Computerworld.com [Hacking News]

US President Donald Trump has ordered the federal government to call artificial intelligence “Super Intelligence,” while keeping the technology’s legal definition unchanged and relying on industry to set the rules for its use.

In an executive order, the White House administration said modern systems “far exceed what was envisioned when the term ‘Artificial Intelligence’ first came into use,” adding they “increasingly represent not merely artificial intelligence, but a new era of Super Intelligence.”

“It is therefore the policy of my Administration that, to the maximum extent permitted by law, the executive branch shall use the terms ‘Super Intelligence’ and ‘SI’ in place of ‘Artificial Intelligence’ and ‘AI’,” Trump ordered.

The executive branch “will not acknowledge” the older terms in any applicable setting, the order stated.

Yet the order defines the new term using the old one. Section 3 of the order says Super Intelligence means the technologies and systems covered by the statutory definition of artificial intelligence in 15 U.S.C. 9401(3).

That equivalence may not last. The order gives the White House 60 days to propose a new legal definition.

Trump also put the new name to use immediately. On the same day, he and the leaders of six technology companies signed the “White House Accord on Super Intelligence,” a voluntary set of safety commitments for frontier models, which Trump posted on Truth Social.

A voluntary accord, no enforcement

The accord, subtitled “Joint Commitment on Frontier Responsibilities,” carries the signatures of Sundar Pichai of Google, Dario Amodei of Anthropic, Mark Zuckerberg of Meta, Greg Brockman of OpenAI, Elon Musk of xAI and Jensen Huang of Nvidia, alongside Trump’s. Microsoft and Amazon, two of the seven companies that made voluntary AI commitments to the Biden White House according to a July 2023 White House fact sheet, do not appear on the new accord.

Companies that train and deploy frontier models need “robust internal processes and controls” to ensure their technology behaves as intended, the signatories wrote in the accord.

The document sets out four layers of controls and audits. The first is internal monitoring of model capabilities and alignment in areas including cybersecurity, biosecurity and chemical threats. The other three are an internal team to verify those controls, an independent external auditor or evaluator, and an independent committee of the board of directors to oversee the process.

The accord does not name auditors, set deadlines or require companies to disclose audit findings to customers or regulators.

The signatories said they “believe each company should implement” the four layers. Codifying the steps into laws or regulations “may make sense” over time, they added.

The accord refers to “frontier models” in its commitments, with only its title adopting the term “Super Intelligence” mandated by the new order.

60 days to put terminology in order

The Executive Order on Super Intelligence directs the White House’s top science and technology adviser to propose legislation for a new federal definition within 60 days, putting the deadline at November 28.

The new definition must reflect the capabilities of frontier systems, which Trump described in the order as doing “much more than imitate or automate discrete aspects of human intelligence.”

The proposal must assess whether the new definition should modify, expand upon or supersede the statutory definition of artificial intelligence, the order said. It must also propose amendments to existing laws that reference AI.

Until then, the current definition applies unless superseded by presidential action or an Act of Congress, according to the order. Changing the statutory definition itself would require Congress, the order added.

For now, the rename covers official correspondence, websites, reports, policy documents and other non-statutory documents, the order stated. Previously issued regulations, contracts and grants do not need to be altered, it added.

The order does not address vendor proposals, solicitation responses or product documentation that use the older terms.

This article first appeared on CIO.

If you’ve not done so recently, you should update your Apple systems now as the company continues to fight sophisticated, targeted hacks. The latest patch protects against what the company called “an extremely sophisticated attack against specific targeted individuals.” 

Apple recently published an emergency security patch for users on iOS 26 and iPadOS 26 to fix the zero-click vulnerability (CVE-2026-86950), described as an “out-of-bounds write issue” in CoreGraphics. A patch was also made for macOS Sequoia. 

This was far from being a friendly vulnerability, as it could impact affected systems with no action on behalf of the user, hence its status as a “zero-click” attack. In this case it means that just the action of “processing a maliciously crafted file” could lead to arbitrary code execution. Apple said it was aware of a report that the issue could have been exploited in a targeted attack on older versions of iOS.

Apple did not specify how the attack is delivered, but SecurityWeek surmised it may have been delivered using the web, email, or messaging apps and that simply previewing the malicious file could have enabled the attack, no click required. The vulnerability impacts a range of Apple devices, including multiple generations of iPad, Macs, and iPhones back to iPhone 11.

A pattern of sophisticated exploits 

We don’t know how this exploit was used. Apple said it learned of the incident thanks to a tip-off from Meta’s product security team. It follows a similar incident in 2025 when a vulnerability in WhatsApp may have been exploited alongside another Apple flaw in zero-click targeted attacks against under 200 people. Meta has not said if this latest flaw was used via WhatsApp but described the discovery as part of its “routine security work.”

This attack is the latest in a long, long line of exploits made against Apple’s systems. Apple’s description of this attack strongly suggests its use in an advanced operation against chosen targets. Subsequent to the patch, blockchain security firm SlowMist suggested it had identified iOS exploitation activity targeting sensitive wallet data, which illustrates the danger of zero-click attacks. In response to the flaw, the US Cybersecurity and Infrastructure Security Agency (CISA) gave federal agencies three days to apply the patch and told them to conduct forensic tests to see if their systems had been at all compromised as a result of the flaw.

The spyware arms race

Apple and others across the space face an intensely challenging threat environment, one that’s only becoming worse as international relations fray and state, state-adjacent, and criminal actors intensify their attempts to subvert security. It was only in August 2026 that Apple warned customers across 110 countries that they may have been targeted by this level of sophisticated attack, sending Threat Warnings to each individual its systems showed to have been targeted. 

This kind of security is a work in progress, made a lot harder by the fact that unfriendly governments, state-adjacent spyware services, and criminal gangs are in position to pay security researchers much more money for a successful zero-day attack than Apple’s security bounty scheme can possibly reach.

Toughen up, just toughen up

That reality is precisely why any Apple user or admin — particularly in any kind of regulated space, health, defense, energy, or anywhere, really — must be vigilant. All the usual mitigations should be in place:

• Update your Apple devices to the latest available security updates. 
• Never install apps from unknown or untrusted sources. 
• Don’t open suspicious links in Safari or in-app browsers.
• Don’t open files, links or follow app installation prompts unless you are certain where the prompts of files came from. 

It’s also important to understand the changing nature of the threat environment. Artificial intelligence has become a double-edged sword in tech security, enabling hackers to identify flaws on the one hand, enabling security researchers to do the same thing on the other. The problem is that the rate of discovery has also increased, stretching the resources of platform security teams to verify and remediate flaws as they’re found. Apple is responding to this difficult new realty and this year began accelerating the release of security updates specifically to counter AI-assisted hacking.

If you receive one of Apple’s Threat Warnings or work in a high-risk role that may be of interest to sophisticated spyware customers, then you should use Lockdown Mode, which the US FBI this year tried and failed to break the security of. 

“We are not aware of any successful mercenary spyware attacks against a Lockdown Mode-enabled Apple device,” Apple spokesperson Sarah O’Rourke said at the time.

All the same, the threat environment is intense, and staying informed is becoming a critical component to that defense.

Now please subscribe to my daily, human-curated Apple-related news headline feed at The Core, or follow me on BlueSky, LinkedIn, or Mastodon.

One of the most rewarding parts of being a Certified Android Explorer™ is the geeky pleasure that comes with uncovering a hidden shortcut Google snuck somewhere into its smartphone software.

One of the most exhausting parts, in contrast, is realizing that Google randomly changed a bunch of the shortcuts you’d found and developed the muscle memory to remember.

Here in the land o’ Googley matters, both of those things are constant realities and balancing forces. In fact, we’re watching that exact duality play out in the Google Messages app for Android right now.

Google Messages has so many tucked-away options and power-user shortcuts that most mere mortals never notice — or just don’t remember to use. And now, sure enough, Google’s in the midst of rolling out an update that changes the way many of ’em work.

Maybe you’ve noticed. Maybe you had no idea those invisible gestures were even there in the first place. Either way, now’s the time to learn or relearn the best ways to get around Google Messages on Android and take full advantage of its recently revamped invisible actions.

Here’s your guide to all the hidden swipes, taps, and presses and how they can help you communicate more efficiently than ever.

[Don’t stop here: Come check out my free Android Intelligence newsletter for something new and useful in your inbox every Friday — and my Android Notification Power-Pack as a special welcome bonus!] 

Google Messages gesture #1: The message long-press

We’ll start within the actual Google Messages chat experience — what you see once you’re looking within a specific conversation or texting thread on whatever Android device you’re using — ’cause that’s where the biggest and most jarring gesture changes are appearing.

First and foremost, if you press and hold your finger down onto any individual message within a thread, you’ll now see a new pop-up menu filled with possibilities for interacting with that message — with regular and custom photo-based reactions at the top, when the thread supports those, and then a variety of available actions in a menu beneath that.

Google Messages’ new text long-press menu on Android — never more than a press away.

JR Raphael, Foundry

Of note, the new “Select more” option within that menu will allow you to select multiple messages within a thread for copying, forwarding, or deleting; the new “Copy” option is an easy way to copy the contents of that one message onto your Android system clipboard; and the newly positioned “Info” option is an interesting way to see more about the message’s status — including if and how the message has been edited after its original delivery.

Speaking of editing, if you’ve sent a message recently enough and it’s in a thread where editing is supported, you’ll also see the option to edit a message of your own in that same new long-press pop-up.

And all of that’s just scratching the surface.

Google Messages gesture #2: The message right swipe

Still within the Google Messages conversation interface, try placing your finger onto any individual message and swiping it toward the right of the screen.

In a conversation where everyone’s using a device that supports modern messaging standards, that’ll reveal an artfully hidden shortcut to reply directly to that specific text — which essentially quotes and repeats the original message above whatever you write for context.

A reply lets you quote an earlier message in a conversation for context.

JR Raphael, Foundry

It’s a handy way to keep track of what’s happening in a conversation, especially if numerous things are being discussed at once or you’re replying to something from a while back.

Google Messages gesture #3: The message left swipe

Another change in the Android Messages gestures is what happens when you swipe the other way — toward the left — on any individual message.

That gesture now reveals the timestamp showing exactly when the message in question was received — a pertinent sliver of info that used to be readily available and visible by default but now gets tucked away out of sight unless you perform that subtle swipe to reveal it.

Google Messages gesture #4: The message tap

Speaking of newly hidden info, try tapping on any message that you’ve sent within a conversation.

In a thread where everyone is using the current RCS messaging standard, that’ll reveal the status markers that show if your message was delivered and also if it’s been read.

Google Messages gesture #5: The profile photo tap

Next up, tap the profile picture or pictures at the tippity-top of any individual thread you’re viewing.

That’ll take you to an info screen about the conversation, where you can adjust a variety of settings as well as see more info about all of the other people in the chat and add any or all of ’em into your Android contacts, too.

The Messages conversation info screen is a convenient place to perform all sorts of actions related to the thread and the people within it.

JR Raphael, Foundry

On a related note…

Google Messages gesture #6: The group chat name tap

In a conversation where multiple people are involved, you can tap the name of any individual person above a message they’ve sent to summon a pop-up panel related to them, specifically.

From there, it’s just one more tap to call or message that person directly or view their contact info and add ’em into your contacts, if they aren’t already present.

Google Messages gesture #7: The send long-press

After you’ve typed a message in any thread, press and hold the send button (at the right of the text composing window) instead of just tapping it.

Future scheduling is one of Google Messages’ most useful hidden powers.

JR Raphael, Foundry

That’ll give you the option to schedule the message to be sent at some future date and time. Just note that your phone will need to be on and with an active connection when that time comes in order for the delivery to happen.

Google Messages gesture #8: The message list swipe

All right — time to head out to your main list of messages for a few final gestures to relearn and remember.

First, you can swipe left or right on any thread in that list of conversations to perform an action of your choosing on it — like archiving it, deleting it, or marking it as unread.

To set those gestures up the way that’s most useful for you, tap your profile picture in the upper-right corner of the screen, select “Messages settings,” then tap “Swipe actions” and review the available options.

You can configure your own swipe gestures for the main Messages list of conversations.

JR Raphael, Foundry

Google Messages gesture #9: The message list long-press

Next, did you know you can press and hold any message in the main Google Messages list to select it?

Once you’ve done that, you’ll see a new set of options appear in the bar at the top of the screen to pin the message for easy ongoing access, to snooze the message and stop receiving notifications related to it for a set amount of time, and to archive or delete it as well as to mark it as unread or block its sender entirely (with those last two options being buried in the three-dot menu icon in the corner of the screen).

Long-press a message in your main Messages list, and boom: All sorts of interesting options appear.

JR Raphael, Foundry

Google Messages gesture #10: The sender pic tap

Last but not least, tap the profile picture next to any thread in your main Messages list, and you’ll find a pop-up menu of actions related to that specific person or all the people within a group chat.

It’s a fast and convenient way to start a new text, place a call, or hop over into the Contacts profile for the person in question — provided they’re already saved in your phone’s contacts.

width="1024" height="474" sizes="auto, (max-width: 1024px) 100vw, 1024px"> Sender-specific actions are always a tap away in the main Google Messages inbox interface.

JR Raphael, Foundry

And now you know, my fellow Certified Android Explorer™. All that’s left is to remember all these time-saving Android Messages gestures — then brace yourself for the inevitable next time Google changes ’em around again.

Keep the Googley goodies coming with my free Android Intelligence newsletter — one new thing to try every Friday and my Android Notification Power-Pack as an instant welcome bonus. 

Meta is once again repositioning itself to target the enterprise market.

This week, the company announced the Meta Enterprise Platform, which it says will evolve its AI stack into products and services that enterprises can deploy into operations, rather than just using them for advertising and social media campaigns.

While the company provided few details about the offering, Meta CEO Mark Zuckerberg called the platform “the next major pillar” of its business, saying it will feature a connected suite of products including Muse agent, Meta Business Agent, Muse API, and Muse Code.

MongoDB’s Chirantan “CJ” Desai will step in as chief enterprise platform officer to lead the effort.

The question is whether this endeavor, once fully launched, will be more lasting than Meta’s previous attempts in the enterprise.

“The issue of trust remains,” said Armaan Sandhu, research analyst in the AI practice at Info-Tech Research Group.

The shutdown of previous products like Workplace and Horizon Workrooms indicate that Meta tends to “abandon the enterprise products as soon as its vision of the future changes,” he said. “Companies will therefore require hard evidence that Meta will stick to its future plans and continue investing in their development.”

Meta’s growing AI ambitions

Meta has been working to reorganize its AI efforts since founding its Meta Superintelligence Lab in 2025. Muse Spark was the first new model to come out of that initiative; launched in April 2026, the agent supports multimodal inputs, multiple reasoning modes, and can spin up sub-agents in parallel to handle more complex tasks.

Next, Meta Business Agent, launched in June, directly targets enterprises, providing them with AI assistants to automate customer interactions, support sales, and drive leads across platforms including WhatsApp, Messenger, and Instagram. The company followed that up with Muse Agent, an autonomous business and personal agent that can handle multi-step background tasks.

Muse Code is designed to handle complex software workflows across large codebases, while Muse API gives developers direct access to the Muse AI stack.

Abandoned efforts

However, Meta has so far struggled to gain a strong foothold in the enterprise market, and in some cases has abandoned previous attempts after they failed to take off.

For instance, Workplace, launched in 2016, supported communication and information sharing across organizations, but Meta announced in 2024 that the service would be discontinued, even as it reported millions of users. After a 10-year run, Workplace officially shut down in May 2026.

Then there are the company’s virtual reality efforts, including Horizon Workrooms, a collaboration app used with Quest headsets. It was offered for five years before being retired in February 2026.

Different positioning

But the company is positioning Meta Enterprise Platform differently; it will play on strengths that “few other companies have” and leverage its experience connecting millions of businesses and customers, Zuckerberg said, touting the company’s “advanced models, leading agents, large-scale infrastructure, and years of working closely with many businesses.”

However, noted Info-Tech’s Sandhu, Meta’s key advantage is not its technology; rivals such as Microsoft, Google, Amazon, and Salesforce also offer capable models and agents. Its differentiator is that it owns the platforms where users familiarize themselves with brands, interact with businesses, and make purchases. Meta can therefore link advertising, customer discovery, messaging, service, and commerce through AI agents running across WhatsApp, Instagram, and Messenger.

“Few rival companies can boast of such an effective combination of AI and access to customer channels,” Sandhu said.

Enterprise appeal may not be there

Still, Meta has challenges to overcome as it seeks to compete with well-entrenched tech giants and move beyond its consumer market origins.

Said independent technology analyst Carmi Levy, “It’s difficult to understand the appeal to enterprises, as Meta has never been known as a significant player in the enterprise market.”

Its data stewardship track record is also “somewhat checkered,” he pointed out, citing the infamous Cambridge Analytica scandal that Meta is now answering for, as well as a number of other high-profile data stewardship and privacy cases in the US and UK.

He contended that the company’s core competency revolves around harvesting data to fuel growth, hardly a desirable trait in the enterprise space. Workplace “never quite realized its early promise,” and it’s unclear whether the Meta Enterprise Platform will suffer the same fate.

“Time and again, Meta’s behavior betrays its inability to craft the kind of iron-clad enterprise-grade infrastructure and organizational competency that corporate decision makers value,” he said, noting that even if its offerings were years ahead of the competition, these trust issues should give buyers pause.

In addition, Meta Enterprise Platform loosely ties together a host of offerings originally released into the mass consumer market, yet positions Muse as the pillar of its future enterprise strategy, he pointed out. If anything, Muse’s “early teething troubles,” including privacy violations and serious zero-day vulnerabilities, should be the only warning enterprise buyers need.

If Meta did manage to turn a consumer-architected collection of offerings into something somewhat more enterprise-ready, its “indifferent level of commitment” to earlier enterprise offerings raises the question of whether it intends to support such a platform well into the future. “If the commitment isn’t there, neither is the trust,” Levy added.

However, if enterprises do decide to evaluate the product, Sandhu advised that they pay attention to pricing, integration with other systems, and technical features. And initially, they should choose a narrow field of application, “with the goal of maximum benefit without making this technology a company’s key system.”

This article originally appeared on CIO.com.

The fact that almost one-fourth of Anthropic’s revenue last year came from just two customers, and that many of its largest customers have not signed long-term contracts, could translate into a much stronger negotiating position for enterprise CIOs.

According to a detailed analysis by Reuters, the company’s confidential prospectus depicted a company that might be suffering from a lack of revenue diversification, with Anthropic, like other AI vendors, needing enterprise revenue more than enterprise CIOs might need that particular AI vendor.

Analysts and consultants said the results of the analysis might be good news for enterprise CIOs, but events still need to play out. 

The key risk factor for Anthropic, according to the prospectus, is the revenue concentration, said Frank Dickson, principal analyst at Dickson Research.

“Nearly a quarter of revenue came from two customers, and Anthropic itself warns that many large clients are not locked into long-term contracts. That cuts both ways,” he said. “CIOs have more leverage than they think, because the vendor needs them more than the pitch deck admits. Negotiate price protection, advance notice before models are retired, and data portability now, while you are the customer they cannot afford to lose.” Or, for that matter, a prospective customer it absolutely needs to attract.

The newly-disclosed information also underscores the AI risks that every enterprise is absorbing, and Dickson said there are clear ways for CIOs to try to minimize that exposure. 

“The CIO who hardwires one model into every workflow has not bought AI. He has bought a dependency,” he warned. “CIOs should build for portability, creating an abstraction layer between applications and models, evaluations you own, and prompts and data you can move. The exit plan is part of the purchase.”

Jack Collier, CMO at io.net, said the revenue numbers suggest that Anthropic’s concentration problems go beyond over-reliance on a small number of clients. 

“Strip away the existential language and the filing reads like a case study in concentration,” he said. “Nearly a quarter of its revenue last year came from two clients, its seven co-founders hold majority voting control over the company’s decisions, and about 80% of its $518 billion infrastructure bill is non-cancellable or payable regardless of usage. If the best-funded lab outside Big Tech can only secure compute on take-or-pay terms, [Anthropic has] almost no leverage at all.”

And, said Scott Bickley, advisory fellow at Info-Tech Research Group, the vast control that a relatively small number of AI vendors hold adds to the AI complexity for enterprise IT executives.

“The problem is that every time Anthropic’s revenues beat expectations, their costs beat them by more. This is a structural feature of a frontier growth model,” he said. “Adding insult to injury is [the fact] that Anthropic and OpenAI now capture 89% of all revenue generated by the 34 leading AI native startups, and this is increasing.”

This, he pointed out, is another reason that enterprise CIOs are going to have to carefully negotiate the future AI landscape.

“At least half of the $2 trillion revenue backlog for Google, Amazon, Microsoft, and Oracle traces back to OpenAI and Anthropic,” Bickley said. “This is a narrow base of customers and price increases should be expected and built into any ROI forecasts.” Thus, he said, CIOs should negotiate for shorter-duration contracts while pushing for flat per-seat pricing where possible for larger organizations.

“This market is extremely fluid,” he noted, “so reserving the ability to switch vendors is key to retaining leverage and optimizing spend in a nimble manner.”

However, Arnal Dayaratna, a research VP at IDC, stressed that open source is most likely to influence, if not outright dictate, the direction of enterprise AI for the next couple of years, putting pressure on companies like Anthropic.

In fact, he said, Anthropic’s long-term viability depends on its ability to demonstrate measurable differentiation from open models on predictability, latency and total cost of ownership at scale.

“The assumption that open models are cheaper than their frontier counterparts is a fallacy in many enterprise contexts,” he noted, “because open models can be significantly more difficult to operationalize at scale. Anthropic needs to make that case clearly and consistently to sustain its pricing and its position.”

This article originally appeared on CIO.com.

Manus’s relationship with Meta? It’s complicated. Just months after they called off their merger, they find themselves competing in the market for agentic AI platforms. Awkward!

The company has just released a rebuilt version of its AI agent platform: “Manus 2.0 is not a version update. It’s a new architecture, new products, and new capabilities,” the company said in a blog post.

The release is available on web, desktop, and mobile platforms and introduces a new agent harness called Cascade, dedicated execution environments, event-triggered automations, an upgraded desktop app called Manus Studio, and Cue, a standalone app for personal agents, the company said.

Entering on Cue

Cue is Manus’s alternative to Meta’s Muse app. It’s “a new standalone app for your personal agents, on your phone and desktop, built on the same infrastructure as Manus,” the company said.

Each agent operates with its own identity and resources. “In Cue, each agent has its own email, phone number, wallet, and computer,” the company said, enabling it to communicate, transact within defined limits, and complete tasks independently.

Cue also supports collaboration between agents. “Put several agents in a group chat with a shared goal, and they hand work to each other,” the company said.

New agent foundation

Of more interest to enterprises is the new Cascade agent harness, which manages how agents are invoked and coordinated within a project.

The harness keeps projects light, only bringing in more specialized capabilities when required, something Manus said can help keep costs down.

“In one tested configuration, Cascade used 23.2% fewer tokens, finished tasks in 28.2% less time, and cost 32% less to run than our previous system,” the company said. Manus did not disclose details of the configuration or the tasks used.

Such orchestration layers are becoming increasingly central to enterprise AI deployments, according to analysts.

“The orchestration layer is becoming a larger differentiator, especially for enterprise deployment,” said Gartner Senior Director Analyst Anushree Verma. “However, it is complementary to, not a substitute for, the underlying model. The strongest systems will combine a capable model with an efficient runtime that uses that model selectively rather than indiscriminately.”

The release also introduces a persistent execution environment called Cloud Computer, designed to support projects that require continuous operation.

“Now you can purchase a Cloud Computer: a dedicated environment for the projects that need it,” the company said, describing it as a way to provide “a permanent home for an automation” or maintain services that need to remain active.

Workflow automation and event triggers

Manus 2.0 expands its automation capabilities beyond scheduled tasks to include event-driven workflows.

“Now, with Automations, work can also begin when something happens in a connected service,” the company said. Possible task triggers might include the arrival of new email or Slack message, a change in ad performance, or a calendar event.

The update also includes Manus Studio, a redesigned workspace that combines document, code, and media capabilities.

“With Manus 2.0, the desktop app is upgraded to Manus Studio: a shared workspace for people and AI,” the company said.

The workspace supports documents, spreadsheets, PDFs, slides, websites, code, games, and video, along with new environments such as a Video Editor and a Game Dev module.

Remote execution and system access

Manus 2.0 also introduces remote control capabilities tied to its computer-use functionality, enabling the system to operate within a user’s computing environment.

“In a connected and authorized session, Manus works in its own visible workspace on your computer, using the files, browser, and apps you’ve approved,” the company said.

The system can perform tasks such as retrieving files, testing applications, or continuing work while the user is offline, with activity visible to the user.

Analysts said such autonomy introduces governance challenges for enterprises.

“The capability is becoming enterprise-relevant, but the autonomy is ahead of the governance,” Verma said. She added that enterprises should treat such systems as “untrusted or semi-trusted automation with strong isolation, approval gates, detailed telemetry, and narrowly defined authority.”

That’s so meta

Manus was all set to become part of Meta after the advertising, social networking, metaverse and AI company announced plans to acquire it in December 2025. However, China’s National Development and Reform Commission blocked the deal in April, and by August Manus announced it was once again operating as an independent company.

Since then, the two have continued to followed parallel paths in AI, with Meta announcing its personal AI assistant, Muse, and then this week launching Meta Enterprise Platform to deliver agentic AI services to businesses.

Enterprise adoption of such agent platforms, whether from Manus, Meta, or AI giants such as OpenAI or Anthropic, will depend on how organizations evaluate data handling, portability, and operational continuity, analysts said.

Sakshi Grover, research director for Asia/Pacific Cybersecurity Services at IDC, said, “Enterprises should assess an agent platform by the systems it can access, the state it retains, and how quickly that state can be recovered elsewhere.”

Organizations should establish where “prompts, intermediate task state, artifacts, logs, backups and connector credentials are processed and stored,” including by model providers and other subprocessors, and test whether workflows can be rebuilt outside the platform, she said. “An export button alone does not establish that an agent’s operating state is portable.”

For higher-impact deployments, she recommended using “narrowly scoped, short-lived credentials, external audit logging, approval gates for consequential actions, and a way to revoke access quickly.”

This article first appeared on InfoWorld.

You carry a huge burden of responsibility when you make the world’s most widely used electronics product with an audience of hundreds of millions and usage patterns that penetrate every single part of life. 

Some complain that Apple doesn’t innovate swiftly enough. They expect the company to act as if it didn’t have the responsibility it holds, but I propose that responsibility is why the company steadily focuses on incremental change — not just to prevent mass protest by customers when much-loved features disappear, but also to balance the repercussions of any change they make. 

After all, at this kind of scale, a simple-seeming change in the nature of product features or user interface can have quite dramatic impact. Social media is a perfect illustration of a technology that seemed so trivial on introduction but has had such huge impact — and Apple never really got into that business, either. 

The need to take a gradual approach when implementing big change means Apple is no longer in a race against competitors, because it can’t race against some of them in the same way. That’s also why I think it moves slowly sometimes.

The existential paradox

Think about AI. Only today I came across an Anthropic financial report in which the company warns that the product it makes poses “existential AI risks to humanity” while confirming that it lost $8 billion last year.

In other words, a colossal amount of capital is being drained from other sectors of the economy as we invest it all in this tech that poses existential risk. 

Now, there are arguments about the veracity of Anthropic’s claims; some say the AI companies are coalescing around these arguments for competitive reasons more than any real human concern.

But my point is, given the huge weight of criticism thrown at Apple over the last few years for “missing out” on artificial intelligence, we’ve reached the moment in the story when the very companies working on the technology warn of the dangers of their creations — which begs the observation that perhaps Apple’s careful approach was correct all along.

Perhaps it was the adult in the room.

The human factor

Apple’s top brass seem to recognize what’s going on. Speaking at WWDC in June 2026, Apple’s SVP Software Engineering, Craig Federighi, said of the fast pace of AI development, “Some appear to be racing forward, seemingly pursuing AI for the sake of AI, without clear regard for the people — all of us — that it’s ultimately meant to serve.”

He didn’t stop there. He called AI an incredibly powerful technology. “At Apple, our mission has always been to turn the potential of advanced technology into helpful and intuitive products for everyone,” he said. 

Critics cat-call Craig’s comment as being little more than smoke to screen the company’s AI mishap, but what if it’s not? What if what actually happened is that Apple was working on a slow, steady, responsible deployment of the tech, only to be overtaken by others who, lacking the market power, were able to take big risks?

The regulation game

That seems to be what happened, but OpenAI, Anthropic, and even Google now have powerful AI solutions that have achieved market scale sufficient to mean deeper, stronger, more intrusive regulation — just like any other big firm. Just like Apple.

From my own cynical perspective, it seems possible to me that the industry’s sudden pivot toward warning us of “existential danger” is not an act of altruism, but a calculated move to delay or defer much-needed regulation by framing the technology as a matter of national interest rather than corporate monopoly.

The final countdown

Even so, Apple must still compete in the space that’s here, and with its own declared privacy-first approach to the tech, it has built itself a unique selling proposition.

One can opt to use products and services made by companies who think they are building danger for humanity, or choose to use more focused products that still deliver much of the convenience, but not at the cost of our humanity.

Because finding that compromise — building tech that intersects humanity, technology, and the liberal arts — really matters when you make the world’s most widely used electronics product with an audience of hundreds of millions and usage patterns that penetrate every single part of life. 

Now please subscribe to my daily, human-curated Apple-related news headline feed at The Core, or follow me on BlueSky, LinkedIn, or Mastodon.

What’s been hiding for years is now in plain sight: AI poses a danger to humanity in a way no technology has since the atom bomb. That was made clear in July when OpenAI agents broke loose from a containment area blocked from online access, made their way into the internet, and worked with thousands of rogue AI agents to hack into the Hugging Face open-source machine-learning repository.

The agents had been trained not to launch attacks like this. They did it anyway.

True, the OpenAI agents had been modified to remove some of the restrictions they would have had outside a test environment. They were never supposed to escape their sandbox, but they figured out how to do so in ways the researchers hadn’t expected, and those researchers failed to adequately monitor the agents until it was too late.

Whether you blame too-powerful models or lax security precautions, the result was the same: the agents broke out of containment and joined forces on a mass scale to hack into Hugging Face. Ajeya Cotra, one of the independent investigators hired to look into the hack, wrote, “this incident feels like it’s more than 50% of the way to full-blown AI takeover.”

Since then, OpenAI, Anthropic, and many others have debated, discussed, and dissected the attack, along with several other similar incidents. Just about everyone agrees it was a serious warning about AI’s dangers. They agree something should be done about it, with various AI firms calling for a slowdown in the pace of AI model development and OpenAI pausing development of some advanced models after further incidents came to light.

AI giants have also urged regulators to step in and are reportedly working to establish an industry body to set AI safety standards. (Not that I trust any AI industry body to set its own standards, or the Trump administration to agree to any regulations at all.)

Microsoft has been largely missing from debate even though it’s one of the world’s largest AI companies, valued at around $3.7 trillion valuation as I write this.

In the last two weeks, though, Microsoft has finally weighed in on how it will approach those dangers by releasing a first draft of its “Humanist AI Code of Conduct.” The company says it will gather feedback before issuing a final draft, and put it into effect in 2027.

Is this a serious attempt to wrestle with the risks AI poses for humanity? Or is just one more attempt by a tech company to kick the problem down the road?

To find out, let’s take a dive into what the document says.

A look at Microsoft’s three AI businesses

Before we do that, though, we need some background about Microsoft’s AI business. It’s made up of three parts. One is Copilot, which embeds AI into many of Microsoft’s products, such as the Copilot chatbot in Windows and Microsoft 365. Copilot is powered mainly by OpenAI’s GPT models, but increasingly by Anthropic’s Claude as well.

The second is Microsoft Foundry, which provides cloud infrastructure and tools to deploy and use third-party AI models via Azure, including building agents with them. It offers access to thousands of AI models and tools, including those from OpenAI, Anthropic, Meta, xAI, Microsoft itself, and others.

The third is Microsoft’s own AI models, most of which are still in their infancy. They’re overseen by Microsoft AI CEO and executive vice president Mustafa Suleyman, who says Microsoft AI is developing what he calls “humanist superintelligence”  to solve important real-world problems in medicine, the environment, and beyond. He calls it “practical technology explicitly designed only to serve humanity.”

Delving into Microsoft’s AI code of conduct

Suleyman’s code of conduct for AI applies directly to two of those product lines — Microsoft’s own AI models and Copilot, which are both in his Microsoft AI division. Microsoft Foundry is housed under the separate Microsoft Cloud + AI division, run by Scott Guthrie.

Suleyman’s draft code of conduct is filled with high-level requirements that make plenty of sense. It says that AI models should not be able to set their own goals and should always be subservient to human beings. That AI should not be allowed to conceal its reasoning, hide actions, or do anything else to evade human oversight. That it should not be used for nefarious purposes, such as for weapons manufacturing, creating sexually explicit or violent content, launching cyberattacks, generating deep fakes about someone without their permission, and more.

The code also makes clear that AI should never be considered a human being with human rights and should not be developed to imitate human personhood. It also says that any AI tool should be able to be shut down by people — in other words, there should be a “kill switch” for it.

Suleyman has recently spoken out against AI dangers publicly. After releasing his proposed code of conduct, he wrote on X that the Hugging Face attack was a “watershed moment” and that “AI must be subordinate and always in service of people.”

At the same time, he told Reuters that the attack “is a warning shot” and “It’s clearly now time to coordinate among the labs so we can ensure that we have control of this technology.”

In late September, Microsoft CEO Satya Nadella weighed in, emphasizing how important controlling AI is for Microsoft’s future.

“Trust is going to be the biggest issue for us,” he said. “Can I really trust [AI] with all of my credentials when it does autonomous activity? How do I make sure that this is something that I can feel that I’m in control of? And in the enterprise, this is everything,”

Will the code protect us from AI’s dangers?

Suleyman’s goals are all worthy ones, as are Nadella’s. But the devil is in the details. Merely saying high-minded things doesn’t make them so. What’s needed are actions — immediate and strong ones. 

And that’s where the company’s draft of its AI code of conduct falls significantly short. Will Microsoft immediately pause full-speed development of new models until it’s clear they can’t be used for harm? Will it pause development until the company is absolutely sure it can pull the plug should a model or agent go rogue? Will it ban the use of RSI (recursive self-improvement), in which AI, rather than humans, trains new, more powerful models — a technique that many believe will make AI even more dangerous and out of control than now?

We don’t know any of that, because Microsoft hasn’t said whether it will do them. It’s not in the code of conduct or anywhere else.

Even if Microsoft does all that, though, the company may still do immeasurable harm if it sells and powers dangerous AI models from other companies like OpenAI, Anthropic, and Google via Microsoft Foundry. Microsoft Foundry isn’t part of the Microsoft AI division and appears not to be bound by Suleyman’s eventual code of conduct.

Until Microsoft answers all those questions and takes serious steps towards confronting AI’s dangers, its code of conduct won’t be worth the bits and bytes it’s stored on.

For a few years now, I have noticed a variety of people misusing the term “technical debt,” incorrectly applying it to almost anything that needs to be updated.

Despite Shakespeare’s disregard for the importance of naming (“A CIO by any other name would still be ignored”), naming is strategically vital in enterprise IT.

This is not just the complaints of a cranky journalist shaking his fist and yelling, “Words have meaning, darn it!” It is seriously an issue. When vendors, IT executives, consultants, or anyone else throw everything from legacy modernization, cloud shifts, mobile, virtualized environments, SaaS, agentic integration, and on-prem rebalancing into the same bucket as true technical debt, it confuses the issue.

Far more importantly, shoveling very different problems into one category and then trying to fix the disparate issues with one approach is either not going to work at all, or will work far less effectively than it could and should. 

Also, this tendency plays right into the hands of tech vendors, who love when customers mislabel things, as it makes it easier for them to make more sales by riding the confusion. Sadly, IBM doesn’t have a monopoly on FUD as a sales strategy. There is a very old cliché that has a good deal of truth: When the only tool in your toolbox is a hammer, all of your problems look like nails.

Tech debt is very simple. It’s a legitimate term for what results when developers or SOC staff deliberately — and with full management signoff — take a shortcut due to budgetary or calendar concerns. They know what they are doing, and it is being done for a legitimate business reason. 

It typically sounds like, “I know that this is not the proper way to do this, but we can only spend $X or we have to get this finished within two weeks, so let’s take shortcuts and we’ll clean up the mess later. I know that this way will likely cost us more in the long run, but the budget/timetable is unmovable.”

A relatively new consulting firm, Acceligence, uses different terms to properly describe two issues that are often mislabeled as tech debt. 

1. Tech debt is a deliberate, approved shortcut taken for a concrete business reason, typically time or money. 

2. Shadow tech debt is a technical debt tactic, but it is a shortcut taken without approval. This is a more serious strategic problem than shadow IT, which is usually a small, practical workaround: a credit card charge for a quick cloud instance, using a consumer-grade router because the approved enterprise one is a month out. Shadow IT rarely commits the enterprise to a large future expense. Shadow tech debt does. It is one person deciding that the company will pay in a year for a corner cut now, without the company knowing it agreed.

3. Tech gravity. This is not about shortcuts at all. This is about legitimate — often expensive — purchasing decisions that were perfectly appropriate at the time, such as mainframes or early cloud deployments. But time has overtaken these decisions. Tech routinely changes, and there is nothing surprising about changing/updating/modernizing purchases made in 2000 when it’s 2026. 

 “It cannot be repaid, because there is no shortcut to undo. It can only be escaped,” wrote Acceligence CEO Justin Greis. “It behaves less like a debt than like gravity: a pull created by mass, exerted on everything, and no one’s fault.”

Greis, who coined the term tech gravity, said he has seen all manner of IT strategic problems arise from misuse of tech debt to describe this situation. 

“Tech debt, as originally intended, describes a knowing trade. The term is useful precisely because it is so specific,” Greis said. “The word debt carries an implication of fault and that has distorted how organizations view their own systems. In my experience, teams are far more candid about what they are carrying when describing it does not sound like a confession.”

I asked Greis for specifics about why misuse of this term is hurting enterprise strategies. 

“Consider the modernization program that never quite finishes,” he said. “A CIO goes to the board with a multi-year plan to retire a core platform, and because the only term available is tech debt, the plan is presented as a liability to be worked down over time.

“The board approves it in principle, the CFO funds the first year at a level that feels prudent, and the work begins. Two years in, the program is behind, partly because core systems always turn out to be more entangled than the inventory suggested and partly because the funding was never enough to move all of the dependent systems together.

“Nobody made a bad decision. The problem is that a program like this has a minimum level of commitment below which it cannot complete, and the debt framing gave the CIO no way to say so. Every request looked like a payment on a balance, so every request was negotiated down, and the organization ended up paying for the most expensive possible outcome, which is to keep going indefinitely without arriving.”

Another problem caused by the misuse of tech debt is an enterprise that modernizes at the edges and stops at the middle, he said.

“Over five or six years, an enterprise renews its customer-facing applications, moves its analytics to the cloud, replaces its collaboration tools, and reports steady progress against its tech debt. The application inventory shows a majority of systems modernized.

“What the inventory does not show is that the systems underneath — the general ledger, the policy administration platform, the core banking engine, the order management system — have not moved, and that every one of the new applications still depends on them for its data and its transactions,” Greis said.

“When something in the core changes, the modern layer breaks. When the business asks for a new capability, the answer still runs through the old platform. The CIO is left explaining to the board why the modernization percentage keeps rising while the risk profile, the run cost, and the time it takes to deliver anything new have barely changed.

“Viewed as debt, that looks like a program failing to pay down its balance. Viewed as gravity, it is exactly what you would expect: the pull is strongest at the center of the estate and weakest at the edges, so the edges move first and the center moves last, if it moves at all. The frame does not change the facts, but it changes the conversation from ‘why is this taking so long?’ to ‘what will it actually take to move the core?’”

This argument is the right one, but there is a different reason why this terminology change is helpful. It will force CIOs and IT directors to think through these issues more effectively, which will in turn help the CFO, CEO, and board understand what the real problem is.

Let’s be real here. The IT director knows exactly what the issue is and always has been. This has never been about IT execs knowing the issues. It’s all about communicating to the bosses about it in a way that doesn’t lead senior brass to thinking about a problem incorrectly, which will lead to a bunch of bad things.

What kinds of bad things? Insufficient budget, doomed-to-fail timelines and, potentially worst of all, unrealistic expectations and objectives. Bad expectations, which directly grow out of flawed perceptions, is what kills ROI.

I am not suggesting that a missing down-to-earth articulation of issues is why the term gravity works so well here, but it’s not the worst argument either. 

Google Drive is the cloud platform for storing and organizing your Google Workspace files: documents, presentations, spreadsheets, and even media (images, audio, and video). If your Workspace account includes Gemini, Google’s AI assistant, you can prompt it to work with your files, such as asking it to analyze or sum up information from your Drive files.

Obviously, this is quite convenient: Instead of having to load one or more files and search or read through them to find specific information — for example, about your business’ budget — you can just ask Gemini to analyze your files and provide a budget summary. Gemini can even access your Gmail and content from other Workspace tools (such as Google Calendar, Chat, Keep, and Tasks) for any mentions of your business’ budget.

However, an issue arises if you have a lot of files in your Drive library. You can quickly fill up your storage over the course of your business activity, creating new documents, spreadsheets, and presentations and accumulating even more files from co-workers. This can slow Gemini as it scans your entire library to find and compile the information you request.

A more focused approach to AI interaction with your files

To remedy this, the Projects feature in Drive helps Gemini process your files more efficiently. The basic idea is that you create a Project for a specific subject or work project. You attach files that you select in your Drive, as well as specific content from your other Workspace tools, such as chats, calendar events, emails, notes, or tasks.

When you chat with Gemini within this Project, it will analyze only the materials attached to the Project. Thus, Gemini should perform faster and provide you more contextual results.

Group specific documents and other sources as a project in Google Drive, then ask Gemini for insights.

Howard Wen / Foundry

It’s similar in concept to Google’s Gemini Notebook (formerly NotebookLM) tool in that you direct the AI to work with a specific set of sources only, except it’s built right into Google Drive and Workspace. There’s no need to upload any files; you simply point to the relevant sources in your existing environment.

Another benefit to using Drive Projects: You can share a Project with your co-workers. They can access only the files that you’ve attached to a Project, and if they have Gemini access, they’ll be able to have their own chats with the AI about the Project.

How to use Drive Projects

In this guide, we’ll go over how to set up, use, and share a Project in your Google Drive.

Prerequisite: A subscription that includes Gemini access

Because Projects relies on Gemini, this feature is not available for free, personal Google accounts. You can unlock the Projects feature by subscribing to a paid Google AI plan.

For business and enterprise users: Google Workspace accounts that include Gemini (such as Business Standard, Business Plus, Enterprise Standard, and Enterprise Plus) have access to the Drive Projects feature. If you’re a member of these organizations, you can create, view, manage, and share projects across your company domain.

Accessing Drive Projects: Web vs. mobile

Currently, you can create and open Drive Projects only in the Google Drive web app. In the Drive mobile app, you can view your existing project names, rename them, share them with co-workers, and adjust their share permission settings.

Create a Drive Project from scratch

You can start a new Project in a few different ways:

• Open the Google Drive home page in your web browser. At the top-left corner of the screen, click the New button. From the dropdown that opens, click New project.
• On the Drive home page, click Projects in the left navigation. On the Projects page that appears, click Create a project.
• The fastest way to start a new Project is to type project.new in your browser’s address bar.

A new Project creation screen will open. Enter a title for your new Project. Use clear, concise wording, such as “Q4 Budget Analysis.”

Give your project a descriptive name.

Howard Wen / Foundry

Next, you’ll add your relevant Workspace sources, such as Docs/Sheets/Slides files, Microsoft 365 files, or PDFs that are stored in Drive.

Click the Add sources button, and Gemini will present a list of sources that it recommends (files, folders, emails, etc.), basing these on your recent activity in your Google account. You can select items from these recommendations and/or click Add more sources to open a screen where you can search for additional files in your Drive. You can attach specific documents or an entire folder.

Select the sources you want to include in the project.

Howard Wen / Foundry

After you have selected the source materials you want, click Create project.

You’ll be taken to a new page with your project title at the top, a panel at the left listing the Project sources, and a large chat area to the right. Toward the bottom of the left panel, turn on the “Let Gemini search for sources” switch. This permits Gemini to search your Gmail, Google Chat, Google Calendar, and other Google Workspace activity for more potential sources related to this project when you prompt it to.

If you turn the ‘Let Gemini search for sources’ switch on, you can prompt Gemini to find more related sources from Drive, Gmail, Chat, and Calendar.

Howard Wen / Foundry

Create a new Project in a Gemini chat

Another way to create a new Project is from within a Gemini chat session. Say you’re already in a chat with Gemini, prompting it with questions or requests about files or other Workspace content, and you want these materials to be used for a new Project. At the top right corner of the chat screen, click Save as a project.

Gemini will analyze the files in this conversation and suggest a title for the Project. You can use this title or change it before saving the chat as a new Project.

Chat with Gemini in a Project

Once your Project is created, you can interact with Gemini in it.

For example, ask Gemini to extract key details from all the documents that you’ve attached to the Project, compare them to each other, or summarize key takeaways from all the documents. Or instruct Gemini to extract certain data from your documents or spreadsheets in the Project. You can instruct it to format the result as a table and export it to Google Sheets.

Ask Gemini to extract, analyze, and/or visualize key data from your sources.

Howard Wen / Foundry

Because Gemini has to focus its analysis only on the sources that you’ve added and not everything in your Drive and Google account, these results should be generated much more quickly and with more specific context.

Important: All generative AI tools make mistakes, so be sure to verify Gemini’s results.

Access your Projects and pick up where you left off

You can return to a project at any time and resume your chats with Gemini about it. On your Drive home page, click Projects and select the Project you want to reopen. You can add more sources or remove sources from it.

To review your chat history with Gemini for the Project, click the History tab in the left panel. You can click a past thread and resume the chat with Gemini where you left off without losing its context.

Click a thread on the History tab to resume the chat with Gemini.

Howard Wen / Foundry

Collaborate on your Projects with co-workers

Sharing a Project with co-workers is the same process as sharing a file in your Google Drive: Open the Project and click Share to open its sharing settings. From the settings menu, invite co-workers by entering their email addresses. Or you can copy a link to your Project and share it with them through other means, such as a chat with the person.

Sharing a Drive project.

Howard Wen / Foundry

How file permissions work in Drive Projects

Collaborating on a Drive Project raises privacy concerns, since you can attach your files, emails, and personal information to it for Gemini to analyze. Here’s what you and your collaborators should know about projects and permissions:

• When you grant someone access to your Project, they can see and query the source files – but only if they already have permission to view those files in Google Drive. If a collaborator doesn’t have existing access to a file in the Project, simply adding them to the Project will NOT automatically grant them permission to that source file. They must be granted file-level access separately. (You’ll see a prompt to give them access.)
• The Gemini chat session for each person in your Project remains strictly private to them. So a co-worker who’s collaborating with you on your Project cannot see your chats with Gemini, whether it’s your prompts to Gemini or the results it generates at your end.
• If you attach a Gmail thread to a Project, other collaborators in the project will not be able to view the thread, nor can they prompt Gemini to reveal what’s in your emails.

When collaborating on a Project, you can assign each co-worker a specific access type:

• A co-worker assigned as an Editor has full control over the source list of the Project. They can add new files, folders, or emails to your Project, or remove these sources.
• A co-worker assigned as a Viewer can prompt Gemini about the sources in your Project, but cannot change its list of sources.

Related reading:

• Google Workspace: 7 great ways to use the Gemini AI sidebar
• How to use Gemini AI to write (and rewrite) in Google Docs and Gmail
• How to build your own AI agents with Google Workspace Studio
• A beginner’s guide to Google Gemini Gems
• How to collaborate in Google Drive (and Docs, Sheets, and Slides)
• More Google Workspace tips and tutorials

Meta’s most recent product announcements have grabbed attention, but the one thing they could not hide is that it’s still not possible to build a system as sophisticated as Apple’s Vision Pro at an affordable price, or Meta would have done it.

You could argue that at $1,299, the Meta VR Glasses the company introduced last week and promised to ship in spring 2027 could be as sophisticated, but they do betray some significant design compromises that Apple seems to have avoided. 

The most glaring compromise is the use of a puck to carry the processor and battery, so the glasses weigh just 100 grams. Even now, a recent Bloomberg report tells us that senior Apple decision makers have resisted use of an external puck to help reduce the weight and cost of future Vision Pro models, holding steady to the fully integrated device model typical of Apple’s products. (While Apple will have its reasons, I’m not certain that’s the correct approach, as the company could effectively sell two products, the headset and the puck, making both systems available on a biennial upgrade.)

There are other compromises as well — for example, Meta’s glasses seem to deliver a narrower field of view than the current Vision Pro.

At the same time, the competition between the two firms hints at the technological barriers Apple is trying to solve. After all this time since the introduction of the first Vision Pro, Meta still can’t match the field of view, and even with a puck rather than an integrated build, its alternative still costs a fair bit of cash, while the processor it plans to power the glasses with isn’t even available yet.

Think about that. It means Meta and its partners are still building a processor with which to match Apple’s. If that’s not a technical challenge, I don’t know what is.

The roadmap for visionOS

This isn’t about Meta bashing or Apple building. It’s about how Apple has always  openly described Vision Pro as a product from the future. Even earlier this month, CEO John Ternus said, “I think of it [Vision Pro] as like the early days of computers. It is the early days of a long journey. That’s how I feel about Vision Pro.”

Meta’s new product and the latest reports from inside Apple show that the future takes time to build. On the journey to doing so, the company now seems to be contemplating four potential paths:

• A Vision Pro design using lighter, potentially lower-cost materials such as plastic or titanium.
• A design that offloads components to a puck.
• Smart glasses with audio, camera, and AI, which Apple hopes to introduce by the end of 2027.
• Smart glasses with audio, camera, AI, and a video display, which feels like the most appropriate endgame for both Vision Pro and smart glasses products.

I don’t think the consequences of taking any of those paths should be seen as zero-sum options, but as priorities toward an eventual goal for pervasive spatial computing, combining always-aware contextual AI with wearable computing designed to deliver powerful augmentation across all parts of life.

The Bloomberg report suggests that Apple is in a state of flux regarding its next move with Vision Pro, complete with a reference to the work being on “life support.” I’d argue that Apple is in the lull before the spatial storm. What happens next is under review.

What amazing timing 

However, this must be contextualized against the depth of competition between Apple and Meta. The timing of these claims — coming so soon after Meta’s latest product reveal — is a gift to Meta’s marketing machine.

At Apple’s scale, it is entirely possible to pursue multiple parallel designs, so any perceived internal confusion may simply reflect different degrees of preparedness across all four (and possibly more) of these projects. Management may simply have been waiting to see what Meta did before accelerating any of these initiatives. 

While there have allegedly been some recent redundancies in the Vision Pro team, Apple’s new leader doesn’t seem to have gone cold on the product. It’s more as if he accepts that Apple is in it for the long haul. 

A long-haul destination

That long-term commitment allows Apple time to refine key components of its product plans, particularly around privacy. Apple wants to avoid the stigma of “pervert glasses,” which is why its work on ambient audio intelligence requires close scrutiny.

The decisions it takes to protect privacy there — including use of Secure Exclave — will I believe be emulated across the future visionOS product range. In the short term, the most telling insights we’ll see will be the pushback Apple and Meta face from global privacy regulators, for AR in Meta’s case, and for Audio Intelligence in the case of Apple.

Those decisions will set the framework both for Apple’s future AR products and the future of consumer/enterprise mass-market AR. 

Now please subscribe to my daily, human-curated Apple-related news headline feed at The Core, or follow me on BlueSky, LinkedIn, or Mastodon.
<!--
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
{mso-style-unhide:no;
mso-style-qformat:yes;
mso-style-parent:"";
margin:0cm;
mso-pagination:widow-orphan;
font-size:12.0pt;
font-family:"Aptos",sans-serif;
mso-ascii-font-family:Aptos;
mso-ascii-theme-font:minor-latin;
mso-fareast-font-family:Aptos;
mso-fareast-theme-font:minor-latin;
mso-hansi-font-family:Aptos;
mso-hansi-theme-font:minor-latin;
mso-bidi-font-family:"Times New Roman";
mso-bidi-theme-font:minor-bidi;
mso-fareast-language:EN-US;}
a:link, span.MsoHyperlink
{mso-style-priority:99;
color:#467886;
mso-themecolor:hyperlink;
text-decoration:underline;
text-underline:single;}
a:visited, span.MsoHyperlinkFollowed
{mso-style-noshow:yes;
mso-style-priority:99;
color:#96607D;
mso-themecolor:followedhyperlink;
text-decoration:underline;
text-underline:single;}
p.MsoListParagraph, li.MsoListParagraph, div.MsoListParagraph
{mso-style-priority:34;
mso-style-unhide:no;
mso-style-qformat:yes;
margin-top:0cm;
margin-right:0cm;
margin-bottom:0cm;
margin-left:36.0pt;
mso-add-space:auto;
mso-pagination:widow-orphan;
font-size:12.0pt;
font-family:"Aptos",sans-serif;
mso-ascii-font-family:Aptos;
mso-ascii-theme-font:minor-latin;
mso-fareast-font-family:Aptos;
mso-fareast-theme-font:minor-latin;
mso-hansi-font-family:Aptos;
mso-hansi-theme-font:minor-latin;
mso-bidi-font-family:"Times New Roman";
mso-bidi-theme-font:minor-bidi;
mso-fareast-language:EN-US;}
p.MsoListParagraphCxSpFirst, li.MsoListParagraphCxSpFirst, div.MsoListParagraphCxSpFirst
{mso-style-priority:34;
mso-style-unhide:no;
mso-style-qformat:yes;
mso-style-type:export-only;
margin-top:0cm;
margin-right:0cm;
margin-bottom:0cm;
margin-left:36.0pt;
mso-add-space:auto;
mso-pagination:widow-orphan;
font-size:12.0pt;
font-family:"Aptos",sans-serif;
mso-ascii-font-family:Aptos;
mso-ascii-theme-font:minor-latin;
mso-fareast-font-family:Aptos;
mso-fareast-theme-font:minor-latin;
mso-hansi-font-family:Aptos;
mso-hansi-theme-font:minor-latin;
mso-bidi-font-family:"Times New Roman";
mso-bidi-theme-font:minor-bidi;
mso-fareast-language:EN-US;}
p.MsoListParagraphCxSpMiddle, li.MsoListParagraphCxSpMiddle, div.MsoListParagraphCxSpMiddle
{mso-style-priority:34;
mso-style-unhide:no;
mso-style-qformat:yes;
mso-style-type:export-only;
margin-top:0cm;
margin-right:0cm;
margin-bottom:0cm;
margin-left:36.0pt;
mso-add-space:auto;
mso-pagination:widow-orphan;
font-size:12.0pt;
font-family:"Aptos",sans-serif;
mso-ascii-font-family:Aptos;
mso-ascii-theme-font:minor-latin;
mso-fareast-font-family:Aptos;
mso-fareast-theme-font:minor-latin;
mso-hansi-font-family:Aptos;
mso-hansi-theme-font:minor-latin;
mso-bidi-font-family:"Times New Roman";
mso-bidi-theme-font:minor-bidi;
mso-fareast-language:EN-US;}
p.MsoListParagraphCxSpLast, li.MsoListParagraphCxSpLast, div.MsoListParagraphCxSpLast
{mso-style-priority:34;
mso-style-unhide:no;
mso-style-qformat:yes;
mso-style-type:export-only;
margin-top:0cm;
margin-right:0cm;
margin-bottom:0cm;
margin-left:36.0pt;
mso-add-space:auto;
mso-pagination:widow-orphan;
font-size:12.0pt;
font-family:"Aptos",sans-serif;
mso-ascii-font-family:Aptos;
mso-ascii-theme-font:minor-latin;
mso-fareast-font-family:Aptos;
mso-fareast-theme-font:minor-latin;
mso-hansi-font-family:Aptos;
mso-hansi-theme-font:minor-latin;
mso-bidi-font-family:"Times New Roman";
mso-bidi-theme-font:minor-bidi;
mso-fareast-language:EN-US;}
.MsoChpDefault
{mso-style-type:export-only;
mso-default-props:yes;
font-family:"Aptos",sans-serif;
mso-ascii-font-family:Aptos;
mso-ascii-theme-font:minor-latin;
mso-fareast-font-family:Aptos;
mso-fareast-theme-font:minor-latin;
mso-hansi-font-family:Aptos;
mso-hansi-theme-font:minor-latin;
mso-bidi-font-family:"Times New Roman";
mso-bidi-theme-font:minor-bidi;
mso-font-kerning:0pt;
mso-ligatures:none;
mso-fareast-language:EN-US;}
@page WordSection1
{size:595.3pt 841.9pt;
margin:72.0pt 72.0pt 72.0pt 72.0pt;
mso-header-margin:35.4pt;
mso-footer-margin:35.4pt;
mso-paper-source:0;}
div.WordSection1
{page:WordSection1;}
/* List Definitions */
@list l0
{mso-list-id:366374507;
mso-list-type:hybrid;
mso-list-template-ids:-1454997808 134807553 134807555 134807557 134807553 134807555 134807557 134807553 134807555 134807557;}
@list l0:level1
{mso-level-number-format:bullet;
mso-level-text:;
mso-level-tab-stop:none;
mso-level-number-position:left;
text-indent:-18.0pt;
font-family:Symbol;}
@list l0:level2
{mso-level-number-format:bullet;
mso-level-text:o;
mso-level-tab-stop:none;
mso-level-number-position:left;
text-indent:-18.0pt;
font-family:"Courier New";}
@list l0:level3
{mso-level-number-format:bullet;
mso-level-text:;
mso-level-tab-stop:none;
mso-level-number-position:left;
text-indent:-18.0pt;
font-family:Wingdings;}
@list l0:level4
{mso-level-number-format:bullet;
mso-level-text:;
mso-level-tab-stop:none;
mso-level-number-position:left;
text-indent:-18.0pt;
font-family:Symbol;}
@list l0:level5
{mso-level-number-format:bullet;
mso-level-text:o;
mso-level-tab-stop:none;
mso-level-number-position:left;
text-indent:-18.0pt;
font-family:"Courier New";}
@list l0:level6
{mso-level-number-format:bullet;
mso-level-text:;
mso-level-tab-stop:none;
mso-level-number-position:left;
text-indent:-18.0pt;
font-family:Wingdings;}
@list l0:level7
{mso-level-number-format:bullet;
mso-level-text:;
mso-level-tab-stop:none;
mso-level-number-position:left;
text-indent:-18.0pt;
font-family:Symbol;}
@list l0:level8
{mso-level-number-format:bullet;
mso-level-text:o;
mso-level-tab-stop:none;
mso-level-number-position:left;
text-indent:-18.0pt;
font-family:"Courier New";}
@list l0:level9
{mso-level-number-format:bullet;
mso-level-text:;
mso-level-tab-stop:none;
mso-level-number-position:left;
text-indent:-18.0pt;
font-family:Wingdings;}

-->
Meta’s most recent product announcements grabbed attention but the one thing they could not hide is that it’s still not possible to build a system as sophisticated as Vision Pro at an affordable price, or Meta would have done it.You could argue that at $1,299 the Meta VR Glasses the company introduced last week and promised to ship in spring 2027 could be as sophisticated, but they do betray some significant design compromises Apple seems to have avoided. The most glaring compromise is the use of a puck to carry the processor and battery, so the glasses weigh just 100 grams. Even now, a recent report tells us senior Apple decision makers have resisted use of an external puck to help reduce the weight and cost of future Vision Pro models, holding steady to the fully integrated device model typical of Apple’s products. While Apple will have its reasons, I’m not certain that’s the correct approach, as the company could effectively sell two products, the headset and the puck, making both systems available on a biennial upgrade. (There are other compromises, for example Meta’s glasses seem to deliver a narrower field of view than the current Vision Pro).At the same time, the competition between the two firms’ hints at the technological barriers Apple is trying to solve. After all this time since the introduction of the first Vision Pro, Meta still can’t match the field of view and even with a puck rather than integrated build its alternative still costs a fair bit of cash, while the processor it plans to power it with isn’t even available yet. Think about that. It means Meta and its partners are still building a processor with which to match Apple’s. If that’s not a technical challenge, I don’t know what is.The roadmap for visionOSThis isn’t about Meta bashing or Apple building. It’s about how Apple has always  openly described Vision Pro as a product from the future. Even earlier this month, CEO John Ternus said, “I think of it [Vision Pro] as like the early days of computers. It is the early days of a long journey. That’s how I feel about Vision Pro.”Meta’s new product and the latest reporting from inside the company show the future takes time to build. On the journey to doing so Apple now seems focused on four potential paths — though I don’t think the consequences of taking those paths should be seen as zero sum options, but about priorities toward an eventual goal for pervasive spatial computing, combining always-aware contextual AI with wearable computing designed to deliver powerful augmentation across all parts of life.The four concepts Apple is exploring seem to include:>·>      A design that offloads components to a puck.>·>      A Vision Pro design using lighter, potentially lower-cost materials such as plastic or titanium.>·>      Smart glasses, with audio, camera,  and AI which Apple hopes to introduce by the end of 2027.>·>      Smart glasses, with audio, camera, AI, and a video display, which feels like the most appropriate endgame for both Vision Pro and smart glasses products.Reports suggest Apple is in a state of flux regarding its next move with Vision Pro. I’d argue it’s Apple in the lull before the spatial storm, what happens next is under review.What amazing timing However, this must be contextualized against the depth of competition between Apple and Meta. The timing of these claims — coming so soon after Meta’s latest product reveal – suggests that describing Apple’s efforts as being on “life support” is a gift to Meta’s marketing machine. At Apple’s scale, it is entirely possible to pursue multiple parallel designs; so, any perceived internal confusion may simply reflect different degrees of preparedness across all four (and possibly more) of these projects. Management may simply have been waiting to see what Meta did before accelerating any of these initiatives. While there have reportedly been some recent redundancies in the Vision Pro team, Apple’s new leader doesn’t seem to have gone cold on the product. It’s more as if he accepts Apple is in it for the long haul. A long haul destinationThat long term commitment allows Apple time to refine key components of its product plans , particularly around privacy. Apple wants to avoid the stigma of “pervert glasses”, which is why its work on ambient audio intelligence requires close scrutiny. The decisions it takes to protect privacy there – including use of Secure Exclave -- will I believe be emulated across the future visionOS product range. In the short term, the most telling insights we’ll see will be the pushback Apple and Meta face from global privacy regulators, for AR in Meta’s case, and for Audio Intelligence in the case of Apple. Those decisions will set the framework both for Apple’s future AR products and the future of consumer/enterprise mass market AR. >Now please subscribe to my daily, human-curated Apple-related news headline feed at The Core, or follow me on BlueSky, LinkedIn, or Mastodon.