×
iPadOS 26 finally made the iPad a true multitasking machine

iPadOS 26 finally made the iPad a true multitasking machine

I genuinely never expected to say this: it appears Apple finally figured out multitasking on the iPad. With iPadOS 26, the company hasn’t completely taken off the guardrails for how you can use and place apps on your tablet, but it came awfully close. With a couple of clever iPad-specific tricks, and better support for a mouse or trackpad, the whole system just makes sense in a way it never has before. I’ve been running the very first developer beta for less than a day on an 11-inch iPad Air, and I can already sense a change in how I use my tablet.

The new multitasking system is mostly pretty easy to understand. You unlock the iPad and you’re dropped onto the homescreen full of app icons, same as always. When you open an app, it opens full-screen by default; at WWDC, Apple executives were careful to note that if you don’t want to encounter the new multitasking system, you never have to, and I think that’s true. (You can even turn the whole windowing system off in settings, if you want to.) But as soon as you tap and drag the little icon in the bottom right corner of the window, the app starts to shrink. You can make the app any size you want — any size the developer supports, at least, which does vary pretty widely — and put it anywhere on the screen. If you grab the top bar of an app and sort of toss it off to one side, it’ll automatically fill half the screen. Tap the new stoplight menu, and you can either close the app or make it full screen again.

Once you’ve resized an app, that is now the size and placement of that app. If you close and re-open it, it’ll open in the same size and place until you move it or resize it again. I’ve come to think of this space on my iPad as kind of an “app screen,” next to my homescreen, organized and laid out however I like. And unlike Stage Manager, everything is always how I left it. You can put as many apps on this screen as you want – I’ve had a dozen on there at a time, which is way more than makes any logical sense and certainly more than you can see at once.

Exposé is the best way to see everything you have open, all at once.
Image: David Pierce / The Verge

At least in this early beta, there are a couple of quirks to the system. For one thing, I absolutely cannot figure out what the yellow “minimize” button does. You tap it, and the app closes, just like an app normally closes, and re-opens in place just like an app you’ve closed. I suspect this is a new app state developers will have to take advantage of, particularly as the iPad’s new background processing comes to fruition, but so far it doesn’t seem to do anything new.

You can only have one app active at a time, but you can scroll up and down in an app in the background — if you’re using a mouse or trackpad. If you’re touching the screen, you have to tap the app to bring it to the foreground before you can do anything with it at all. I’ve been using the CMD-Tab app switcher much more to quickly jump between things on the app screen, but it doesn’t always show every app I have open. The only way to get to everything is through the Exposé feature, which works well but is just a slightly slower way to move around.

By and large, though, it all just works. You can combine multitasking with Stage Manager, to turn your one organized app screen into a handful of them. You can run a bunch of windows on your iPad and a bunch more on an external monitor, or at least you can in theory — there’s a bug in the first beta that makes the whole system crash every time I open a second app on an external display. (I’m not super concerned about the functionality in the long run, but as always, install beta software at your own peril.)

A photo of a dog, with the iPad’s menu bar expanded over top.

The menu bar goes a long way toward making the iPad more Mac-ish.
Image: David Pierce / The Verge

There are a few other Mac-style things about iPadOS 26 that help make multitasking work. First is that stoplight menu, which shows up at the top left corner of every app and makes it easier to close or maximize windows. There’s also the new pointer, which appears when you connect a Magic Keyboard or other mouse or trackpad. I think I underestimated how much better it would feel to have a small, more exact arrow pointer rather than the old circle. And there’s the new menu bar, which appears at the top of the screen and offers all the settings and options for whatever app you’re currently using.

In just a few hours with iPadOS 26, I’ve noticed a couple of big changes to how I use my tablet. I’m spending much less time on the homescreen, which almost doesn’t need to exist anymore — I’m launching apps with Spotlight, resizing them on my app screen, and getting back to them with CMD-Tab or Exposé. I’ve been using the dock more, too, as a quick way to get to the apps I use most when they’re not already front and center. It’s all a lot like the way I use my Mac, and it’s amazing how quickly that muscle memory kicked in.

Several apps open, in multi-window view, on an iPad.

Can I interest you in a completely unusable number of apps on your screen at once? Yes!
Image: David Pierce / The Verge

Whenever I’ve tried to use multiple apps on the iPad, I’ve always felt like I was fighting the system. It would let you look at multiple things at a time, but it wanted you to see everything full screen. Now, and especially once developers update for even more fluid window sizes, multitasking feels like a first-class tablet citizen. It feels like a Mac in the sense that it feels like everything is happening in the same place, on the same screen, rather than constantly bouncing you between different full-screen experiences. Is it busier and occasionally more chaotic than the traditionally focused iPad vibe? You betcha! But I love it so far.

Ultimately, the answer to “can your iPad replace your laptop” still depends largely on your use case, and even more on the state of iPad apps. Until the iPad can run truly desktop-class browsers, I’m reluctant to say it can be your primary device. But now, maybe for the first time ever, the operating system itself doesn’t feel like it’s the problem. This is the most flexible version of Apple’s most flexible device. It can be a quiet, simple, single-tasking tablet or it can be a mess of windows put together just about any way you’d like. I’ve been waiting years for Apple to truly unleash the iPad, and I think it may have just done so.

Source link
#iPadOS #finally #iPad #true #multitasking #machine

In its decision, the NHTSA says it determined that Zoox robotaxis “have an equivalent or greater level of motor vehicle safety” compared to vehicles compliant with federal vehicle safety standards. Under the exemption, Zoox will be subject to “an enhanced oversight condition” that “may update and expand” as the company’s self-driving technology evolves. Zoox issued a software recall for its vehicles earlier this month over concerns that they may not detect smoke.

#Zoox #charge #rides #steeringwheelfree #robotaxisAmazon,Autonomous Cars,Electric Cars,News,Tech,Transportation">Zoox can now charge for rides in its steering-wheel-free robotaxisZoox just got permission to charge for robotaxi rides in its boxy, steering-wheel-less vehicles. On Thursday, the National Highway Traffic Safety Administration announced it has granted the Amazon-owned Zoox a temporary exemption, allowing it to deploy up to 2,500 vehicles annually over the next two years, as reported earlier by Reuters.In its decision, the NHTSA says it determined that Zoox robotaxis “have an equivalent or greater level of motor vehicle safety” compared to vehicles compliant with federal vehicle safety standards. Under the exemption, Zoox will be subject to “an enhanced oversight condition” that “may update and expand” as the company’s self-driving technology evolves. Zoox issued a software recall for its vehicles earlier this month over concerns that they may not detect smoke.#Zoox #charge #rides #steeringwheelfree #robotaxisAmazon,Autonomous Cars,Electric Cars,News,Tech,Transportation

announced it has granted the Amazon-owned Zoox a temporary exemption, allowing it to deploy up to 2,500 vehicles annually over the next two years, as reported earlier by Reuters.

In its decision, the NHTSA says it determined that Zoox robotaxis “have an equivalent or greater level of motor vehicle safety” compared to vehicles compliant with federal vehicle safety standards. Under the exemption, Zoox will be subject to “an enhanced oversight condition” that “may update and expand” as the company’s self-driving technology evolves. Zoox issued a software recall for its vehicles earlier this month over concerns that they may not detect smoke.

#Zoox #charge #rides #steeringwheelfree #robotaxisAmazon,Autonomous Cars,Electric Cars,News,Tech,Transportation">Zoox can now charge for rides in its steering-wheel-free robotaxis

Zoox just got permission to charge for robotaxi rides in its boxy, steering-wheel-less vehicles. On Thursday, the National Highway Traffic Safety Administration announced it has granted the Amazon-owned Zoox a temporary exemption, allowing it to deploy up to 2,500 vehicles annually over the next two years, as reported earlier by Reuters.

In its decision, the NHTSA says it determined that Zoox robotaxis “have an equivalent or greater level of motor vehicle safety” compared to vehicles compliant with federal vehicle safety standards. Under the exemption, Zoox will be subject to “an enhanced oversight condition” that “may update and expand” as the company’s self-driving technology evolves. Zoox issued a software recall for its vehicles earlier this month over concerns that they may not detect smoke.

#Zoox #charge #rides #steeringwheelfree #robotaxisAmazon,Autonomous Cars,Electric Cars,News,Tech,Transportation
Microsoft is in a unique position as AI overtakes the tech industry. It’s one of the world’s largest cloud providers and software-as-a-service companies, while also holding valuable stakes in the two biggest AI labs, OpenAI and Anthropic.

Those incentives are starting to clash as Microsoft posts blockbuster financial results. The company just reported an extremely profitable quarter with $90 billion in revenue and net income of $35.8 billion. For the fiscal year, which ended June 30, Microsoft reported $331.8 billion in revenue with a net income of $133.7 billion for the year.

And CEO Satya Nadella is not about to let the trajectory of Anthropic and OpenAI — which are expanding into applications and agentic infrastructure that could ultimately let them own customer relationships — derail that kind of cash.

Nadella has been preaching to enterprises to use multiple models and to stop relying on the frontier AI labs for the agentic harness/app layer.

Doing so is dangerous, he’s been saying, because it requires companies to share too many of their internal secrets with model makers of dubious trustworthiness. He knows his customers. Enterprise IT fears both data leaks and being locked into a vendor.

Now he has openly told Wall Street analysts during the company’s quarterly conference call Wednesday that this is an opportunity for Microsoft to sell customers its own homegrown models, alongside agents, AI security and more, while promising lower costs.

In other words, he’s pitching Microsoft as an alternative to many of the upscale services that OpenAI and Anthropic are developing for their own growth.

When UBS analyst Karl Keirstead specifically asked Nadella to weigh in on the open vs. closed-sourced debate roiling the AI industry, and how Microsoft will benefit from it, Nadella came out swinging.

“The goal is to have the firm be in control of their own destiny,” the CEO said of enterprises. “We are very, very clear about the architectural sort of design of the platform, which is you got to keep your harness separate from the model … that means any model at any given time is swappable.”

Microsoft, of course, sells a menu of harnesses (aka AI agents), too, under the Copilot name, including its coding agent GitHub Copilot. Coding agents are where much of the AI dollars are being spent today.

And he used the high-profile incident from last week as proof of his warnings.

“If you look even at the Hugging Face incident, the biggest thing that we should take away from that is you can’t sort of depend on any one model,” Nadella said. “You will maybe need multiple models to even remediate some challenges that get caused by one model. Like that’s the way to think about it, right? Which is you can’t be subject to a refusal of one model.”

The incident involved an unreleased model from OpenAI breaking out of its sandbox and successfully mounting a full-scale hack on Hugging Face, all in pursuit of besting a benchmark. Trying to understand what happened, Hugging Face at first tried to use a private frontier model (which it hasn’t named) that refused to help it. So it turned to the Chinese open-source model Z.ai GLM 5.2 to analyze logs and defend its infrastructure. The incident has so shocked the industry that even Sam Altman is now saying that maybe AI development should slow down a bit.

Nadella also made clear that Microsoft is happily selling its own homegrown models, the MAI family, on its own homegrown AI chips, Maya, and pitching them as cheaper alternatives.

“Every customer wants the right model for each task based on quality, latency, cost, and compliance. We offer the broadest model catalog in the cloud with over 11,000 models, including the leads from OpenAI, Anthropic, Mistral, xAI, as well as our own MAI family,” he said.

He added: “We’re also accelerating our own model development. We announced more than a dozen new models across image, voice, transcription, coding, security, including our first reasoning model, MAI thinking one, all with cost-efficient inference at the core for the enterprise use cases. We are co-designing these models with our silicon, and we are seeing 40% better performance per watt when running MAI models on Maya 200.”

As for Mythos? Nadella pointed to Microsoft’s new Mythos competitor announced earlier this week, MAI Cyber One Flash. It “achieves better performance than the much larger Mythos model, but at half the cost when combined with our multi-agent security harness,” he said.

Sure, the Microsoft CEO says that enterprises should use the frontier models that OpenAI and Anthropic offer in their mix. But his bigger message is: don’t trust them enough to rely on them.

When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.

#Microsoft #openly #competing #OpenAI #Anthropic #TechCrunchMicrosoft">Microsoft is openly competing with OpenAI, Anthropic more than ever | TechCrunch
Microsoft is in a unique position as AI overtakes the tech industry. It’s one of the world’s largest cloud providers and software-as-a-service companies, while also holding  valuable stakes in the two biggest AI labs, OpenAI and Anthropic.

Those incentives are starting to clash as Microsoft posts blockbuster financial results. The company just reported an extremely profitable quarter with  billion in revenue and net income of .8 billion. For the fiscal year, which ended June 30, Microsoft reported 1.8 billion in revenue with a net income of 3.7 billion for the year.







And CEO Satya Nadella is not about to let the trajectory of Anthropic and OpenAI — which are expanding into applications and agentic infrastructure that could ultimately let them own customer relationships — derail that kind of cash.

Nadella has been preaching to enterprises to use multiple models and to stop relying on the frontier AI labs for the agentic harness/app layer.

Doing so is dangerous, he’s been saying, because it requires companies to share too many of their internal secrets with model makers of dubious trustworthiness. He knows his customers. Enterprise IT fears both data leaks and being locked into a vendor.

Now he has openly told Wall Street analysts during the company’s quarterly conference call Wednesday that this is an opportunity for Microsoft to sell customers its own homegrown models, alongside agents, AI security and more, while promising lower costs.

In other words, he’s pitching Microsoft as an alternative to many of the upscale services that OpenAI and Anthropic are developing for their own growth.


When UBS analyst Karl Keirstead specifically asked Nadella to weigh in on the open vs. closed-sourced debate roiling the AI industry, and how Microsoft will benefit from it, Nadella came out swinging.

“The goal is to have the firm be in control of their own destiny,” the CEO said of enterprises. “We are very, very clear about the architectural sort of design of the platform, which is you got to keep your harness separate from the model … that means any model at any given time is swappable.”

Microsoft, of course, sells a menu of harnesses (aka AI agents), too, under the Copilot name, including its coding agent GitHub Copilot. Coding agents are where much of the AI dollars are being spent today. 







And he used the high-profile incident from last week as proof of his warnings.

“If you look even at the Hugging Face incident, the biggest thing that we should take away from that is you can’t sort of depend on any one model,” Nadella said. “You will maybe need multiple models to even remediate some challenges that get caused by one model. Like that’s the way to think about it, right? Which is you can’t be subject to a refusal of one model.”

The incident involved an unreleased model from OpenAI breaking out of its sandbox and successfully mounting a full-scale hack on Hugging Face, all in pursuit of besting a benchmark. Trying to understand what happened, Hugging Face at first tried to use a private frontier model (which it hasn’t named) that refused to help it. So it turned to the Chinese open-source model Z.ai GLM 5.2 to analyze logs and defend its infrastructure. The incident has so shocked the industry that even Sam Altman is now saying that maybe AI development should slow down a bit.

Nadella also made clear that Microsoft is happily selling its own homegrown models, the MAI family, on its own homegrown AI chips, Maya, and pitching them as cheaper alternatives.  

“Every customer wants the right model for each task based on quality, latency, cost, and compliance. We offer the broadest model catalog in the cloud with over 11,000 models, including the leads from OpenAI, Anthropic, Mistral, xAI, as well as our own MAI family,” he said.

He added: “We’re also accelerating our own model development. We announced more than a dozen new models across image, voice, transcription, coding, security, including our first reasoning model, MAI thinking one, all with cost-efficient inference at the core for the enterprise use cases. We are co-designing these models with our silicon, and we are seeing 40% better performance per watt when running MAI models on Maya 200.”

As for Mythos? Nadella pointed to Microsoft’s new Mythos competitor announced earlier this week, MAI Cyber One Flash. It “achieves better performance than the much larger Mythos model, but at half the cost when combined with our multi-agent security harness,” he said.

Sure, the Microsoft CEO says that enterprises should use the frontier models that OpenAI and Anthropic offer in their mix. But his bigger message is: don’t trust them enough to rely on them.
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.#Microsoft #openly #competing #OpenAI #Anthropic #TechCrunchMicrosoft

valuable stakes in the two biggest AI labs, OpenAI and Anthropic.

Those incentives are starting to clash as Microsoft posts blockbuster financial results. The company just reported an extremely profitable quarter with $90 billion in revenue and net income of $35.8 billion. For the fiscal year, which ended June 30, Microsoft reported $331.8 billion in revenue with a net income of $133.7 billion for the year.

And CEO Satya Nadella is not about to let the trajectory of Anthropic and OpenAI — which are expanding into applications and agentic infrastructure that could ultimately let them own customer relationships — derail that kind of cash.

Nadella has been preaching to enterprises to use multiple models and to stop relying on the frontier AI labs for the agentic harness/app layer.

Doing so is dangerous, he’s been saying, because it requires companies to share too many of their internal secrets with model makers of dubious trustworthiness. He knows his customers. Enterprise IT fears both data leaks and being locked into a vendor.

Now he has openly told Wall Street analysts during the company’s quarterly conference call Wednesday that this is an opportunity for Microsoft to sell customers its own homegrown models, alongside agents, AI security and more, while promising lower costs.

In other words, he’s pitching Microsoft as an alternative to many of the upscale services that OpenAI and Anthropic are developing for their own growth.

When UBS analyst Karl Keirstead specifically asked Nadella to weigh in on the open vs. closed-sourced debate roiling the AI industry, and how Microsoft will benefit from it, Nadella came out swinging.

“The goal is to have the firm be in control of their own destiny,” the CEO said of enterprises. “We are very, very clear about the architectural sort of design of the platform, which is you got to keep your harness separate from the model … that means any model at any given time is swappable.”

Microsoft, of course, sells a menu of harnesses (aka AI agents), too, under the Copilot name, including its coding agent GitHub Copilot. Coding agents are where much of the AI dollars are being spent today.

And he used the high-profile incident from last week as proof of his warnings.

“If you look even at the Hugging Face incident, the biggest thing that we should take away from that is you can’t sort of depend on any one model,” Nadella said. “You will maybe need multiple models to even remediate some challenges that get caused by one model. Like that’s the way to think about it, right? Which is you can’t be subject to a refusal of one model.”

The incident involved an unreleased model from OpenAI breaking out of its sandbox and successfully mounting a full-scale hack on Hugging Face, all in pursuit of besting a benchmark. Trying to understand what happened, Hugging Face at first tried to use a private frontier model (which it hasn’t named) that refused to help it. So it turned to the Chinese open-source model Z.ai GLM 5.2 to analyze logs and defend its infrastructure. The incident has so shocked the industry that even Sam Altman is now saying that maybe AI development should slow down a bit.

Nadella also made clear that Microsoft is happily selling its own homegrown models, the MAI family, on its own homegrown AI chips, Maya, and pitching them as cheaper alternatives.

“Every customer wants the right model for each task based on quality, latency, cost, and compliance. We offer the broadest model catalog in the cloud with over 11,000 models, including the leads from OpenAI, Anthropic, Mistral, xAI, as well as our own MAI family,” he said.

He added: “We’re also accelerating our own model development. We announced more than a dozen new models across image, voice, transcription, coding, security, including our first reasoning model, MAI thinking one, all with cost-efficient inference at the core for the enterprise use cases. We are co-designing these models with our silicon, and we are seeing 40% better performance per watt when running MAI models on Maya 200.”

As for Mythos? Nadella pointed to Microsoft’s new Mythos competitor announced earlier this week, MAI Cyber One Flash. It “achieves better performance than the much larger Mythos model, but at half the cost when combined with our multi-agent security harness,” he said.

Sure, the Microsoft CEO says that enterprises should use the frontier models that OpenAI and Anthropic offer in their mix. But his bigger message is: don’t trust them enough to rely on them.

When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.

#Microsoft #openly #competing #OpenAI #Anthropic #TechCrunchMicrosoft">Microsoft is openly competing with OpenAI, Anthropic more than ever | TechCrunch

Microsoft is in a unique position as AI overtakes the tech industry. It’s one of the world’s largest cloud providers and software-as-a-service companies, while also holding valuable stakes in the two biggest AI labs, OpenAI and Anthropic.

Those incentives are starting to clash as Microsoft posts blockbuster financial results. The company just reported an extremely profitable quarter with $90 billion in revenue and net income of $35.8 billion. For the fiscal year, which ended June 30, Microsoft reported $331.8 billion in revenue with a net income of $133.7 billion for the year.

And CEO Satya Nadella is not about to let the trajectory of Anthropic and OpenAI — which are expanding into applications and agentic infrastructure that could ultimately let them own customer relationships — derail that kind of cash.

Nadella has been preaching to enterprises to use multiple models and to stop relying on the frontier AI labs for the agentic harness/app layer.

Doing so is dangerous, he’s been saying, because it requires companies to share too many of their internal secrets with model makers of dubious trustworthiness. He knows his customers. Enterprise IT fears both data leaks and being locked into a vendor.

Now he has openly told Wall Street analysts during the company’s quarterly conference call Wednesday that this is an opportunity for Microsoft to sell customers its own homegrown models, alongside agents, AI security and more, while promising lower costs.

In other words, he’s pitching Microsoft as an alternative to many of the upscale services that OpenAI and Anthropic are developing for their own growth.

When UBS analyst Karl Keirstead specifically asked Nadella to weigh in on the open vs. closed-sourced debate roiling the AI industry, and how Microsoft will benefit from it, Nadella came out swinging.

“The goal is to have the firm be in control of their own destiny,” the CEO said of enterprises. “We are very, very clear about the architectural sort of design of the platform, which is you got to keep your harness separate from the model … that means any model at any given time is swappable.”

Microsoft, of course, sells a menu of harnesses (aka AI agents), too, under the Copilot name, including its coding agent GitHub Copilot. Coding agents are where much of the AI dollars are being spent today.

And he used the high-profile incident from last week as proof of his warnings.

“If you look even at the Hugging Face incident, the biggest thing that we should take away from that is you can’t sort of depend on any one model,” Nadella said. “You will maybe need multiple models to even remediate some challenges that get caused by one model. Like that’s the way to think about it, right? Which is you can’t be subject to a refusal of one model.”

The incident involved an unreleased model from OpenAI breaking out of its sandbox and successfully mounting a full-scale hack on Hugging Face, all in pursuit of besting a benchmark. Trying to understand what happened, Hugging Face at first tried to use a private frontier model (which it hasn’t named) that refused to help it. So it turned to the Chinese open-source model Z.ai GLM 5.2 to analyze logs and defend its infrastructure. The incident has so shocked the industry that even Sam Altman is now saying that maybe AI development should slow down a bit.

Nadella also made clear that Microsoft is happily selling its own homegrown models, the MAI family, on its own homegrown AI chips, Maya, and pitching them as cheaper alternatives.

“Every customer wants the right model for each task based on quality, latency, cost, and compliance. We offer the broadest model catalog in the cloud with over 11,000 models, including the leads from OpenAI, Anthropic, Mistral, xAI, as well as our own MAI family,” he said.

He added: “We’re also accelerating our own model development. We announced more than a dozen new models across image, voice, transcription, coding, security, including our first reasoning model, MAI thinking one, all with cost-efficient inference at the core for the enterprise use cases. We are co-designing these models with our silicon, and we are seeing 40% better performance per watt when running MAI models on Maya 200.”

As for Mythos? Nadella pointed to Microsoft’s new Mythos competitor announced earlier this week, MAI Cyber One Flash. It “achieves better performance than the much larger Mythos model, but at half the cost when combined with our multi-agent security harness,” he said.

Sure, the Microsoft CEO says that enterprises should use the frontier models that OpenAI and Anthropic offer in their mix. But his bigger message is: don’t trust them enough to rely on them.

When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.

#Microsoft #openly #competing #OpenAI #Anthropic #TechCrunchMicrosoft

Post Comment