My personal journal and blog. Subscribe via RSS
The brilliant folks at Sandwich have debuted their new app for immersive video on the Vision Pro: Theater (App Store link). It’s so cool.
What was cooler this week was Sandwich live streaming The Talk Show from WWDC in spatial video. I tuned in for the first half of the show, and it was a stunning experience. I can’t wait to see more of this type of thing in the future.
This is a really nice addition to Apple Pay coming in iOS 18: users will be able to pay with Apple Pay on Windows, and in Chrome browsers. Very nice implementation. Excited to implement this for Air Mail later this year.
A thorough and beautiful post from Louie Mantia about the new “dark mode” icons available in iOS 18.
Some (seemingly) good news after the surprise announcement last week that Alamo was closing its doors here in Dallas. Alamo is such a great place, and I really hope it survives.
Casey Newton, writing at Platformer:
It was a moment that has seemed inevitable since November 2022, when ChatGPT launched and catalyzed global interest in how AI can enhance products. In the 18 months since, impatient investors have worried that Apple might be letting the moment pass it by. Savvier observers have noted that this is how Apple has worked for decades now: approaching new technologies deliberately, and on its own time; developing its distinctive take on the product; and releasing it only when polished to the company’s quality standards.
Judging from the preview, Apple Intelligence was created in just this way. The company took time to develop principles around what AI should do on its devices. It landed on a suite of AI features for the operating system, designed to make its devices more valuable by leveraging the massive amount of personalized data on your devices. (Sensitive to the implications of such an invasive technology, Apple also took pains to develop a more private approach to data processing for AI apps as well.)
The privacy-based approach to AI is such a breath of fresh air in the industry.
The headlining feature of Monday’s WWDC Keynote was clearly the introduction of “Apple Intelligence”, Apple’s newly minted name for all-things-AI at the company.
The first mention of “Artificial Intelligence” by Apple was over an hour into the keynote, when Tim Cook called the company’s initiatives “personal intelligence” and “the next big step for Apple.”
I love the framing of a “personal intelligence system”. It strikes a bold contrast between other AI tools that “use world knowledge” versus something that understands you.
Apple didn’t just roll out a bunch of models and developer jargon– they showed real features that use AI to enhance what you can do by yourself with these devices.
A few of the highlights for me were:
As mentioned in the keynote, I really dig how Apple Intelligence is being framed as an AI assistant that “context and meets you where you are.”
WWDC begins today and the annual keynote featuring new software updates for Apple platforms will inevitably bring a ton of new AI features. I’m sure those will be interesting, but my desires for the platforms are focused elsewhere.
Here are a few other items I’d love to see this year. First a few developer-oriented updates:
Credit to Cihat Gündüz for this idea, but it’s a brilliant one. With Apple’s introduction of a very nice sports app earlier this year it seems that this one is within the realm of possibility. Maybe not this year, but eventually would be amazing. I’d still love to build a purpose-built sports scores app with support for a few niche features for folks like me that constantly are watching games in delay after my kids go to bed. ⚾️
Developers should not need to ship their own AI models and optimizations to use them on Apple platforms. Let’s get some first-party support for various AI models for different purposes that can be used with a simple abstracted API.
And some end-user features:
Mail.app is an amazing app, but its features still seem very dated. How about Snoozing, Sending Later, a Screener, and “better” handling of newsletters.
AirDrop is great, when it works. Which is a rare thing for me. And even when it does work it’s incredibly slow, even for devices that are within inches of each other. I’d love to see AirDrop capable of sending a link or photo between devices within seconds, not minutes.
As noted last week, let’s focus on a Screen Time feature that actually does what it says. Or at a minimum, doesn’t allow kids to easily bypass.
“Messages” is a great app, but compared to its modern competitors (WhatsApp, et al) it’s too simplistic. More filters, better support for SMS and green-bubble group chats would be welcome additions. How about a way to filter all of the 2FA and confirmation codes out of the conversations flow? Group chats that can be summarized by AI. Delayed sending of messages, so I don’t bother people in other time zones late at night. Drafts syncing. Typing indicators for group chats. Expiring “secure” thread. Lots of opportunity here.
Siri needs a complete reboot, if not a complete rebrand too. It’s so bad, I almost never use it. And when I do and hope for the best, it rarely succeeds. Siri needs a new (likely AI/LLM-based) engine. It’s time.
I’m not sure where to go with the Vision Pro, but I’m hoping for a very nice 2.0 on visionOS.
→ See also, Michael Tsai’s excellent roundup of WWDC wish lists. I agree with all of Ryan Jones’ ideas for control center:
Hopes for a better Control Center:
- Big clear single tap audio output
- Pick home controls
- Any shortcut
- Hide less in long presses
- Rotation lock except video
- All buttons are customizable
- No double button in Focus Modes
- Mini TV Remote at first level
Looking forward to watching the keynote later today. I’ve avoided the spoilers and I’m just looking forward to enjoying the presentation.
Speaking of cool new devices, the Daylight DC-1 looks very cool as well.
Nice new app from Leica (App Store Link) with lens simulation styles. A few that are free, and a few more for an upgrade.
Transform your Apple Watch into a playful, functional handheld device that provides connectivity without distraction, encouraging you to leave your phone at home.
I don’t quite get it, but I love it. More cool hardware devices coming soon..
Joanna Stern, with another excellent piece for the Wall Street Journal. (Apple News+) This time about Apple’s Screen Time, which I’ve been trying to use successfully for years. Glad I’m not alone in this thinking.
Stern’s tweet:
Apple’s Screen Time parental controls are broken, and it feels like an afterthought for the company.
The latest example? Two security researchers have been reporting a bug to Apple since 2021 that lets kids visit blocked sites.
Only after I called did Apple say it would be fixed in the next software update.
These seem like more substantial bugs that can be easily fixed in the “next software update.”
Nick Heer summarizes the problem perfectly:
One could reasonably argue nobody should entirely depend on software to determine how devices are used by themselves or their children, but I do not think many people realistically do. It is part of a combination of factors. Screen Time should at perform the baseline functions it promises. It sucks how common problems are basically ignored until Stern writes about them.
Sean Hollister, writing at The Verge:
Before he brought VR headsets out of the dark ages with Oculus, before he got fired by Facebook following his confusing and controversial political donations, before he went on to found a defense contractor named Anduril that builds lethal drones (!), Palmer Luckey was a Nintendo Game Boy modder.
It’s true. There’s photo evidence — and today, he’s revealing his own take on the ultimate cartridge-playing Nintendo Game Boy, one that you can actually buy for $199.
This thing looks really cool. $199, and shipping by Christmas. I love the resurgence of niche and creative hardware products.
Yesterday, current and former members of OpenAI and Google posted an open letter about the risks of AI and the companies developing it:
We are current and former employees at frontier AI companies, and we believe in the potential of AI technology to deliver unprecedented benefits to humanity.
We also understand the serious risks posed by these technologies. These risks range from the further entrenchment of existing inequalities, to manipulation and misinformation, to the loss of control of autonomous AI systems potentially resulting in human extinction. […]
That escalated quickly.
AI companies possess substantial non-public information about the capabilities and limitations of their systems, the adequacy of their protective measures, and the risk levels of different kinds of harm. However, they currently have only weak obligations to share some of this information with governments, and none with civil society. We do not think they can all be relied upon to share it voluntarily.
So long as there is no effective government oversight of these corporations, current and former employees are among the few people who can hold them accountable to the public.
Holding out hope that the government, especially here in the US, is going to sweep in and do a great job regulating this industry seems like a fools errand. What in our recent history shows that the government would be able to do this?
I agree with Casey Newton, on Mastodon:
There’s yet another open letter from the AI safety crowd. If they want more people to take them seriously, they need to get more specific
Rand Fishkin writing at the SparkToro Blog with a bombshell leak of policies from Google on how its search algorithm works:
On Sunday, May 5th, I received an email from a person claiming to have access to a massive leak of API documentation from inside Google’s Search division. The email further claimed that these leaked documents were confirmed as authentic by ex-Google employees, and that those ex-employees and others had shared additional, private information about Google’s search operations.
Many of their claims directly contradict public statements made by Googlers over the years, in particular the company’s repeated denial that click-centric user signals are employed, denial that subdomains are considered separately in rankings, denials of a sandbox for newer websites, denials that a domain’s age is collected or considered, and more.
These documents are really sending the SEO industry into a tailspin this week. This is a fascinating look into one of the most closely guarded secrets in tech.
Mike King also received the documents and has a great breakdown on iPullRank:
I have reviewed the API reference docs and contextualized them with some other previous Google leaks and the DOJ antitrust testimony. […]
You’d be tempted to broadly call these “ranking factors,” but that would be imprecise. Many, even most, of them are ranking factors, but many are not. What I’ll do here is contextualize some of the most interesting ranking systems and features (at least, those I was able to find in the first few hours of reviewing this massive leak) based on my extensive research and things that Google has told/lied to us about over the years.
“Lied” is harsh, but it’s the only accurate word to use here. While I don’t necessarily fault Google’s public representatives for protecting their proprietary information, I do take issue with their efforts to actively discredit people in the marketing, tech, and journalism worlds who have presented reproducible discoveries.
At first, there was silence from Google on this leak. Yesterday, The Verge received email confirmation with a statement:
“We would caution against making inaccurate assumptions about Search based on out-of-context, outdated, or incomplete information,” Google spokesperson Davis Thompson told The Verge in an email. “We’ve shared extensive information about how Search works and the types of factors that our systems weigh, while also working to protect the integrity of our results from manipulation.”
Sara Fischer writing for Axios:
The Atlantic, one of the oldest magazines in the U.S., and Vox Media, one of the nation’s largest digital media holding companies, have both inked separate licensing and product deals with ChatGPT parent OpenAI. […]
The deals give OpenAI added momentum in its quest for credible content to train its algorithms and inform its chatbots — and could also protect the Microsoft-backed company further from future copyright liability. […]
Deal terms weren’t disclosed, but it’s safe to assume both publishers are being compensated for their content — that’s how previous deals between publishers and OpenAI have been structured.
The land grab for licensed content to feed into AI continues.
Also interesting:
Internally, Vox will leverage OpenAI’s tech for its first party data platform, Forte, to bolster advertising creative optimization and audience targeting capabilities.
The Vox Media Union’s response:
Today, members of the Vox Media Union, Thrillist Union, and The Dodo Union were informed without warning that Vox Media entered into a “strategic content and product partnership” with OpenAl. As both journalists and workers, we have serious concerns about this partnership, which we believe could adversely impact members of our union, not to mention the well-documented ethical and environmental concerns surrounding the use of generative Al. We demand that Vox Media engage with us on this issue transparently - and address our many unanswered questions about this partnership — instead of continuing to fail to include our voices in decisions like these. We know that Al is already having a monumental impact on our work, and we demand a seat at the table in discussions about its future at Vox Media.
Nilay Patel, Editor in Chief of The Verge, responds on Threads:
I don’t have a lot to say about this – our newsroom is independent of the company’s business dealings as it’s always been. We’ll figure out some disclosure language and do a disclosure when it’s appropriate, we are pretty good at those ;)
[…]
To me it’s the same as anything – there is a firewall between editorial and the commercial side of the business. They don’t get to tell us what to do, and we don’t get involved in how they generate revenue
It’s been a wild week here in Dallas. From the Dallas Morning News:
Ferocious winds and heavy rain swept through North Texas early Tuesday, flooding roads, downing trees and power lines and knocking out power for hundreds of thousands of people.
Many in the Dallas area woke to the wail of tornado sirens when the National Weather Service issued a tornado warning.
“Take cover now!” the weather service in Fort Worth said at 6:13 a.m. on X. The storm packed nearly 80 mph winds, golf ball-sized hail and the threat of flash floods.
Dallas County Judge Clay Lewis Jenkins issued an emergency declaration and warned some residents will be without power for days. At a news conference Tuesday, Lewis Jenkins said the extent of the damage and number of people affected make this storm unusual, and he urged patience in the days ahead.
This is the worst storm I’ve been a part of since living here. No power for us still, but hopefully by end of week. Stay safe out there.
The Eruption and the Pinto - by Patrick Witty:
Then, at 8:32 am, the mountain erupted with fury. Lasher skid to a halt as the gigantic ash cloud barreled towards him, jumping out of his Pinto and taking the now-infamous photo. Normally, it’s a Cardinal sin for a photographer’s vehicle (or camera bag) to appear in the frame. But not this time.
It was too late for Lasher to turn around his car, so he jumped on his Yamaha and fled the plume of ash. “By the time he unhooked his dirt bike his Pinto was on fire,” Smith told me.
via @stop on threads
Sonos has made a lot of headlines lately.
First, the good. Chris Welch, reviews the new Sonos Ace headphones for The Verge:
These look like what you’d get if you put Sony’s WH-1000XM5 and Apple’s AirPods Max into a blender. The pleather ear cushions are magnetic and easily removable, though Sonos tosses in some thoughtful touches of its own; the insides are color-coded so you can easily tell which goes on what side. There’s a fingerprint-resistant coating on the exterior of the headphones to reduce smudges — particularly helpful for the black pair. And the memory foam headband has varying levels of padding to avoid putting too much pressure on any one section of your head.
Mercifully, the Ace are far lighter than the AirPods Max. There’s not quite as much metal throughout, but they still feel very well put together. And on my ears, they felt wonderfully comfortable. […]
Some very impressive details and considerations here.
Try as I might, I couldn’t find any obvious first-generation hardware flaws in my brief time with them. Maybe they’ll reveal themselves as I review the Ace, but on first impression, it’s clear that Sonos sweated the small details. (One more example: inside the fabric carrying case is a pouch for the USB-C and headphone cables that also attaches magnetically.) The controls are done right too, with physical buttons for everything and no tap or swipe gestures to memorize.
The Sonos hardware and design appear superior to the AirPods Max in every way. Will be curious to see if Apple revises the Max later this fall and fixes many of the issues with the first generation. Not that they’re worried about Sonos, but it sure would be nice to see them keep up.
Second, the bad. The new and “improved” Sonos app seems to be a hot mess. I disabled automatic app updates on my phone to avoid this. My iPad did update automatically so I’ve been trying to fight through the new UI there. It’s not great. And as noted in another Reddit comment the app update removed all accessibility support. Yikes.
Last week, Microsoft held a Windows event keynote prior to its Build Conference. Normally this is not the sort of thing I’d pay much attention to, or highlight here, but the entire thing was very compelling and interesting.
From the Microsoft Blog:
Today, at a special event on our new Microsoft campus, we introduced the world to a new category of Windows PCs designed for AI, Copilot+ PCs.
Copilot+ PCs are the fastest, most intelligent Windows PCs ever built. With powerful new silicon capable of an incredible 40+ TOPS (trillion operations per second), all–day battery life and access to the most advanced AI models, Copilot+ PCs will enable you to do things you can’t on any other PC. Easily find and remember what you have seen in your PC with Recall, generate and refine AI images in near real-time directly on the device using Cocreator, and bridge language barriers with Live Captions, translating audio from 40+ languages into English.
The keynote focused several times by name on the MacBook Air line, and how these new Copilot+ PCs are out performing the best-selling Apple laptops.
Tom Warren writes more at The Verge:
Microsoft is confident that it finally nailed the transition to Arm chips — so confident that, this time around, the company spent an entire day pitting its new hardware against the MacBook Air.
On a recent morning at its headquarters in Redmond, Washington, Microsoft representatives set out new Surface devices equipped with Qualcomm’s Snapdragon X Elite chips inside and compared them directly to Apple’s category-leading laptop. I witnessed an hour of demos and benchmarks that started with Geekbench and Cinebench comparisons, then moved on to apps and compatibility.
Geekbench scores aren’t everything, for sure. And it’s not a fair comparison when one device is passively cooled without fans and the other sports an active cooling system with, at least, one fan.
The fan issue was noted by Andrew Cunningham at ArsTechnica:
One caveat that I hadn’t seen mentioned in Microsoft’s presentation or in other coverage of the announcement, though: Microsoft says that both of these devices have fans. Apple still uses fans for the MacBook Pro lineup, but the MacBook Air is totally fanless. Bear that in mind when reading Microsoft’s claims about performance.
The new Surface Pro looks very nice. To me this isn’t a MacBook Air competitor, but rather one for the iPad itself. A detachable keyboard, touch screen, pencil support, and a decent screen that runs a desktop-class operating system is a good thing. Microsoft seems to understand this. For the first time in over 20 years I’m very tempted to buy a Windows PC! Wow.
These announcements are a bold bet and are very compelling. In a time where passionate support for Apple products may be waning, this is a solid step for Microsoft to take. I am continuously impressed with Satya Nadella and his leadership.
Happy Memorial Day and official start to summer. Phew! I’m coming up for air after an intense few working and personal life weeks to end the school year.
☀️🇺🇸⚾️
A nice breakdown of the track at Imola by Madeline Coleman for The Athletic.
Enzo Ferrari was on hand when construction began in 1950, and later described in his book how the circuit could become similar to a famed German track: “A small Nürburgring – I repeated that day looking around – a small Nürburgring, with equal technical resources, spectacular and an ideal path length. This belief has been achieved through the decades that have passed since then.”
Glad this race is back for 2024 and I’m excited for the European schedule to begin this weekend in Italy.
Wow. Someone was selling Apple Employee #10’s employee badge?! What an incredible piece of Apple history! Sure, it’s not Steve Jobs’ badge (despite the auction title), but there are only so many of these in the world – especially from one of the first ten employees.
At first, it looked good. The plastic was scuffed with age, the tape on the map was yellowed, the logo was (mostly) correct, and Sherry Livingston really was Employee #10.
But it also felt a little off. The scuffing looked… sandpapery. The splotches on the map felt overcooked. And I couldn’t stop looking at the “typewritten” part…
Cabel is the best. This is so cool.
Federico Viticci reviews the state of iPadOS:
If you’ve used iPadOS long enough (the iPad has been my primary computer for 12 years now), I’m sure you’ve run into these: the small bugs, annoyances, and missing features that don’t seem like much in isolation. Considered as a whole, however, they paint a not-too-rosy picture for an operating system that, 14 years into its existence, still lags behind macOS in terms of basic functionalities and problems that have never been addressed. […]
You know what’s equally the best and worst part of all this? That I still love the iPad.
The iPad is the only Apple computer that genuinely feels made for someone like me – a person who loves modularity, freedom, and the mix of touch and keyboard interactions. I share my frustrations because I care about the platform and want it to get better. But at the same time, we need to face reality: the iPad’s operating system isn’t improving at the speed the hardware deserves – that iPad owners who spent thousands of dollars on these machines deserve.
Something needs to change.
John Gruber has a slightly different take on Daring Fireball:
iPadOS has never been a workstation-style OS. The obvious truth — reiterated in recent weeks by the EU calling bullshit (or perhaps, conneries) on Apple’s claim that iPad and iPhone are separate platforms — is that iPadOS is a souped-up tablet-oriented variant of iOS.
This has never been more true than now — the M4 iPad Pros are, by some practical measures, the fastest computers Apple makes. But iPadOS is not the sort of system that the typical power user would think to run on super-powerful hardware.
But let’s invert our thinking on this. Instead of starting with the hardware and pondering what the ideal software would be like to take advantage of its power, let’s start with the software. A concept for simplicity-first console-style touchscreen tablet computing. A metaphor for computing with smartphone-style guardrails, with tablet-specific features like stylus support and laptop docking. A tablet OS that is unabashedly a souped-up version of iOS, not a stripped-down version of MacOS. What type of hardware should Apple build to instantiate such a platform?
Nicely produced website and list of the 100 Best Albums. Curated by Apple Music with “the help of artists and experts.”
The list will be complete on May 22nd, so just the bottom 20 are available now. I have many gripes on the ordering of this list so far, but will reserve my judgement until the entire thing is ready. 🤣
(Built in part with Svelte, it appears.)
It’s tech conference season. Yesterday was the main Google I/O keynote and, in case you haven’t heard, Google is working on AI.
As usual, The Verge has a nice video breakdown of the most important announcements. (I find sitting through Google keynotes tedious so this was helpful.)
Just a few overall thoughts from the presentation:
Google Lens gets video. Taking a video of a computer screen of code, and having Google explain the code is very interesting.
The Gemini features within Google Workspace look incredible. Creating sheets from a list of emails and and analyzing data across many sources will be very useful.
So many announcements, naming conventions, and code names. Astra, Veo, Gemma, Gemini, Gems, SynthID, yikes. Hard to decipher and remember what each is for, and what the difference is unless you work at Google on one of these teams.
I find it very interesting to hear the launch timing around each of the announcements. The “later this year” and “in the coming months” timelines really speak to unfinished and potentially reactionary features. The features that are shipping now are the most interesting to me and shows what Google has actually prioritized over the past year.
Nice bit at the end where they use Gemini to count the number of times the word “AI” has occurred in the keynote. (121)
In contrast to the OpenAI Spring Update from Monday, it seems like all of the demos from Google were very scripted and clearly pre-recorded. I understand why given the nature of this tech and its unpredictability, but the pre-recorded demos feel less real and more contrived.
Google has a long history of announcing unfinished work at I/O that often doesn’t even ship to users. I’ll be excited to see which of these announcements make it into shipping products for customers this year. Some very cool ideas and features here, now it’s time to deliver them.