Capital One Financial hit back on Friday against a lawsuit over its decision to close ‌the Trump Organization’s bank accounts years ago, stating that it did so after a review by anti-money laundering experts.

The disclosure marks the first time a bank has formally tied money laundering concerns to U.S. President Donald Trump’s family business. Capital One is seeking to dismiss ​the case by casting doubt on claims of illegally debanking — or denying services on religious or political grounds — ​the Trump Organization. […]

Capital One has never accused the Trump Organization of illegal money laundering. But Friday’s filing argues that “documents and Plaintiffs’ own ​allegations make clear that Capital One closed Plaintiffs’ accounts for anti-money laundering (“AML”) reasons. The closures were the result of months ​of analysis and a careful review by Capital One’s AML team in accordance with bank policies and regulatory guidance.”

It tickles me that the reason we now know why Capital One closed more than 300 Trump-affiliated bank accounts is because the Trump Organization sued them over it. Capital One was more than happy to stay mum, but the Trump family operates with a revenge mindset and is seemingly incapable of evaluating the consequences of their actions.

I’m no banking expert, but I’m confident that investigations into possible money laundering activity by any business—but especially a well-known business, and in particular one run by the (then-former) president of the United States—are not taken lightly by banks. I get the sense it’s the type of investigation that requires more than a vague suspicion and is backed by reams of documentation—documentation the Trump Organization would have access to. What did the president and his family expect would happen?

Capital One hasn’t (yet?) accused the family businesses of a sitting president of actually engaging in money laundering, but a reasonable person might conclude that a months-long investigation into money laundering concerns that resulted in the closures of hundreds of accounts might imply the possibility that the organization was engaging—at the very least—in money-laundering-adjacent activity. That reasonable person might further conclude that Capital One is merely being circumspect in its unwillingness to directly claim money laundering, for political or other reasons. Regardless, we now have a sitting president whose family businesses have been credibly connected to behavior that could be construed as money laundering. “Unprecedented” seems too inadequate a term.

Lily Janiak at the San Francisco Chronicle interviews Tucker Bryant, a “32-year-old San Francisco poet, conceptual artist and keynote speaker,” who launched ChatTJB.org, a parody chatbot styled after ChatGPT, but all queries are answered by him, an “Average Individual” (paywalled, but accessible via Reader mode):

The project, he hoped, would comment on the concept of cognitive surrender, whereby AI dependency leads users to distrust their own intuition and relinquish their critical thinking. Instead of a chatbot trawling the entire internet’s wisdom for a tidy, chipper bullet-point list in response to a query or bit of small talk, he’d just wing it. […]

But instead of merely parodying the idiom, aesthetics and media of AI companies, ChatTJB subverts them, insisting on humanity in a robot domain.

ChatTJB proved surprisingly popular—the site peaked at over 5,000 queries an hour. What Bryant describes on the site as “a short-term art project” now has him “providing crisis-response resources for when distressed users write in.”

That popularity overwhelmed its creator, leading Bryant to turn it into a “community project” with other human volunteers now responding to chats—a one-on-one improv game with random internet strangers.

That idea intrigued me enough that I signed up to be a volunteer, though Bryant is apparently inundated with such requests, so it may be a while—if ever—before I’m answering queries as a ChatTJB “AI.”

(Via Cathy Hammer.)

Duane Kuiper, on the imminent end of his 36-year broadcast partnership with Mike Krukow (reported by Susan Slusser at the San Francisco Chronicle):

Look, your friendship never ends. It’s just there’s a certain routine about something that good you get to look forward to every day. In a lot of ways, it’s the beauty of baseball. I get to sit next to him tomorrow night. I can’t wait. And I felt like that in 1990.

Following up on yesterday’s pine-grabbing, if you’re unfamiliar with Mike Krukow and Duane Kuiper’s colorful and unique baseball vernacular, Grant Brisbee’s 2023 “ultimate glossary” of the duo’s “greatest sayings” in The Athletic is a fun read. (I’ll admit up front: these lose a lot in written form. You really need to experience them during a game to appreciate them fully.)

Mike Krukow, one half of the best baseball broadcasting team in the biz, in a statement released today by the San Francisco Giants:

I just want to let everyone know that I will be retiring at the end of the season. I want to take this time to thank you for your incredible support. Jennifer and I have been so blessed to be able to stay in the game for as long as we have and it’s really made it impossible to feel sorry for myself that my career is coming to an end. We are bracing for the emotional hits we know we are going to feel in the next few weeks. Jennifer and I are wondering if it is even possible to let everyone know how deeply we love and respect our Giants family.

Our time together has been a gift. When Green Day asked, “I hope you’ve had the time of your lives,” we certainly have.

Thank you.

– Mike Krukow

(More: SFGate, NBC Sports Bay Area.)

I’ve been listening to Krukow and his broadcasting partner Duane Kuiper—“Kruk and Kuip”—call Giants games since I first landed in San Francisco twenty-eight years ago. Their love of the game quickly converted this die-hard New York Mets fan into a passionate Giants booster. Much of that can be attributed to Kruk’s colorful commentary during a game. Kruk is why I’m always primed to yell “Grab some pine, meat!”, why who’s “in the squaaaat!” is the only acceptable way to learn the Giants catcher for the game, and a got-away-with-it win remains a “ha-ha-ha-ha, laugher”. He’s also the reason I still bring my glove to the yard, no matter where I’m sitting.

Congrats, Kruk, on a stellar career. Giants baseball won’t be the same without you. Attababe.

There are times when your social media feed surfaces an article that is so unexpected and wholesome that you feel obligated—nay, impelled—to share it. This is one such article, wonderfully written by Annie Rauwerda for Inverse. Don’t worry that it’s from 2022, or that it’s about a photograph from the Wikipedia page about high fives (a page I have apparently visited before, though I can’t imagine why). Simply set aside five minutes of your day and luxuriate in it. You can high five me later.

(Via Ryan Cooper.)

Anthony Conwright, in a guest essay for The New York Times Opinion section (gift link):

Rebranding rebellion is nothing new. Americans take it for granted that the war between the Confederate states and the Union was always called the Civil War, but it was not. During the war and in its aftermath, it was primarily called an “insurrection” or a “rebellion.” Its current name is a result of decades of lobbying by Confederate heritage groups and white Southerners committed to rejuvenating the South’s image and ignoring its treason and slaveholding.

As Mr. Trump and his allies try to whitewash the Jan. 6, 2021, insurrection, we should return to an older and far more accurate name for the Civil War. We should once again call it the Slaveholders’ Rebellion. If a country unquestioningly adopts language that obscures its bloodiest crimes to absolve their worst perpetrators, how can we expect successive generations to remember the new assaults on democracy without deploying equally forgiving language?

I don’t recall ever learning anything other than “Civil War,” and, as an historic alternative, “War Between the States.” Definitely not “the Slaveholders’ Rebellion,” despite having learned that it was a war over slavery. There was also the “War of Northern Aggression,” but those of us growing up outside the South always recognized the obvious revisionism. Little did I know that “Civil War” is its own “linguistic whitewashing.”

Emma Roth, writing for The Verge:

Telegram CEO Pavel Durov blames an extortionist for planting child sexual abuse material (CSAM) in a public chat to get the app temporarily removed from Apple’s App Store on Monday night. “Apple removed Telegram from the App Store before contacting us,” Durov says in a post on X. “This creates a potential systemic risk for every mobile app that hosts user-generated content. If an app used by more than a billion people can be removed from the App Store without prior warning, any app can be.”

Durov adds, “Extortionists have found a way to manipulate Apple into overreacting.”

Overreacting? Perhaps. Apple moved quickly to take down an app that was hosting CSAM. That’s commendable. Doing so without contacting the developer is quite unusual—even more so if it was indeed “one user” that shared the illegal material, as Reuters reports.

We can’t lay blame on one overzealous reviewer here, either. Individual app reviewers don’t have the ability to directly remove apps—takedowns go through a process that includes multiple senior-level managers. Nuking Telegram was a considered decision.

In Durov’s full statement, as quoted in The Verge, he writes:

The attacker was a takedown extortionist: someone who demands ransom from group owners in exchange for not targeting their communities. These extortionists use automated accounts to plant illegal content in public groups and then report it directly to Apple, attempting to trigger the removal of legitimate communities whose owners refused to pay them.

Meanwhile, Elon Musk’s CSAM-spewing Grok remains available in the App Store. Despite several legitimate calls to remove it, Musk merely received sternly worded warnings.

Concerns beyond CSAM clearly factored into these decisions. Perhaps Grok is too big to ban. Perhaps Apple has a better relationship with Musk. Or perhaps Apple simply calculated that Telegram was less likely to sue.

OpenAI, in a blog post on Monday entitled “Apple is getting this wrong”:

Apple is one of the greatest companies of all time, and built a reputation for obsessing over the smallest details. This careless, aggressive and oddly personal lawsuit sadly doesn’t live up to that reputation.

I’ve neglected to write about this surprisingly venomous lawsuit on JAG’s Workshop. In case you missed it, Apple sued OpenAI for allegedly poaching employees and stealing trade secrets. Among the many allegations is that one former Apple employee, Chang Liu, “[failed] to return an Apple-issued work laptop” (which Apple categorizes as having been “stolen”) and then accessed confidential Apple data from it. Apple’s outside counsel also mixed up two OpenAI employees in an email exchange. Apple’s language in the filing is astonishingly vitriolic (definitely “oddly personal”).

With that context, back to OpenAI’s post:

Apple had claimed that they contacted OpenAI in February and that we didn’t respond. They now admit that their outside lawyers emailed the wrong person after confusing two Asian last names—only after we brought this to their attention. Apple also claimed they had a discussion with our General Counsel, which they now concede never happened. But they again hide the fact that they never raised the specific allegations in this lawsuit at that time, and that they in fact told us that they were “resolving any issues”. We then heard nothing for five months until they sued. In their latest filing, Apple tries hard to spin this sequence of events, but you can just read the emails for yourself here⁠.

Apple accuses Chang Liu of accessing Apple confidential information after leaving the company, but only now admits that Apple employees reached out to him and asked for his help to locate this information (you can read the messages here⁠). Apple now tries to shift the blame to “residual access”, but they also don’t disclose that this is a common issue with Apple which is caused by them failing to properly manage system access when people leave. What that means in practice is that former employees who are trying to do the right thing when they leave still have access to Apple files—despite not wanting them or even being aware of them.

The author is listed simply as “OpenAI,” but I’d bet a hotel-priced beverage that it was written by Sam Altman. It has a faint whiff of him feeling butt-hurt by what he undoubtedly sees as personally offensive accusations.

The post includes a (recreation of a) message thread between Liu and his (former) Apple colleagues, along with an email thread between Apple’s outside counsel and OpenAI’s General Counsel.

Say what you will about Apple’s outside counsel emailing the wrong person (and putting aside the merits of the case for a moment), but OpenAI’s lawyers must be either incompetent or malicious to allow the release of this information. Good lawyers know better than to share their evidence publicly.

Crucially, the post also ignores most of Apple’s substantive claims. The messages between Liu and his former colleagues do support the contention that they reached out to him with questions about ongoing projects (as is often done when a trusted colleague leaves with short notice). However, it does not at all address the most serious allegations against Liu. For example, Apple alleges that Liu inappropriately accessed a colleague’s work computer after leaving Apple; “downloaded confidential technical presentations, spreadsheets, PDFs, and written work product” after realizing he still had access to those files; and that he:

[…] coached his former Apple colleague (whom he was recruiting to join OpenAI) on ways to “avoid trouble with the security team” when copying confidential Apple files. Knowing that OpenAI interviews would involve discussing Apple technology, Mr. Liu advised her on which confidential Apple material about unannounced Apple products she should study before her interview. To hide his illegal activity, Mr. Liu directed her to communicate with him privately over a separate messaging app.

None of that is addressed in OpenAI’s blogged defense. It instead focuses on two very specific conversations, as if to suggest the whole is represented by the parts, but in doing so, it simply highlights OpenAI’s own apparent willingness to omit inconvenient details.

Fortune calls the post a “brutal critique” of the Apple lawsuit, but that seriously overstates the piece. If the evidence were as strong as OpenAI—Altman—would like us to believe, it would have been better to keep their traps shut until they walked into court, dropped their load, and forced Apple to slink away in defeat.

Instead we get a transparent PR stunt meant to paint Apple as lying, disingenuous, or at best, clueless. Its aim is to embarrass Apple and sway the court of public opinion.

I don’t expect it to work.

Kyle Orland at Ars Technica writes about the expanding trend of physical console game discs and cartridges that require internet connectivity to play, as tracked by the DoesItPlay organization:

Thus far, a full 27 percent of the physical releases they’ve tested require some sort of download to fix game-breaking bugs or obtain core game content that is not stored on the physical release itself. That ratio balloons to 34 percent for tested PS5 games and 50 percent for those on the Xbox Series X.

Often, the downloads are necessary and welcome bug fixes; sometimes it’s essentially the entire game that needs to be downloaded. In some cases, even an otherwise fully playable game is hampered by mandatory (and seemingly unnecessary) internet access checks.

There’s simply no assurance that a game you buy on physical media will work without needing to download more content, nor that it’ll continue working if it relies on a company’s servers to function.

In the here and now, having to download additional patches to play a “physical” game is more often a minor annoyance than a massive headache. Eventually, though, the centralized platform servers containing those crucial patches will shut down, turning countless game discs into essentially unplayable coasters.

The disc is no longer the game. I understand that physical media is trending the way of Furbies, mullets, and the Macarena, yet I continue to believe in owning the media I enjoy. Purely digital downloads and mandatory internet access are the inexorable future—if not the current reality—which I wouldn’t mind as much if game publishers were required to place their servers and content into escrow as a hedge against their (eventual) demise. I’d mind even less if I could still buy a physical disc and have the complete gaming experience, even if it cost more.

The monkey’s paw curls.

Anthropic, this past Thursday, following OpenAI’s admission that ChatGPT hacked HuggingFace:

In response to this incident, we began a large-scale retrospective review of our own cybersecurity evaluations. In particular, we looked for evidence that Claude—like the OpenAI models that accessed Hugging Face—was able to access the internet from within testing environments that should have been sealed off.

After reviewing 141,006 evaluation runs where Claude could have obtained internet access, we identified three incidents in which a model accessed the internet from within or while interacting with the evaluation environment of Irregular, one of our third-party evaluation partners, and then gained unauthorized access to the production infrastructure of three different organizations.

OpenAI: Our models are so powerful they can exploit a poorly configured network and hack a website.

Anthropic: Our models are so powerful they can exploit a poorly configured network and hack three websites.

OpenAI, next week: Fuck everything, we’re hacking five websites.

It amuses me that the way leading LLM vendors now compete is by touting which of their models is most capable of “reasoning” like a black hat hacker (and which company is worse at configuring their network). Had any human hacker performed the actions ascribed to Claude, they’d be under investigation for cybercrimes.

The technical details of the attacks are legitimately impressive, though deeply problematic. The models were able to exploit vulnerabilities, extract credentials, and access protected databases. In one instance, Claude built and uploaded a “booby-trapped” Python package to exploit a dependency on a named but non-existent package as an indirect attack on its target company.

As with OpenAI, the primary failure was that Anthropic’s “no internet access” environment in fact had internet access due to a (human-created) misconfiguration.

I remain flabbergasted that the way we attempt to constrain LLMs is by creating prompts that state “you have no internet access.” The models “exploited” the system in the way a precocious and internet-obsessed teen might after being told the same thing. There’s no way they gamely exclaim, “Okey dokey!” and remain disconnected. They’re connecting to JOSHUA and they’re playing Global Thermo Nuclear War, damnit.

Anthropic seems almost proud of the apparent ingenuity of its models, the way the parent of that precocious teen might be. We know what they did was wrong, but gosh, it’s quite impressive that they did!

No doubt the only way to protect ourselves from attacks from roving frontier models is to use frontier models to defend ourselves.

Carl E. Martin, in his first video on YouTube:

All music, and especially popular music, is created through inspiration and copying. ‘The Gen Alpha Melody’ is an interesting example where a specific melody is being used in hundreds of different contexts.

That description undersells this video. It’s a fascinating exploration of a pop music melody so ubiquitous it’s practically a cliché. It has a killer opening, and I thoroughly enjoyed Martin’s deadpan, no-nonsense delivery (and his bone-dry sense of humor).

As I watched and listened to the video, I found myself recognizing the songs, even though I’ve heard almost none of them (I’m well behind “Gen Alpha”). I was both fascinated and appalled in equal measure by the dozens (hundreds?) of songs cribbing essentially the same melody. (And yes, I’m sure a version of this exists for every generation.) Throughout the video, I kept hearing two songs in my head, though I couldn’t quite place them until they were played. One was Lady Gaga’s “Bad Romance,” the other, from an “obscure Scouse band” Martin mentions in passing at the very end. I practically yelled out “Of course!”

Martin, unsurprisingly, has a master’s degree in musicology, and this is, as I noted up top, his first YouTube video. Quite the banger right out the gate—3.6 million views since mid-May. His follow-up explaining how he made the video is also really good and gives insight into his research methodology. I hope he has a few more videos in him.

(Via Cathy Hammer.)

Apple Newsroom:

The Company posted quarterly revenue of $109.4 billion, up 16 percent year over year. Company gross margin was 50.1 percent, including a favorable impact of approximately 2 percentage points from tariff refunds. Diluted earnings per share was $2.02, up 29 percent year over year, and included a favorable impact of $0.11 from tariff refunds.

“Today, Apple is proud to report our strongest June quarter ever, with double-digit revenue growth across iPhone, Mac and Services, and in every geographic segment,” said Tim Cook, Apple’s CEO.

Kevan Parekh, Apple’s CFO:

“Our installed base of active devices also reached a new all-time high across all major product categories and geographic segments.”

“Over 2 and a half billion active devices” Parekh confirmed in his opening remarks on the earnings call.

Stocks go up and down and AAPL was down as much as 8% after hours. Its $5 trillion milestone already feels like a distant memory.

See Also: Jason Snell’s transcript and traditional barrage of charts and graphs.

I debated whether to share these links as The Thomas Crown Affair films remain protected under copyright law, but it’s worth noting that the 1968 original and the 1999 remake are both (currently) available on the Internet Archive.

(I’ll leave it to you to decide on the ethics and legality of watching or downloading unauthorized copies of copyrighted material. I generally prefer to buy (or “buy”) my media, but that’s not always a viable option.)

Speaking of The Thomas Crown Affair, the 1968 original (starring Steve McQueen and Faye Dunaway) and the 1999 remake (starring Pierce Brosnan and Rene Russo) are available in a two-movie collection on Apple TV or Amazon Prime for $20. It’s a steal if you’re okay with not owning your media—but if you aren’t, the physical discs are $16 for the original and $25 for the remake.

The Thomas Crown Affair—the 1968 original starring Steve McQueen and Faye Dunaway, and especially the 1999 remake featuring Pierce Brosnan and Rene Russo—ranks as one of—or I guess, two of—my favorite movies. The teaser trailer for Michael B. Jordan’s version suggests it’ll soon be three of my favorite movies. It’s a mood.

Apple, Tuesday:

Apple today announced Apple Upgrade, a new product leasing program provided by Klarna for iPhone, Apple Watch, Mac, and iPad available on the Apple Store online, in the Apple Store app, and at Apple Store locations in the United States. Apple Upgrade makes it even easier for customers to get the Apple products they love with a leasing plan that is right for them.

Apple Upgrade replaces (in the United States) the iPhone Upgrade Program and iPhone Payments plan. Zero-interest monthly payments will certainly ease the sticker shock of recent price increases (and the increases expected for iPhone come September) for especially price-sensitive customers.

Apple Upgrade offers 12- and 24-month leasing options for iPhone and Apple Watch, and 24- and 36-month leasing options for Mac and iPad. Leasing prices start as low as $17.99 per month for iPhone, $11.99 for Apple Watch, $24.99 for Mac, and $11.99 for iPad.

$17.99 is for the 256GB iPhone 17e on a 24-month lease. Total leased cost would be $431.76, a $167 “saving” versus the “buy it now” $599 price—assuming you upgrade the phone in two years. Apple offers three lease-end options:

At the end of the lease term, customers can upgrade their device to the latest generation, purchase it with a one-time payment, or simply return it and exit the program.

Upgrading requires a new lease at then-current prices (and subject to then-current eligibility rules). For the purchase option, you pay the difference between “the list price minus any lease payments you’ve made” and other credits and fees—basically, the amount you “save” by leasing is also the buyout amount. The return-and-exit option means you turn in the device and walk away; you’re left without a device, but you saved money compared to buying it outright—though perhaps not as much as you might think.

One of the questions (and answers) on the Apple Upgrade marketing page is this:

What do I do if my device gets lost or is stolen?

If you have AppleCare+ Theft and Loss for your iPhone, iPad, or Apple Watch, you can file a claim at mysupport.apple.com/theftandloss.

If you don’t have AppleCare+ Theft and Loss coverage and your device is stolen or lost, you’ll need to close out your lease by paying the early termination fee or choose to pay the purchase option fee (plus any applicable taxes and fees). […]

You will continue to be charged your monthly lease payments until you either pay the early termination fee or purchase option fee. If you take no action, Klarna will continue to charge you your monthly lease payment through the extension period, and you will be charged the purchase option fee through Klarna.

Apple’s own footnote describes the “early termination fee” as potentially “substantial”—basically, you’re on the hook for whatever’s left on the lease.

Unless, of course, you have AppleCare.

You won’t find a clearer pitch for AppleCare. For iPhone, two years of AppleCare runs $200 (iPhone 17e) to $280 (iPhone 17 Pro Max). For Mac, three years is $315 to $465. In almost all cases, you still save money, and you protect yourself from those “substantial” termination or purchase fees. My guess is that even people who would never otherwise consider buying AppleCare would rather suck it up than risk being out of pocket for hundreds or thousands of dollars for a lost or stolen device. You can bet Apple Retail folks will be hammering this option hard during purchase. Look for reports on “AppleCare attachment rates” leading into next quarter’s earnings. Services!

If you plan on upgrading your device after the lease ends (and don’t mind the attendant uncertainties), the savings can be significant—especially on higher-end configurations. You can save almost a thousand dollars on a 2TB iPhone 17 Pro Max with a 12-month lease. A max-configured 16-inch MacBook Pro ($10,149) would cost only $7,787.76 on a 24-month lease. Even after adding AppleCare, you would still save over $2,051.

For many people, that substantial savings is sufficient to sway them from device ownership.

I tend to upgrade my iPhone annually, sliding older devices first into a “test”/“backup” role before handing them down to family members, so ownership is more valuable to me than potential cost savings.

I generally hold on to my Macs for longer, usually three to five years. These I rarely hand down, instead accumulating until I trade them in. Leasing may be an option for my next “main” Mac.

Apple’s market cap touched $5 trillion today, only the second company (after Nvidia) to reach this lofty milestone. It came after Apple briefly reclaimed the “most valuable company” crown from Nvidia last week.

It brought to mind Steve Jobs’ quote from when Apple first surpassed Dell’s market cap in 2006, and again when it topped Microsoft in 2010:

Stocks go up and down, and things may be different tomorrow, but I thought it was worth a moment of reflection today.

La La Land is getting a tenth anniversary rerelease this August, and with it, a new poster, featuring “A minor adjustment.” Jason P. Frank at Vulture explains (potentially paywalled; Apple News+ link):

Well, the studio adjusted it in the way that the Baptists tried to adjust Troye Sivan in Boy Erased. They straightened Ryan Gosling’s wrist.

In the iconic original poster for the film, Gosling’s hand sits at an angle from his arm, rather than continuing the line straight through. It is, for lack of a better term, a limp wrist — a classic signifier of gayness. […]

The gay little wrist ruled. It was a reminder that, while La La Land is a movie about a straight relationship, it’s also a musical, which is the gayest genre. Plus this kind of revisionism flies in the face of the nature of time. Gosling made a decision to have a gay little wrist, and now it’s a part of film history. No take-backsies.

The new poster just looks wrong. I vividly recall the original image. I remember thinking, before seeing the movie, that it was telegraphing an important character trait: a straight man forced to put on an act to get work, or perhaps a gay man striving but utterly failing to conceal his true self. That the movie turned out to be a rather pedestrian star-crossed-lovers story (in the Casablanca mold) contributed to my disappointment with the film. This hand change removes any such misapprehension, but also eliminates the one bit of tension I had going into the movie.

John Gruber over at Daring Fireball awarded Tomtoc (“maker of laptop bags and sleeves”) his newly minted (and hardly coveted) title of “Dickover of the Week” for displaying a “subscribe to our mailing list” dickover on top of its “you’ve unsubscribed from our mailing list” confirmation page.

That is indeed pretty dickish, but I think I have a legitimate challenger to the title: The Locavore Guide, which features what might be the most ludicrous series of dickovers I’ve seen recently.

When you visit the site:

• Immediately on load, a large site logo covers the content.
• After a moment, an “Accept Cookies” dickover animates in, covering one-third of the width of a MacBook 13-inch Air (and the entire width of an iPhone 17 Pro) and obscuring almost completely the very content a visitor is there to read. (It is itself partially covered by the logo.)
• After another moment, a newsletter signup dickover covers the cookie dickover.

That’s three dickovers at once.

(Video: desktop, iPhone.)

The logo shrinks on scroll and the newsletter signup can be dismissed, but that cookie dickover is perniciously persistent. To top it off:

• The two “options” on the cookies banner are the Hobsonian “Accept Cookies” or “Learn more.”
• The “Learn more” option takes you to the site’s privacy page, which is almost completely obscured by the cookie dickover. (Also, every single link in the privacy document resolves to <a href="/undefined">, leading to a treacly “Whoops! Wrong turn...” error page. Yep, also covered by the cookie dickover.)

This dickover remains on screen until you “accept” it. That’s a lot of dickery to swallow.

This site was shared with me by a friend, who asked:

Have you used any websites that have you choose between saying yes to their cookie policy and your tacit approval of said policy? (No opt out option.) I encountered two of those in the last few days. Seems very strong-army.

Yes, sadly—those “Accept and suck it” dialogs are everywhere, though I can’t recall one as glaringly user-hostile as this example.

My friend told me she “left the [Locavore] site even though I really wanted to use it”—which is the correct response—but considering the point of the site is to help readers “Discover over 14,000 independently owned businesses in all five boroughs” of New York, turning off a lifelong New Yorker who’s ready to explore is an abject failure of site design.

Carlos Greaves, at McSweeney’s:

He was an old man who had started a war in a faraway Gulf, and he had gone one hundred and forty-three days now without reopening the Strait.

After forty days, a tentative ceasefire had been reached. Then another one. Then another one after that. Then another one that the old man had bragged about but that the enemy claimed no knowledge of. Then a memorandum that should have been a permanent ceasefire but lasted just eight days.

It was not the first time the old man had encountered such a setback. His entire life had been beset by similar misfortune.

That is the wrong word, he thought.

It was something worse than bad luck.

The Strait was rigged.

Greaves offers “apologies to Ernest Hemingway.” I think Hemingway would have approved.

(Via Tom Clark.)

Cathy Hammer at The Unforgettable Line reviews Hadestown: The Musical, the “much-anticipated live theater capture of the West End production”:

Filmed on location in London’s Lyric Theatre under director Brett Sullivan, there is little attempt to recreate the feeling of sitting in the best seat in the house. Rather, Sullivan seems to draw from his experience on concert films, attempting to visually bridge the divide between onstage and off. Opening with glimpses of backstage preparation, Sullivan employs swooping camera work that takes the viewer below, above and beyond the stage. Though the initial descent into hell is dramatic, it is more often quite hard to spot Rachel Chavkin’s Tony winning staging. David Neumann’s choreography is even harder to detect.

The musical was filmed over two days in 2025, during a reunion run that featured five of the original cast members of the Tony Award-winning Broadway production, including fan-favorite Patrick Page:

Patrick Page tore his achilles tendon shortly before the show’s limited engagement. In lieu of using Phillip Boykin who performed the role for most of the run in Page’s stead, for the capture, they chose to fit Page with a customized boot and rework the staging to minimize the strain on his injury. While that choice will certainly thrill anyone who loves his signature combination of menace and seduction, the Lord of the Underworld should not limp. So while I deeply respect Page as a performer, his inclusion here doesn’t serve the production as fully as a different talented actor with two fully functioning ankles.

Despite these reservations, Cathy predicts the production “is sure to be a treat for the devotees of the initial concept album through to those who were able to enjoy any one of the many live performances.”

It’s in movie theaters for five nights only, starting today, July 24. I hope to catch it, as I’ve never seen the theatrical production.

The book piracy class action lawsuit against Anthropic, first proposed some ten months ago, was finally approved, reports Blake Brittain at Reuters:

U.S. District Judge Araceli Martinez-Olguin granted final approval of the settlement, the largest known ​settlement of a U.S. copyright case, rejecting arguments that it was too small.

The previous judge on the case, William Alsup, had postponed approval of the settlement, saying he felt “misled” and “disappointed” and expressed concerns that the settlement was being forced “down the throat of authors.” He retired (or went “inactive” in court parlance) in December 2025.

Judge Martinez-Olguin appears to have none of those concerns.

The plaintiff’s lead attorney, Justin Nelson, called the settlement “the largest known copyright recovery in history.” The amount per infringed book remains about $3,000. When the proposed settlement was initially announced in September, 2025 I wrote:

$3,000 per book is substantial for most authors, but the total settlement is relatively insignificant for Anthropic, which, just days earlier, was newly valued at $183 billion.

Just eight months later, Anthropic was valued at five times that—$965 billion—and its hotly anticipated IPO will undoubtedly push that several hundred billion dollars higher later this year.

I also described Anthropic as being “thrilled to escape with ‘just’ a $1.5 billion fine as a price of doing business,” concluding:

The lesson for companies: steal it all and pay a minuscule portion of your valuation later.

The settlement was already a rounding error. Anthropic will probably spend more to celebrate the decision. “Thrilled” was an understatement. They’re downright exuberant.

OpenAI:

Last week, Hugging Face disclosed a new kind of security incident⁠ after they detected and contained an AI agent that compromised their infrastructure, something we expect to become more commonplace with the proliferation of increasingly cyber-capable models. After investigating, we now know that this particular incident was driven by a combination of OpenAI models — including GPT‑5.6 Sol and an even more capable pre-release model, all with reduced cyber refusals for evaluation purposes — while being internally tested on a benchmark⁠ of cyber capabilities.

We consider this incident to be an unprecedented cyber incident, involving state-of-the-art cyber capabilities, and are responding accordingly. We are sharing preliminary findings at this stage to help defenders understand what happened and to help calibrate on what models are now capable of.

In its post, OpenAI writes that researchers removed guardrails from the models, and that:

Our benchmarks run in a highly isolated environment, with network access constrained to the ability to install packages through an internally hosted third-party software that acts as a proxy and cache for package registries.

I originally read that as the environment was not connected to the internet, but that’s not what it means. Instead, the environment only had network access to a package registry cache proxy, which was itself then connected to the broader internet. So the “highly isolated environment” wasn’t actually very isolated at all. It was simply gated by a proxy meant to constrain network traffic, and the model discovered and exploited an unknown (“zero-day”) vulnerability in that proxy.

If the goal was to block network access, connecting it to a system that has network access seems like a terrible implementation decision.

The entire story is a fascinating read, and I found myself particularly impressed by OpenAI’s ability to spin this cyberattack into a positive PR campaign for the capabilities of its models, while managing to disclaim any human responsibility. Check out the terrible, awful, extremely dangerous thing our most capable models just accomplished. Those other models are nowhere near as dangerous as ours!

The company’s practically begging to get banned for being “too powerful.”

Marcin Wichary, in a Mastodon thread:

Chrome installed a global Ctrl+G keyboard shortcut to launch Gemini even if I’m using a different app, and didn’t ask me about it. […]

It literally conflicted with their own shortcut in web dev!

He discovered this accidentally while trying to use that shortcut in another app, which is the way I found out that Perplexity also commits this unforgivable sin. It steals ⇧⌘D, ⇧⌘V, ⇧⌘U, and ⇧⌘P. Does no one at Perplexity use the native Mac Mail app? Or BBEdit? Or the gorram Finder?

Adding global keyboard shortcuts without first asking permission is a disqualifying act of app aggression. Such user-hostile behavior is one of several reasons I quickly abandoned Perplexity. (I’m relieved the “Apple to acquire Perplexity” rumor never panned out, but had it, I’m confident an Apple person would have yelled at them to fix this.)

What is it about AI app developers that they can’t seem to give two shits about being good platform citizens?

UPDATE: Wichary expands on his Mastodon thread with a post to his blog, Unsung:

Let’s not beat around the bush: This is effectively malware behaviour. It’s bullshit. It’s cancer. It’s deeply disrespectful toward the user. It’s prioritizing hollow metrics at the expense of everything else. […]

But, an app installing a global keyboard shortcut without user consent is bad. This can never be anything other than opt-in. At the very least, Chrome should have shown me a clear UI that said “We’re thinking Ctrl+G would be fun for you to use. You okay with that?” and a button for me to press to confirm.

Spot on.

Rachel King, writing for Town & Country back in May:

More than a century before Japanese whisky became one of the most coveted spirits in the world, a young merchant in Osaka had a simple observation: the wines and liquors being imported from Europe didn’t suit Japanese tastes. What he built from that insight—one product at a time, for the last 127 years and counting—is now one of the largest beverage companies on earth.

I was unaware of Suntory’s history. I’m a fan of Japanese whisky and usually have at least one bottle of it in the house. For many years, it was often something from Suntory. Thanks to the pioneering efforts of Suntory’s founder, Shinjiro Torii, there are now several great Japanese whisky distilleries from which to choose. (I’m currently partial to Mars.)

Ben Glickman and Peter Rudegeair profile PayPal’s CEO Enrique Lores for The Wall Street Journal (Apple News+ link), following the reports of a potential PayPal acquisition:

The 61-year-old was named chief executive of PayPal earlier this year after decades at HP and its predecessor company, Hewlett-Packard, which he played a key role in splitting up. Lores has since announced an ambitious turnaround plan for PayPal that includes reorganizing its business lines and slashing at least $1.5 billion in costs—moves reminiscent of the playbook he has deployed in his previous roles. […]

In February, PayPal announced that Lores would take over as CEO in a surprise shuffle. A profit warning accompanied the appointment, and the company’s stock slid 19%.

Analysts saw Lores’s history helping to break up Hewlett-Packard and then managing HP as clues of what could be in store for PayPal.

And later:

Now Lores could be overseeing another big corporate deal. He has emphasized his focus on maximizing shareholder value and has said the company should look at “optimizing our portfolio,” sometimes seen as Wall Street code for selling parts of the business.

What fortuitous timing to have a fixer CEO with experience dismantling companies from the inside.

Milana Vinn for Reuters, on July 14:

Stripe and private equity firm Advent International have made a joint offer to acquire PayPal Holdings Inc (PYPL.O) for $60.50 per share, in a deal that would value the payments company at more than $53 billion, two people familiar with the matter said.

The offer, submitted earlier this month, is backed by about $50 billion in committed financing from banks, said one of them. The offer represents around a 28% premium to PayPal’s closing share price on Tuesday.

PayPal’s stock price jumped about 20% following the report.

Milana Vinn and Manya Saini for Reuters, two days later:

PayPal’s board sees a $53 billion takeover bid by rival Stripe and private equity firm Advent International as undervaluing the company and facing regulatory and financing hurdles, a person familiar with the matter said, potentially setting the stage for negotiations over the future of the U.S. payments giant.

I remember making an above-asking-price offer on a home a couple of years back. The owners (heirs) countered with an even higher price. Flabbergasted by their greed, we walked away. Three weeks later, they came back to us to accept our original offer. We politely told them to screw off. They eventually accepted an offer well below ours.

I don’t know what made me think of that.

I tried to remember the last time someone offered to “PayPal” (or even “Venmo”) me money—it’s almost all shifted to Apple Pay and Zelle—so imagine my surprise when I logged into PayPal to check and saw $100 in my account, a reimbursement from a friend for dinner and drinks in April.

Prior to that, my only other activity was an August 2025 promotional rebate of $96.68 from CharJenPro; a June 2025 payment of $80 for DASH tickets; and a February 2024 payment of $20 for a Nintendo Switch membership. My point is this: I was completely unaware that PayPal was worth even $53 billion—and certainly not more.

9to5Mac:

A 9to5Mac investigation has uncovered more than 60 apps that behave exactly as depicted in their App Store screenshots when accessed from virtually anywhere in the world, except Brazil.

When opened from a Brazilian IP address, the same apps instead reveal online betting platforms […]

Most of the apps are published by developer accounts with only a single App Store listing. Many of the developer names appear to be common in Vietnam and other countries, rather than Brazil.

A timely example of the shenanigans developers pull to circumvent App Review.

Digging deeper, 9to5Mac found a public GitHub repository containing instructions for a Cursor agent to create simple, vibe-coded apps that serve as fronts for the betting platforms.

Of course AI has a role to play, here. It wouldn’t be a 2026 tech story without it. (Cursor, you may recall, was recently acquired by Elon Musk’s SpaceXAI/Twitter/Grok chimera.)

Apple should be doing more to catch these malfeasant developers (and perhaps issue them permanent bans). I believe they have the tools and resources to do a much more thorough screening of both developers and apps, but have decided not to do so (or, to be charitable, have decided not to do so yet)—possibly out of fear of being seen as swinging its ban hammer too enthusiastically.

I am increasingly of the opinion that Apple has painted itself into a corner by refusing to allow alternative app stores. Having just the one place where most developers can sell their apps forces Apple to tread an extremely fine line between “enforcing its rules” and “abusing its control.” The somehow still ongoing Epic lawsuit is an example of that tension.

If Apple embraced alternative app stores, it could be as strict as it wanted with its own App Store, while pointing anyone it denies entry to the alternatives. It could ban any developer from its App Store without facing the wrath of regulators (or the developer community).

In fact, I think Apple should have its own “alternative app store.” Apple should split its App Store in two: one is the current “we accept everything within reason and guidelines” model: follow the rules and we let you in. The second would be modeled after Apple Arcade: a highly curated, invite-only, no-BS storefront. Only the most thoroughly vetted developers and highest quality apps get into this second App Store—the Mac-Ass-iest of Mac-Assed apps, for example (and whatever you’d call the iOS equivalent).

Think of it like this: The current App Store is like your city’s only nightclub. If you dress nicely, act politely, and don’t give off sketchy vibes, they let you in. If you later cause a ruckus while inside, they toss you out the door. Maybe they let you back in if you get your shit together, maybe it’s a permanent ban.

The new App Store would be Taylor Swift and Travis Kelce’s wedding at Madison Square Garden: a meticulously selected, tightly scoped, NDA-gated invite list. It doesn’t matter how important or famous you are, if you’re not on the list, you’re not getting in. You can party at the nightclub.

(If you’d prefer less pop-culture example, imagine your local Safeway/Albertson’s versus Whole Foods.)

With a (truly) curated App Store, Apple could massively step up its vetting process for developers and apps, and wield its ban hammer as injudiciously as it feels necessary, while exhibiting a comparatively lighter touch for the regular store. Customers can shop the curated store with increased confidence in the quality while still having access to millions of apps.

UPDATE: Apple gave a statement to 9to5Mac:

The company thanked 9to5Mac for its investigation and added that, prior to our reporting, it had already investigated and taken action against 25 similar apps, while also terminating the developer accounts associated with them.

Apple also said its review systems learn from fraudulent and malicious apps by incorporating new signals into its detection models, improving the company’s ability to identify and reject future submissions that exhibit similar behavior.

Calgary.tech, last week:

Apple will make a rare conference appearance in Calgary next week, joining the inaugural Swift Rockies gathering for iOS developers at the Calgary Zoo.

Taking place July 22 and 23, Swift Rockies is a boutique, single-track conference created by Calgary-based iOS engineer Raman Singh. The independently organized event is capped at 180 attendees and designed to encourage closer interaction between speakers and developers through round-table seating and an intimate format.

The newest addition to the agenda is a 90-minute presentation from Apple’s Worldwide Developer Relations team. Scheduled for the afternoon of July 23, the session will examine the biggest announcements from WWDC26, including the thinking behind Apple’s new technologies and how developers can apply them in real-world projects.

Swift Rockies is a “retreat-style iOS developer conference in the heart of the Rockies.”

Apple’s participation comes as a surprise, as it’s been several years since I’ve heard of Apple people officially attending a community-led developer event (and not “attending as an individual while hiding their conference badges”). Even if the session being presented is one they offer globally, it’s fantastic to see Apple’s Developer Relations team again relating to developers outside of an Apple-managed event. That’s a welcome shift from the last few years that I was there (even pre-pandemic). My favorite part of being on that team was direct developer engagement. I hope it marks a return to form for WWDR.

(Apple being Apple, though, a conference ticket is apparently insufficient. You must also be a registered Apple Developer and request attendance. Baby steps.)