We're less than a week away from the debut of Apple's first foldable iPhone, and rumors have been picking up as the event date approaches. Here's the latest.
Crease - Some of the early language about the foldable iPhone's crease suggested Apple might have found a way to make it invisible, but later rumors are more realistic. The crease is minimal, but Apple hasn't found a magic way to get rid of it.
Touch ID - Touch ID is still expected for the foldable iPhone. Apple didn't have room for the Face ID hardware because the device is rumored to be just 4.5mm thick. There will be volume buttons at the top of the iPhone Ultra, with a Touch ID side button for unlocking.
Software - iOS 27makes it clear Apple is making sure apps will work with the foldable iPhone. There's a big push for app resizability and layout flexibility. The iPhone is expected to support multitasking, sidebars, and side-by-side app usage.
Cameras - Camera rumors haven't changed. We're expecting a two-lens camera setup with Wide and Ultra Wide lenses.
Chips - Early rumors said Apple would transition to the C2 and move away from Qualcomm, but it looks like Apple might still use Qualcomm modems in the United States. It hinges on whether the C2 supports mmWave 5G. The iPhone Ultra will also adopt Apple's A20 Pro chip, the same as the iPhone 18 Pro models.
Colors - The foldable iPhone is expected in white/silver and dark blue.
Apple Pencil - Apple tested an Apple Pencil with the foldable iPhone, but it's probably not going to launch.
Launch Timing - Rumors on iPhone Ultra launch timing and production have been mixed, with some sources reporting delays. Gurman says he will be "very surprised" if the device doesn't launch at the same time as the iPhone 18 Pro models or "very close to it." It's still possible supplies could be constrained, leading to long waits for those who don't pre-order early.
Price - The iPhone Ultra is expected to cost over $2,000, possibly up to $2,500 depending on storage capacity.
Apple will introduce the iPhone Ultra and the iPhone 18 Pro models at its "Surprise and Shine" event that will be held on Wednesday, September 9.
The Interop 2027 call for proposals is open, submit your ideas now!
Over the past six years, the Interop project has helped bring key web platform features to a state where developers can rely on them across browsers. Technologies such as CSS layout and utilities, View Transitions, anchor positioning, popover, and shadow DOM have all benefited from the focused effort that Interop makes possible.
This year, browser vendors have continued to work across a broad set of focus areas including View Transitions, Navigation API, CSS Anchor Positioning, WebRTC, WebTransport, Dialogs and Popovers, custom highlights, container style queries, and more. These efforts are helping move the web platform toward a future where developers can spend less time working around browser differences and more time building great experiences.
Interoperability remains one of the most important challenges facing the web. New capabilities are constantly being standardized and implemented, but they're most valuable when they work consistently everywhere. That's why the Interop project continues to evolve each year, guided by input from developers, framework authors, browser vendors, and standards communities.
Today, we're opening the proposal process for Interop 2027.
If you've encountered a web platform feature that behaves differently across browsers, a standards-based capability that's difficult to rely on today, or an area where better interoperability would unlock new opportunities, we want to hear from you.
The proposal submission period is open now and will remain open through September 23, 2026. The Interop team will review submissions and evaluate them using the Interop selection process, with the goal of publishing the final Interop 2027 scope early February 2027.
What makes a strong proposal?
The strongest Interop proposals share a few common traits:
They solve a real developer problem. Make sure you explain how inconsistent browser behavior or missing support affect developers in practice.
They have ecosystem impact. Evidence from developers, frameworks, libraries, issue trackers, surveys, or real-world sites helps demonstrate that a problem is worth solving.
They're about standards-based features. Features with a stable specification and broad standards support are better candidates for Interop.
They can be measured. Interop relies on Web Platform Tests (WPT) to track progress, so proposals should either have strong existing test coverage or a clear path to adding it.
They are scoped for meaningful progress. A focused, concrete proposal is often easier to evaluate and more likely to succeed than a broad request such as "improve CSS" or "fix forms everywhere."
Submit your ideas now, even if you're unsure they qualify.
Even proposals that aren't selected can help us understand where you're experiencing pain and where future interoperability work is needed. On the Microsoft Edge team, we're always listening to developer signals, and unselected Interop proposals often constitute clear needs from developers.
We track these needs, and others, on our Top Developer Needs dashboard to influence progress on the highest-impact, long-running issues. We'll be refreshing this dashboard early next year.
A new version 0.1.11 of the RcppExamples
package is now on CRAN, and has
been built for r2u.
RcppExamples
provides a handful of short examples detailing by concrete working
examples how to set up basic R data structures in C++. It also provides
a simple example for packaging with Rcpp. The package provides (generally
fairly) simple examples, more interesting, compelling (and generally
longer) examples are at the Rcpp
Gallery.
This releases updates a few Rd files to adhere to a stricter standing
of checking by R. The NEWS extract follows:
Changes in
RcppExamples version 0.1.11 (2026-09-03)
Add now-checked-for missing sections to manual pages
Apple is facing a £2 billion ($2.7 billion) lawsuit in London accusing it of imposing stricter App Tracking Transparency rules on third party developers than on its own advertising services (via Reuters).
The case, filed at London's Competition Appeal Tribunal, follows a similar suit in which the tribunal ordered Apple to pay around £1.5 billion over App Store commissions last October, a ruling Apple is appealing. The new claim still needs to be certified before it can proceed.
Amazon today has introduced a big discount on the Apple Studio Display with Nano-Texture Glass and Tilt-Adjustable Stand, available for $1,684.88 with an on-page coupon, down from $1,899.00. This is the first time in nearly three months we've seen any notable discounts on the 2026 Studio Display lineup.
Note: MacRumors is an affiliate partner with some of these vendors. When you click a link and make a purchase, we may receive a small payment, which helps us keep the site running.
Amazon's discount today beats the previous all-time low price by over $100. There are a few other Studio Displays with markdowns on Amazon for Labor Day, but none reach the steep discount of this model. Remember that you will need to clip the on-page coupon in order to see the final sale price.
Note: You won't see the deal price until checkout.
Apple refreshed the Studio Display lineup back in March, featuring two Thunderbolt 5 ports, built-in 12-megapixel Center Stage camera with Desk View, a six-speaker sound system, and more.
If you're on the hunt for more discounts, be sure to visit our Apple Deals roundup where we recap the best Apple-related bargains of the past week.
Deals Newsletter
Interested in hearing more about the best deals you can find in 2026? Sign up for our Deals Newsletter and we'll keep you updated so you don't miss the biggest deals of the season!
Ahead of Apple's event on Wednesday, September 9, where it is expected to unveil its first-ever foldable iPhone, Taiwanese research firm TrendForce today outlined the device's estimated price range, features, and some other key details.
In the U.S., TrendForce estimated that the foldable iPhone will have a starting price of $2,099 to $2,299 and said the highest-end configurations could surpass the $3,000 mark. In other words, the foldable iPhone would easily be the most expensive iPhone ever, coming in well above the iPhone 17 Pro Max's starting price of $1,199.
The foldable iPhone will be named iPhone Ultra, according to a previous report from Macworld's Filipe Espósito, who now writes for Engadget.
TrendForce expects the foldable iPhone to have the following specs:
5.5-inch outer OLED display
7.8-inch inner OLED display with ProMotion for up to a 120Hz refresh rate
Touch ID
A20 Pro chip manufactured with TSMC's 2nm process
12GB of RAM
256GB, 512GB, 1TB, and 2TB storage capacities
24-megapixel front camera
48-megapixel rear Wide camera
48-megapixel rear Ultra Wide camera
With just six more days until Apple's event, the foldable iPhone will finally be revealed soon.
It is apparently not a good day for AI chatbots, as ChatGPT, Claude, and Grok are all down or experiencing issues for many users.
OpenAI's status page acknowledges issues across ChatGPT and Codex, while Grok's website notes that it is currently experiencing issues as well. Likewise, Anthropic's status page indicates Claude is experiencing elevated errors.
All of the companies are investigating the outages.
Update: The chatbots are in the process of coming back online.
Apple has canceled plans for a foldable MacBook and a larger OLEDMacBook Pro model, according to research firm Omdia.
In a new research note, Omdia said Apple has scrapped development of a large foldable MacBook that would have measured between 16.1 and 18 inches when unfolded. It was previously reported that this device had already slipped out of its original 2026 production window, with Omdia among the firms tracking it at the time, though under that earlier reporting it was generally described as a large-screened foldable iPad rather than a MacBook.
Apple has also apparently called off a separate, previously planned MacBook with an OLED display between 14.4 and 16.1 inches in size, which would have sat in its own tier between the standard MacBook Pro sizes and the canceled foldable. With both of those product lines scrapped, the 14.3 to 16.3 inch size range will seemingly remain exclusive to the MacBook Pro series going forward.
Omdia said Apple's broader strategy of bringing OLED displays to its entire iPad and MacBook lineup has not changed, but the company has reworked which products will get OLED and what specific OLED technology each one uses, a shift that lines up with the changes detailed in Apple's rumored 2026 and 2027 product roadmap.
As part of that same reshuffle, Apple is said to be adding a new 13.8 inch MacBook positioned between the LCD MacBook Air and the OLED MacBook Pro, using a new generation of OLED technology that includes an LTPO+ backplane, under-panel camera, and color-filter-on-encapsulation technology, among other features. That new 13.8 inch model appears to fill the gap in Apple's lineup that the now canceled 14.4 to 16.1 inch MacBook OLED would have occupied, positioned below the Pro rather than above it.
The 14.3- and 16.3-inch MacBook Pro OLED models, also referred to as the "MacBook Ultra," are expected to be announced in the second half of 2026 or early 2027. Omdia added that Apple will not discontinue the existing mini-LED MacBook Pro models in those sizes when the high-end OLED versions arrive, and the two display types are expected to sell in parallel for several years, with the mini-LED version likely continuing until at least 2028.
Apple has reportedly revised its plans for adopting OLED displays across its MacBook lineup, and its existing LCD-based MacBook Pro models with mini-LED backlighting are now expected to remain available for longer than anticipated.
According to a new OLED adoption roadmap published by research firm Omdia, Apple still intends long-term adoption of OLED panels across MacBooks and iPads, but the transition pace has recently changed, owing to higher memory and component costs.
Previously, Omdia had expected Apple to discontinue its existing LCD-based models sooner, once its 14-inch and 16-inch OLED MacBook Pro models arrive – either this year or in early 2027.
However, Omdia now predicts that Apple will continue to sell both 14-inch and 16-inch LCD MacBook Pro models in parallel with the OLED MacBook Pro product line into at least 2028, a product strategy arrangement that is said to be based on cost factors.
Previous reports had already suggested Apple will position the OLED-based "MacBook Ultra" above the existing MacBook Pro models in its lineup, but there was little information about how long Apple would keep the latter models around for. Omdia's roadmap is the first to suggest they could remain available for another year at least.
Apple reportedly plans to release MacBook Pro models with OLED displays by early 2027, and Omdia analyst David Hsieh has predicted that the laptops will have an M5 Ultra chip option.
Given that these MacBook Pro models might actually be branded as "MacBook Ultra" models, it would be fitting if there were configurations with the M5 Ultra chip available. But keep your expectations low, as there has never been a MacBook with an Ultra chip, and a previous report from Bloomberg's Mark Gurman only mentioned M5 Pro and M5 Max chips for these laptops.
Apple may introduce an entirely new 13.8-inch OLED MacBook model in 2029, potentially creating a new product tier between the MacBook Air and MacBook Pro, according to an updated product roadmap from research firm Omdia.
The firm's latest roadmap sandwiches a new 13.8-inch MacBook between the projected time line for the Air and Pro models. The surprise model is said to use a single-stack hybrid OLED panel, rather than the more sophisticated tandem OLED technology that is expected in the upcoming OLED MacBook Pro models, which are set to launch either later this year or early next year.
It seems this prediction hasn't come out of nowhere, either. According to a June report from ZDNet Korea, Apple sounded out Samsung Display, LG Display, and BOE for information regarding a 13.8-inch OLED MacBook panel earlier this year. At the time, the general assumption was that it was likely to be part of Apple's work on the first OLED MacBook Air, expected by 2028 at the earliest.
However, looking at Omdia's latest roadmap, the OLED MacBook Air is conspicuously absent, having been replaced by the 13.8-inch OLED machine in a separate "new" product category. The new model is expected to incorporate an LTPO+ backplane, integrated touch sensing, an under-panel camera, and a color filter integrated into the OLED encapsulation layer.
Omdia believes it is still too early to know for sure which display manufacturer would supply panels for the device, but the firm believes that it could potentially become an important mainstream MacBook product, positioned between the LCD MacBook Air and the higher-end OLED MacBook Pro.
Meanwhile, the MacBook Air line is shown as continuing to use LCD displays right into 2030.
Of course, Apple's wholesale OLED adoption plans could change, and Omdia itself caveats that its roadmap is based on current market expectations. On that basis alone, it could be adjusted at any time according to OLED manufacturing capacity, component costs, and Apple's wider product strategy. Additionally, this is the first we've heard of a completely new MacBook line, and it's still three years out. So with all due caution, file this one under "wait and see."
An alleged solder joint diagram for Apple's A20 Pro chip suggests the processor will deliver a number of significant upgrades, including a seven-core GPU, a wider memory interface, and a larger L2 cache for its efficiency cores.
The claims come from a Weibo user who says the layout was reconstructed by mapping the chip's solder ball positions and overlaying them with die level images to sketch out where each block of the processor sits. According to the post, the resulting outline points to a two and four core CPU configuration and a seven core GPU, with the top of the die housing the GPU and the CPU occupying the center.
Based on the layout, the leaker expects the A20 Pro's efficiency cores to share an 8MB L2 cache, up from 6MB on the A19 Pro, while the performance cores retain their existing 16MB L2 cache. The size of the shared system level cache is not yet known, but the leaker believes it is unlikely to shrink given the chip's new image signal processor and neural engine.
With regards to the GPU, the seven core configuration would mark an increase from the six cores on the A19 Pro, a jump of roughly 16%. The leak also points to a 96-bit memory interface for LPDDR5x RAM, up from the 64-bit interface on the A19 Pro, a 50% increase that would meaningfully boost memory bandwidth. Graphics performance tends to scale with memory bandwidth, so the wider bus could have an additional impact on top of the additional GPU core.
The A20 Pro is highly likely to debut the iPhone 18 Pro and iPhone 18 Pro Max next week. It could also be used in Apple's first foldable iPhone.
Less than a week before Apple debuts new iPhone 18 Pro models, research firm TrendForce has estimated the prices of the devices based on the latest hardware cost environment.
TrendForce's price estimates compared to last year's models look like this:
Model
Starting Price
Model
Starting Price
Change
iPhone 17 Pro
$1,099
iPhone 18 Pro
$1,249-$1,299
+$150 to +$200
iPhone 17 Pro Max
$1,199
iPhone 18 Pro Max
$1,349-$1,399
+$150 to +$200
Apple is likely to raise iPhone 18 Pro and iPhone 18 Pro Max prices by around 10-20% compared to last year's Pro models, according to the firm. Taking TrendForce's June forecast of a 14-18% increase, which fits within its latest broader 10-20% range and translates more neatly to Apple's typical pricing scheme, we could see the iPhone 18 Pro starting price jump to between $1,249-$1,299 and the iPhone 18 Pro Max between $1,349-$1,399.
Meanwhile, TrendForce expects prices for Apple's first foldable iPhone (aka "iPhone Ultra") to start anywhere between $2,099-$2,299, with the highest storage option potentially exceeding $3,000.
The regular iPhone 18 model is not expected to debut until spring 2027. Nevertheless, TrendForce expects the same general 10-20% price increase to apply to that device as well.
The overriding reason for the price increases is memory costs. TrendForce says the cost of a 256GB memory chip has risen nearly 400% year-on-year, while its separate analysis puts the overall bill of materials (BOM) for a 256GB Pro model at around 38% more over the same period. That said, Apple is expected to absorb some of that increase via lower margins, rather than passing the full cost onto consumers.
Last year, Apple managed to avoid raising the price of the iPhone 17 Pro Max, and only marginally raised the iPhone 17 Pro price in the U.S. year-on-year by $100. This year, however, the insatiable AI infrastructure buildout has made memory costs even more expensive, effectively holding the rest of the electronics industry to ransom. And things don't look like they're going to get any better anytime soon.
We will find out the real prices during Apple's "Surprise and shine" event on Wednesday, September 9.
Anker today announced the launch of Anker MindBase, an AI home hub meant to work with almost 100 Anker devices and Matter-enabled smart home devices from other companies. It works alongside HomeKit, and is equipped with an AI Core compute card and the on-device Anker TrueSmart AI Agent.
In addition to working as a smart home hub, the MindBase supports up to 48TB of local storage, and it functions as a NAS. It can be used for storing files, photos, and videos locally. MindBase integrates with Anker security cameras, and Anker plans to sell a MindBase Security Kit that includes a new TrackLight Cam S1, S4 Video Doorbell, and Smart Security Shield sensor.
Anker says on-device AI evaluates events in real time to assess risk and respond through observation, lighting, verbal warnings, or active deterrence. The sensor can be placed at the edge of a property, using radar to detect strangers from up to 15 meters away.
Anker is also debuting several new home security products.
Eufy Video Doorbell S4 - The Video Doorbell S4 is a wired doorbell and security camera combo. It has a 9-megapixel lens with 3K resolution and a 180-degree field of view.
Eufy Window Camera E10 - The E10 is a camera that mounts on the inside of a window for use in rentals. It has a 25-degree adjustable tilt with anti-glare optics, and it records in 2.5K resolution.
Eufy Fall Detection Sensor - The Fall Detection Sensor is a non-wearable sensor that uses mmWave radar sensing to detect falls without using a camera. It can be mounted on the wall and supports two-way communication.
Eufy NVR Security System E50 - The E50 pairs an 8-channel network video recorder with four hybrid bullet pan-tilt-zoom dual-lens camera. There's a 2K resolution wide-angle bullet lens and a motorized 4K resolution PTZ lens.
The E50 system is available on the Eufy website starting at $999. The Video Doorbell S4 is priced at $300 and will be available in late September. The Window Camera E10 will be available later this year for $70. Availability and pricing have not yet been announced for the MindBase and Fall Detection Sensor.
Aqara today introduced five new Matter-enabled smart lights compatible with Apple Home. The lights support Thread and Zigbee as well, and can be paired with Aqara sensors and switches.
The Floor Lamp T1 is a tall, rod-shaped light with a wall-wash effect. It has 18 controllable zones that can be set to different colors, and it supports 2000K to 9000K white light at up to 1,000 lumens.
Aqara's LED Downlight T2 has a thin profile for low ceilings or retrofitting. It supports 16 million colors and 2000K to 9000K white light. The trimmable LED Strip T2 has 108 LEDs per meter and is ideal for use under cabinets and along shelves. Aqara says it delivers even light without visible hotspots. The light strip has 18 controllable zones every three meters for smooth color gradients.
The Outdoor String Lights H1 are weatherproof for use on patios and decks. The lights have individually controllable zones with the same white light range as the other lights, and can adjust based on time of day, environment, and triggers from other smart home devices.
Aqara also announced the Permanent Outdoor Lights H1 that can be used all year. The lights feature independently controlled LEDs and can be extended or cut to fit any roofline.
Aqara is showing off the new lights at IFA. Pricing and launch timing are not yet available.
Govee today introduced three new lighting products, including theSky Ceiling Light, COB Strip Light 2 Pro, and Permanent Outdoor Lights 2 Pro.
Govee says the $400 Matter-enabled Sky Ceiling Light is a "mood-responsive" light that uses AI Lighting Bot 2.0 to set the color of the light based on the user's mood or request. The light has a 478nm sky-blue spectrum for use indoors when daylight is limited, which Govee claims supports the body's "natural vitality and rhythm." It also offers white light across a 1,000K to 10,000K temperature range, with up to 7,000 lumens.
The $170 COB Strip Light 2 Pro displays custom light scenes using AI text prompts. Govee says users can curate ambient lighting for gaming, parties, movie nights, and more. The light strip has 1,344 LEDs and 16 ICs per meter, plus a triple-layer groove design for smooth output in single-color and multi-color modes.
Govee's Permanent Outdoor Lights 2 Pro feature a dual-layer design with two independently controlled ICs per light head, which allows the top and bottom zones to show different colors and effects at once. Like Govee's other lights, the Outdoor Lights 2 Pro can be controlled with AI Lighting Bot 2.0 and are priced starting at $500.
Another minor maintenance release version 0.1.5 of package RcppClassicExamples
arrived earlier today on CRAN,
and has been built for r2u. This package
illustrates usage of the very old and otherwise deprecated initial Rcpp
API which no new projects should use as the normal and current
Rcpp API is so much better.
This release follows one from six months ago, and is even smaller. We
just update a few Rd files to adhere to a stricter standing of checking
by R.
No new code or features. Full details below. And as a reminder, don’t
use the old RcppClassic – use Rcpp
instead.
An original Apple–1 Computer could sell for between $500,000 and $800,000 when it is auctioned off this October.
There are fewer than 75 Apple–1 Computers known to exist, and a smaller number that are operational. The "Hatfield" Apple–1 is number 48 on the Apple–1 registry, and it is in working condition.
The machine was purchased by Fred Hatfield, a computer enthusiast and employee of Computer Data Systems. It is signed by Steve Wozniak and comes with a letter written to Hatfield and signed by Steve Jobs. Jobs authorized a trade-in for the newly released Apple II after Hatfield complained about the Apple–1 software, but Hatfield declined the offer because he would have needed to pay $400 for the upgrade.
Hatfield sold the Apple–1 in 2013 for $40,000. At the time, it was non-functional, but the buyer fixed it, got it signed by Wozniak, and sold it at auction for $671,400. It's now going up for sale again.
Steve Jobs and Steve Wozniak produced 200 Apple–1 computers in 1976. The device was originally offered as a bare circuit board to be sold as a kit and completed by hobbyists, but Jobs sold 50 fully assembled Apple–1 machines to California-based retailer The Byte Shop. The Byte Shop resold the assembled machines for $666.66.
The Hatfield Apple–1 will be available for public and private viewing at locations in Los Angeles, Boston, San Francisco, and New York, with dates and information listed on the auctioneer's website. Bids will be accepted from October 4 to October 14.
Functional Apple–1 computers have sold for over $900,000 at auction in the past, though pricing varies based on condition and provenance.
Apple has added a new "iPhone Handoff" feature to iOS 27 that will allow you to switch between two iPhones while using the same phone number on each device.
This functionality was briefly mentioned during the WWDC 2026 keynote in June, on a slide that listed hundreds of new features coming in iOS 27 and corresponding software updates, but Apple never shared any further details at the time.
MacRumors forum member pdfu today shared a video that demonstrates how the feature will work, using two simulated iPhones running in Xcode 27's Device Hub.
"iPhone Handoff" can be set up from the Settings app, under the Cellular menu. During the setup process, you can choose which iPhone you want to be the main device that you use most often and which one you want to be the companion device.
Apple says a user's main eSIM will remain on the main iPhone, while the companion iPhone will receive a companion eSIM, and this will enable you to switch between the devices back and forth while using the same phone number.
All cellular settings are managed on the main iPhone, and location sharing is based on whichever iPhone is actively using the phone number.
In addition to needing two iPhones running iOS 27, "iPhone Handoff" will seemingly require carrier support. The forum post mentions that T-Mobile in the U.S. and Deutsche Telekom in Germany might be among the first carriers to offer the feature.
Following three months of beta testing, iOS 27 is expected to be released later this month, so we will likely hear more about this feature soon.
Uploaded xdg-desktop-portal-wlr 0.8.4-1 to unstable
Uploaded swayimg 5.5-1 to unstable
Uploaded fyi 1.0.4-2 to unstable
Uploaded labwc 0.20.2-1 to unstable
Uploaded yambar 1.11.0-2 to unstable, but that got removed because it FTBFS;
given that upstream has a big warning saying “This project is not developed
anymore” it is probably for the better
Closed #1133660
which was a FTBFS bug on usbguard, but neither I nor another use could
reproduce the buil failure
Created
ITP#1145583 for
miru which is a nice
little screen magnifier for wlroots based compositors
I did not partake in the flamewars on debian-vote about the LLM situation. I am
not sure how anyone can find this style of “discussion” productive. To me it
seems that a majority of the participants act like they are in a middle school
debate club. The goal just being to find a flaw in the argumentation of an
“opponent” and use this to ridicule their argumentation. Basically what
politicians do.
The good thing is, that most Debian members did not stoop on that level.
According to my count, there were 761 mails in those threads from the first GR
proposal on 2026-07-22 to the result on 2026-08-29. Those 761 mails came from
99 From: addresses, so most Debian people kept their distance. Given that
according to nm.debian.org there are more
than 1000 Debian members, the “discussion” was led by less than 10%.
The distribution of who wrote how many mails is also interesting. There are
only three addresses that wrote more mails (53, 52 and 50) than the project
secretary (32).
I think the most fitting approach to Debian mailinglists is a quote from
WOPR:
A STRANGE GAME. THE ONLY WINNING MOVE IS NOT TO PLAY.
DH Related Work
I released version 0.66.0 and 0.67.0 of the APIS framework as well as a couple
of bugfix releases for the 0.67.x version. In 0.67.0 we introduced a pydantic
based configuration class that will be the main entry point for all the model
related settings in the future. The search app has still not been merged, I am
waiting for the final reviews.
Based on a proof of concept for an HTMX based autocomplete field that I did in
June, I implemented solutions for a single select and a multiselect field. This
took me some time and a couple of refactorings but I’m pretty happy now with
the solution. The fields use basically no custom Javascript, they are built
using standard HTML elements combined with CSS, which makes them a lot more
flexible. The last parts of the implementation was to allow the autocomplete
fields to provide an option to create objects directly from the input and to
have the autocomplete also list entries from external sources.
Apple today shared a short video featuring an Apple Watch user named Amanda who says her device saved her life.
In 2024, Amanda began receiving alerts about an abnormal heart rate. Her watch alerted her about changes to her resting heart rate and her overnight vitals, prompting her to visit her doctor.
She ended up being diagnosed with acute myeloid leukemia, an aggressive cancer. After treatment, she is in remission.
"I absolutely believe that if it wasn't for my Apple Watch, I wouldn't be talking to you," she said.
The Apple Watch regularly monitors heart rate and can send alerts when a high heart rate is detected during rest. It also notes low heart rates and irregular heart rhythms, plus the ECG app can detect atrial fibrillation.
It worked, I wore it. Years later I saw a blog post on Pour La Victoire
on making a corset cover based on the same book, but with completely
different results, and thought that it would have been nice to make
another one to publish instructions for my take on it.
However, I didn’t have any embroidery flouncing on hand, nor did I have
a need for a new corset cover, and the project remained on the list, on
low priority (although I did buy some beading lace for it, when I
stumbled on it).
Then, after finishing my vampire shirt, I noticed that I had just enough
fabric left for a corset cover, and by just enough I really mean just
enough, as I discovered when laying the pattern on the fabric.
So I dug in my files to get the original pattern I used, brought it up
to date, and added the missing details such as the pleating guides that
I had skipped when making the pattern just for myself. Doing so I
realized that on my old cover I had done the fake pleat in the front
wrong, making just a single pleat instead of a box pleat. Also, I
originally directly gathered the sleeves in the armscyes, but watching
the book again I realized that the sleeves were made up of a gathered
ruffle plus a straight band.
Both issues were fixed and I could cut the fabric and start sewing. By
machine, including using a narrow hem foot instead of sewing rolled hems
by hand as my instinct kept reminding me would have looked neater.
But this is a garment from a sewing machine time, and probably one that
in many cases would have been bought from a mass producer, and it’s
underwear, so there is no real need for the hems to be perfect, as it’s
going to be hidden anyway. But most importantly, I wanted to write
instructions for machine sewing, for a change, and so I had to machine
sew all steps that I had to take pictures of.
I did do the buttonholes by hand, because I hate the buttonhole
attachment on my machine, and the buttonhole attachment hates me.
I used a lighter weight fabric for the sleeve ruffles, both because I
didn’t have a big enough piece of main fabric not to have to piece them,
and because I felt that it looks better, as it’s the same voile I used
for the ruffles on the vampire shirt.
When it came to the beading lace, I had two that I had bought more or
less thinking about this project: the earlier one was narrow and
suitable to do its job, but the one I had bought more recently was
taller, with an edge that made it suitable to give more fullness to the
bust when gathered up.
I contemplated for a short while, and then decided to go for fullness
and use the taller border for the top edge, but the smaller one at the
waist, where fullness is not wanted.
The book claimed that this pattern required little labour, and indeed it
did: even when taking step by step pictures it only took a few hours
spread over a week, plus the time to make buttonholes by hand over the
next week.
Recently, Daring Fireball's John Gruber resurfaced an old tip that lets you quickly access the settings menu for the currently opened app on your iPhone. If you find it a hassle digging through the Settings app to find the associated app menu, this solution saves you the bother.
It uses a simple Shortcut – made conveniently downloadable courtesy of Quinn Nelson – that you can add to Control Center (or better yet the Action button) by using the "Run Shortcut" control. We couldn't not highlight it here because it's so useful.
Amazon has introduced the first cash discounts on Apple's brand new Mac mini, with discounts including two of the M6 models as well as the M5 Pro model. All of these deals are pre-order discounts on the 2026 Mac mini, which officially launches on September 22.
Note: MacRumors is an affiliate partner with Amazon. When you click a link and make a purchase, we may receive a small payment, which helps us keep the site running.
You can get the 16GB RAM/256GB SSD Mac mini for $879.99, down from $899.00, and the 16GB RAM/512GB SSD model for $1,069.99, down from $1,099.00. Both of these devices come with the new M6 chip, and Amazon also has $29 off the M5 Pro Mac mini right now.
In terms of upgrades, Apple said the Mac mini with the M6 chip delivers up to 40% faster CPU performance, up to 4× faster performance for AI tasks in particular, up to 2× faster graphics performance, and up to 2× faster storage speeds compared to the previous-generation model with the 10-core M4 chip, 32GB of unified memory, and 2TB of storage.
If you're on the hunt for more discounts, be sure to visit our Apple Deals roundup where we recap the best Apple-related bargains of the past week.
Deals Newsletter
Interested in hearing more about the best deals you can find in 2026? Sign up for our Deals Newsletter and we'll keep you updated so you don't miss the biggest deals of the season!
Starting today, Oklahoma residents can add a digital version of their driver's license to the Apple Wallet app on the iPhone and Apple Watch. The digital ID can be used to display proof of identity or age at TSA checkpoints in select airports, at select businesses, and in select apps, but the number of places where it is accepted is still quite limited.
Oklahoma is the 16th state to implement support for Apple Wallet IDs in the U.S., joining Arizona, Arkansas, California, Colorado, Georgia, Hawaii, Illinois, Iowa, Maryland, Montana, New Mexico, North Dakota, Ohio, Virginia, and West Virginia. The feature is also available in Puerto Rico and works with Japan's My Number Card.
Apple has also been preparing for the feature to go live in North Carolina and Utah in the coming months, and years ago it announced that Connecticut, Kentucky, and Mississippi were committed to launching the feature at some point.
To set up the feature, open the Wallet app on the iPhone and tap on the plus sign in the top-right corner. Next, tap on Driver's License and ID Cards, select Oklahoma from the list, and follow the on-screen steps to complete the process. Then, you can present your digital ID where it is accepted without handing over your device.
You can present an Apple Wallet ID where accepted by simply tapping your iPhone or Apple Watch on an identity reader. Then, you will be able to view the specific information being requested by the TSA or business. Only after authorizing with Face ID or Touch ID is the requested identity information released from your device.
As mentioned, you do not need to unlock, show, or hand over your device to present an Apple Wallet ID in person, ensuring user privacy.
Apple Wallet IDs are generally not accepted by law enforcement, so carrying a physical ID is still legally required for traffic stops. In addition, the number of businesses that accept Apple Wallet IDs is still quite small. At least for now, Apple Wallet IDs are designed to be a convenient alternative where they are accepted.
Apple Wallet IDs are accepted at TSA checkpoints in more than 250 airports across the U.S. for domestic travel, according to Apple.
Apple last week announced two new Mac Studio models, with one built around the M5 Max chip and one around the all-new M5 Ultra chip, replacing the M4 Max and M3 Ultra models that had been on sale since March 2025.
The gap between the two is unusually wide this time. Mac Studio with M5 Max starts at $2,499, holding the price Apple set when it raised Mac prices in June, while Mac Studio with M5 Ultra starts at $5,499, a further $200 increase on the post-June M3 Ultra price and $1,500 above what the M3 Ultra cost at launch.
That makes the premium $3,000 before any configure-to-order options, and memory and storage upgrades are more expensive than usual because of the memory chip shortage. Our guide helps to answer the question of how to decide which of these two Mac Studio models is best for you. The shared features include:
An identical enclosure measuring 3.7 inches tall and 7.7 inches wide
Neural Accelerators built into every GPU core, and hardware-accelerated ray tracing
Hardware-accelerated H.264, HEVC, ProRes, and ProRes RAW, plus AV1 decode
A next-generation SSD architecture that Apple says is up to twice as fast as before
Thunderbolt 5 with up to 120Gb/s of bandwidth, and clustering of multiple systems using RDMA (remote direct memory access)
10Gb Ethernet, HDMI, and an SDXC card slot (UHS-II)
Wi-Fi 7 and Bluetooth 6 via the N1 chip, both new to the Mac Studio
Genlock over USB-C for synchronization with professional camera capture
A 3.5mm headphone jack with support for high-impedance headphones
Here is everything that differs between the two models:
Mac Studio (M5 Max)
Mac Studio (M5 Ultra)
Starts at $2,499
Starts at $5,499
M5 Max chip
M5 Ultra chip
18-core CPU with 6 super cores and 12 performance cores
Up to 36-core CPU with 12 super cores and 24 performance cores, from 30 cores
Up to 40-core GPU, from 32 cores
Up to 80-core GPU, from 64 cores
16-core Neural Engine
32-core Neural Engine
Up to 614GB/s memory bandwidth
1.2TB/s memory bandwidth
36GB, 48GB, 64GB, or 128GB of unified memory
96GB, 256GB, or 512GB of unified memory
512GB, 1TB, 2TB, 4TB, or 8TB of storage
1TB, 2TB, 4TB, 8TB, or 16TB of storage
Up to 3.9x faster AI performance than M4 Max
Up to 4.3x the peak AI compute performance of M3 Ultra
Up to 50% faster graphics than M4 Max
Up to 1.8x faster graphics than M3 Ultra
One video decode engine and two video encode engines
Two video decode engines and four video encode engines
Two ProRes encode and decode engines
Four ProRes encode and decode engines
Four Thunderbolt 5 ports, two USB-C ports, and two USB-A ports
Six Thunderbolt 5 ports and two USB-A ports
Support for up to five 4K displays, four 6K displays, or two 8K displays
Support for up to eight 4K displays, eight 6K displays, or four 8K displays
6.0 pounds (2.7 kg)
8.0 pounds (3.6 kg)
For the overwhelming majority of buyers, including most professionals, the M5 Max is the more appropriate choice. An 18-core CPU, up to 40 GPU cores with Neural Accelerators, up to 128GB of unified memory, and up to 614GB/s of bandwidth cover workflows such as multi-stream 4K and 6K editing, large Logic sessions with real-time plugins, compiling substantial codebases, and most 3D and motion work.
Apple's own figures support this, with up to 3x faster Magic Mask performance in Blackmagic Design DaVinci Resolve Studio and 1.4x faster scene rendering in Maxon Redshift compared to M4 Max, and graphics that are up to 50% faster than the previous generation. Since memory cannot be added later, the sensible approach is to spend more on unified memory at the point of purchase rather than on the higher chip tier.
You should only choose the M5 Ultra if you anticipate running into a performance ceiling the M5 Max cannot clear. The most obvious use-case is local AI inference, where the whole model needs to be resident in memory. 256GB and 512GB configurations paired with 1.2TB/s of bandwidth allow models with hundreds of billions of parameters to run on the machine itself, which keeps sensitive material on-device and eliminates the recurring cost of paying for cloud inference.
The other workflows that can take advantage of the M5 Ultra include studios finishing 8K material from several cameras at once, where the doubled Media Engine and its four encode engines earn their keep, research and simulation work that needs an enormous memory pool, and anyone who specifically needs six Thunderbolt 5 ports or more than five displays.
There is a notable catch in how the two memory ladders overlap. The M5 Ultra begins at 96GB, which is less than the 128GB you can configure on the M5 Max, so paying the premium at the entry configuration buys you more cores and more bandwidth, but less memory. The Ultra only separates itself at 256GB and above, and Apple limits the 512GB option to the 80-core GPU model, which means the configurations that justify the machine sit far above its starting price.
This is where the price proximity argument becomes noticeable. Configuring an M5 Max with memory and storage narrows the gap considerably, and at that point the Ultra becomes worth pricing alongside it, because it roughly doubles GPU cores, Neural Engine cores, memory bandwidth, and Media Engine capacity in one step. If your workload sits comfortably inside 128GB, though, that money is better spent on storage and displays than on the chip.
Both models support clustering over Thunderbolt 5 and RDMA, and Apple says four Mac Studio systems working together deliver up to 3x faster AI inference than a single machine, so scaling inference does not necessarily mean buying the most expensive configuration outright.
If the M5 Max still looks like more performance than you need, the new Mac mini with the M5 Pro chip starts at $1,699 with up to a 20-core GPU and 307GB/s of memory bandwidth, which is enough for a great deal of professional work.
Both new Mac Studio models are available to pre-order now, with availability beginning on September 22. The 512GB configuration arrives in late October, and its pricing is expected to start well above $10,000.
The Elec reports that Google, Oppo, and Vivo plan to adopt ultra-thin glass (UTG) in their foldable phones next year, following Apple's lead. Apple's book-style iPhone Ultra, expected to be unveiled next week, reportedly uses two UTG layers, with the display sandwiched between them, isolating it from direct contact with the hinge.
The method, pioneered by Apple, is meant to spread mechanical stress across multiple layers to improve durability and reduce visible creasing over time. The UTG layers will be mass produced for Apple's rivals by Samsung, for wide adoption in 2027.
Longtime leaker Sonny Dickson has shared an image of differently colored SIM card trays that he claims are for Apple's upcoming iPhone 18 Pro models, suggesting Apple's flagship lineup will again be available in a trio of colors: Silver, Dark Cherry, and Sky Blue.
Apple currently offers the iPhone 17 Pro in Cosmic Orange, Silver, and Deep Blue.
Dickson shared the image in response to a MacRumors report on Tuesday covering a Chinese leaker's claim that the iPhone 18 Pro models could come in Dark Cherry, Sky Blue, and Black.
Dickson told MacRumors the colors in the shared image are the only ones he's come across in recent months. However, the leaker did share iPhone 18 Pro dummy units back in May that included a fourth Black color. When that was pointed out to him, Dickson argued that black dummies are made every year, but it's not a color Apple prefers to use anymore.
iPhone models sold in the U.S., Canada, and several other regions use eSIMs and don't come with SIM card trays, but devices sold in Europe, Japan, China, and Australia still include one (with the exception of the iPhone Air, which lacked a SIM tray everywhere it was sold).
Chinese leaker Instant Digital in April said there would be no black color for the iPhone 18 Pro. Last year was the first time Apple's high-end iPhones have not been available with a black or dark gray color option in any way, so it's possible Apple will continue that trend. Macworldreported in the same month that Apple would offer Light Blue, Dark Cherry, Dark Gray, and Silver colors.
We'll know either way soon enough. Apple's iPhone 18 Pro and iPhone 18 Pro Max are expected to be unveiled at the company's September event next week, alongside the company's first foldable iPhone, aka "iPhone Ultra."
Apple last month announced a new Mac mini and a new Mac Studio. Both machines start shipping on Tuesday, September 22, which means they will launch in the same month as new iPhones, Apple Watches, and AirPods.
Apple is expected to introduce the unannounced products at its annual iPhone fall event, which is this year carrying the tagline "Surprise and shine." The event takes place on Wednesday, September 9, making this month one of the busiest Apple has had in a while.
We should therefore see eight new Apple products go on sale before the end of September:
Mac mini (M6 and M5 Pro): Announced last week, the new models start shipping September 22, with prices starting at $899.
Mac Studio (M5 Max and M5 Ultra): Also announced last week, the new models begin shipping September 22, with prices starting at $2,499.
iPhone 18 Pro: The flagship device set to be unveiled at Apple's fall event, with a smaller Dynamic Island and a A20 Pro chip.
iPhone 18 Pro Max: The device is rumored to gain a camera upgrade exclusive to the larger model, alongside a bigger battery.
Foldable iPhone: The company's first foldable is expected to have a ~5.5-inch display when closed, and a ~7.8-inch display when open, plus a super thin design with minimal crease and a durable hinge. It could be called "iPhone Ultra."
Apple Watch Series 12: Apple's upcoming smartwatch is expected feature fitness tracking improvements and a faster S12 chip.
Apple Watch Ultra 4: Apple's flagship rugged watch is expected to be updated alongside the Series 12, gaining the same S12 chip.
AirPods 5: The new earbuds were referenced in macOS 26.7 and are reportedly ready for release, suggesting a launch alongside the new iPhones.
Apple isn't going to release a standard iPhone 18 in September 2026. The lineup will instead focus on more expensive iPhones. The iPhone 18, iPhone 18e, and iPhone Air 2 will come out in spring 2027 instead. Apple does not plan to refresh the Apple Watch SE this year.
Apple has held its fall event in the first two weeks of September in each of the last several years, with new iPhones going on sale the following Friday.
Apple's "iPhone Ultra" was recently believed to have a delayed launch running into October, even though it's expected to be announced at the same time as the iPhone 18 Pro models. However, Bloomberg's Mark Gurman said this week he would be "very surprised" if the foldable did not come out at the same time as the Pro iPhones or "very close to it."
"Either on the same launch date, a dual launch, or coming out within weeks later. This idea of a late October, November launch, massive delays, supply chain issues, I don't believe to be the case," he said.
Note that the Mac Studio configuration with 512GB of unified memory is coming in late October rather than on September 22. Every other Mac Studio configuration ships in September. This article, "Apple Launching Eight New Products This Month" first appeared on MacRumors.com
As we enter the next era, what will be the defining measure of our progress?
Every industry has a word that shapes how it thinks. For pilots, it’s safety. For insurers, it’s risk.
For the semiconductor industry, it’s yield.
Yield does not ask how elegant the solution is, how many years it took or what the roadmap promised. Rather, it asks one simple question: What useful output did we produce?
For more than 60 years, the semiconductor industry has asked that question, relentlessly maximizing the number of usable chips produced from every wafer. Generation after generation, wafer after wafer, it is precisely that discipline that turned the transistor from a laboratory curiosity into the foundation of modern life.
Today, we need to apply the same principle to the unprecedented resources that the world is pouring into AI: capital on a scale once reserved for nations, gigawatts of power and record-breaking fabs and datacenters.
The question that will define this decade is the same one this industry has always asked: What actually comes out? Not just chips and tokens, but as affordable intelligence, as work that matters and as outcomes that improve lives.
That is the yield imperative.
YouTube Video
The limit of more
AI has proliferated with remarkable speed, at a rate of adoption faster than the internet, the PC or even the smartphone. Yet, global penetration still stands at just 18% of the working population, and the vast majority of that usage is chat-based. As systems move from answering individual prompts to reasoning, planning, using tools and executing longer agentic workflows, the infrastructure equation changes dramatically. A single agentic task can use more than 3,400 times as many tokens as a typical chat interaction.
Source: Microsoft AI Diffusion Report
We are only in the early innings of agentic adoption, and the infrastructure is already strained. Power is setting the limits on what we can build and when. Packages and racks are growing larger and denser. Memory is becoming an even tighter constraint.
For years, the industry’s rational answer to each new requirement resulted in more: more silicon in the package, more memory beside it, more power to feed it and more fiber to connect it. Each generation delivered meaningful progress. But when each new gain requires more input than the one before it, we are on a treadmill. It moves only as long as we keep adding to it.
I believe we need to pursue two paths forward. The first is evolutionary: we continue improving the architectures we have today, driving incremental efficiency, utilization and economics within each generation. The second is transformational: changing the curve itself with innovation in new architectures, new materials and new approaches to system and model design.
The history of our industry is defined by transformations like these. When increasing CPU clock speeds ran into the power wall, we moved to multicore processors. When planar NAND reached its limits, memory went vertical.
And now, once again, we have an opportunity to challenge our assumptions and rethink the fundamentals. Because the next chapter of AI won’t be defined simply by how much infrastructure we build, it will be defined by how much intelligence we can create from it.
Engineering useful yield
For decades, the computing industry has optimized yield in the context of manufacturing. Today, that discipline has to extend across layers, from datacenters and silicon through models and the agentic harnesses that orchestrate them. And the work does not stop once the technology is built. We must then deploy and scale it faster, while developing new tools and systems to maximize utilization across our fleet.
From development through execution, each layer has a yield of its own, and losses and gains compound across them. Capacity at any one layer is only a starting point. The real measure is how effectively those layers work together to produce useful output from the system as a whole.
Our experience at Microsoft building and operating AI infrastructure at scale has reinforced two key lessons. First, the biggest constraints are rarely solved in the layer where they appear. Second, when we attack a constraint across the whole stack, tradeoffs that seemed inherent to the problem often turn out to be artifacts of the architecture.
The greatest advances often come when we apply these learnings through co-design, working across layers to turn apparent limits into solvable system constraints. Innovations in memory, networking and power show what this approach looks like in practice.
Memory: More intelligence from every byte
Today, memory is viewed as a supply problem or a component problem. In reality, it is a system problem.
In AI inference, memory is now setting the limits on system performance. It must hold larger models, preserve longer contexts and deliver data fast enough to keep the compute fed. And agents raise the bar even further. Generation, retrieval, tool use and persistent memory run together in loops that can last minutes or hours. The result is a much longer memory horizon, with far more information kept close to the compute and available across an expanding sequence of turns. Doing that efficiently at scale will define the next generation of AI infrastructure.
Our experience building the Azure Maia platform demonstrates that memory bottlenecks are not resolved by a single layer.
Model architecture, data science and compression can reduce the amount of KV cache, which stores the model’s working context during generation. Software can manage memory hierarchies more effectively, silicon can be optimized for data movement efficiency and compilers can place data closer to compute. No one change removes the constraint. Together, they increase the useful intelligence the system can deliver from the same memory resources.
That is useful yield: not simply adding bytes but getting more useful intelligence from every byte we already have.
Networking: Designing across layers
As we zoom out to the cluster level, we see that intelligence does not come from one chip. It comes from thousands of chips operating as one system. Faster links matter, but the productivity of the system also depends on congestion management, failure recovery, workload placement, programming complexity and the boundaries between silicon, system and software. Together, they determine whether expensive compute is producing intelligence or sitting idle.
When architecting the platform for Maia, we did not begin with an existing networking design. We began with the outcome we wanted to deliver: efficient inference at fleet scale, designing across silicon, networking and system software. Instead of separate scale-up and scale-out fabrics, we built a two-tier scale-up network, integrated the NIC functionality directly into the chip and developed a custom transport layer.
The result is scalable, consistent performance across dense inference clusters, with a unified fabric that simplifies programming, improves workload flexibility and makes better use of available capacity. And with less network hardware needed to deliver this performance, we also lowered the cost of running the entire system.
Our objective is not merely to move data faster. It is to keep more compute productive and deliver more tokens from every watt and every dollar.
Power: Co-designing for efficiency, from grid to chip
Moving from the cluster to the grid, AI has introduced new challenges around power availability, distribution and utilization. Racks have gone from tens of kilowatts to hundreds of kilowatts, and datacenter campuses can operate on the scale of gigawatts. Power used to be something the system simply plugged into. Now, it is something we design around, from the grid to the chip.
That is why the industry is rethinking power across the system. Solid-state transformers and 800-volt direct current power delivery can reduce distribution losses as power moves through infrastructure. Power and cooling are no longer downstream of the design, they are part of the product definition from the start. And increasingly, that co-design is needed all the way into the silicon.
Azure Cobalt 200, our Arm-based server CPU, shows what this looks like in practice. We designed Cobalt so that every core has its own voltage and frequency controls, paired with software-based, per-virtual-machine power capping. This finer-grained control enables targeted power adjustments while protecting the performance of critical workloads, allowing us to run more servers within the same power envelope.
As Cobalt demonstrates, hardware-software co-design enables us to more effectively turn every megawatt into customer value.
The breakthroughs between boundaries
The pattern we see across memory, networking and power extends throughout the system: start with the useful output, then optimize the whole rather than any one layer. This first requires us to be precise about the output we are optimizing for and which design constraints are truly fixed. Are we optimizing for peak performance or sustained system throughput? Would the workload benefit from significantly more capacity with marginally less redundancy? What creates more value: a broader set of capabilities or significantly earlier customer deployment?
Not every constraint in today’s systems is a law of physics. Some are inherited from decisions made elsewhere in the system and can change only when we work across traditional boundaries. Evolution comes from the steady gains each company drives within its own domain, but transformation comes when we challenge those assumptions together and redesign the system as a whole.
The breakthroughs ahead will emerge from collaboration across the ecosystem, spanning hyperscalers and silicon providers, equipment makers and materials innovators, utilities and datacenter operators, model builders and software developers.
Toward full yield
But tokens and intelligence are not the finish line. What we produce becomes the input for someone else’s work.
What matters next is how broadly that input translates into productivity across the economy and value in people’s lives, whether it helps a scientist accelerate discovery, a clinician identify a signal earlier, a student get help at the right time or a small business find a new path to growth.
That happens as AI becomes part of everyday work across industries and around the world. People build on it, new uses emerge, the tools improve and the value compounds.
For that cycle to spread, AI must be broadly accessible. And at scale, accessibility depends on efficiency. It is the only way to deploy enough intelligence, and at a cost that allows it to reach everyone.
When every person and every company can access intelligence, build on it and create value of their own, that is full yield.
We will continue to build capacity because the world will need it. But our defining measure of progress must be what comes out: not only chips or tokens, but useful intelligence translated into empowerment, opportunity and human achievement.
That is the yield imperative. And it is work our entire industry must take on together.
Rani Borkar leads the core organizations responsible for planning, architecting, developing and deploying hardware and infrastructure for Microsoft’s leading cloud computing platform — from silicon, to systems, to supply chain.
Learn more about Microsoft’s silicon to systems approach to Azure infrastructure.
Log in
Apple is facing a £2 billion ($2.7 billion) lawsuit in London accusing it of imposing stricter App Tracking Transparency rules on third party developers than on its own advertising services (via 
Note: MacRumors is an affiliate partner with some of these vendors. When you click a link and make a purchase, we may receive a small payment, which helps us keep the site running.
Apple reportedly plans to release MacBook Pro models with OLED displays by early 2027, and Omdia analyst David Hsieh has 
Recently, Daring Fireball's John Gruber 
Note: MacRumors is an affiliate partner with Amazon. When you click a link and make a purchase, we may receive a small payment, which helps us keep the site running.
