Month: November 2022

Matt Novak, Gizmodo:

FTX filed for bankruptcy on Friday, leaving reasonable people to wonder how a cryptocurrency platform founded in 2019, which reached a valuation of $32 billion in 2021, could plummet to zero in such a short time. There’s a new piece in the New York Times which gained exclusive access to FTX founder Sam Bankman-Fried, but if you’re looking for answers, you’re not going to find it there. In fact, the interview with SBF, as he’s often called, is presented with such a gauzy lens that you have to start wondering what the hell is going on with crypto reporting at the Times.

The referenced article reminds me a lot of the coverage from financial sectors after the housing bubble burst at the turn of the century. The Times reporter, David Yaffe-Bellany, acknowledges Bankman-Fried’s guarded responses, but it fails to meaningfully interrogate many of his claims, as Novak documents, while being overly deferential to this industry in general. Perhaps Yaffe-Bellany was unable to get a substantial response which, given the legal jeopardy faced by Bankman-Fried and FTX, makes sense. But its failure to describe how the interview subject is dodging meaningful questions and attempting to control the narrative is effectively helping him do both of those things.

The Times article ends on this curious exchange:

Shortly before the interview, Mr. Bankman-Fried had posted a cryptic tweet: the word “What.” Then he had tweeted the letter H. Asked to explain, Mr. Bankman-Fried said he planned to post the letter A and then the letter P. “It’s going to be more than one word,” he said. “I’m making it up as I go.”

So he was planning a series of cryptic tweets? “Something like that.”

But why? “I don’t know,” he said. “I’m improvising. I think it’s time.”

Again, Yaffe-Bellany did not explore this further. It makes absolutely no sense. Or, rather, it made no sense — until a few Twitter users realized that accounts which archive the deleted tweets of notable cryptosphere personalities would not be triggered if a new tweet was posted at the same time another was deleted. Bankman-Fried’s free jazz tweets is more likely an attempt to cover up his removal of over one hundred posts.

Remarkably, for a country that is as large and as loyal to Apple’s ecosystems as Canada,1 this new Vancouver store is only the third location in the country which is not a boring indoor mall store. Apple’s longstanding spot on Rue Sainte-Catherine is its only true “flagship” store; its DIX30 spot is in an outdoor mall in a Montreal suburb. Canada has six metro areas with over a million residents. Yet not even Toronto has a flagship store, though it is not for a lack of trying.


  1. Apple has a nearly 60% market share here in smartphones, and a 26% share of desktops. ↥︎

Josh Eidelson, Bloomberg:

Workers say that whereas the focus of an Apple Genius used to be to impress customers with a high level of service, they and other employees are now increasingly pressured to upsell. They’re pushed to prioritize “ownership opportunities,” the company’s euphemism for persuading people to buy new gadgets instead of repairing their old ones. They’re also evaluated based on how many customers pay for an extended warranty through the AppleCare program and how many people they deal with per hour. Some stores email workers’ stats to colleagues or post them on the wall in employee-only areas, with those of lower sellers highlighted in red.

Bloomberg published a similar story three years ago — I linked to it at the time — and several others have reported declining satisfaction with the Apple Store for the past decade. These stories feel like the moves of people in management seeking short-term gains rather than compounding long-term returns on investments. I say good for the employees unionizing and creating standards for themselves and their work.

David Roth, Defector:

It’s a cycle. People create something, together, that reflects their energy and weird work; that thing becomes compelling as a result, and that makes it valuable, and at some point someone puts a price on it and someone else pays that price. It is at that moment that the thing begins to change. The new owner will almost always decide that what is most interesting about this thing is not the human essence that gave it value, but The Owner Himself, and will act accordingly. People will come back for the valuable stuff until the owner succeeds in crowding it out; when that crowding is done, the owned thing dies. Until then, what’s left is just what’s valuable—the humanity and brilliance and unpredictability and fun that all that cynical and idiotic and self-serving wealth is always and everywhere busy replacing with itself. There’s nothing to do but look for the good stuff until the looking becomes too challenging, or until it’s gone.

You may disagree with Roth’s headline thesis — “everything is Silicon Valley now” — or his tie-in with the story du jour, Twitter, or his analysis of baseball’s problems. But the paragraph above? That is something to keep pinned in your brain. For most of us, it is a reminder to be wary of how things are changed in exploitative ways; for those in power, it should be seen as a cautionary pattern.

Jason Del Rey, Recode:

The sponsor-ification of the Amazon shopping experience is just the latest twist. If you’re looking closely enough, a quick search on Amazon for, say, “iPhone screen protector” or “youth soccer socks” will only turn up paid product listings carrying a “Sponsored” label at the top of the results before scrolling.

[…]

In recent years, Amazon has found new ways to toe the line on how clearly it labels its ads. In one case, it tucked a smaller, lighter “sponsored label” underneath a bigger, bolder section label of “Highly Rated.” Amazon previously had a section for organic results called “Top Rated” but launched the “Highly Rated” widget as a section for sponsored listings only. Graham, the Amazon spokesperson, said the company has since released a second type of Highly Rated section that includes both organic and sponsored results.

Along the way, Amazon has essentially given up on effectively surfacing the best products to customers in an organic fashion, and has outsourced that key function to advertisers.

The sponsor listings run deeper and sketchier than you might think. Take a close look at a search for cloth napkins, and you will see:

  • a sponsored banner at the top from Cotton Clinic, with two listings,

  • sponsored products filling the first four positions of the search results grid,

  • the “Highly Rated” sponsored section just a little further down the page,

  • a section which involves some kind of co-branding deal where Amazon links to a Daily Mail article about the “best cloth napkins” and the Mail gets a commission from purchases,

  • another four sponsored listings mixed into the organic-looking results grid,

  • another four sponsored listings at the end of the organic-looking results grid,

  • and then a sponsored video banner just above the page controls for bamboo-based paper towels.

That sure is a lot of ads for one page of about sixty total results — about one-quarter of everything shown. All this comes on top of accusations of self-preferencing, copycat products, the often low quality of the products it labels “Amazon’s Choice”, and the fly-by-night brands that frequently pollute search results. Shopping on Amazon means navigating a labyrinth of crap, paid listings, and paid listings for crap.

Del Rey:

The Amazon spokesperson said that advertising on Amazon only works well if the company makes ads useful for customers, and that a variety of factors determine when and where ads are placed throughout the shopping experience.

I get what is being said here, but it feels like something of a self-fulfilling prophecy. If increasing amounts of the homepage are ads and many of those ads reflect actual products that people may be shopping for, then they likely “work” by some metric Amazon is using. For a start, Amazon seems to go out of its way to make the “sponsored” label as close to invisible, so many people probably do not know they are clicking on an ad. Del Rey’s reporting reveals many sellers are having to increase prices to pay for ads so they remain high in search results, too. How well they actually work for sellers and customers is hard to say given how high-quality organic listings are squeezed between advertising and drop-shipped junk.

I agree that labelling paid spots is better than supermarkets selling shelf space and not acknowledging it.

“Annoying” is probably not a word a developer would like his app to be associated with.

I certainly don’t mind.

Thankfully, users have also thought of my app as “indispensable” and a “game changer”.

It has helped people living with ADHD and early dementia. It has saved marriages. And it seems to have even saved lives. I’d love to say I had these noble goals in mind when I made Due in 2010.

The truth is, I created Due for myself.

The first version tackled just one thing — quickly capture what needs to be done, and when it needs to be done.

When I realized it was easy to miss notifications, I added Auto Snooze. It repeatedly notifies me of overdue reminders until I act on them.

Then, I found myself needing to reschedule reminders frequently. So I made it easy to postpone a reminder — even without launching the app.

Someone who had used Due for the past 9 years said the person who created Due knows his stuff.

I guess that’s because there isn’t a day that goes by where I don’t use Due myself.

Learn more about what Due can do for you. Perhaps you’d find Due as useful as my customers1 and I do.


  1. Links to App Store reviews used in post. ↥︎

The awkwardly named “semaform” article format that is the main gimmick of Semafor is supposed to ensure information and analysis remain in separate buckets. So, how is that working out for them?

Here is an article from Friday by Diego Mendoza, which carries the following headline:

A fake Eli Lilly tweet drove its stock down and reignited anger over insulin prices

The opening paragraph:

Pharmaceutical giant Eli Lilly’s stock price plunged Friday following a fake tweet from a Twitter Blue-verified account that claimed the company would be making insulin free.

And here is the third paragraph under the subsequent “Know More” section, which was not in executive editor Gina Chua’s breakdown, but which someone may reasonably conclude based on the way it is written that it is less of an analysis section and more about presenting context and extra information:

While it’s unclear whether the fake tweet was the culprit behind the dive, it reignited public anger against pharmaceutical companies over the high cost of insulin. Eli Lilly’s competitors also saw drops in their share prices with Denmark’s Novo Nordisk’s dropping by about 5.1% and France’s Sanofi tumbling by about 4.5%.

So the headline says this tweet “drove [Eli Lilly’s] stock down”, the first paragraph says only that the stock price fell on Friday, and just a few paragraphs later, Mendoza acknowledges the tweet may not have been related in any way since other health-related stocks dropped in value too.

The absence of clear understanding of App Tracking Transparency’s material effect has produced a cottage industry of writers explaining how it is responsible for all kinds of trends in advertising. Some of those theories are more believable than others; here is an example of the latter kind.

Thomas Germain, Gizmodo, with an excellent single-paragraph summary of his thesis:

ATT kneecapped the Facebook ad targeting systems, motivating advertisers to look for new places to spend their money. Even though ATT hurt TikTok in the same ways, the short-form video app was in the perfect position to offer an alternative to Meta: its popularity was exploding, its newness meant ad prices were low, and it had designed novel advertising models built for the new privacy world order.

So as the most compelling demographic for advertisers flocked to TikTok, the cool new social app, advertising dollars followed suit? Well, I am shocked.

When I read this kind of story, I always wonder the counterfactual: would the overall thrust of the story still make sense if App Tracking Transparency were not in the picture? In this case, it is hard to see how spending would not be directed toward TikTok regardless of what was happening over at Meta’s side, where usage has stagnated especially among the audiences most desired by advertisers. Perhaps App Tracking Transparency added an extra sprinkle of salt in the wound, but all the qualifiers and acknowledgements in Germain’s article make it hard to believe it is of even tertiary relevance to TikTok’s fortunes.

Adi Robertson, reviewed Meta’s Quest Pro at the Verge:

But, again, the current use is limited. Meta has one confirmed app that uses foveated rendering, the game Red Matter 2, which renders at a high resolution in the Quest Pro — but is still held back by that fundamental graininess. Face tracking is primarily useful in Horizon, the Meta social platform that includes a Roblox-like recreational “metaverse” called Worlds and a personal office dubbed Workrooms. Worlds and Workrooms are available for the Quest 2 and Quest Pro alike, but Workrooms is particularly aimed at Pro users. And — there’s just no nice way to put it — it’s one of the worst apps I’ve ever used.

Some of the more enthusiastic supporters of Meta’s virtual reality efforts have pointed to meetings and Workrooms as the biggest attraction to working within a goggle-based world. But Robertson describes using the Quest Pro as a chore, writing that “it was just too painful for my head and eyes” to use for long stretches of time because of its weight and fuzzy displays.

On the one hand, even though Meta bought Oculus eight years ago, it is still early days in the world of virtual reality. On the other, Meta is selling this as a finished product, not a prototype, for $1,500 to end users. Also, while Workrooms has been available for a year, Robertson says it is so unreliable it is “like spinning a roulette wheel designed by Franz Kafka”. Meta may feel comfortable building this stuff in the open, but it means reviews which paint a bleak caveat-riddled picture.

Mike Masnick, Techdirt:

As for the background on all this, some of you youngsters might not remember this, but back in 2011 Twitter signed a consent decree with the FTC over its failure to safeguard user info. Now, almost every big tech company these days has a consent decree with the FTC after they royally screwed up something and effectively leaked users’ private data. Most of the consent decrees last for 20 years. That might make you think such consent decrees are meaningless, but the opposite is true. While under these consent decrees, the FTC now has tremendous power to cause a world of hurt to the company for screwing up.

[…]

All of this is kinda important right now, as Elon tries to roll out features in record speeds. Because… the consent decree has some requirements for rolling out new products and making sure they’re secure. The original consent decree says that any new product or service must be rolled out with a written plan […]

Alex Heath, the Verge:

The company’s chief privacy officer Damien Kieran, chief information security officer Lea Kissner, and chief compliance officer Marianne Fogarty have all resigned, according to two employees and an internal message seen by The Verge. Kissner confirmed their departure in a tweet on Thursday.

In a note posted to Twitter’s Slack and viewable to all staff that was obtained by The Verge, an attorney on the company’s privacy team wrote, “Elon has shown that his only priority with Twitter users is how to monetize them. I do not believe he cares about the human rights activists. the dissidents, our users in un-monetizable regions, and all the other users who have made Twitter the global town square you have all spent so long building, and we all love.”

Casey Newton and Zoë Schiffer, Platformer:

Just after lunchtime on the West Coast, Musk held an unannounced all-hands meeting with staff — his first as CEO. Musk had given employees just one hour’s notice; he arrived 15 minutes late.

Over the next hour Musk shared more bad news with the company. Depending on the length and severity of the recession, he said, the company could lose several billion dollars next year. He would not speculate how much runway Twitter had left. “Bankruptcy isn’t out of the question,” he said, as we were also first to report.

Around that time, someone put a poll into Twitter’s Slack channel: “Is it too early for vodka?” the poll asked. (It was a little after 1PM PT.)

Three more high-level Twitter staffers resigned after this call, in addition to the three which resigned this morning (Correction: they resigned the night before, though most staff found out in the morning). Because of the departures of legal and privacy leaders, Twitter is telling engineers to confirm compliance with FTC laws themselves. Do not do that.

In case it has not been obvious, I actually want Musk to succeed. I like Twitter. It is the only social network that has stuck for me; I have been using it daily for fifteen years. Days like today are worrisome for me, and I sympathize with the employees who are left and are understandably anxious about the future of the company and the demands now placed upon them.

There are aspects of this which should be humbling for Musk and others like him who are certain they can figure it all out. He made a purchase which he quickly regretted but was legally required to finish. In doing so, he burdened the company with a heavy debt load while winking at trolls. He says he wants to save Twitter, but it is entirely possible he is content with crashing it into the ground on his own terms and saving himself the embarrassment. There is a third way: without leadership and fractured engineering support, Twitter could simply fall apart.

I hope Musk succeeds and makes Twitter an exciting, vibrant, and safe place to spend time. But his two weeks in charge — and, in particular, the past couple of days — have done nothing to make me believe he can, nor have his weak answers to predictable questions. This experiment is not going well and investors in Musk’s other companies are taking notice. What a shambles.

Update: Robin Wheeler, one of the executives who resigned today, un-resigned, and Twitter has un-undone the “official” badges.

Alexander George, Wired:

Feeling his engineers have lost touch with the common folk, Herb employs Homer because he is, Powell says, “an average schmo.” Homer is given free rein to design his dream car, a futuristic creation filled with conveniences that make driving easy and pleasant. The episode comes with nods to automotive styling history: “Some things are so snazzy that they never go out of style, like tail fins and bubble domes. And shag carpeting!” Homer says.

The result of his efforts is The Homer, “powerful like a gorilla, yet soft and yielding like a Nerf ball.” It’s uglier than anything on the real-world market and, at $82,000 ($128,000 in 2014 dollars), it instantly puts Herb’s company out of business. (There are echoes here of real world mega-flops like Ford’s Edsel and the Tucker Torpedo.)

Not sure why I thought about this episode over the last few days, give or take, but I will remember soon enough.

Dougall Johnson dove deep into the nitty-gritty of Apple’s Rosetta 2 x86 emulator for its ARM-based Macs, and I cannot comprehend almost any of it. But the conclusion? That is something even I can understand:

Engineering is about making the right tradeoffs, and I’d say Rosetta 2 has done exactly that. While other emulators might require inter-instruction optimisations for performance, Rosetta 2 is able to trust a fast CPU, generate code that respects its caches and predictors, and solve the messiest problems in hardware.

I remember early skepticism, even from before Apple announced it was switching to its own silicon, about the ability to make Intel Mac apps work on ARM based on Microsoft’s inconsistent attempts to emulate 64-bit x86 on ARM. Rosetta 2 was a relief. I did not buy an ARM-based Mac until earlier this year and, so, most apps I use had long ago released fully native versions, but there are still a couple of emulated applications I use, and the process feels completely transparent to me.

Taylor Hatmaker, TechCrunch:

Now when tweets appear in Twitter’s timeline it’s impossible to visually distinguish the two categories of blue check accounts from one another. Doing so would require clicking through to examine a user’s follower count, which isn’t necessarily a reliable way to tell, or searching for whatever clues might be found in their other tweets. Clicking on the check mark itself from a profile page apparently displays different copy too, but like everything on Twitter right now, that is subject to change.

Casey Newton:

Also seems notable that the only way users would know that something changed about verification today is if you … read news coverage? Or Elon’s tweets?

This is not surprising you, right? Of course people would take advantage of a paid verification status to sow comedy and chaos. Kudos must go to Twitter’s probably overworked moderation team for suspending accounts or removing their verification checkmark. On a typical day, once all the chaos of the past week settles down, it might not be this bad.

But what happens when there is a breaking news event? Imagine the already fraught way Twitter users disseminate news during a crisis, and add a bunch of “verified” accounts to the mix. Or what about enabling a phishing attack?

I still think clearly labelling accounts as “official” makes a lot of sense — which, incidentally, is a plan Twitter has put on hold — but the current state of blue badges is unnecessarily fraught. Anyone should be able to look at a tweet and know if the information in it is coming from an official source, but even the account of Twitter’s new owner briefly displayed that it was verified because Musk pays for Twitter Blue, not because he’s a public figure.

It is all going about as well as can be expected. Sure is a good thing the current U.S. president is not as big a fan of making major policy announcements on Twitter.

Wang Boyuan of PingWest on Twitter:

In the recently released iOS 16.2 beta2, AirDrop to everyone feature is found to last for 10 minutes instead of forever. However, this change only appears in iPhone’s China models. […]

Well, it turns out to be also a “hot fix” in today’s iOS 16.1.1 update (should be widely discussed in the next several hours). But Apple didn’t specify this change in update description, only saying it contains “bug fixes and security updates and is recommended for all users.”

Wang appears to have broken this news, and I appreciate Lawrence Li’s heads-up. Wang — and others — speculate this change is driven by the use of AirDrop by protesters in Beijing and, in his tweet, points to a New York Times story from Li Yuan:

Encouraged by the Beijing protester’s extremely rare display of courage, young Chinese are using creative ways to spread the banners’ anti-Xi messages. They graffitied the slogans in public toilets in China. They used Apple’s AirDrop feature to send photos of the messages to fellow passengers’ iPhones in subway cars. They posted the slogans on university campuses all over the world. They organized chat groups to bond and shouted “Remove Xi Jinping” in front of Chinese embassies. This all happened while the Communist Party was convening an all-important congress in Beijing and putting forth an image of a country singularly united behind a great leader.

It once again appears that Apple is caving to political pressure in China. There is diplomacy, and there is bending to the demands of the powerful and capricious; this appears to be a case of the latter.

A weird quirk of this change is that, absent the above context, adding a timeout to the “Everyone” setting for AirDrop is actually a good idea. Some people have reported receiving unwanted AirDrops in public, a story which CNBC illustrated with a stock photo of a “senior man surprised at tablet”. Indeed, Apple told Mark Gurman of Bloomberg that it will be rolling out the feature for all iPhones — but it would not say why this change was added to a routine security update only for users in China.

Esther Crawford, project manager at Twitter:

A lot of folks have asked about how you’ll be able to distinguish between @TwitterBlue subscribers with blue checkmarks and accounts that are verified as official, which is why we’re introducing the “Official” label to select accounts when we launch.

Not all previously verified accounts will get the “Official” label and the label is not available for purchase. Accounts that will receive it include government accounts, commercial companies, business partners, major media outlets, publishers and some public figures.

The verified badge was supposed to indicate accounts which were legitimate and high-profile, and could potentially be a source of confusion or impersonation. But it morphed into something more like a badge of approval by Twitter. It got really messy and I wonder if most people know what the badge actually means at this point.

This new version seems like an improvement. It clearly labels accounts as “official”, and sort of makes the familiar blue badge an indicator of whether a user pays for Twitter. These are not terrible ideas. But Twitter will not be applying this policy universally or consistently, as already-verified accounts will not currently be required to pay up or lose their badge, nor will they necessarily receive the new “official” label. If I am reading this right, it is also possible that “official” accounts may not necessarily have the blue badge — just the little “official” label. It is all a bit of a shambles right now.

That is kind of the story of the whole rollout, too — a bit of a shambles. I get that Twitter’s new owner wanted to shake things up at a clearly mismanaged company, but there was no reason to make people sleep on the floor for a premature rollout of a new subscription service. Overthinking a project can lead to decision paralysis, but this appears to have been executed without much thought at all. It is going to be a bumpy ride.

Update: Things are going about as well as can be expected.

Patrick McGee and Ryan McMorrow, Financial Times:

Yet when it comes to selling its devices to Chinese consumers, business has boomed. Operating profits in greater China — which includes Hong Kong, Macau, Taiwan and mainland China — have shot up 104 per cent over 24 months to $31.2bn in the financial year to September, eclipsing the $15.2bn earned by Tencent and the $13.5bn from Alibaba in their most recent 12-month period, according to S&P Global Market Intelligence.

The record profits underscore the bargain Apple has struck with Beijing, allowing the iPhone maker to sail through President Xi Jinping’s crackdown on homegrown tech groups while reaping the rewards from US sanctions, which are helping to damage its only real competitor in the country — national champion Huawei.

This article builds modestly upon Wayne Ma’s reporting last year for the Information about Tim Cook’s efforts to ingratiate himself and Apple among officials in China. But the deal struck by Cook was due to expire last year, according to Ma, and the company has begun to diversify its manufacturing dependency. I bet Apple wishes that diversification could happen much faster, as it just announced supply would be reduced of iPhone 14 Pro and Pro Max models because of COVID-19 restrictions at its main factory for those phones.

As I wrote when I linked to Ma’s story last year, it is notable that this diplomatic trade-off has required Apple to comply with an increasingly authoritarian regime. In some ways, these compromises have permitted Apple’s customers a greater level of security and privacy than is offered in products from other companies. At the same time, it has also made Apple complicit in implementing those laws. But if it did not compromise, it is likely Apple would be required to be wholly compliant with just about any law since it is so dependent on the country for manufacturing.

Last month, Apple’s Craig Federighi and Greg Joswiak were interviewed by Joanna Stern at the Wall Street Journal’s Tech Live event. Stern asked some pretty good questions; Federighi and Joz are too media trained to answer as comprehensively as any of us would like. But there was one exchange I was surprised to see:

Stern: Speaking of — we’re talking a little bit about ecosystems, Craig, and I wanted to read you back an email of yours from 2013. You might have read it before, or written it before. It was to Apple’s Eddy Cue — he’s a colleague of yours — and you said “I’m concerned that iMessage on Android would simply serve to remove an obstacle to iPhone families giving their kids Android phones.” Whatever happened to iMessage on Android?

Federighi: I’m not aware of it shipping.

Stern: Yeah. You sent this email, though.

Federighi: I did.

Stern: You sent this email. You felt like, maybe we shouldn’t release this product because other people will buy Android phones.

Federighi: My feeling — and, I think, if one read the whole email, [it was] clear the back-and-forth with Eddy — was if we’re gonna enter a market and go down the road of building an application, we have to be in it in a way that’s gonna make a difference. That we would have a lot of customers, that we would be able to deliver great experiences; this comes at a real cost. And my fear was we weren’t in a position to do that. And, so, if we just shipped an app that really didn’t get critical mass on other platforms, what it would have accomplished is it would have held us back in innovating in all the ways we wanted to innovate in Messages for our customers, and wouldn’t really have accomplished much at all in any other way. And, so, we just felt, you know, pick where you can make a difference. Pick where you’re gonna invest and do it where you’d make a difference. And this seemed like a throwaway that wasn’t going to serve the world honestly.

Today, that email exchange was republished in full in the Internal Tech Emails newsletter, and I was surprised to see Federighi’s memory of that nine year old email was pretty close, albeit less comprehensive. Here is the email before the part Stern quoted:

Do you have any thoughts on how we would make switching to iMessage (from WhatsApp) compelling to masses of Android users who don’t have a bunch of iOS friends? iMessage is a nice app/service, but to get users to switch social networks we’d need more than a marginally better app. (This is why Google is willing to pay $1B — for the network, not for the app).

Cue’s reply:

Keep making a better app. Our app is secure and private today. [Theirs] is not. We also need to be adding features to iMessage to keep it a leader – better group messaging, status, location, payments, etc.

In the nine years since this exchange took place, it is interesting to reflect on Google’s actual performance in messaging and wonder if, maybe, a universally-accessible service from Apple could have found a more welcoming market than Federighi seems to believe. A real glass of ice water to somebody in Hell kind of moment. I appreciate the restraint and the questioning of whether a cross-platform product is worth Apple’s efforts, but I get the feeling many Android users would welcome it.

Charles Edge (via Michael Tsai):

The net result is that when doing the last few upgrades, they have required 12+GB for the installer itself (which can be run from a USB drive) and up to 44GB for the installer to do the work it needs to do, so a total of up to about 56GB. Therefore, scoping policies to run an updater without causing undo issues to end users it’s entirely appropriate to make sure they have the amounts of free space indicated per version. Given that drives can be a terabyte in size, this doesn’t seem wildly inappropriate; however, many organizations still buy devices with 256GB drives (thus going from an eighth in the 64GB drive era to a quarter of common drive space required to be free for certain upgrades on smaller drives today). […]

If users upgrade to each version in sequence rather than skipping over versions, they will have a lower disk space requirement.

Even so, these space requirements make my eyes pop. In just a few years and with broadly incremental updates for users between versions, the MacOS installer is about twice as large and needs about twice as much space to do its job. MacOS is not alone: the iPhone X shipped with iOS 11 — released in 2017, the same year as the oldest MacOS version in Edge’s comparison — which had a 3 GB installer, about half the size of the current iOS 16 installer for the same iPhone model.

This is not limited to operating systems, either. Users are disrespected by increasing and surprising bloat in applications. For work, I need to run the Microsoft OneDrive client on one of my Macs, and I was surprised to see that it recently crossed the 1 GB threshold. This is a file syncing utility. For comparison — and you can check this for yourself — it was 70 MB just four years ago. I am not being dramatic when I call this sort of behaviour disrespectful; it shows contempt for users to eat up a full gigabyte of disk space for a background application that uploads and downloads files. OneDrive is obviously not the only offender — all Electron apps are needlessly bloated by bundling a discrete copy of Chromium, of course, and Adobe’s creative applications are several gigabytes each. But OneDrive is a grotesque example of needless bytes.

It is sort of miraculous that modern video codecs like H.265 have made it possible to fit videos of increasing quality and resolution into shrinking amounts of space. But when it comes to software, the pattern is exactly backwards.