Information Density: Daring Fireball – Signal Evidence & AI Readability

[IMG: Daring Fireball]
By John Gruber
Archive
The Talk Show
Dithering
Projects
Contact
Colophon
Feeds / Social
Sponsorship
[IMG: exe.dev]
Build anything with exe.dev. It’s just a computer.
Awarding Jay Haynes His Being Right Points for Predicting Apple Hitting $3 Trillion in Market Cap ★
Here’s a fun one. Back in 2014 I linked to a post by Jay Haynes in which he projected that with a very reasonable level of annual growth, Apple ought to reach a $3 trillion market cap within 10 years. At the time of his writing, Apple’s market cap was “just” $450 billion, and no company had hit the $1 trillion market. So projecting a $3 trillion valuation in 10 years was a bold prediction.
Apple hit $3 trillion in just 8 years.
Haynes’s original blog went belly-up, alas, but he republished the piece on Medium, with a bit of additional commentary up front, in 2016. Re-reading Haynes’s piece today, it holds up extremely well, including his case that the iPhone and iPad are almost textbook examples of Clayton Christensen’s disruption theory (yet Christensen himself got it wrong).
(Thanks to Nathan Peretic, longtime DF reader and owner of a perfect personal homepage, for prompting me to revisit this and award Haynes his well-earned Being Right Points.)
Thieves Are Texting Threats to Victims of iPhone Theft in London ★
Lizzie Dearden and Amelia Nierenberg, reporting for The New York Times (gift link):
The crime Alex Pikula reported to the police was one they had
heard before: An e-bike rider had zoomed past as Mr. Pikula left a
theater in London’s West End, ripping his phone from his hands. It
was frustrating, Mr. Pikula thought, but that was that.
He was wrong.
His mother soon started receiving strange texts, claiming to have
her son’s emails and bank information. Then she received a video
of a man brandishing a gun. Then came threats of sexual assault
and death.
“I know who you are and where you live,” read one, full of
obscenities and typos. “I’ve killed or [sic] far less than a
phone before,” it went on. “We will see if you value your life
over this phone.”
All of the messages wanted her to do one thing: unlink her son’s
Apple ID from his stolen phone.
The story only mentions the word iPhone twice, but phone appears over 30 times. “Apple ID” appears four times. There’s zero mention of Android or Google. It’s just implicitly assumed that the only phones worth stealing or threatening victims about are iPhones. The story makes no mention of Apple’s Stolen Device Protection, which Apple recently began turning on by default when users install iOS 26.4.
Dearden and Nierenberg filed a previous report in October about organized iPhone crime rings in London. And in November I linked to a story where a thief, after stealing an Android phone, turned around and handed it back, explaining to the victim, “Don’t want no Samsung.”
Trump Mobile Website Exposed the Number of Pre-Orders — Both Completed and Abandoned — and the Associated Customer Information ★
Catie McLeod, The Guardian:
Trump Mobile said in a statement that it was investigating the
issue — “with the assistance of independent cybersecurity
professionals” — in which the full names, addresses and phone
numbers of people who filled out preorder forms appeared to be
exposed. [...]
Jonathan Soma, a programmer and professor at New York’s Columbia
University, reviewed the code that the Australian had uncovered
and copied from the Trump Mobile website. Soma said the website
used a common e-commerce model, in which every potential order
added another “1” to a list, the total of which had reached 27,224
possible pre-orders on the available information.
But he said the code reflected the last step before payment,
meaning those who didn’t proceed with the purchase were also
recorded in the data, even those people who have abandoned their
carts without paying the deposit, so the true number of preorders
was likely to be even lower.
“I probably started three phone purchases and didn’t buy any of
them,” he said.
Auric Goldfinger is surely rolling over in his grave.
The History of ‘OK’ ★
Merriam-Webster:
The 1820s and 1830s shared another linguistic fad with today: an
appreciation for deliberate misspellings. (Kewl, rite?) This
trend, which had humorists adopting now-cringey bumpkin personas
with ignorance manifested in uneducated spellings, turned no go
into know go and no use into know yuse (lol). Abbreviations
were not immune, and no go became K.G.. So too all right
became O.W., as an abbreviation for oll wright. And all
correct became o.k., as an abbreviation for oll korrect.
Although OK became one of the more commonly used initialisms, it
might have passed into oblivion when the linguistic fad had passed
if not for the presidential election of 1840, when Martin Van
Buren was given the nickname of “Old Kinderhook” because of his
hometown of Kinderhook, NY. The Van Buren stans who joined “OK
Clubs” nationwide were themselves, they proclaimed, “OK.” Their
campaign was memorable enough to have both popularized the word
and to have hijacked the story of its origin: there are today
still those who believe that “Old Kinderhook” is the original
meaning of OK.
I have a strong preference for OK (perhaps infused by the classic Macintosh Human Interface Guidelines’s adamance on the spelling). Okay is OK in prose, but never as a UI button label. Ok and ok are not OK.
WorkOS: ‘Agents Need Context. Ship the Integrations That Give It to Them.’ ★
My thanks to WorkOS for once again sponsoring DF last week. The context that actually matters isn’t in your database. It’s in the tools your users live in every day. Multi-stage agents stall the moment they hit a step they can’t see. And every missing integration is a different OAuth flow, a different token lifecycle, weeks of plumbing before the agent reads a single record.
WorkOS Pipes connects your agent to the tools your users live in. Pre-built connectors for GitHub, Slack, Salesforce, Google Drive, and more. Pipes handles OAuth, token refresh, and credential storage. You call the real provider API with a fresh token, every time. Your agent pulls context at every step, for as long as the task runs.
Give your agent context.
[H2] Sunday, 24 May 2026
Why Steve Kerr Stayed With the Warriors ★
Terrific, poignant profile of Warriors head coach Steve Kerr by Wright Thompson for ESPN:
Kerr doesn’t want the Warriors to end up like the New England Patriots, marred by grudges and grievances. He watched Michael Jordan retire, then unretire, then retire, then unretire. His friends used to grill him about MJ.
“Why doesn’t he go out on top?”
“Because he can’t,” Kerr told them.
For the past few years, Kerr has watched his mentor, San Antonio Spurs coach Gregg Popovich, struggle through this same decision. Pop once called Steve to tell him he’d finally decided to retire. Steve congratulated him on a Hall of Fame career. A week later Pop signed an extension with San Antonio. Popovich finally officially quit six weeks before our lunch, six months after a stroke diminished him physically. People who loved him had to show him the door, as gently as possible. That hurt Steve. He respects Popovich so much. He loved playing for him and coaching with him. He once told Gregg he was the finest man he’d ever known and thanked him for all he’d done for him. Pop smiled and said his feet were made of clay like everyone else’s. Steve didn’t believe it then. Now he does.
“I realized he couldn’t do it,” Kerr said. “He couldn’t walk away.”
I asked how he’d avoided the trap. He laughed.
“I’m sitting here wondering,” he said.
It sounds so easy to go out on top. But it very seldom happens.
[H1] The Fonts of the U.S. Federal Courts
[H6] Friday, 22 May 2026
The 13 circuits of the U.S. federal courts of appeals operate with a fair amount of independence, including their typographic choices. I was reminded of this today while reading the aforelinked decision from the Ninth Circuit in Epic v. Apple, because the Ninth Circuit sets their decisions in Times New Roman — a font that came up back in December in the context of the Trump State Department.
Long argument short, Times New Roman isn’t bad, but it isn’t good. It is the median choice. But most of the circuit courts use it: the Third, Fourth, Sixth, Eighth, Ninth, Tenth, and Eleventh. It could be worse: the First circuit not only uses Courier New (the worst version of Courier, so of course it’s the one Microsoft shipped with Windows), but fully justifies their text — contrary to the nature of a monospaced font. (The Fourth circuit only recently switched from Courier New to Times New Roman — an upgrade, to be sure, but a disappointingly mediocre one.) It could be better: the Second and Seventh use Palatino. (Note how much better that Seventh Circuit decision looks than the Second’s, with its wider margins creating a narrower column of text.)
But it can be much better. The Fifth Circuit was long typographically superior to its peers, using Century Schoolbook — a highly legible font with great tradition and the right vibe. But in 2020, the Fifth Circuit upgraded, switching to Equity, Matthew Butterick’s excellent type family (which, of course, is used throughout Butterick’s own web book, Typography for Lawyers). Here’s a before and after tweet noting the change. The results are typographically sublime (including improved margins).
The gold standard is the U.S. Supreme Court, which uses Century Schoolbook. Yes, I just praised the Fifth Circuit’s change from Century Schoolbook to Equity as an upgrade, but tradition and consistency have their place. The Supreme Court’s typographic style has been stunningly consistent for — no pun intended — well over a century. (If only that were true of their recent decisions. Rimshot.) Here is last month’s Louisiana v. Callais decision — the gerrymandering / redistricting case. Here is 1954’s Brown v. Board of Education. I’d give the nod to the older one, which made better use of proper small caps, but the overall consistency is obvious.
Here is the 2026 edition of the Rules of the Supreme Court. Not only does the Court use Century Schoolbook for its own decisions, it requires submissions to the Court to use the same (p. 44):
The text of every booklet-format document, including any appendix
thereto, shall be typeset in a Century family (e. g., Century
Expanded, New Century Schoolbook, or Century Schoolbook) 12-point
type with 2-point or more leading between lines. Quotations in
excess of 50 words shall be indented. The typeface of footnotes
shall be 10-point type with 2-point or more leading between lines.
The text of the document must appear on both sides of the page.
Every booklet-format document shall be produced on paper that is
opaque, unglazed, and not less than 60 pounds in weight, and
shall have margins of at least three-fourths of an inch on all
sides. The text field, including footnotes, may not exceed 4⅛
by 7⅛ inches.
Why the extra one-eighths of an inch instead of just 4 × 7? I don’t know. But 4⅛ × 7⅛ is exactly the size of the text field in the court’s own decisions.
Now compare the current 2026 rulebook to this edition printed in 1910 (with rules adopted in 1884). The consistency is striking — but, once again, the older version makes better use of small caps and just has a bit more vim and vigor to it. Just look at page 44, for example. It’s perfect. The current Court’s document formatters should aspire only to more closely ape the confidence and sturdiness of this older one. A century from now, U.S. Supreme Court decisions should look as similar to today’s as today’s do to those from a century ago.
The various circuit courts using lesser typefaces, looser margins, and lazier formatting should follow the Fifth’s lead and get their shit together. Tuck your shirt in, comb your hair, straighten your tie, and pop a mint in your mouth. If you’re a United States federal court, your typographic style should reflect that.
Back in 2020, Butterick took a well-deserved victory lap when the Fifth Circuit adopted Equity.1 He quoted Fifth Circuit Judge Don Willett, a typography fan who spearheaded the restyling project, on its rationale. Willett wrote:
[Why] did the circuit devote finite judicial energy to swapping
typefaces and widening margins? Simple answer: Our job is not
just to present clear opinions, but to present our opinions
clearly. Getting the law right is, of course, our tip-top
priority. Nothing matters more. ... But good enough is never good
enough. Our work is consequential, impacting the lives and
livelihoods of real people walloped by real problems in the real
world. The stakes are high, and we must present our best opinion,
not merely a passable one. And that presentation begins before
the first word is ever read. ★
In the very same post, Butterick sings the praises of the Apple Extended Keyboard II, and notes that he has several spares in reserve. I do keenly intend to take Butterick up on his standing offer to dine when next I’m in Los Angeles, but I worry that if we meet, we’ll trigger some sort of calamitous singularity of aligned taste. ↩︎
[H2] Friday, 22 May 2026
The Ninth Circuit Appeal Ruling in ‘Epic v. Apple’ That Apple Is Seeking to Overturn at the Supreme Court (PDF) ★
Following up on yesterday’s item re: Apple’s petition to the Supreme Court, here’s the Ninth Circuit ruling. It starts with a “Summary” that is specifically intended for the convenience of the reader. Page 50 is where it covers Apple’s argument regarding Trump v. CASA as precedent that an injunction on commissions should apply only to Epic Games, not to all developers in the U.S. App Store.
Zero Sum Problems and Apple Sports ★
Kieran Healy kindly accepted my implicit homework assignment yesterday, and wrote a piece on Apple Sports’s bizarre “zero sum” team stats visualization:
It also doesn’t do away with the core problem. That problem is
principally one of information design rather than data
visualization. What I mean is that what we’re trying to organize
is, in effect, fifteen pairs of related but fundamentally distinct
numbers. If we had fifteen cases and two variables things
would be simple. But with fifteen variables and two cases … well,
this is not the kind of thing you can make a single effective and
non-confusing graph out of. That’s why I kind of sympathize with
the designer. In a constrained space they have to show thirty
numbers (thirty two, including the score). Lots of information. A
straight table seems like it would be boring. Surely there’s some
way to thematically integrate the numbers in a visually appealing
manner that brings out some of the relationships across the rows.
That’s what graphs do; it seems like the right thing to reach for.
But at its heart this information is not a graph. It just sort of
looks like one, and that ends up confusing people.
Just a crackerjack explanation for why this presentation in Apple Sports is confusing, and for why it is a difficult problem to solve. The problem is

[IMG: Daring Fireball]
By John Gruber
Archive
The Talk Show
Dithering
Projects
Contact
Colophon
Feeds / Social
Sponsorship
[IMG: exe.dev]
Build anything with exe.dev. It’s just a computer.
[H1] AI Is Technology, Not a Product
[H6] Saturday, 16 May 2026
Steven Levy, writing for Wired last month after Apple’s CEO transition was announced, under the provocative headline “Apple’s Next CEO Needs to Launch a Killer AI Product” (News+ link to get around Wired’s miserly paywall):
Much more recently, I quizzed Ternus and global marketing
head Greg Joswiak about Apple’s future, specifically its plans to
get ahead of the AI transformation. Ternus acknowledged that AI is
“an immense kind of inflection point,” but couched it as one of
many leaps that Apple has navigated. Each hit product — the Apple
II, the Mac, iTunes, the iPod, the iPhone, iPad — piggybacked on
a previous product. “We never think about shipping a technology,”
he said. “We want to ship amazing products, features, and
experiences, and we don’t want our customers to think about what
[underlying] technology makes it possible. That’s the way we think
about AI.”
That’s fine, but I look back to the mid-2000s when everybody was
waiting for Apple to come out with a phone. When Jobs finally
delivered in January 2007, the product defined the mobile era.
It’s a big ask for Ternus to do something similar for the AI age — but it’s an opportunity that must be seized. AI threatens to
disrupt the entire iPhone ecosystem. By the end of this decade,
it’s unlikely that people will swipe on their phones to tap on
Uber or Lyft. They will just tell their always-on AI agent to get
them home. Or that agent will have already figured out where they
need to go, and the car will be waiting without the friction of a
request. “There’s an app for that,” may be replaced by “Let the
agent do that.”
I’m a huge longtime Steven Levy fan, but this is nonsense. It’s hard to read this and not worry that he too has lost his mind to the AI snake-oil hypesters. What Ternus told him is exactly right. The Apple way is never to ship a technology. The iPod wasn’t about MP3 files. It wasn’t about 1.8-inch hard drives. It was about music. The iPhone did define the mobile era (which we’re still very much in), but Apple doesn’t need to capitalize on every single market the mobile era opened up. Social media is a defining component of the mobile era. It comprises the entirety of Meta’s value and a sizable slice of Google’s (via YouTube). Apple doesn’t have a social network business. It’s fine — because the way people consume and create social media is using their phones.
Does AI “threaten to disrupt the entire iPhone ecosystem”? It’s possible, but it doesn’t seem nearly as likely to me as Levy asserts. Changing the iPhone ecosystem? Sure — that’s already true. Obviating the iPhone ecosystem? I don’t see it. Levy’s argument reminds me of the hype around “the cloud” when that first became a term. It’s so meaningless when used broadly (e.g. “Everything will soon be in the cloud”) that it could mean anything. It’s step #2 in the gnomes-stealing-underpants master plan.
The idea that AI agents “will have already figured out where [we] need to go, and the car will be waiting without the friction of a request” strikes me as pure fever dream high-on-the-hype fantasy. I’m just going to step outside a restaurant when I’m done eating a meal and a ride-share is going to be there, waiting for me, without my having hailed it? Every time? And I’m going to find this pleasing, not creepy? And ride-share drivers are going to respond to all these requests, because the requests will never be wrong? And this is going to happen, somehow, without my carrying a phone with me? And this is going to happen in the next four years? I don’t think I’d want this even if it were plausible, but it doesn’t sound plausible.
Actual products have to be real. Actual experiences have to rely on actual products. How exactly in Levy’s end-of-this-decade scenario will we tell our “always-on AI agent” to get us home? What microphone is listening to the command? What speaker is telling us the request was understood and acted upon? What screen do we look at to see how far away the hailed car is? I’d bet a pretty large sum of money that in 2030, when someone hails a ride-share vehicle to take them home, the most common product they’ll use to do that will be their phone. Whether they’re doing it via a verbal command issued to an “always-on AI agent” or good old tapping and swiping, it’ll be a phone.
If you think that people will buy smaller devices to replace their phones, and use those to talk to “always-on AI agents” instead, you have to answer some questions. What company is the best in the world at making smaller-than-phone personal computing devices? What device will people use as their camera? What device will people use as their screen, for watching videos, playing games, texting, and (one hopes) reading? My answers to those three questions: Apple, phone, phone. Why would smaller devices — you know, like watches, earbuds, and, say, glasses — work independently rather than pair with the phone that you’re almost certainly still going to be carrying with you?
Only a fool would argue that Apple can stand on the sidelines and ignore AI. It’s very different from, say, social media that way. Social media doesn’t pervade everything in technology. You can ignore social media as a user. (And you’re probably more productive, and happier, if you do.) A company can eschew social media as a business. AI, on the other hand, is pervasive. It can’t be ignored. But it’s just technology.
Wireless networking is pervasive too. But Apple doesn’t have “a killer wireless networking product”.1 Wireless networking simply pervades everything Apple makes. I’m hard pressed to think of a single product Apple makes that doesn’t use some combination of Wi-Fi, cellular, Bluetooth, and proprietary wireless protocols. There was a time, not too long ago, when Apple didn’t make a single product with wireless connectivity. Now it’s pervasive in all their devices. That’s more what AI is going to be like. There’s not going to be one “killer AI device”. Everything is going to be an AI device, to some extent, just like how everything today is a wireless connectivity device, to some extent.
Postscript: “Existing Stakeholders Have a Say in the Future”.
AirPort qualified, arguably. But Apple walked away from it, alas. ↩︎
Previous:Nextpad++
Next:The Fonts of the U.S. Federal Courts

[IMG: Daring Fireball]
By John Gruber
Archive
The Talk Show
Dithering
Projects
Contact
Colophon
Feeds / Social
Sponsorship
[IMG: exe.dev]
Build anything with exe.dev. It’s just a computer.
Apple Sports Expands to More Than 90 New Countries on Cusp of World Cup
Apple Newsroom:
Apple Sports — the free app for iPhone that gives fans access to
real-time scores, stats, and more — is now available to download
on the App Store in more than 170 countries and regions around the
world, including more than 90 newly added markets. Designed for
speed and simplicity, the app delivers a personalized experience,
putting fans’ favorite teams and leagues front and center with a
simple, intuitive interface designed by Apple.
Apple Sports is helping fans get ready for the World Cup by
allowing them to explore tournament groupings and customize their
scoreboards simply by following the entire tournament or their
favorite national teams — making it easier to stay on top of key
moments when the tournament kicks off in June. Following a team
also enables Live Activities on a user’s iPhone Lock Screen or
Apple Watch, letting them follow every moment of a match with just
a quick glance.
I’ve got some gripes about certain specific aspects of Apple Sports. Like, where does one even start to explain how much is wrong with their zero-sum visualization of team stats? Has anyone ever even seen a presentation like that before? Anyone?
But overall it really is a good app. I don’t love the UI layout but I don’t hate it, either, and it is interesting. It’s a very modern layout. Apple Sports is fast to load — the primary reason Eddy Cue wanted the app in the first place — and its Live Activities are very good. It remains my go-to for “checking scores” for every sport except baseball, for which I have a much better dedicated app.
Yes, Apple promotes some of its own sports-related properties in the app occasionally. Just now I had a promotion for the F1 Canadian Grand Prix at the top. But the ads that do appear are always sports-related and never obscure content. That’s a fair deal.
I was glad when Apple Sports debuted two years ago and it’s lived on my first or second home screen ever since, depending on which sports are in season. I’m really glad Apple has stuck with it, shipping steady improvements on a regular basis. Expanding now to nearly the entire world is a big step. If you’re new to it, it might take some getting used to, but give it a shot. It stuck with me.
Still kind of curious that Apple Sports remains iPhone-only — not even an iPad version — but in a way I find that charming too. Maybe Apple is tight on money?
★ Thursday, 21 May 2026

[IMG: Daring Fireball]
By John Gruber
Archive
The Talk Show
Dithering
Projects
Contact
Colophon
Feeds / Social
Sponsorship
[IMG: exe.dev]
Build anything with exe.dev. It’s just a computer.
Jay Haynes: ‘Apple’s $3 Trillion Valuation’
Jay Haynes:
If we assume that Apple will grow its owner earnings at 5% for the
next 10 years, and then 2% for all years after that (with
adjustments for cash and debt), Apple’s market cap wouldn’t be
$453 billion. It wouldn’t even be $1.2 trillion. It would be $3
trillion. This is a share price of $3,275 in contrast to today’s
share price of $506. At just 5% annual growth for Apple.
That doesn’t mean the market is wrong. It just means the market thinks Apple’s growth is over, that it won’t even manage single-digit growth for the next decade.
Update, May 2026: Haynes’s original blog went belly-up (perhaps when Typepad did?), alas, but he republished the piece on Medium, with a bit of additional commentary up front, in 2016. This came up in the context of my revisiting Haynes’s prediction that Apple would reach a $3 trillion valuation.
★ Wednesday, 29 January 2014