Link Log

And Another Thing About Squircles ⇥ mjtsai.com

If you cannot get enough squircle talk, Michael Tsai has a whole roundup of different takes, some of which you may not have seen already.

Here is another thing: robbing icons of a variable for creative excellence and flexibility is just another form of context collapse. You know how Instagram and YouTube are host to professionals and amateurs alike? There are benefits to creating an impression of similar legitimacy, but the rigidity of social media platforms’ formats also collapses the difference between legitimate information and absolute nonsense.

Losing some of the artistry in an icon makes it more difficult to distinguish between legitimate and well-crafted Mac apps, and everything else. It is not a perfect proxy, to be sure, and there are some very nice squircle icons. But it is nevertheless a unfortunate change that makes apps of varying levels of quality look more similar.

Amazon Will Print Book-Length A.I. Gibberish on Demand ⇥ nytimes.com

Kashmir Hill, of the New York Times (gift link), learned about an A.I.-generated biography of her for sale on Amazon for $27. It is just one of many wholly-generated books available there:

Amazon does not mind if people hawk A.I.-generated books on its platform, unless they are truly and deeply terrible. “Charlie Kirk: An Inspiring Journey of Young Political Conservative and Activist Who Fights for America,” published in February 2025, became an Amazon best seller after Mr. Kirk was killed last September — which means it probably sold thousands of copies. But after dozens of scathing reviews called it “mind-numbing,” “a scam” and “a disgrace,” Amazon took it down.

Too bad about all the trees killed for this print-on-demand nonsense, though at least it fits with Amazon’s decline into one of the world’s biggest retailers of sketchy, counterfeit, and knock-off products.

Morgan Stanley Says the ‘Space’ in SpaceX Is Worth About $8 Per Share ⇥ ft.com

Bryce Elder, on the Financial Times’ Alphaville blog:

Here’s paragraph one of the Morgan Stanley’s SpaceX initiation note:

With an ‘X of 1’ position in space infrastructure, we believe SpaceX can convert energy into intelligence at scale with optionality to monetize through a range of consumer and enterprise solutions for the next era of AI… the final frontier.

[Adam] Jonas — formerly the Wall Street bank’s Tesla and occasionally other auto companies analyst — was last year reallocated to the sort of free-radical futurologist role we’d thought had been cornered by Shingy.

Today’s news that SpaceX has already started dipping below its IPO valuation reminded me of this piece. Remind me — is being compared to Shingy good?

At any rate, Morgan Stanley estimates that the combination of X and Grok will grow this year to generate only a little less revenue than Twitter did in 2021, its last full fiscal year as a public company. Fear not, investors, as the bank also says SpaceX will make $17 billion in “enterprise A.I.”, and then nearly triple that next year. I guess everyone wants to give money to the CSAM and misogynist fantasy generator for business.

Apple Updates Its Advertising Policies ⇥ ads.apple.com

Sarah Perez, TechCrunch:

In a newly published Apple Advertising Services policy, effective as of July 14, 2026, the iPhone maker shares its rules for advertising on Apple Maps. Notably, it prohibits the broad category of home services businesses, like plumbing, electrical, locksmith, HVAC, pest control, roofing, and general contracting services, among others.

If Apple is interested in updating its advertising policy further, I suggest none. I spent lots of money on its nominally premium products, and I pay a monthly fee to use the company’s services. Alas, here we are.

Apple is also prohibiting ads for cryptocurrency ATMs. Also, there is this:

2.7.2 Ad content that directly or indirectly promotes or facilitates the sale of products or services that compete with Apple hardware products (e.g., mobile phones/smartphones, tablets, notebook/desktop computers, and smart watches) is reviewed on a case-by-case basis.

I do not see this in its previous restricted or unacceptable ad guidelines.

Eric Benjamin Seufert, who writes Mobile Dev Memo, noticed the phrase “on the relevant Apple software applications or Apple devices” has been changed to remove references to Apple-specific devices or software:

The new language could simply accommodate the availability of Apple-owned services on the web and through third-party devices and operating systems; the Apple TV app, for instance, is available on smart TVs, streaming devices, and game consoles. But the addition of “other properties” is conspicuously broad and appears to give Apple the contractual latitude to distribute ads beyond its own services entirely. This would allow for a material expansion of the company’s advertising surface area.

Apple Maps is also available on the web and is used by DuckDuckGo, so this could simply be covering that inevitable broader placement. But the reality is that Apple is now operating an advertising business, and putting more ads in more places is one way to make the numbers go up. Not the user satisfaction numbers, of course — I cannot imagine a single user who wants more ads in their life, unless they own shares in this company. But it will print money at no cost, and that is what the people running a huge corporation with little competition want to do. It was not inevitable until Apple made it so by opening the door.

The Shape of Apps ⇥ parakeet.co

Paul Kafasis, writing on the Rogue Amoeba blog:

With last year’s release of MacOS 26 (Tahoe), Apple made a mess of app icons. In the first betas of MacOS 27 (Golden Gate), however, there are signs of a turnaround. We’re urging Apple to continue making improvements, by restoring the ability for MacOS app icons to have distinct shapes.

Kafasis reignited my simmering frustration with the mandated squircle in MacOS. My Dock contains three of the apps shown in the collection in this post: MarsEdit, NetNewsWire, and Sketch. I like their current more-uniform icons fine enough, but they are less distinguished than the ones these applications used to have.

But maybe that is the whole point?

Louie Mantia, writing on the Parakeet blog:

The shape of apps is a squircle. And it has been proven to work for everyone. Companies can use their logo as an app icon. Designers can create something specifically for the platform. And both of these get to look like an app. Whether people consider the squircle a container or a canvas, this uniform appearance communicates its function: a squircle represents an app, just like how a piece of paper represents a digital document, or a folder represents, well, a folder.

Semantically, there’s something really beautiful about that. As an icon designer, I appreciate that different types of things have visual distinction.

Mantia touches on all the pragmatic reasons to unify the shape of icons in an operating system, all of which I have considered, and then drops the above paragraphs — and things started to make more sense to me. This is a different way to think about it. This is not a situation where a cleaner looks like a zesty beverage with toxic consequences. The shared general function of these icons does help communicate something and makes them less ambiguous in that sense.

But a broad category of functionality is only part of the story of an icon and — with respect to Mantia’s long and illustrious history of work in this area, and that of co-Parakeeter Luka Grafera — taking away a difference of shape also limits what an icon can communicate. It may not be a cleaning product that looks and is packaged like juice, but imagine if every consumable liquid was in identical bottles with only a different label. You might go for something refreshing after a workout and end up drinking soup. Sure, you can argue the label for a beverage should not look similar to the one for soup, but the two would be far easier to distinguish if they were not in the same package.

I still think constraining designers to a singular shape, while more clearly defining specific objects as apps, has made it harder to distinguish between them, particularly when combined with the glassy and contrast-killing layer effects of Tahoe. Happily, though not retreating on the squircle, the shapes within icons in MacOS Golden Gate are at least more clearly defined.

Meta’s ‘Activity from Other Businesses’ ⇥ yashgarg.dev

Yash Garg:

A friend pointed me to this option in Meta Accounts Center called “Activity from other businesses”, which shows “activity sent from other businesses or organizations to show you relevant content.” Their main help page doesn’t even work in my region.

If you dig around in your Meta account privacy settings, you can turn this feature off. While you are there, though, you might take a scroll through the audience-based advertising list. These are companies that have, most often for me, “uploaded or used a list” of email addresses or phone numbers. In my case, this is hundreds of businesses — some of which I recognize or can see why they would have my contact information, and many of which I have never heard of. Advertisers are supposed to have permission to use this information, of course, but there is basically no way for me to confirm whether I gave permission or report that I did not. Restricting further use is also comically unfriendly: it is a button two levels deep and, for all but the first four advertisers, you must click a “See more” button each time to display the full list.

Update: Rodrigo Ghedin counts four separate privacy-hostile things Meta has done or is rumoured to be working on since the beginning of June.

Lies Told About Data Centres ⇥ karlbode.com

Karl Bode:

One lie that companies have been telling local municipalities is that if they greenlight a massive local AI data center, it will immediately bring a flood of savvy innovators to your podunk-ass town.

The promotional materials for Kevin O’Leary’s still hypothetical “data centre park” imagine a futuristic campus full of bright young minds doing complicated A.I. stuff on-site in north-central Alberta. But why would they be there — fifty kilometres from the nearest city and 500 kilometres from Edmonton, the nearest major city? Why would they not be in Vancouver, or Silicon Valley, or anywhere else with an internet connection?

Also:

These companies aren’t coincidentally aiming construction at states and municipalities that are too broken and corrupted to put up meaningful regulatory opposition. […]

That is one reason why we are seeing a bunch of these proposals here. And these companies benefit from a lack of transparency.

Perhaps Meta Should Not Have Spent Decades Being Creepy ⇥ pivot-to-ai.com

Meta, in a press release called “Meta’s A.I. Glasses: Your Questions Answered”:

Can’t people just cover up or disable the LED?

The camera is disabled when people try to do this. Beginning with our second generation of glasses, the camera is automatically disabled if we detect that the capture LED has been blocked. No photos or videos can be taken until we detect that the light is unblocked.

Since the introduction of this safeguard, we’ve seen some people go beyond using tape to sophisticated efforts to modify or destroy the capture LED. We are continuously improving our ability to detect tampering, and now we’re updating the glasses to disable the camera if they detect the LED was physically tampered with or destroyed. No other kind of camera has done this and we’re proud to lead the industry forward.

Meta is not being entirely honest here. For many, many years, Apple’s laptops have contained a camera indicator light with among the highest security protections possible. Over ten years ago, iSight cameras were not adequately secured. I cannot find a more recent example showing a similar vulnerability, at least suggesting a better level of protection in today’s cameras. Other computers also have built-in cameras with in-use indicators, with various approaches to security, some of which have vulnerabilities. In general, though, it is not new for an indicator light to resist tampering as long as the camera remains functional.

What is different is the threat. Cameras built into computers need protection mostly from remote attacks, while Meta’s glasses need protection from an owner deliberately altering them. Meta is selling creep glasses and hoping it can outsmart everyone buying them — and that the rest of us similarly trust Meta to protect our privacy.

David Gerard, Pivot to A.I.:

But it gets better! Meta’s planning a new version of the glasses that records continuously! [FT, archive]

a new hardware line of smart glasses that would continuously record audio while taking photos every few seconds.

And they won’t have a recording light. […]

This product is still rumoured, and perhaps the shipping version will have some kind of external recording indicator. But given the way Meta would intend a device like this to be used, I doubt it will, otherwise it would be indicating basically all the time.

Meta is in the business of asking for forgiveness instead of seeking permission. It will release these regardless of public approval, and with the belief it can control how that continuous recording is used. But determined people will surely find workarounds and vulnerabilities, and Meta surely believes it can trust itself to fix them. But I do not. Meta has not earned the right to ship anything like this without incurring deep suspicion about the product and anyone using it.

ShinyHunters Leaks Madison Square Garden Surveillance Records ⇥ 404media.co

Joseph Cox, 404 Media, last month:

ShinyHunters published the [Madison Square Garden] data on Tuesday. The full file download is nearly 45GB. A spokesperson of the group sent 404 Media a smaller sample of the data. One file includes what appear to be emails sent by customers to MSG and sometimes MSG’s response. One email is a man complaining about potentially being flagged by MSG’s facial recognition systems (MSG owner Jim Dolan has long spied on people inside his arenas, with MSG deploying various surveillance technologies, WIRED reported.)

Noah Shachtman and Maddy Varner, Wired:

The talent database also tracks some celebrities’ race, gender identity, and sexual orientation; 93 entries are marked as “LGBTQIA.” Why MSG felt the need to label Ricky Martin or Phoebe Bridgers or Geese’s Emily Green in this way is unclear.

“I’ve never met James Dolan. I don’t know the higher-up leadership at Madison Square Garden. But, like, there does seem to be a bit of a pattern here,” says Evan Greer, director of the digital rights group Fight for the Future, citing WIRED’s reporting on the Garden’s minute-by-minute surveillance of a trans woman. “They just seem overly interested in queer and trans people in their venue,” Greer adds.

I cannot imagine what it would be like to be the person who put this together, stepped back from their computer, and thought yes, this seems fine. How surveilling people like this is not criminal, I have no idea.

Update: Madison Square Garden is suing Wired for its story, calling it false and “with reckless disregard for the truth” while also acknowledging that, according to Travis Bland at Consequence, it kept “information on celebrities’ sexual orientations” — but for good reasons. Bland’s article appears to have a link to the lawsuit, but it is a document in Madison Square Garden’s internal SharePoint, so it is password-protected. If you are interested in reading the suit, there is a valid link in the press release (PDF) announcing this legal action.

We Need Modern Equivalents to Those Old ‘Twin Cam’ Badges on Cars ⇥ theautopian.com

Jason Torchinsky, the Autopian:

You know what I miss that cars often had up until the ’90s or so, and are almost extinct today? Badges. Not the usual make/model badges, though, we still have plenty of those, I mean the highly-specific feature badges that cars used to proudly display. Actually, not even just features — also technical details.

Torchinsky’s ideas for modern versions of this are very good.

Apple Sues OpenAI Alleging Trade Secret Theft ⇥ cnbc.com

Kif Leswing and MacKenzie Sigalos, CNBC:

Apple on Friday sued OpenAI in federal court in Northern California, alleging trade secret theft, saying that the artificial intelligence lab took the iPhone maker’s intellectual property in order to develop its own consumer hardware.

The docket is on CourtListener, along with Apple’s complaint (PDF). What is immediately notable about this whole thing is the myriad conflicts of interest. Apple, of course, has a current partnership with OpenAI to power Apple Intelligence features, something repeatedly mentioned on the previous version of its marketing webpage, and conspicuously absent on the current one.

But the history runs far deeper. Among the defendants in the suit is io Products — do not blame me for that capitalization choice — founded by Jony Ive and acquired by OpenAI last May. Total mentions of “Ive” in the body of this lawsuit? Zero. Also, one of the investors in Ive’s company was Emerson Collective, founded by Laurene Powell Jobs, and which now has equity in OpenAI. Neither Emerson nor Powell Jobs are defendants.

In May, Kalley Huang of the New York Times reported OpenAI was “weighing legal action such as sending Apple a notice claiming breach of contract” because it felt Apple was not doing enough to promote ChatGPT through its Apple Intelligence integration. I would be surprised if that modest tie-in exists for much longer. When you update to Siri A.I. in iOS 27, for example, ChatGPT is turned off by default, though it is possible to re-enable it. It is also listed in Settings under an “Extension” sub-section — currently singular, though perhaps not at some point in the future.

Serial Fraudsters Are Responsible for an Offensive Cybersecurity Startup ⇥ krebsonsecurity.com

Brian Krebs:

The X/Twitter account IRIS C2 (@C2IRIS) has gained more than 4,000 followers since its creation in January 2025, posting frequently about security vulnerabilities, AI and software exploits. IRIS C2 says it is a company in McLean, Va. that sells offensive cybersecurity capabilities.

[…]

A search on the Arlington, Va. address listed in the incorporation records for Calvexa Group LLC finds the property is occupied by Jack Burkman, the 60-year-old founder and managing partner of the lobbying firm Burkman & Associates. When approached with questions about IRIS C2, Burkman referred further inquiries to his longtime associate, 28-year-old Jacob Wohl.

Now there are a couple of names I have not heard in a while, probably because their previous scheme was established under fake names. In an interview with Krebs, Wohl claims to “know more about tech than anyone”, which sounds about as delusional as anything else he has ever said.

Anyway, if you want to risk getting scammed, the company claims to be hiring.

Meta Announces First Canadian Data Centre to Be Built North of Edmonton ⇥ about.fb.com

Meta:

Once complete, our Sturgeon County data center will represent an investment of more than CAD $13 billion. We anticipate approximately 3,000 construction workers will be onsite at the peak of construction, and the data center will support more than 300 operational jobs.

[…]

As with all of our data centers, this data center’s electricity use will be matched with 100% clean and renewable energy.

As for the data centre itself, however…

Chris Varcoe, Calgary Herald:

Last week, Calgary-based Pembina Pipelines, Morgan Stanley Infrastructure Partners and Kineticor announced they will build a new 932-megawatt (MW) gas-fired power generation facility in Sturgeon County to power the data centre, although they didn’t identify a customer at the time.

Chalk up another questionable outcome for the environmental record. Meta makes a lot of big promises in this news release and in an interview for Varcoe’s column, but I question whether the company can be adequately monitored, particularly by this provincial government.

Flock and Ring Are Champions of the Privatized Surveillance State ⇥ engadget.com

Tim Cushing, Techdirt:

Even if you truly believe the company you work for is capable of doing this, perhaps read the room a bit before offering up this sort of insane assertion to a journalist:

Langley offers a prediction: In less than 10 years, Flock’s cameras, airborne and fixed, will eradicate almost all crime in the U.S. 

That would be Flock Safety CEO (and co-founder) Garrett Langley speaking to Thomas Brewster of Forbes. Flock Safety has grown a lot over the past few years, following paths paved by Amazon’s doorbell surveillance camera acquisition, Ring, and other upstarts in the public/private surveillance mesh network field.

Ring founder Jamie Siminoff made a similarly unsubstantiated claim to Jennifer Pattison Tuohy of the Verge. The reality, however, is that cameras from companies like Flock and Ring are giving a sheen of authority to false accusations made by law enforcement.

In September, for example, Chrisanna Elser of Colorado was falsely accused of theft by a police officer using Flock’s cameras. The entire exchange was recorded on her doorbell camera. 404 Media obtained body camera footage which showed the officer refused to see exonerating evidence, in the form of cameras on her Rivian truck and a Ring camera at her tailor’s house. This is just a mass surveillance race to the bottom. The officer in question was required to complete additional training for politeness as a result.

Joel Feder, the Drive:

On an otherwise normal Sunday afternoon in late June, I’d decided to take the $155,000 Range Rover I was testing that week out to run some errands with my wife. Little did I know that choice would complete a technological chain linking surveillance cameras, AI, and law enforcement that led to me and my wife being surrounded by police, hands on their guns, in a Kohl’s parking lot in suburban Minnesota.

Without spoiling this story too much, this was the result of a typo that affected several more vehicles than the one Feder was driving.

Max Miller, Engadget:

Although Flock cameras are often referred to as license plate readers, that’s reductive. Reading license plates is their primary task, but they can be used to track just about anyone or anything. Even without a license plate, law enforcement officers can search for things such as, hypothetically, “green sedan with American flag bumper sticker,” or, “pickup truck with paint scratches on left side and dirt bike in truck bed.” Reducing Flock ALPRs to license plate readers is a bit like calling your own eyes “Engadget article readers” simply because that’s what you’re using them for at this particular moment. The company also offers AI surveillance cameras which do track individuals.

I keep thinking about Elser’s story. The way she was implicated was thanks to a network of cameras surveilling her every public move, and the way she was exonerated was because of documentary evidence from a bunch of cameras surveilling her every public move. It was not very long ago that U.S. media was freaking out about the number of CCTV cameras in the U.K., but the U.S. has quickly caught up. There are hundreds across Canada, too.

To state the obvious, this ubiquitous surveillance has not eradicated crime. What it has done is give police the false confidence to accuse random people of crimes they did not commit. It has also allowed police to stalk people for personal reasons, despite major investors Andreessen Horowitz claiming critics “overlook the vigilant protections in place to ensure that Flock cannot be used for surveillance or to violate privacy”. That is nonsense — perhaps a lie, similar to those told by Flock. But it is not a lie any greater than the idea that we can eradicate crime if we just have more cameras with A.I. features.

Google Discontinues Google Earth Pro for Desktop ⇥ support.google.com

“Aamir F.”, on behalf of the Google Earth Team:

We’re continuing to make Google Earth on web and mobile (Android | iOS) the best place for people to get helpful geospatial insights. While you can continue using the legacy Google Earth Pro desktop app, it will no longer be available for new downloads beginning on June 25, 2027. We encourage using web and mobile for the best Google Earth experience.

This is a two-factor loss: the MacOS app is Intel-based and, thus, will no longer be supported by the system come next year.

Christoph Grützner, researcher at the University of Jena:

Discontinuing Google Earth desktop does not come unexpected, but it’s terrible news. The web tool is utterly useless for me and many geo folk.

There’ll be workarounds for most features, but not for easy 3D view of historical imagery and for sharing placemarks.

Carlos Moffat, faculty at the University of Delaware:

Still one of the most approachable and efficient tools for planning fieldwork. I used GEarth in our latest Antarctic cruise just a few months ago.

On the Google Earth community thread, you will find comments from people with all sorts of interesting use cases for the desktop app. I opened the web version today to see if it supports my limited use cases — it does — and I was greeted by a dialog advertising Gemini features, of course.

An Age-Gated Internet Is an Admission of Defeat ⇥ jamesrball.com

I am a tiny bit sorry for all the links I am posting regarding age gating. As I have written before, it is something I am still trying to work out for myself. After all, it seems straightforward why age verification is an easy way to reduce the risks to children of today’s platforms, the operators of which marketed their products directly to kids. But the trade-offs are numerous, whether from a predominantly U.S.-centric view of individual freedoms, or from the way such restrictions fail to address product safety concerns.

James Ball, Techtris:

In practice, of course, a social media ban would be widely circumvented. If we think the internet is bad for teenagers now, imagine a world in which all online access is illicit. Social platforms would have a new defence against online harms against that group: they’re just not meant to be there.

Provided Facebook, X, TikTok or whoever could show it had some legally compliant age verification system, harm befalling a teenager wouldn’t be their fault. They weren’t supposed to be there, after all. Instead of being the negligent owners of a space marketed to teens, they’re the blameless victims of trespass.

Age gates are such a procedural and mediocre response. Instead of reining in whatever broader risks may be invented or exacerbated by these companies’ products — and, in particular, the unique problems of each — we are throwing up our hands and pretending it is all too complicated.

Meta Laid Off Staff in Favour of Agentic Coding, and It Is Not Going Well ⇥ reuters.com

Katie Paul and Courtney Rozen, Reuters:

Meta Chief Executive Mark Zuckerberg acknowledged shortcomings in the company’s sweeping restructuring at an internal town hall on Thursday, saying the systems known as AI agents had not progressed as quickly as he had expected, according to a recording heard by Reuters.

[…]

In retrospect, he said, the “trajectory of the agentic development over at least the last four months hasn’t really accelerated in the way that we expected,” and ​that the company’s bets on the new structure “haven’t come to fruition yet.” Zuckerberg was referring to AI agents, automated systems that can ​execute tasks on behalf of a user.

A quintessentially Zuckerbergian premise: he pivots the whole company around whatever is the new thing, says oops, then reminds himself that nothing really matters as long as people keep looking at ads on Instagram.

For all Meta’s power and its massive base of users and ad buyers, the company’s YouTube channels are fascinating places. The main channel has a bunch of promo videos for the headlining products Meta wants people to associate with a world-changing company. There are corporate presentations hosted by Zuckerberg, ads for A.I. and glasses products, and little behind-the-scenes things, and all of them have hundreds-of-thousands to millions of views. That is what you would expect for an account with 456,000 subscribers and a household name. It really wants you to believe its leadership is made of visionary stuff.

But the way it actually makes its money — advertising — is nowhere to be found on that channel. For that, you need to go to the Meta for Business channel, which has a respectable 181,000 subscribers and lots of tips for how to use the company’s ad tools more effectively. But the view counts on those videos are, frankly, terrible. Most are in the dozens-to-hundreds; again, this is a channel with enough subscribers to get a famous silver plaque.

Meta can tell the world it is a revolutionary company while being internally honest about what it actually does. Many companies do that. But it is fairly troubling that Meta seemingly tricked itself into believing the external picture it paints.

Indonesia Is Also Facing Problems With Age-Gated Social Media ⇥ asia.nikkei.com

Marcel Thee, Nikkei Asia:

Following Australia’s move in late 2025, Indonesia became the first non-Western country to announce restrictions on social media access for users under 16. Rules unveiled on March 6 require platforms such as TikTok, Instagram, YouTube and Roblox to introduce age verification checks and remove underage accounts. The restrictions were supposed to take effect on March 28, but enforcement has been patchy, with tech companies largely ignoring the requirements and underage users continuing to access social media simply by lying about their ages.

As Michael Geist put it recently, “the better the privacy protection, the less effective the ban”.

Maestral, My Favourite Dropbox Client for MacOS, Is No Longer Being Maintained ⇥ github.com

Sam Schott:

As of 2026-07-28, this project is archived. It’s been a fun challenge to develop a syncing client, but unfortunately, I find too little time to invest in Maestral these days. I’ve also moved away from using Dropbox myself.

Maestral will still remain usable in the medium term, but will no longer be actively maintained or receive updates.

This is a bummer — understandable, but a bummer nevertheless. The official Dropbox app is hundreds of megabytes because, of course, it is a website masquerading as a native application with the help of Electron. Maestral is comparatively svelte, has a low memory footprint, and does not use Apple’s questionable File Provider API.

I use Dropbox only for some relatively basic but necessary things, like hosting PDFs for this very website, which is why Maestral has been the perfect client for me. It is a testament to the effect of a good third-party app that I have stuck with Dropbox despite its corporate-focused strategy pivot.

Anyway, many thanks to Schott for creating such a good little utility. Maestral is open source so I have hope somebody will take the baton.

A.I. and Liability ⇥ schneier.com

Bruce Schneier:

AI agents are agents of the person or organization that deploys them—and should be treated by the law as such. If a company hired human writers to write its summaries, that company would be liable for inaccuracies in those summaries. If a company’s human agent signed contracts in the company’s name, that company would be bound by those contracts. And if a doctor gave dangerously wrong medical advice, they would be liable for malpractice.

To allow businesses to hide behind the excuse of faulty AI in those same circumstances would be a massive handout to companies, and would introduce disastrous incentives for corporate misbehavior. Why hire human writers, lawyers or doctors when AIs are not only cheaper, but also absolve employers whenever they make a mistake?

In his video essay nominally about artificial intelligence, economist Cahal Moran repeatedly references the book “The Unaccountability Machine” by Dan Davies. I am only a couple of chapters in, but I think I am going to get a lot out of it, and I feel confident you should check it out from your local library.