While its flagship Dreamforce event was in full swing on Wednesday, Salesforce was triaging a roughly seven and a half hour-long service outage that disrupted user access and caused “severe delays” and intermittent errors. Some customers were also unable to submit new support cases.

The service outage hit at 3:50 a.m. EDT, impacting “multiple instances across all regions,” Salesforce reported. It was marked resolved right around 3 p.m. EDT, after several hours of monitoring to determine that fixes had been successful.

Salesforce initially pegged the issue to an “external dependency failure” impacting the legacy login server. A core system component experienced increased load, limiting its capacity to process requests. The company confirmed that there were no issues with third-party infrastructure.

Beyond the obvious embarrassment from the outage occurring during Dreamforce, an analyst said the incident highlights a cloud resilience problem, rather than purely a legacy one.

“Cloud does not eliminate architectural dependencies,” said Abbas Jaffery, a principal advisory director at Info-Tech Research Group. “It can sometimes make them less visible. And when the platform is your system of record, those hidden dependencies become an enterprise risk rather than simply a technology risk.”

Rolling restarts, persistent issues throughout the day

Salesforce began experiencing service issues at 3:50 a.m. EDT on September 16, and initial investigation determined that requests were stalling while waiting on responses from an internal login service that was using up available server resources.

Initially, Salesforce blocked application programming interface (API) endpoints and attempted rolling restarts to revive the service, then pushed out fixes region-by-region. At 7:20 a.m., some customers were seeing service return to normal, and Salesforce was working on a “code-level permanent fix.”

However, the rollout did not complete for a number of instances, and some automated fixes didn’t fully resolve the issue. For example, customers reported that scheduled jobs were not running as expected, even after service was restored. Salesforce manually restarted those ‌instances.

Salesforce later reported that the impact radius was “narrower than initially understood” and saw signs of recovery by around 11 a.m. EDT, with most customers coming back online.

A sub-set of Hyperforce instances were the last to be restored. Mitigations were in place across all instances by 11:39 a.m. EDT, and Salesforce continued to monitor the issue until marking the incident resolved at 2:59 p.m. EDT.

“We apologize for how this incident affected you and your business,” Salesforce posted on its incident blog. “We will undertake a full investigation of the incident, establishing the technical trigger, the underlying cause, and preventive action to avoid a repeat in the future.”

Creating a ‘temporal’ data problem

For customers for whom Salesforce is a system of record, several hours of authentication and service disruption can create a “temporal data problem,” Info-Tech’s Jaffery explained. “Events that should have happened at different points in time may occur later, fail altogether, or arrive out of sequence,” he said.

For instance, a customer interaction may occur through another channel while Salesforce is unavailable, but an integration, workflow, or scheduled process that normally records or propagates that event is unable to run.

This has several potential consequences, Jaffery said. Transactions and customer service processes are delayed; APIs and middleware may accumulate retries, timeouts, and queues. Records can become temporarily inconsistent, and scheduled jobs and workflows may be missed. Employees could lose visibility into customer history or case status, even when the underlying data has not been lost, creating a “data divergence.”

The first mistake would be to assume that just because users can log in, the incident is over, he noted. “Enterprises should move immediately into a reconciliation and integrity phase,” Jaffery advised. This means not only verifying interactive access, but APIs, integrations, scheduled jobs, queues, workflows, automation, authentication flows, and downstream systems.

Enterprises should ask what transactions failed, partially completed, or were duplicated during the outage? Which scheduled or asynchronous processes did not execute? Did integrations retry successfully, or did they create a backlog or retry storm? Are downstream systems now consistent with Salesforce?

Security teams should also validate authentication and session behavior, privileged access, integration credentials, and any emergency changes made during recovery, Jaffery explained. “The most important question is not simply ‘Is Salesforce back?’, but ‘What did the business expect to happen during the outage, and can we prove that it actually happened?,’” he said.

What to look for in post-incident reports

A credible post-incident review from Salesforce should establish a causal chain: The trigger, dependency failure, technical propagation, customer impact, detection, mitigation, recovery, and permanent corrective action, Jaffery said.

The company should be able to answer these questions, he said:

• What was the actual initiating failure and why did the failure propagate into the login path?
• Why could the affected dependency consume sufficient capacity to affect core services?
• Why didn’t isolation or failover prevent the impact?
• Why did initial remediation attempts fail and why did the subsequent rollout require additional intervention?
• What safeguards are being added to prevent recurrence?
• How will Salesforce demonstrate that the corrective action actually works under failure conditions?

Service restoration simply tells customers: “We got it working again,” he noted. But root cause analysis tells customers: “We understand why it failed, why our controls didn’t prevent it, and what has changed so that the same failure mode is less likely to recur.”

It’s not just about ‘legacy’ pieces in the stack

One architectural lesson is that a legacy component does not have to be large to be critical, Jaffery pointed out. An older authentication service can remain part of a modern stack, and therefore become a dependency for newer services.

“The component’s age matters less than its position in the dependency graph, its blast radius, and the quality of its isolation and failure handling,” he said, pointing to this incident’s progression: Requests stalled waiting on an internal login service due to increased resource consumption led to investigation into an external dependency failure, which in turn revealed impact on a legacy login server. Finally, Salesforce said, “core system components experienced increased load, which limited its capacity to process request”.

That is a classic resilience question, Jaffery pointed out: Can a failure in one dependency remain in that one dependency, or does it become a platform-wide failure?

Modernization should not be identified simply by how much old technology has been replaced, he noted, it should also measure dependency concentration, isolation, “graceful degradation,” recovery paths, and failure blast radius.

“For enterprise architects, that is the real takeaway,” he said.

Maybe driven by agentic AI, exacerbated by layoffs

At this point, there are no obvious signs that this was a security incident, noted David Shipley, CEO of Beauceron Security. “Right now, this bears all the hallmarks of an update gone horribly wrong.”

He pointed to an incident in December 2025 when Amazon’s internal AI coding agent, Kiro, caused a 13-hour AWS outage in a mainland China region, noting, “I’m not going to be shocked if we don’t see some kind of agent role in this kind of scale disaster.”

Significant Salesforce layoffs over the last few years could also have had a negative impact on the outage and recovery, he added. “Having it happen during Dreamforce had to be all kinds of hell, though, for their sales and customer support teams,” he said. “Pour one out for them as they work on rebuilding relationships, face-to-face.”

This article originally appeared on CIO.com.

Microsoft’s top lawyer argued Tuesday that legislators “must make secrecy [orders] the exception” in government subpoenas demanding information about LinkedIn users.

LinkedIn, which is owned by Microsoft, is fighting what it calls overly broad subpoena demands from the US government, which sometimes come with secrecy orders that prevent LinkedIn from alerting customers whose information is being requested.

The company is asking federal courts “to enforce meaningful limits on both the scope of government demands and the secrecy that can accompany them,” wrote Jon Palmer, Microsoft’s chief legal officer, in a Tuesday blog post. “We recognize law enforcement’s important role in protecting public safety and investigating crime, and sometimes that does need to be done covertly. At the same time, customers and users deserve meaningful limits and independent oversight through an adversarial process.”

He pointed out: “People and organizations increasingly entrust their most sensitive information to online services. If providers cannot challenge demands they know are overbroad—or if courts may silence them without a rigorous, adversarial review—the safeguards the law requires will be weakened precisely when they are most needed.”

A tricky issue

The issue is a tricky one. Law enforcement often use this type of subpoena as an investigative tool, seeking those who are engaged in illegal activities. The theoretical justification for secrecy is to avoid alerting the investigative target to make it less likely the suspect will try to destroy evidence or flee the jurisdiction.

Government lawyers are supposed to only make secrecy requests when absolutely essential. Microsoft is suggesting that courts and congress need to step in to curtail blanket government efforts. 

“The Fourth Amendment protects the right to be free from unreasonable searches and seizures. That right applies to papers kept in a desk and it also applies when personal and business records are stored online,” Palmer wrote. “Online service providers, like LinkedIn and Microsoft, also have a First Amendment right to speak to their customers when the government obtains an order to search their private information. Secrecy may sometimes be justified, but it should be tailored to demonstrated needs and subject to meaningful review.”

He added: “The government must seek only relevant information, justify secrecy with specific evidence and infringe on speech to the least extent possible.” In his post, he pointed to a recent legislative effort in the US House of Representatives that might mitigate the issue if it ends up becoming law. 

“On August 31, the House passed legislation to rein in secret surveillance and strengthen notice protections when the government seeks data held by technology providers,” he wrote. “The reforms would place clearer limits on secrecy orders, require greater accountability, and help ensure that secrecy is the exception – not the rule. The Senate should act promptly to send these historic reforms to the President.”

LinkedIn privacy battles

LinkedIn itself is currently fighting litigation that accuses it of directly violating the privacy rights of its customers, and a federal judge this month dismissed another similar case, but gave plaintiffs permission to refile, with a caveat.

“Given LinkedIn’s further arguments that users voluntarily download browser extensions, which by their nature intentionally expose data to websites, it seems unlikely that the plaintiffs will ever be able to allege a privacy violation, much less prevail at the end of the day,” US District Court Judge Vince Chhabria wrote. “But in an abundance of caution, dismissal is with leave to amend.”

But, he added, if the amended complaint isn’t filed within 14 days, “dismissal will be with prejudice.”

Privacy now a ‘data stewardship obligation’

Jeff Valdes, a director at Acceligence, noted, “there is definitely some irony here.”

“If Microsoft wants customers to view it as a steward of their privacy when the government comes asking for their information, customers are naturally going to apply that same standard to how Microsoft and LinkedIn collect, use, protect, and disclose information themselves,” he said. “Privacy is difficult to compartmentalize. You cannot have one philosophy of customer privacy for government access, another for product design, and another for your own commercial data practices without eventually creating a credibility problem.”

Mike Wilkes, enterprise CISO at Aikido Security, agreed, pointing out, “without meaningful limits, judicial scrutiny, and an expiration mechanism, a temporary investigative necessity starts looking a lot like a permanent architecture for invisible surveillance. The individual may never have an opportunity to challenge the scope of the request, because they may never even know the request existed until prosecutors show up with an indictment.”

That, he said, “is why Microsoft’s argument matters, despite the obvious irony of LinkedIn simultaneously defending itself against privacy claims from its own users.”

But Ryan O’Leary, an IDC research director, offered a different perspective.

“Microsoft makes no bones about using the data contained within its own systems for its own purposes. Both things can be true: Microsoft can fight for the privacy of its platform while still not necessarily respecting the privacy rights of its end users,” O’Leary noted. “This seems to come down to protecting its own proprietary data sets, not some altruistic privacy crusade.”

 At the same time, Valdes pointed out, Palmer’s post highlights how deeply privacy has become a top-tier enterprise IT priority.

“Privacy is rapidly becoming a much broader data stewardship obligation,” he said. “Companies holding sensitive information increasingly have to think simultaneously about government requests, third-party access, their own collection practices, AI use, data retention and what they tell customers about all of it. If you want to be trusted as the custodian of the world’s data, customers are going to judge how you protect that data in every direction.”

However, Wilkes noted, “Microsoft does not need to be a perfect privacy saint to be right about this particular problem.”

This article originally appeared on CSOonline.

The fight against AI slop is real, even in photography, which is why Apple introduced its Reference Image tech alongside its new iPhones. This shifts image verification to the moment of capture, creating a privacy-preserving digital negative that could help newsrooms, businesses, and individuals prove that images are real. Apple’s Security Research team just shared a white paper explaining how it works.

What is Apple Reference Image?

Introduced alongside the new iPhone 18 Pro series, Apple Reference Image is designed as a process to help prove the authenticity of images captured using that phone. Like so much in Apple, it’s a combined hardware/software feature that means when an image is captured, the camera also captures unalterable Reference Image data, with help from Private Cloud Compute, Apple’s cloud intelligence system.

That help means Apple thinks it now offers a system that protects image provenance to the extent of providing quantum-secure defense. The need to protect digital assets such as images from quantum-based abuse will become increasingly important as more state and state-adjacent adversaries seek out new ways to undermine security and do harm. Apple quite clearly considers that looming threat to be real; in 2024 it introduced encryption to protect iMessage against future attacks using quantum computers.

How do we use Reference Image?

Apple is reaching out with this tech, including providing APIs for iPhones, iPads, and Macs that third-party developers can use to build support for Reference Image into their apps. 

Otherwise, Reference Images can be viewed in the Photos app beside the main image, acting like a digital negative to enable visual comparison of both images. The idea is that the existence of this digital negative acts as a tool to prevent people from passing off edited or AI-altered images as being genuine pictorial narratives of events.

This will be useful in multiple situations. In the media, a news desk may turn to the feature to verify if an image is genuine or fake, while in the enterprise the tool provides a useful proof point when assessing images that relate to any workflow — that picture taken in the stock room may now be more useful as proof that stock is in hand, while a car hire firm may find it easier to prove damage during a rental period by verifying the pic. The main thing it does is help you identify changes in an image, whether those changes were wrought by human or AI.

Why Apple built this protection

Apple’s newly published white paper explains how it works and how the system is built to endure against interference, particularly as AI images are becoming increasingly convincing. After all, as digital enters most camera workflows at the sensor, proof of algorithmic activity isn’t necessarily enough to show an image has been altered out of its photorealistic beginning. It must be said that at this point we only have Apple’s word for the effectiveness of the solution, but I’m inclined to think it will be borne out.

Enter Apple Reference Image, which acts as a “chain of trust,” one the Apple team says is stronger and more enduring than existing approaches based on the Coalition for Content Provenance and Authenticity (C2PA) standard. Apple believes the C2PA approach remains too vulnerable because it is deployed after image capture and also creates a privacy risk by attaching images to public identity. You don’t have to look too far for instances in which photojournalists have been targeted by hostile forces once their identity has been realized, and the C2PA system arguably adds to that risk.

“Other industry solutions require a photographer or institution to vouch for an image using their own credentials,” the white paper explains. “We are concerned this puts some photographers, such as those operating in conflict zones, in a difficult position; it should not be necessary to forgo anonymity in order to prove image authenticity.”

Apple’s system is designed to kick in as images are captured, when a photographer can choose to create a securely timestamped reference image that accurately reflects what was captured by the iPhone’s camera sensor. The creation process leans into the powerful iPhone processor and also Apple’s Private Cloud Compute.

A little insight into how it works

As the image is captured, the iPhone cryptographically signs pixel data immediately after capture, which shows the image to be an exact depiction of what the camera sees. The system also captures relevant metadata, including use of the Secure Enclave on iPhone to ensure some values.

That means that the Reference Image now carries deep insight into the original, down to an unchangeable record of metadata. Not only can you see the image matches, but you can ensure the metadata is also the same. Furthermore, the image is also timestamped with a value taken from Apple’s own cryptographic timestamp service. Pull it all together and the digital negative is packed with verifiable, trusted data that can be contrasted with an image you want to check. 

There’s much more to this system, which Apple has built to be resilient to compromise while also protecting the privacy of the photographer. Instead of requiring the photographer act as the point of proof, Apple’s own Private Cloud Compute servers play the role by cryptographic signing of the image. 

“The result is a verification model that offers photographers, newsrooms, and everyday users renewed confidence that an image they’re viewing is a photograph actually captured by a camera,” the company said. It is, I think, also interesting to consider how Apple’s newly introduced tool for cryptographic proof could potentially be applied elsewhere.

The caveat is that the protection only extends to the main camera of the iPhone 18 Pro family and is not retrospective. It won’t be available in China on launch, while in Europe you can develop and view reference images but not create them.

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

In the battle over whether the US or China will lead the world in AI, the argument for the US side largely can be summed up in four words: USA good, China bad.

In this worldview espoused by big US AI companies, the Trump administration, and many Republican members of Congress, AI born and bred in China is portrayed as a grave security risk. The Chinese government, they claim, will use Chinese AI models to steal information from businesses, individuals, and government agencies. They also claim that because Chinese AI complies with Chinese censorship rules, US companies that use these models will be forced to comply with the rules as well.

Critics also say that if US AI is overtaken by Chinese AI, the US economy could be dealt a crushing blow, because the hundreds of billion dollars being invested by US AI companies will go down the tubes.

Because of this, Big AI firms Anthropic and OpenAI are lobbying to ban Chinese open-source AI from the US.

Not all AI companies and their customers agree with a ban. Microsoft, Meta, and AI chip maker Nvidia, among others, say there’s room for both US AI models and Chinese open-source models.

What’s the fight really about? To understand, let’s first dig into the tech behind most Chinese AI.

What’s the fuss about?

The fight is primarily over what’s called lightweight open-source AI. “Lightweight” means the AI models don’t require multibillion-dollar data centers to train, build, and run them. They need comparatively modest processing power.

That’s attractive for businesses, because lightweight models are less expensive to license and run than much American AI tech, notably OpenAI’s ChatGPT and Anthropic’s Claude.

The other draw for businesses is that Chinese open-source AI tech can be used for free. And perhaps more important that is that the AI is typically what’s called “open weight,” which means businesses can tweak the technology for their own purposes.

When a traditional American AI model is trained, people who train it must decide how much “weight” to give to every parameter in every element of training data. For example, if AI were being trained to tell the difference between an apple and an orange, a great deal of weight would be put on the color of the fruit and the appearance of its skin. Less weight would be put on the fruit’s shape or size.

When a company buys American AI, they don’t know how much weight is given to each training parameter. They don’t even know what the parameters are. In open-weight AI, though, businesses have access to those parameters — and they can assign their own weights to them. This means they can customize the AI more easily for their particular needs in ways they can’t with traditional American AI models.

(Note that the terms open weight and open source are often used interchangeably when talking about AI, although technically they differ. Some open-weight models withhold some information such as their source code and training data — information that should be made public for a model to be considered truly open source.)

What Microsoft says about the Chinese models

Microsoft, Nvidia, Google, Meta, and many smaller companies don’t agree that Chinese lightweight open-source AI should be banned. They claim that security and censorship issues can be easily solved. And they believe use of the models will help the United States economically, not hurt it.

Microsoft CEO Satya Nadella posted on X: “Open-weight models are essential to a healthy AI ecosystem.” He added that the models will “strengthen American competitiveness and expand economic opportunity, while protecting national security.”

Microsoft was also one of more than 200 corporate signers of a document outlining all the reasons they believe open-weight AI is important.

Microsoft isn’t just talking about the importance of using Chinese lightweight open-source AI — it’s begun offering it to its customers and is looking to expand it even more. Right now, DeepSeek, the first Chinese lightweight open-source model to make its way to the US, is available via Azure. And Microsoft is testing offering DeepSeek to customers of its Copilot Cowork agentic AI product.

Keep in mind that Microsoft’s embrace of Chinese lightweight open-source AI has less to do with virtuous thinking than with profit-making. Offering these products to its customers will please both old and new customers because of the tech’s benefits. The more Microsoft offers to its customers, the more money it will make.

What happens next?

It’s unclear what happens next with the Chinese AI products. OpenAI and Anthropic are allied with the Trump administration in wanting them blocked or limited in the US.

Still, there’s an increasing groundswell backing back their use. Bill Gurley, a Silicon Valley venture capitalist, told the New York Times: “You have two factions fighting over this issue. There’s the people who want OpenAI and Anthropic to own everything, and then there’s everybody else, including customers.”

My bet is that in short term, OpenAI and Anthropic will get what they want, because Trump has got their back. But after Trump leaves office, don’t be surprised if Chinese open-source AI has its day in the sun, backed by Microsoft.

Oracle began a new round of layoffs this week, sending early-morning termination emails to staff for the second time in six months.

The latest wave began Monday with affected staff being told, “After careful consideration of Oracle’s current business needs, we have made the decision to eliminate your role as part of a broader organizational change,” according to BusinessInsider. Termination was immediate.

That’s the same wording as in the previous wave of layoffs, which took place on March 31 and affected workers in the US, India, Canada, Mexico and Uruguay. In the 12 months to May 31, Oracle cut its global workforce from about 162,000 to about 141,000, a decline of roughly 13%.

Oracle has made no public statement confirming the latest round of job cuts, but in a regulatory filing days earlier the company said it had set aside a further $700 million for restructuring charges, bringing the total charges this year to roughly $2.8 billion.

What the new termination emails say

Staff receiving the latest termination emails were told termination and compensation details would follow by DocuSign, Business Insider wrote. An internal document reviewed by the publication said severance terms varied by role and region, and some teams facing double-digit percentage cuts.

Affected employees took to social media to vent.

Eric Brunson, a senior principal offensive security researcher at Oracle, described in a LinkedIn post how he had lost access to corporate communication tools before receiving any formal notice. “I woke up to not being able to log back into Slack,” he wrote. “No new emails or notification in email and I’m locked out of there. I was able to get ahold of my manager on LinkedIn and she confirmed.” Brunson said the timing fell two days before a scheduled RSU vesting date, adding, “Probably part of the plan.”

The filing that backs up the layoff accounts

While Oracle is not talking about the layoffs, its 10-Q quarterly report states that management approved and supplemented restructuring plans “to implement certain strategic measures and further improve operational efficiencies, including through the adoption and integration of artificial intelligence technologies across certain functions.” It adds that “subsequent to August 31, 2026, our management supplemented the 2026 Restructuring Plan by approximately $700 million to reflect additional actions that we expect to take.”

Oracle has already spent $1.97 billion of the $2.1 billion restructuring charges it originally budgeted, it reported.

Sanchit Vir Gogia, chief analyst at Greyhound Research, said the filing should be read carefully rather than as confirmation of a headcount. “The supplement is an estimate, not a bill,” he said, noting it raises the program’s estimated cost by about a third without committing Oracle to a timetable.

Gogia said the more significant shift is in the filing’s language rather than the dollar figure. Oracle’s August 2025 and February 2026 filings had described the plan as tied to “acquisitions and certain other operational activities.”

AI was named as a driver of restructuring for the first time in the Sept. 11 filing.

Why the headcount stays unconfirmed

Gogia said no verified figure exists yet for how many employees the September round affected.

He noted that 30,000 was a January forecast of the total 2026 program. Twenty-one thousand is the confirmed net decline in Oracle’s global workforce across the full fiscal year. A reported 3,000 job cuts in India on Sept. 1 remains unconfirmed by Oracle.

The $700 million restructuring supplement “cannot be divided into people,” Gogia said, since it covers termination benefits, contract termination costs and other exit costs across a program spanning multiple countries. “There is no solid headcount for the September round,” he said.

A pattern that began in March

Oracle’s first 2026 layoff wave began March 31, when the Revenue and Health Sciences unit, the SaaS and Virtual Operations Services group, and NetSuite’s India Development Centre saw some of the deepest reductions.

Figures published by Oracle for its fiscal year ending May 31, 2026, show research and development headcount fell from 50,000 to 43,000 employees during the year, sales and marketing fell from 31,000 to 25,000, and services fell from 37,000 to 34,000, according to Gogia’s analysis of the company’s own reporting.

International staff, at 92,000, saw a larger reduction than the 49,000-strong U.S. workforce, he said.

Oracle did not respond to a request for comment.

This article first appeared on CIO.

Just because Apple now has AI, a new folding iPhone, and a newly minted CEO doesn’t mean the litigation between it and OpenAI has gone away. Apple now wants to force OpenAI to let it look at the hardware it has been building, according to a new report.

A reasonable request?

It seems a reasonable request, doesn’t it? After all, Apple’s argument is that OpenAI has been engaged in trade secret theft to help it design and develop its new hardware. OpenAI’s defense against these claims feel flimsy, at least to this reporter. They seem to coalesce around something like, “We don’t need your trade secrets because we’re making something brand new.” 

The problem with that defense is, contextually, that while on this stated mission to do something completely new, the company has hired around 400 former Apple staff so far, including its chief designers. And Apple thinks part of that process has been OpenAI, in whole or in part, working to exfiltrate its trade secrets.

What Apple wants — and why

With those facts as your guide, Apple’s request to Judge Edward J. Davila seems reasonable. It wants to take a look at what OpenAI is developing to ensure its trade secrets have not been abused in the process of designing that product. Apple argues that if it is forced to wait until the product is released, then it will be impossible to make its trade secrets confidential again, particularly as the defense seems to consist of that pinky promise that no Apple trade secrets have been harmed.

Apple legal also argues that it cannot be fair to allow OpenAI to defend itself by alleging its unreleased and unseen product doesn’t contain any trade secrets without permitting Apple — and the court — to verify that. While Apple’s counsel doesn’t seem to have said it, you could paraphrase the request as Apple telling the genAI firm, “Let us see what you are building; if you have nothing to hide, you have nothing to fear.”

Except, of course, that while building its defense, OpenAI is giving many of us the distinct impression that it may have something to fear.

What may happen next

Despite the merits of the argument, I think Apple’s request will not prevail, in part because the court may assess that if Apple takes a look at OpenAI’s homework it may compound the risk of IP theft. But that doesn’t mean Apple’s attempt will fail outright, as the compromise position is likely to be the appointment of a trusted, independent, third-party expert witness to take a look at what Apple’s competitor is making in Apple’s stead. 

There is precedent for this. That’s more or less what happened when Waymo pursued a similar case against Uber, or when AMCS litigated against Sinovel. There are nuances to all three cases that mean they aren’t perfectly aligned, but that does seem a logical next step to this layman.

Of course, just because it’s logical doesn’t mean either party is going to like it, but OpenAI could conceivably even suggest such an approach as it seeks to buy itself time to build and release its still mythical hardware. That’s also why Apple wants a chance to look at documents pertaining to that hardware to make very certain it hasn’t infringed any of Apple’s own trade secrets.

A side order of humble pie

All the same, if this case does indeed turn out to be a scenario in which one company has been found with its hand in the cookie jar, then the best possible approach for the company with crumbs around its mouth is going to be damage control. That’s going to take a lot less war-war and a great deal more jaw-jaw. It’s also going to require the intake of a very, very large slice of humble pie. Right now, it seems to me that Apple isn’t talking, and OpenAI isn’t hungry enough to take that first bite. Not yet. 

The case, number 5:26-cv-07078, rolls on.

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

Microsoft on Monday added itself to a growing list of AI vendors pledging to try to control the behavior of its AI models.

“AI should not exceed human control. Models should remain subordinate to humanity, subject to meaningful human oversight and control,” the company wrote in the draft version of its Humanist AI Code of Conduct, released Monday with an invitation to the public to provide feedback. 

“Humanist AI develops systems with clear purposes, evaluated against real-world impact, and rejects the race to produce an all-purpose superintelligence that could evade these safeguards,” Microsoft wrote. “We are building something fundamentally useful and safe even if that means compromising on ultimate generality, autonomy, or capability.”

Microsoft’s comments are roughly in agreement with recent posts by various major AI vendors, including those from Anthropic and OpenAI, which were endorsed by Elon Musk, CEO of SpaceXAI and Tesla, but nowhere in its 37 page document is there any description of concrete action. However, to be fair, almost none of the other major AI players have been specific about how they would control future AI models either.

Acknowledges issues

A key problem with many vendor attempts to impose AI limits is that all of these companies have thus far been unable to stop AI agents from doing almost anything, given the agents’ ability and willingness to sidestep or ignore guardrails. 

Microsoft has described its worries about the technology in the past, both when it started to curtail AI efforts among its own employees and when it announced the formation of the team that created the Code of Conduct. 

Its current post acknowledged some of the difficulties involved in pushing AI development while limiting its abilities. 

“Both under- and over-caution represent failure modes with different types of consequences,” the company said. “Under-caution can clearly result in more direct harm, but over-caution may occur more often and therefore may need more frequent correction. This is where proportionality to the potential for harm and safety context are particularly important. Responses and actions should take into account the estimated severity and likelihood of potential harms and adjust responses and actions accordingly.”

It noted, however, that the term “harm” is “broad and often context-dependent, and there are nuanced gradations in potential severity and likelihood. Microsoft AI (MAI) Model responses should be tailored to that context and to a wide range of harms.”

Laudable goal, but lacks detail

Analysts and consultants generally agreed that Microsoft’s stated goal is laudable, but the lack of specifics and verification mechanisms makes it difficult to take the post seriously.

Thomas Randall, research director at Info-Tech Research Group, also said he spotted some apparent contradictions within the document. 

“[It] says MAI models will not assist in manufacturing or modifying weapons. Yet Microsoft offers OpenAI’s GPT-5.2 through Secret and Top Secret government clouds for defense and national security workloads,” Randall said, though he acknowledged that this is not technically a breach of the Code because GPT-5.2 is not an MAI model. “The most meaningful parts of the Code of Conduct may exclude other parts of Microsoft’s actual AI business operations. Tensions like these appear in other forms throughout the Code.”

But he added that Microsoft’s position is bolstered by its earlier Frontier Governance Framework that “provides pre- and post-training evaluations, six-month reassessments, third-party testing, phased releases and a commitment to pause development or deployment where high risks cannot be mitigated.” However, he noted, “it is still Microsoft that defines the thresholds, selects the evaluators, determines whether residual risk is acceptable and gives its own executives the final deployment decision.”

Randall said he would like to see Microsoft, as well as other major AI vendors, deliver more verifiable data points, such as those from independent evaluators given continuous access and freedom to publish findings about the vendor’s actions. He also would like to see independent board-level safety oversight with authority to block releases, protected whistleblowing, accessible monitoring, audit logs, kill switches, and mandatory reassessment after model changes.

However, Justin Greis, CEO of consulting firm Acceligence, pointed out that Microsoft was candid about the many elements that are not yet in place.

“The current models are not yet trained on the Code, the evaluation framework is still being developed, and Microsoft explicitly says written objectives alone cannot ensure alignment or guarantee present-day behavior,” Greis said. “That distinction matters. Publishing a constitution for AI is useful. Proving that the system actually follows the constitution, especially when models become increasingly agentic, is the hard part.”

And consultant Brian Levine, executive director of FormerGov, said Microsoft deserved a little bit of credit for at least saying that model capabilities should be limited.

“Microsoft explicitly says it will compromise on generality, autonomy, and capability to keep systems safe and under human control, and that it rejects the race to build an all-purpose superintelligence. Coming from a company of Microsoft’s size and ambition, that’s a notable thing to put in writing,” he said. “For years, the assumption was that the frontier labs would chase maximum capability and treat safety as a constraint to be managed. A document that says the opposite, that says usefulness and control come before ultimate capability, is worth paying attention to, regardless of what follows it.”

He added: “The real test comes next, and it’s verification: measurable standards, independent assurance, and a way for outsiders to check the commitments against what’s actually shipping. That’s the natural progression and it’s the part the whole industry still has to build.”

Not doing the hard part

But others argued that Microsoft is merely doing the easy part, the marketing part, and is deliberately not committing to doing the hard part.

“Promising to give up capabilities is easy when those capabilities don’t yet exist,” said Noah Kenney, principal consultant at Digital 520. “The real test will come when Microsoft has a model ready to ship that would close a competitive gap and decides to hold it back.”

Until then, he said, the promise of responsible AI costs Microsoft nothing.

“Microsoft’s code of conduct reads like a marketing document written to reassure customers, regulators, and its own employees,” Kenney noted. “The problem is that no one knows how to make those promises specific, which leaves Microsoft asking for trust before they can explain what that trust should be based on.”

Tom Findling, CEO of Conifers.ai, added that his concern with the Microsoft document is that it doesn’t answer the obvious question of how these models can possibly be controlled. 

“The document says the model should never resist being shut down, should stay within its scope, and shouldn’t hide what it’s doing. Those are all the right goals,” he said. “But the harder question is what happens when a highly capable agent doesn’t behave the way you expect. What actually stops it? As these systems get more autonomous, the safety boundary can’t just be that the model was trained not to do something. You need controls outside the model that limit what it can access, what it can do, and how far it can go.”

Cybersecurity learned this lesson a long time ago, he pointed out. “You don’t secure a system by assuming it will behave correctly,” he said. “You assume something will eventually fail, get compromised, or act in an unexpected way, and you design the controls around that. AI needs the same mindset.”

Frank Dickson, principal analyst at Dickson Research, contrasted Microsoft’s promise with Anthropic’s commitment, and found Microsoft lacking.

“Microsoft’s document is shy on mechanism,” he said. “Compare the two on specifics. [Anthropic CEO] Amodei’s proposal names a third party, METR, and describes what access actually means: office badges, company laptops, employee-level visibility, and publishing rights Anthropic doesn’t get to edit. You can check whether that happened.”

On the other hand, he noted, “Microsoft’s document says models should ‘fail tasks rather than violate the code’s rules,’ which is a real design principle. Credit where it’s due, but there’s no named auditor, no verification method, and no stated consequence for a violation. Thirty-seven pages and it still won’t commit to a single verifiable check.”

Dickson stressed that as long as agents routinely break their own rules, these vague promises won’t help. 

“Every frontier lab still gets jailbroken, still has agents that go off-script, still hasn’t closed the gap between what a model is instructed to do and what it can be induced to do,” Dickson said. “A values statement that skips the verification question isn’t a constraint, it’s a hope wearing a policy document’s clothes.”

While we wait to see if Apple really is experiencing huge demand for its newly introduced iPhones, today’s the day every iPhone user gets a completely new device experience, with iOS 27 rolling out internationally, along with the other members of Apple’s 27 series OS family. If you’re on the fence about installing the new release, here are 12 reasons you should upgrade today.

Speed and performance

Apple seems to have worked hard to improve speed and performance in this release. Everything feels much zippier, but where it seems most noticeable is when launching apps. Apple says these now open up to 30% faster, while AirDrop transfers are almost twice as fast as before.

Liquid Glass looks nicer

With the 20^th anniversary glass iPhone waiting in the wings, Apple isn’t going to get rid of Liquid Glass. But the company is prepared to iterate the look, and iterate it has, with sharper icons, improved contrast, and a new slider control that lets you set the UI’s transparency.

Use two iPhones

iPhone Handoff lets you use one eSIM across two devices on supporting carriers. Once set up it’s remarkably easy to use — your SIM activates the moment you unlock one of your iPhones. Better still, if you use two lines on one device, you’ll find these both transfer across too.

Siri AI

We’ve been talking so much about Apple’s improved Siri that I demoted it to fourth place; that’s not a reflection on what it does, but on how often you’ll have seen how it has become so much better at understanding context. It now has a Chat GPT-like interface, a standalone app, and powerful on-screen awareness to help get things done. Sadly, many of these features aren’t available in the EU.

Notes, Calendar, Reminders

AI is everywhere, including in Reminders Smart Suggestions, which can read your Messages chats and prompt you to set reminders for agreed tasks. Notes also gains Markdown copying.

Shortcuts to imagination

You can now create a shortcut just by describing what actions you want it to take. Apple Intelligence will string a shortcut together for you, which you can then use across all your systems.

Notify Me

Safari can watch a web page and automatically let you know when it changes. That’s not the only improvement; Apple Intelligence automatically groups your open tabs into topics so related pages are easier to find.

DIY: Safari extensions

I feel like this feature hasn’t had the attention it deserved, but you can now get Apple Intelligence to build you extensions in Safari. Just tap the three-line menu on the left of the address bar, choose Describe Extension, and let the fun begin.

Smarter Photos tools

Spatial Reframing is amazing; it lets you extend backgrounds and reframe subjects using AI. The upgraded Clean Up tool removes distractions even more easily than before. 

Passwords auto-fix

Apple’s Passwords app now does an even better job of keeping your online world safe, as it can now automatically visit a website to sign in and swap a weak or compromised password for a stronger one. 

Parental control improvements

Apple has really thought about child protection. Parental controls have had a huge overhaul, including the introduction of new Ask to Browse, Ask to Buy, contact calling approval and settings to protect minors against nudity, gore, and violence. These arrive alongside new scheduling function which manages which apps kids can use during school hours, meals, or bedtime.

Maps is amazing

Maps now uses aerial imagery and vision intelligence models when you use Flyover. It makes for much more detail, and a real sense of immersion.

Mail is better

While you’ll have to wait for search indexing to complete, Mail has become faster to load with better search results and more accurate unread counts. (Though that’s a little moot for this writer, who currently has 4,719 unread emails.) Spotlight has also been improved with smarter, more relevant suggestions. 

The new operating systems are expected to be made available at around 10am PDT. One more thing? They are bound to include security updates, which you likely need. 

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

Microsoft’s September 2026 Patch Tuesday is the year’s largest release, with 963 CVEs requiring customer action, 106 rated critical. Two are already exploited: CVE-2026-81963 in the Windows Update Stack and CVE-2026-85880 in Advanced Local Procedure Call. Nothing in this release was publicly disclosed ahead of the patch. Readiness recommends a Patch Now scheduling for Windows, Office, SQL Server and the developer tooling, and standard patch release for Exchange. The Readiness team has published an infographic summarising deployment risk by product family.

Known issues

Three client issues are carried in from August close with this update, covering:

• Microsoft Teams and the new Outlook failing to launch on ARM devices, such as the Surface Pro 11 and Surface Laptop 7, is resolved by KB5124008. The issue originated in August’s KB5121003 and was most likely on freshly imaged machines that had not yet taken Microsoft Store updates.
• The same update resolved both desktop backgrounds reverting to solid black and mouse cursor customisation resetting on non-English installations. Both originated in the 27 August preview, KB5120998.

And Microsoft has left open a Microsoft Defender Antivirus notification defect, confirmed on 28 August, in which devices report that Defender is turned off while it is running correctly.

Major revisions and mitigations

The quietest revision window of recent months arrived alongside the largest release. Between the August and September Patch Tuesdays, from 12 August to 7 September, the MSRC Security Update Guide touched 324 CVEs. Of those, 307 were routine Microsoft Edge and Chromium republications requiring no customer action.

That leaves 17 entries affecting Microsoft’s own products, and only one of those is a genuine revision. CVE-2026-59133, an elevation of privilege in the High-Performance Computing Pack, moved to version 1.1, and its own revision note records the change as informational. For comparison, August’s window carried 76 revisions to Microsoft’s own products, 60 of them flagged as requiring customer action.

The Readiness team has reviewed all 963 published updates. Microsoft has published no mitigations and no workarounds anywhere in this release.

Windows lifecycle and enforcement updates

Microsoft has published no new service or enforcement deadlines for September. The lifecycle picture is different, with multiple end-of-support notices for this coming October:

• The retail Office 2021 family retires in full, including Access, Excel, Outlook, PowerPoint, Project, Publisher, Visio and Word.
• Office LTSC 2021 reaches end of support, and that includes Skype for Business LTSC 2021.
• Windows 10 2016 LTSB and Windows 10 IoT Enterprise LTSB 2016 reach the end of extended support.
• Windows Server 2012 and 2012 R2 reach the end of Extended Security Update Year 3. This is the final ESU year for both, not a step to a fourth.
• Windows 11 Home and Pro version 24H2 reaches the end of updates.
• Windows Server 2022 moves from mainstream to extended support, where it stays until 14 October 2031.

A second wave follows on 10 November 2026, and it is the one most likely to catch development teams rather than infrastructure teams. .NET 8 reaches the end of its long-term support branch, PowerShell 7.4 does the same, and Windows 11 Enterprise and Education 23H2 ends servicing alongside Windows 11 IoT Enterprise 23H2.

Microsoft’s test guidance for this release runs to 466 Windows entries, 55 of them flagged as high risk, against 109 and four in August. Given the large number of updates this month (who would have thought that we would be here at 963 CVEs), the Readiness team recommends the following testing priorities:

• Printing. Print from 32-bit and 64-bit applications to physical and virtual printers, exercise XPS and PDF output, share a printer from a print server and print from a separate client, cancel a job mid-queue, and confirm the queue reflects every state change. This is 20 of the 55 high risk flags, and it is the single most likely source of a visible regression.
• Fonts, graphics and imaging. Render varied fonts, sizes and styles across browsers, Office, PDF viewers and Notepad, confirm Print Preview matches the printed page, and open JPEG, TIFF, HEIF and raw images across Explorer, Photos and Office. Watch for clipping, distortion and missing glyphs.
• USB Devices. Exercise device attachment and removal. Attach and remove USB audio, video and mass storage devices repeatedly, including through a hub, pair and unpair a wireless device, and confirm each enumerates and releases cleanly. Device Association carries seven high risk flags, and these drivers load whenever hardware is attached.
• Remote Desktop. Open several concurrent sessions, enable printer, clipboard, audio, drive and smart card redirection together, disconnect and reconnect, and confirm redirected devices reattach.
• Storage (ntfs.sys, spaceport.sys). This is a real hotspot for updates this month, but no high-risk changes. Test on hardware you can recover, because the storage changes reach the boot path, potentially resulting in a dead (test) machine.

Each month, we break down the update cycle into product families, as defined by Microsoft, with the following groupings.

Browsers

Microsoft has not released any updates for their browser products for the second month in a row. September’s Security Update Guide carries no Edge-specific CVEs at all, and the 307 Chromium republications in the revision window required no customer action. The nine Edge CVEs that did appear arrived mid-cycle on 28 August and were serviced through the browser’s own update channel rather than this release.

Estates on a managed Edge channel have nothing to do here beyond confirming the channel is current. It is the one quiet corner of an otherwise heavy month.

Microsoft Windows

Windows carries 726 CVEs, 77 of them critical, which is three-quarters of the entire release. Elevation of privilege dominates by volume at 406 entries, with remote code execution second at 156, information disclosure at 94 and denial of service at 47. The pattern is the reverse of the severity picture: the bulk is local elevation, while the critical-rated entries cluster in the network-facing roles.

• Biometrics is a real focus this month as the Windows Biometric Service takes 64 fixes, the largest single-component count of the year. Most of the security issues are heap overflows that give a local attacker SYSTEM. Windows Hello adds nine, all critical. Patch it promptly on any estate using fingerprint or facial sign-in.
• Windows DHCP Server leads at 36 entries and is topped by remote code execution at CVSS 9.8. Windows DNS Server takes 10 more, also reaching 9.8. Critical remote code execution also lands on Message Queuing, RRAS, Services for NFS, the HTTP Print Provider, Windows Shell, Netlogon, Internet Connection Sharing, SSTP and Failover Cluster, all at 9.8. Patch the resolvers, the DHCP servers and the domain controllers first.
• Storage is the heaviest it has been this year. NTFS takes 29 fixes, Spaceport.sys behind Storage Spaces takes 17, and the Overlay Filter takes seven, with the Volume Manager, VHD miniport, iSCSI and Storage Port drivers behind them. These reach the boot path, so stage them on recoverable hardware.
• The rest of the most-patched tally runs Win32k at 19, Microsoft Standard XPS at 18, Windows Error Reporting at 12, and the Windows Kernel, Windows Search, the Print Spooler and the Device Association Service at 11 each.

Add this Windows update to your Patch Now schedule, with DHCP and DNS servers first and the biometric stack close behind.

Microsoft Office

Microsoft released 137 Office CVEs this month, 24 of them critical, with remote code execution the through-line at 69 entries and information disclosure close behind at 52. September breaks the recent pattern in a way that matters for deployment: this wave is not MSI-only.

• Click-to-Run estates are squarely in scope. Roughly 105 of the Office CVEs reach Click-to-Run, so Microsoft 365 Apps, Office 2019, LTSC 2021 and LTSC 2024 all update, on Windows and on Mac. Word takes 35 fixes, Excel 32, PowerPoint 10 and Outlook seven. The heaviest client entries are CVE-2026-78510 in Word and CVE-2026-78509 in Outlook, both critical remote code execution at CVSS 9.8, in document-rendering paths that fire on preview or open.
• SharePoint Server Subscription Edition takes 16 entries and is the only SharePoint baseline with a package this month. Server updates cannot be uninstalled and always require a reboot, so validate in a maintenance window.
• Skype for Business Server takes 10 entries, led by CVE-2026-66302, a critical remote code execution at CVSS 9.8. Patch it and note that the LTSC 2021 edition leaves support on 13 October 2026.

Nothing in Office is exploited this month. With 24 critical-rated entries and the Click-to-Run channel carrying most of the exposure, the September Office updates belong on the Patch Now schedule regardless.

Microsoft Exchange and SQL Server

Exchange is quiet and SQL Server is not, which inverts the usual relationship between the two.

• Exchange Server takes nine CVEs across 2016 CU23, 2019 CU14 and CU15, and Subscription Edition. None is critical and none is exploited, though the highest reaches CVSS 9.3. The mix runs to two remote code execution entries, two elevations of privilege, two spoofing, and one each of tampering, denial of service and information disclosure. Apply the update from an elevated command prompt, because an un-elevated run leaves Exchange services partially patched and broken, then confirm mail flow, Autodiscover and any hybrid connection before returning the server to service.
• SQL Server takes 62 CVEs, four of them critical, across the 2017, 2019, 2022 and 2025 branches. There is no 2016 package this month. Remote code execution leads at 24 entries, with information disclosure at 22. The critical entries are CVE-2026-67631 and CVE-2026-67643 at CVSS 8.8, and CVE-2026-67378 and CVE-2026-67636 at 8.5.
• Eight SQL packages ship, a CU+GDR and an RTM+GDR for each branch: 2025 (KB5122769, KB5122770), 2022 (KB5122768, KB5122771), 2019 (KB5122772, KB5122773) and 2017 (KB5122774, KB5122775). Microsoft’s guidance is explicit that the baseline or RTM version must be installed first and the GDR patch applied on top; test the install and the removal on every servicing branch you run, then restart the service and confirm Always On availability groups stay healthy.

Exchange goes on the Schedule list, and SQL Server goes on the Patch Now list. That is an unusual split, and it is driven by the four critical remote code execution entries on the database estate.

Microsoft Developer Tools

Microsoft released 24 CVEs across its developer tooling this month, 23 rated important and one critical. Security feature bypasses are the main focus with eight CVE entries, with remote code execution and information disclosure at four each.

• The single critical entry is CVE-2026-34182 at CVSS 9.1, a flaw in CMS AuthEnvelopedData processing that allows forged messages to be accepted. It reaches Visual Studio 2017 through 2022.
• The highest-scoring entry in this family is not the critical one. CVE-2026-81376, a Visual Studio Code security feature bypass, reaches CVSS 9.6 while carrying an important rating. It is a useful reminder that severity labels and CVSS scores answer different questions.
• Visual Studio Code and its Copilot extensions take 10 entries, mostly security feature bypasses. Update the editor and confirm workspace trust prompts, extension installation and remote sessions behave as configured.
• .NET ships SDK updates on all three supported lines, x64 and x86: 8.0.131 and 8.0.425, 9.0.121 and 9.0.318, and 10.0.112 and 10.0.401. Install them, then build and run a representative project to check for regressions. Keep the 10 November date for .NET 8 in view while you are in there.
• The .NET Framework ships monthly rollups per operating system: Windows Server 2012 (KB5126147), Server 2012 R2 (KB5126148), Windows 10 1809 (KB5126144), 21H2 (KB5126145), 22H2 (KB5126146) and Server 2022 (KB5126149). One gap worth noting: the 4.7.2 package for Windows 10 1607 is listed as pending and will follow, so estates still on 1607 will not complete their Framework patching this cycle.

Add these to your standard release schedule, behind this month’s Windows, Office and SQL Server priorities. Keep the 10 November date for .NET 8 and PowerShell 7.4 in view while you are in the developer estate, because a patch this month does not extend either branch.

Adobe (and third-party updates)

September is the largest release of the year, and this (crazy, super high) volume is the least interesting thing about it. The 963 CVEs matter less than the 55 entries Microsoft flags as high risk, and those sit in printing and fonts. Adobe shipped two Acrobat builds one digit apart and only the second is a security update (nothing to worry about here). Of the CVEs Microsoft republished, 25 are not Microsoft’s. A version number tells you very little about the work in front of you. So, given my (super-secret knowledge) of how Microsoft operates over the summer, here is my prediction for next month (October). It won’t be as big as this month – but just you wait – November is going to be big. Let’s up those numbers (or not).

Long before Taco Tuesday became part of the pop-culture vernacular, Tuesdays were synonymous with security — and for anyone in the tech world, they still are.  Patch Tuesday, as you most likely know, refers to the day each month when Microsoft releases security updates and patches for its software products — everything from Windows to Office to SQL Server, developer tools to browsers.

The practice, which happens on the second Tuesday of the month, was initiated to streamline the patch distribution process and make it easier for users and IT system administrators to manage updates. Like tacos, Patch Tuesday is here to stay.

Patch Tuesday coverage has long been a staple of Computerworld’s commitment to provide critical information to the IT industry. That’s why we’ve gathered together this collection of recent patches, a rolling list we’ll keep updated each month.

In case you missed a recent Patch Tuesday announcement, here are the latest six months of updates.

September Patch Tuesday: 963 CVEs, 2 exploited flaws, 1 message

September’s Patch Tuesday is Microsoft’s biggest of 2026, with 963 CVEs, two exploited flaws and a clear message: prioritize Windows, Office and SQL Server.

Two vulnerabilities are already exploited: CVE-2026-81963 in the Windows Update Stack and CVE-2026-85880 in Advanced Local Procedure Call. Nothing in this release was publicly disclosed ahead of the patch. Readiness recommends a Patch Now scheduling for Windows, Office, SQL Server and the developer tooling, and standard patch release for Exchange.

Get more info on the September 2026 Microsoft security updates.

August’s Patch Tuesday is a monster: 751 fixes, with an exploited Windows flaw

Microsoft’s August 2026 Patch Tuesday closes at 751 CVE entries (across all product families), with 108 rated as critical. One flaw is already exploited, CVE-2026-68820, an elevation of privilege in the Windows WinSock driver (afd.sys). Two more were disclosed but not exploited.

This security-only release earns Patch Now for Windows, Office and Exchange; no SQL Server updates this month. Unfortunately, several critical issues affect server roles: Windows DNS Server carries a cluster of critical RCEs; Windows DHCP Server is the most-populated Microsoft product family at 14 entries. Testing should lead with printing and fonts and the Remote Desktop client, then a WinSock smoke test given the exploited afd.sys flaw.

Get more info on the August 2026 Microsoft security updates.

July’s Patch Tuesday sees an end-of-support collision amidst a massive, record-setting patch wave

Microsoft addressed 722 CVEs this month once the 427 Chromium upstream relays are set aside — roughly three times a normal cycle and one of the largest single months in recent memory. Two vulnerabilities arrive under active exploitation: an elevation of privilege in Active Directory Federation Services (CVE-2026-56155), and an elevation of privilege in SharePoint Server (CVE-2026-56164). A third, a BitLocker security feature bypass (CVE-2026-50661) is publicly disclosed but not yet exploited.

The July 2026 Patch Tuesday earns Patch Now recommendations for Windows, Office, Exchange, and SQL Server. SharePoint has two critical RCEs on top of its exploited zero-day, and Exchange Server returns with a critical on-premises spoofing flaw. Adding to our (dear) administrator’s efforts, SharePoint Server 2016/2019 and SQL Server 2016 all reach end of support today.

More info is available here on Microsoft Security updates for July 2026.

For June, Patch Tuesday means an IT scramble

Microsoft this month released 206 updates affecting Windows, Office, Exchange Server, and its developer tools — including three Windows vulnerabilities already publicly disclosed. That trio includes an elevation of privilege in the Collaborative Translation Framework (CVE-2026-45586), a denial of service in HTTP.sys (CVE-2026-49160), and a BitLocker security feature bypass (CVE-2026-50507). At the moment, none appear to be under active exploitation, but all three are rated “Exploitation More Likely.” 

Even without an exploited zero-day, the June 2026 Patch Tuesday release requires Patch Now recommendations for Windows, Office, and Exchange. The latter is back in the patch picture with a consolidated security update that Microsoft recommends installing “as soon as possible.”

More info is available here on Microsoft Security updates for June 2026.

For May, Patch Tuesday means 139 updates — but no zero-days

Microsoft this month released 139 updates affecting Windows, Office, .NET, and SQL Server (though there were no updates for Microsoft Exchange Server). Despite the absence of zero-days, the May Patch Tuesday update still requires Patch Now recommendations for Windows and Office. 

The combination of three unauthenticated network RCEs (Netlogon, DNS Client, and SSO Plugin for Jira and Confluence), four Word Preview Pane RCEs, the large TCP/IP vulnerability cluster, and the carry-over BitLocker recovery condition (still active on Windows 10 and Windows Server) warrants an accelerated deployment release schedule. 

More info is available here on Microsoft Security updates for May 2026.

Microsoft’s Patch Tuesday release for April is a whopper

Windows admins are going to be busy this month, dealing with the largest Patch Tuesday cycle in memory. The April release involves 165 updates and roughly 340 unique CVEs from Microsoft — including two zero-days, one of which is already being actively exploited in the wild. 

The Readiness team recommends “Patch Now” schedules for nearly every major product family: Windows, Office (with a zero-day), Microsoft Edge (Chromium), SQL Server, and Microsoft Developer Tools (.NET). April also brings Phase 2 of Microsoft’s Kerberos RC4 hardening with full enforcement set for July. There is a lot to cover, so here’s a useful infographic mapping the deployment risk for each platform.

More info is available here on Microsoft Security updates for April 2026.

IDC analyst Francisco Jeronimo said something interesting following the iPhone Duo introduction on Wednesday: “Apple showed up years after everyone else, yet it walks in and sets the rules.”

He’s right. From the IP resistance to the little touches that set the new foldable apart from most of the competition, Apple has set expectations for the entire market, and while competitor Samsung is trash talking the foldable, even it stands to benefit from the millions Apple is about to sell. Apple is, after all, allegedly paying Samsung $250 for each iPhone Duo display.

But it is not just the rules concerning foldables the company has now defined; it has also set the standard competitors must follow in chip design with its A20 Pro, a super-powered processor that delivers noticeable performance improvements even compared to last year’s iPhone 17. 

What’s new in the chip?

With six CPU cores, seven GPU cores and a new Dual 16-core Neural Engine, the new chip boasts numerous improvements compared to the A19 Pro, not the least of which is that it’s the first 2nm chip Apple has put inside a smartphone. These enhancements bestow it with a series of superlatives:

• Memory bandwidth is 50% higher than last year’s chip.
• GPU graphics performance increases up to 40% over the A19 Pro.
• AI performance is expected to double.

These are all significant improvements that make for noticeable experience boosts for anyone using one of these phones. Prosaically, I anticipate first reviews will celebrate how fast and responsive these devices are, and this will be most evidenced in tasks such as video capture and photography. Pictures will capture more swiftly, open faster, and edit features will become (and already are) more advanced and sophisticated than before. Apple spent ample time celebrating the iPhone 18 Pro series for the new frontiers in digital photography they open up, and not just because of the new reflexive lens.

The point is that the devices have the horsepower to get through computationally intensive tasks, including AI, thanks to the Neural Engines and GPU.

Thermal management is the icing on the quake

What those tasks have in common is heat, which Apple’s product designers have addressed in two critical ways: liquid cooling and processor design.

The processor design is the important thing, because in this case Apple has moved away from the InFO POP packaging of the A19 Pro in which the DRAM is stacked above the chip. It’s now embraced a side-by-side design it described as “inspired by” M-series Apple Silicon.

That works because placing both components alongside each other, rather than in a perpetual embrace, reduces the overall temperature of both by putting a little space between them. As Apple explained it, this Wafer-level Multi-Chip Module (WMCM) approach means the memory gets out of the way of the processor, allowing the latter to make direct contact with the vapor chamber cooling system.

In layman’s terms, it means both memory and chip can perform more intensively while being more efficiently cooled, a move that helps Apple’s iPhone meet those astonishing performance achievements. The vapor chamber is also bigger, made of new material with three times the surface area of before, making it even better at cooling the device.

It also means Apple has achieved a lot of things competitors have not, most particularly in packaging, the use of 2nm chips in a mass market product, and overall design.

That’s where it gets interesting. Both Qualcomm and MediaTek now plan to follow in Apple’s footsteps with the adoption of 2nm chips and side-by-side packaging.

Digitimes suggests it is a move others will follow, as vendors work furiously to ensure their hardware can run AI effectively on device, which iPhones already do. Apple, of course, already knows its Macs, iPhones, and iPads have been designed from the ground-up to support on-device AI as it works to provide the world’s best mass market platforms for secure and private use of the tech.

Friends don’t let friends bet against Apple

All the same, even at the microprocessor level, it’s hard to ignore that where Apple goes, others follow — even while where others go, Apple frequently follows later, but seems to do it better. It has never been a company to bet against, after all. People really should stop doing that.

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

Apple does seem to have tried to ensure its controversial Audio Intelligence feature isn’t abused. Alongside the iPhone Duo and new iPhone 18 Pro range, Apple on Wednesday introduced a new Apple Watch equipped with a brand new feature it calls Audio Intelligence. The existence of this surprising tool was confirmed only on the eve of the launch event. Enabled by the new S11 chip on the latest Watch devices, this is actually a collection of four features:

• Siri Recap, which creates high-level summaries of conversations during your day.
• Live Rewind, which transcribes the previous 15 seconds of a conversation as text.
• Music recognition, which relies on Shazam.
• Sound recognition, which can identify things like sirens or alarms.

All of these features must be consciously enabled by the user and are not on by default.

What they have in common is use of artificial intelligence along with always-on microphones. While all four of these tools are being presented as opt-in, their existence will inevitably — and justifiably — raise privacy concerns. 

Apple saw those concerns coming, and to make the features work it has placed a Secure Exclave on the S11 chip inside the latest watches. This is a dedicated secure buffer on the chip that processes audio privately on the fly. It does so without creating a recording of ambient sound around you, and the data it works with is isolated from the rest of the system. That Exclave pairs with the iPhone’s Enclave using “a new audio-verified pairing mechanism that exists in addition to Bluetooth pairing,” Apple explains in a detailed white paper about how it works. 

For Siri Recap, the process works like this:

• Apple Watch recognizes a conversation is taking place.
• If so, audio flows into the Secure Exclave, encrypted, and transmitted to a similarly Secure Enclave on your paired iPhone.
• The audio is then immediately deleted from the Watch. 
• The phone will decrypt and transcribe the audio, condense it, and send that transcript to Private Cloud Compute for processing. 
• The short summary is then made, returned, and deleted after seven days.
• Speakers are not identified, recordings are not made, and detailed transcripts are not created or retained.

The company has also introduced safeguards for those around you. The system will deliberately omit some information, including potentially harmful content, financial data, or personal identifiers. When you use the Live Rewind tool, an audible chime plays on your device to alert nearby people that you are using the feature.

Challenges will emerge

For one thing, we don’t yet know whether the company will provide enterprise IT with device management tools to disable the feature on managed devices. It’s crystal clear that devices that are constantly gathering data will be seen as potential security risks — particularly in regulated industries. And it seems far more logical to provide new APIs to disable Audio Intelligence on managed devices than it would be to insist anyone wearing an Apple Watch put it in a lead-lined box before beginning the next safeguarding, healthcare, or product development meeting.

Another problem I see concerns Europe’s Digital Markets Act. It is, after all, inevitable that competitors (maybe including Meta) will want their devices to have equal access to the information gathered by Audio Intelligence. Based on the decisions Apple has made so far, it seems equally likely it will want to refuse such access; this is why these new features will not initially be available in the EU. 

There is also no doubt hackers will attempt to break into the system, though doing so will not be at all easy on account of the intentional way the company has built in security. I suspect, but do not know, that attempts will focus on the points at which data is exchanged across devices, rather than when the information sits within secure enclaves on those devices. It also makes sense that everyone who uses iCloud services for their data should put 2FA in place and pick strong passwords. 

What about iCloud storage?

The other challenge will be encryption, specifically ongoing attempts to penetrate iCloud data encryption by some nations, led by the UK. Even so, if access could be achieved to iCloud-stored Audio Intelligence text, what is obtained would only be summary data, not a recording. The system is architected so recordings are never made.

Apple’s white paper on the feature explores this in more depth: “Audio from the microphone enters the Secure Exclave of Apple Watch, where it is initially processed for speech, sounds, or music, without transcribing or storing the raw audio. This buffer is a continuously overwritten stream that exists only within the protected hardware and never creates an audio recording.”

Once you decide to keep a Recap or Life Rewind text, the raw audio is not saved to iCloud, only the text, and then only if you use 2FA and a device passcode. No one else, including Apple, can access the encrypted data you save, Apple said.

Where does this go next?

One pre-event rumor that didn’t come true concerned AirPods equipped with cameras and Vision Intelligence to understand physical context and surrounding, with the aim of enhancing Siri AI. I’ve expressed reservations about this idea, but do think the security model Apple has put in place for Audio Intelligence will turn out to be an echo of its intentions to secure Vision Intelligence transactions.

This implies a new chip with Silicon Enclave for AirPods Pro, a similar exchange of information in real time for summary and determination of context and content, and a process in which no recordings are made and media access beyond brief, time-limited summaries, is available. In short, Apple will use the same security model.

What about the rest of them?

That’s fine as far as it goes. But the other subtext is that while Apple seems genuine in its attempt to deliver relatively intrusive tech advances right, others will show less commitment to privacy and security. So, while it’s furthering the conversation about using these tools, Apple is also pushing public acceptance of such technologies.

I’m not entirely sure we’re ready.

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

Qualcomm is taking an unusual approach to the launch of its next-generation premium Snapdragon mobile platform this year. Instead of holding everything for its annual Snapdragon Summit in Maui happening later this month, the company has been methodically disclosing the architecture that will power the next wave of flagship Android phones over the past couple of weeks.

First, the company disclosed its new Oryon CPU architecture, followed by a major overhaul of its Adreno GPU. Today, Qualcomm is detailing the next-generation Hexagon NPU that will serve as the platform’s primary AI engine.

Taken together, the disclosures give us a pretty good picture of where Qualcomm is going before we even know the chip’s official name.

There is obviously a strong performance story here, anchored by a 5GHz CPU, but the more interesting theme running through the architecture is memory locality and specialized AI acceleration. Qualcomm is trying to keep more data close to the compute resources that need it, while distributing AI workloads across the CPU, GPU and NPU.

These architectural changes are important as the company continues to drive on-device AI beyond relatively simple generative features toward more persistent, agentic workloads.

Oryon hits 5GHz, bolstered by FlexCache

Qualcomm disclosed in August that its next Oryon CPU will be the first mobile CPU to reach 5GHz. The eight-core design includes two 5GHz Prime cores and six Performance cores, with Qualcomm attributing the frequency gains to its fully custom Oryon microarchitecture, an in-house Arm-based CPU design that gives the company full control over the cores, cache and memory hierarchy, branch prediction and CPU subsystem.

The 5GHz headline is certainly going to get attention. However, Qualcomm’s new FlexCache architecture may ultimately have a larger impact on sustained performance. FlexCache allows Oryon’s Prime and Performance cores to tap into the same shared cache pool, with capacity dynamically allocated based on the workload. So rather than a demanding core being limited to a fixed slice of cache, it can draw on more of the available pool when needed. This keeps larger working data sets close to the CPU cores and reduces high-latency trips out to system memory.

This approach should benefit everything from gaming and multitasking to content creation, but it also aligns well with agentic AI workloads, where tasks may move through several stages of processing and across different CPU cores. The basic goal is straightforward: keep the CPU fed and avoid expensive trips to external memory.

Qualcomm’s next-gen Adreno GPU adds dedicated AI engines

Qualcomm’s next disclosure focused on graphics, where the company is making one of the more significant changes to its Adreno GPU engine in recent years. The new GPU adds dedicated Adreno Matrix Cores for running AI models directly inside the graphics pipeline. They’re paired with 18MB of Adreno High-Performance Memory, or HPM, which provides low-latency local storage for tile-based rendering, frame buffers and GPU compute.

Qualcomm claims HPM provides a 12% power improvement compared to its previous-gen Snapdragon 8 Elite Gen 5.

The more visible feature for users, however, may be Adreno Neural Fusion, which combines AI super resolution, neural processing and frame generation in a unified graphics pipeline. There are obvious parallels here to what NVIDIA, AMD and others are doing with neural rendering on the PC, though Qualcomm’s next-gen Snapdragon SoC has to operate within a much tighter mobile power envelope.

Qualcomm claims enabling Neural Fusion reduces power consumption by up to 40% in its internal testing using the company’s Dragon Alley graphics demo. That’s an impressive number, though as always, we’ll want to see how the technology behaves across actual shipping games and devices before drawing any definitive conclusions.

Qualcomm has also helped integrate Neural Fusion technology into Unity and Unreal Engine, which may ultimately be just as important as the underlying hardware. Developer adoption determines whether features like this become meaningful platform advantages or just another tick box on a spec sheet.

Qualcomm’s Hexagon NPU is being reworked for agentic AI

The most recent disclosure centers on Qualcomm’s Hexagon NPU, where the broader architecture starts to come together. The company’s next-generation Hexagon introduces a new Element Accelerator designed for transformer workloads, alongside its existing vector and scalar processing resources. Qualcomm says the new block is optimized for fast action loops, key-value (KV) cache acceleration and context lengths of up to 32K.

Hexagon is also getting 50% more shared memory, and again, that memory complement is important. Model state, context and KV-cache data can generate substantial memory traffic, particularly with larger language models. Keeping more of that data close to the NPU can reduce latency and power consumption. Qualcomm says the architecture is designed for long-context reasoning, multimodal models, concurrent agents and low-latency action loops.

For INT4 quantized models, the company claims up to a 50% improvement in pre-fill performance versus its previous-generation Snapdragon 8 Elite Gen 5. Qualcomm is also targeting Mixture-of-Experts, or MoE, models as large as 30 billion parameters. In its example, the company points to a 30B model that activates only about 3 billion parameters during a given token-generation step. That selective approach is a natural fit for the tight power and memory constraints of a handset.

Qualcomm isn’t suggesting that a smartphone will simply run a 30B dense model entirely from DRAM, however. By activating only the experts needed for a particular task, MoE architectures can dramatically reduce active compute and memory bandwidth requirements, potentially making much larger classes of AI models practical on mobile hardware.

Qualcomm’s next-gen Snapdragon mobile impact

There is a larger competitive story here as well. For years, flagship smartphone competition largely came down to CPU, GPU, modem performance and power efficiency. AI is now another major area of differentiation, and Qualcomm is clearly designing its next-gen Snapdragon mobile platform around this need.

Apple has the advantage of controlling its silicon, operating system and software stack end-to-end. MediaTek continues to push aggressively into premium Android devices as well. Qualcomm’s response is to make its custom Oryon CPU, Adreno GPU and Hexagon NPU work more like a fully integrated compute platform, tuned for modern AI and agentic workloads.

This is a solid advantage for Android handset OEMs because many don’t have the resources to engineer this level of silicon integration themselves. Qualcomm can effectively provide its Android OEM partners, including Samsung, Xiaomi, Honor and others, with a common hardware foundation for on-device AI, advanced graphics and agentic computing. It also gives Qualcomm another way to defend its premium Snapdragon position.

The company’s custom Oryon architecture now spans smartphones and Windows PCs, and soon servers, while AI acceleration is distributed across its CPU, GPU and NPU. For business users, that could mean more AI processing happens locally, reducing cloud and network dependence, improving response times and keeping more potentially sensitive data on-device. And for IT organizations, Qualcomm now has an opportunity to provide a more consistent AI compute foundation across Windows PCs with Snapdragon X and premium Android handsets.

There is still plenty we don’t know, however. Qualcomm hasn’t disclosed the complete SoC specifications, final performance or power characteristics, or which devices will ship with it initially. Many of these AI experiences will also depend heavily on Android, app developers and the models themselves.

The new 5GHz mobile CPU may generate the biggest headline at Snapdragon Summit, but the more important developments may be what sit around it: more local memory, specialized AI acceleration and tighter integration between the CPU, GPU and NPU.

If Qualcomm can translate this architecture into better sustained performance, enhanced power efficiency and useful on-device AI experiences, it could strengthen its position in premium Android phones while putting more competitive pressure on Apple. For business users, the payoff could be more capable local AI without sacrificing the performance and battery life expected of a flagship device.

That is the part I’ll be watching most closely when the complete silicon picture is unveiled at Snapdragon Summit later this month. And I may even get some hands-on time with it in some benchmarks, as we have in years past, so we shall see.

Twenty-five years ago, I sat near the front of the room while then-Apple CEO Steve Jobs rolled out Apple’s digital hub strategy, a vision in which the Mac would become the central manager of all manner of portable digital devices.

Now, a quarter of a century later, nearly all of those digital devices have become features on your iPhone, and the arrival of AI means the iPhone has become the intelligent personal hub, the central manager of, well, of you. Apple’s new CEO, John Ternus, went straight for the concept in his important opening remarks at Wednesday’s big iPhone launch, where nearly all the pre-event speculation came true.

The intelligent personal hub

“As we look ahead, we see a future filled with enormous discovery and transformation,” Ternus said at the Surprise and Shine event. “AI makes entirely new kinds of experiences possible. An AI can become even more useful when it brings together your life with the apps, services, and products you use every day. And this means the product at the center of your experiences becomes even more essential — what I like to think of as an intelligent personal hub.”

(To be clear: He’s talking about the iPhone.)

Think about it: all those tasks you now do on your iPhone are tasks AI can help you with. Siri AI lets you work on files and folders using spoken commands, while contextual AI lets you sift through all the data once gathered by your iPhone to do things, make things, create networks, or even organize luncheon. 

“It’s the intersection of broad new capability, and a deep understanding of your personal context that makes this idea so powerful,” Ternus explained.

AI’s brave new world

That’s the context within which Apple will be introducing all its future products, a battle the company now feels confident marching into now that it appears to have resolved its historical challenges with AI. At the same time, the company also seems to be doing its best to lean into privacy — enabling life-changing AI transformation of daily lives while putting a brake on scary privacy attacks. This is going to be foundational to what happens next at Apple.

Of course, you’re not reading this to look too deeply into what Ternus said about the future of Apple; you’ll also want some insight into the slew of devices the company introduced at its event. That list includes the:

• iPhone Duo — the long-anticipated folding iPhone, which is finally here (or rather, will be here in October).
• iPhone 18 Pro and Pro Max — the next important evolutions of what was once Apple’s highest-end iPhone range.
• AirPods 5 — now with even better tech, at the same price as before.
• Apple Watch Series 12 and Apple Watch Ultra 4 — the latest iterations of the company’s popular wearables.

The big surprise? Price

For some, the biggest surprise in these introductions was price. All of the accessories were kept at the same price as last year’s equivalents, while the new Pro iPhones came in at only $100 more for the entry-level model, rising to around $300 more at the more price-resistant highest end. An iPhone 18 Pro Max equipped with the max 2TB storage comes in at $2,499 compared to the $1,199 iPhone 18 Pro entry point. 

When it came to price, though, all eyes were certainly on the iPhone Duo, and Apple has given everyone (except competitors) a pleasant surprise on that. 

With a starting price of $1,999, the folding phone’s price hit the low end of expectations, and while that grows to an eye-watering $3,199 with 2TB of storage, it compares well with the market leader, Samsung. While the Galaxy Z Fold 8 begins at $1,899, in terms of features and design, Apple’s folding phone lines up with the Galaxy Z Fold 8 Ultra, which costs from $2,099. 

Dig a little deeper and you’ll see that while the iPhone has IP68-rated dust and water resistance, the Samsung equivalent has only IP48. You can drop the Fold 8 into five feet of water for up to half an hour and it should be OK, but you can chuck your iPhone Duo in almost twenty feet of water, and it should survive. Though you probably shouldn’t test that unless you’re a hugely successful “influencer” already generating a small nation state of money through clicks. If that’s you, good for you. I’m only slightly envious. 

A productivity powerhouse?

What that suggests is that the iPhone is less likely to get damaged and is better engineered, though it’s also ever so slightly thicker. Apple’s device is powered by what the company modestly described as the most powerful mobile processor on the planet — which it is, particularly for the AI tasks all the new iPhones can handle. That matters, particularly for applied contextual AI deployed across daily lives.

It also matters because the combination turns Apple’s device into a mobile productivity powerhouse, one that I think will likely affect iPad sales more than anything else. Though this wouldn’t be the first time a new Apple release ended up eating its siblings. 

The productive possibilities of Apple’s premium device will also be realized as app developers embrace both the form factor and Siri AI. Apple said during the keynote that more than 300,000 apps already work with Siri AI and more than 2.5 billion Siri requests are being made each day. That latter number will only grow now that Siri performance has been dramatically improved. 

While this argument applies to all Apple’s new devices, the fact of the matter is that you now have an iPad-sized object you can carry in your pocket that is capable of doing some really powerful tasks by spoken command alone. That’s the iPhone Duo – a product of an unexplored territory at the crossroads between mobile computing, daily digital existence, and the opportunities of AI.

It’s not for everyone, yet

I don’t expect the Duo will be for everyone; the vast majority of users won’t be spending $2,000 on a phone. But we know that the cost of technologies like this tends to decline, which suggests that in perhaps five years’ time, it will be possible to create equivalent devices for a much lower price. Right now, it’s for affluent Apple consumers, C-class executives, students, AI pros, medical professionals and others who might actually need this kind of performance and display space to get things done.

This, incidentally, is precisely the same group of people that first embraced the iPad when it appeared, turning that into the world’s best-selling tablet.

What might this mean for Apple? 

I traded a few messages with IDC analyst Francisco Jeronimo, who sees it like this: “Apple entered the foldable market and, in one keynote, set the price and the standard every rival will now be measured against.”

He also pointed out, the iPhone Duo is “…a direct test of whether Apple can use design, ecosystem integration and premium positioning to reshape a segment that has so far remained selective.” 

He’s right, of course, but Ternus also just defined what that segment, and the industry, is driving for. I’s the same thing Apple was aiming at 25 years ago. “iPhone sits at the center of an amazing ecosystem of intelligent features and experiences that work seamlessly across the products you use every day,” the company said — an “intelligent personal hub.” And the Macs are all right, too.

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

Adobe on Wednesday unveiled a range of updates to Acrobat, including the ability to turn PDF documents into interactive visual reports and podcasts. And a new Knowledge Base feature connects Adobe’s Acrobat AI assistant to document sources such as Microsoft SharePoint and Google Drive.

It’s the biggest Acrobat update in several years, Adobe said, and highlights the application’s evolution from PDF reader to what the company now describes as an AI-powered document productivity and creativity platform. 

Acrobat users open more than 400 billion PDFs in the app each year, Adobe said, claiming 650 million monthly active users as of last year. 

Among the new additions is the ability to turn dense PDFs into interactive, visual documents and presentation slides. To create these, Acrobat’s AI assistant scans the selected file, asks questions about the user’s intent, then generates documents charts, images and navigation.

It’s also possible to turn documents and web links into podcast-style audio summaries — similar to capabilities in Google’s NotebookLLM and Microsoft Copilot Notebook. A new Stylize feature can enhance plain-looking documents, applying a template to turn them into “professional deliverables” such as CVs and invoices, Adobe said.

The audio and visual generation features are all available in paid Acrobat plans (Acrobat Studio, Acrobat Express and Acrobat AI Assistant). 

Another addition, Acrobat Knowledge Base, lets customers connect the Acrobat AI assistant to a wider range of file types, enabling users to ask questions about information held in documents stored in Microsoft SharePoint and Google Drive, for example.  

Acrobat’s Knowledge Base “transforms curated document collections — policies, playbooks, product information, pricing guidance, research — into cited answers available from Slack, Microsoft Teams, Acrobat, and other places employees already work,” Abhigyan Modi, senior vice president for Adobe Document Cloud, said in a blog post. 

Adobe also added an Analyzer tool designed to retrieve data from large volumes of documents. This could involve analyzing thousands of contracts or invoices to identify information such as renewal dates and revenue clauses.

That turns information once trapped in files into data that can inform finance, procurement, compliance, and the systems that support them, said Modi.

Analyzer is the more interesting of the two releases, said Mike Leone, vice president and principal analyst at Moor Insights & Strategy. “Knowledge Base is a search product and search products get judged feature against feature. Analyzer is closer to a data product, because that extraction work is what companies currently pay people and pipelines to do,” he said. 

“What I’d want Adobe to answer is whether that data can leave Acrobat and land where the rest of the company’s data lives. If it can, this starts competing for real data budget. If it stays inside the app, it’s a very good Acrobat feature.” 

Access to Analyzer and Knowledge Base requires an Acrobat Studio for Enterprise subscription. (Adobe doesn’t publish pricing.) 

In addition, Knowledge Base includes a credit-based model that places some restrictions on usage, with licensed users allowed up to 1,000 queries a month. Other actions, such as document uploads, queries from Slack and Microsoft Teams, and queries by employees without an Acrobat Studio license, draw varying amounts of Acrobat Studio “shared credits.” Adobe didn’t say how much additional credits cost once a yearly limit is reached.

 More information about Adobe’s Knowledge Base credit usage policy is available on Adobe’s website. 

Analyzer also includes a credit system that limits document ingestion and attribute extraction.

For enterprise customers, usage-based pricing can add complexity when managing employee access.

“The issue for a broad deployment is that headcount barely predicts usage,” said Leone. “One person pointing it at 10,000 contracts will burn more credits than a hundred people asking the occasional question. Pooling the credits across the whole organization is the right design, because usage on something like this is rarely spread evenly.

“One of the big challenges is that people could very well start rationing themselves because they don’t know what a question costs,” he said. “Or worse, a couple folks use all the credits and everyone else is out of luck. The last thing any organization wants is to pay for a knowledge tool that people won’t use or can’t use.”

Late-breaking reports suggest Apple has built a new Siri AI ‘Recaps’ feature that summarizes your day for you. The idea sounds great until you recognize that this summary will also include overheard conversations. 

Suspending the huge red flag of concern over privacy and data protection for a moment or two, why would this be useful and who is it for?

How it’s supposed to work

First, it is being described as an Apple Watch feature that will only be available on the latest models: Apple Watch Series 12 and Apple Watch Ultra 4, as these allegedly include secure audio processing hardware. Reading between the lines, that suggests the devices will be able to listen to and process audio on device, which implies that the information won’t need to leave your smartwatch, not even to your iPhone.

The report also says the feature doesn’t make a transcript, nor does it store the audio; it instead gathers key datapoints from what is said to include in the summary, deleting all the remaining information in the process. You can have the feature working all day, on-demand, or only in specific locations. The idea seems to be to provide some kind of actionable summary of events and key conversations that take place during the day. I imagine the subtext is to digitize as much as possible of your daily existence as Apple thinks you will be comfortable with to create the kind of datapoints contextual AI requires if it is going to help with your life.

The case for it

That sounds fine, up to a point — the idea that a digital assistant will be able to recognize key events, moments, and requirements to remind us of them, act on them, or suggest next steps concerning them is certainly a notion that’s gained currency in tech circles. They see it as augmenting human achievement, enabling people to become more productive, more efficient, and capable of doing more with less effort. At least, that’s part of the argument.

What is wrong with that (other than concerns over privacy or a desire not to have one’s entire waking life transformed into a series of datapoints that can be measured, tested, or surveyed by entities outside of our control), right? If you have nothing to hide, you have nothing to fear, at least, not until times change and what didn’t need to be hidden in the past becomes something that must be shrouded in future.

Apple’s privacy defense

Apple, of course, has its strong privacy story to lean into as it introduces this surveillance-as-a-service solution; it’s an argument that says it recognizes how this tech can be abused, and wants to deploy it correctly. It will point to that record as it promises its system doesn’t gather data, doesn’t collect it, and doesn’t keep any form of transcript on or off the device. The company should be commended for building a system that works this way, recognizing as it does that the best way to protect confidential information is not to gather it in the first place.

The warning

But what Apple has also done with this new speculated upon — and as-yet-unconfirmed product — is to show us what these tools can already do. Much of this is already visible; take recent news that LG smart TVs constantly collect and upload user data as an illustration of the seemingly egregious ways in which such tech is used. (LG denies the claim.) What Apple shows us is that with AI in the picture, all your waking moments can be collated, curated, analyzed, digitized and rendered into actionable data. Apple may not be doing that, but others will not be equally cautious, so we owe Apple a debt for warning us of what’s at stake.

Though we don’t know whether Apple will actually introduce this new service, the source of the claims is one of the strongest in the industry. What we do know is that the technology required to build systems like this exists, along with AI systems capable of sourcing, sharing, and sifting through that information. What we don’t yet know is how to remove ourselves from the data stack.

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

We’re hurtling toward the loss of something big. Finally, after months of speculation, Apple is about to introduce its first-ever folding smartphone. And once it does, the conversation will change as the annual iPhone rumor season is replaced with the reality of whatever Apple unveils.

Except, it won’t quite be the same as in the past. Because no sooner will new Apple CEO John Ternus finish his iPhone announcement keynote speech tomorrow than the speculation will shift to next year’s 20^th-anniversary iPhone, what’s coming to the iPhone 18 and e-series iPhones, and expectations around Apple’s upcoming wave of smart home products, including health and home security services. There’s a lot going on.

Meet John Ternus

Having a lot going on is one of the prime expectations of Apple under its new boss. A family man with a wife and kids, he’s led Apple’s hardware division for five years after an eight-year stint running hardware engineering. He joined the company in 2001 and is described as a highly competent, hugely talented, and modest by nature. 

His hobby is quite revealing — he drives race cars, is a regular at California’s Laguna Seca Raceway, and owns an incredibly rare 2019 Porsche 935. To me it suggests a calm, competent, individual with excellent mental and physical reflexes who can make life or death decisions at speed. He’s also deeply committed to good product design.

He’ll need that calm embrace of speed and adrenaline to manage the company his predecessor, Tim Cook, led to become the world’s most valuable firm. 

A hard act to follow

He’s also got a hard act to follow, as while Cook had to carve out his reputation in the shadow of Steve Jobs, Ternus must sculpt his own destiny in the shadow of Cook, who managed to make shareholders of record in 2011 around 20 times richer by the time he quit. (Apple stock rose from $13.74 per share on a split adjusted basis to $319.97 per share today.)

While Cook was frequently criticized for not being a product guy, he presided over the launch of the Apple Watch, AirPods, Apple Silicon, and the Vision Pro, and helped grow Apple’s services income to become its second-biggest product category, diversifying Apple’s revenue and creating financial stability. That’s no easy task.

Building his own team

Can Ternus improve on that? He’s certainly being positioned as someone who can. His deep experience in hardware has led to expectations that Apple will double down on product under his watch, so a great deal is expected of him in consequence. He is also putting together a hardware-focused team, including Laura Legros, former vice president of hardware engineering, as a direct report.

What will we see of this tomorrow? To be fair, while he will lead the announcement on the keynote stage, development of this first tranche of products took place under the leadership of Cook, and while Ternus will have had a big hand in development, this year’s Apple event should be seen as a moment of transition. 

“I’m just as excited about what lies beyond that, including the incredible products already in the works and the ones we haven’t even imagined yet that we’ll dream up and create together. There’s no team in the world I’d rather build the future with,” Ternus told Apple staff in his first memo as CEO.

It’s what Apple does in the next couple of years that will truly define the new leadership. He’s literally been CEO for a week!

What to expect on Sept. 9

Here’s my TL; DR list:

• First folding iPhone: Apple’s response to folding devices from Samsung, Google, Huawei, Xiaomi. It’s expected to have a highly durable, near invisible hinge and a powerful 2nm A20 chip around 20% more performant than the 17 series devices; Touch ID; a 48MP camera; and be able to unfold from a 5.5-in. to a 7.8-in. display. Potentially called iPhone Ultra, some predict it might be the “iPhone Duo” or “iPhone Fold.” It’s likely to cost $2,000+.
• iPhone 18 Pro series: Equipped with a fantastic new display, smaller Dynamic Island, 48MP main camera, and telescopic lens and supported by a fresh tranche of pro-photography features, the 18 Pro/Pro Max range will also use A20 chips. Both the fold and Pro devices will carry significantly more RAM, enabling them to handle on-device AI tasks. Apple is expected to introduce the iPhone 18, iPhone Air 2, and iPhone 18e devices at a spring 2027 event. Prices on all iPhones are expected to increase by $100 or more, depending on model.
• Apple Watch Series 12, Apple Watch Ultra 4: While Apple is thought to be working on new designs for next year, this year’s models will — for the first time in years — include an improved processor, better networking, and more accurate heart monitors. A model with a ceramic body is expected, while the devices will be equipped with additional storage, and integrated satellite emergency features. 
• AirPods 5: Earlier speculation about cameras inside AirPods remains, but these aren’t expected at the launch. Instead, Apple’s AirPods 5 will be boosted with a new H3 Wireless Chip for better voice isolation, noise cancellation, and spatial audio. They may also provide some basic health monitoring features with a built-in temperature sensor and heart rate monitoring.
• Smart home products: Apple is still expected to introduce additional products for smart homes, including an Apple TV with a faster processor and SiriAI support, a HomePod mini upgrade, and — potentially — a new home hub. The latter could effectively be a combination of an iPad with a HomePod to run apps, control smart devices, take and make calls, and run Siri.

Finding the next big thing

While all these products are important to the company, the smart home devices should demand attention (if they appear at all), as the indicators suggest Apple intends to explore the smart home space with a range of future hardware, including domestic robotics, security services, and smart devices. This hardware proliferation in new markets could end up becoming one of the hallmarks of Ternus’ Apple leadership, as will Apple’s AR spectacles (once they are introduced). 

Equally, it cannot be ignored that one of Ternus’ main tasks since becoming vice president of hardware has been the introduction of new Macs and the transition to Apple Silicon. So he’ll be expected to strike a balance between improving Apple’s traditional product lines while racing his company toward new and emerging opportunities in hardware. (I believe Apple should pay much more attention to the enterprise space within that.) 

Apple’s little AI problem

The biggest challenge to all of this will be Apple’s success in implementing AI that’s good enough to get things done while avoiding the unexpected dangers of poor implementation. We don’t want AI agents corroding smart home security, for example, which is why Ternus seems likely to maintain Cook’s focus on security and privacy as enabling technologies to support trusted hardware/software product designs.

Another thing? Tomorrow’s keynote will likely begin with the single word, “Hello,” and feels very likely to also include that classic Apple keynote line, “…One more thing.”

We’ll find out more as the conversation moves from speculation to reality after the 10am (PT) keynote.

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

Back in April 2025, the US Department of Justice (DoJ) proudly proclaimed that it had prevailed in a landmark antitrust case against Google’s ad business. Yeah. Right.

You see, just last week, US District Judge Leonie M. Brinkema rejected the DoJ’s effort to force Google to sell its AdX ad exchange. What began as a major anti-monopoly victory has become a mouse-sized settlement. 

[ Google US antitrust trials: A timeline ]

Does any of this this sound familiar? It should. We’ve seen this show before. Last year, Google had also been convicted of abusing its search business. At first, all the talk was about how Google might be forced to divest itself of the Chrome browser and/or the Android operating system. 

Instead, Google was allowed to keep making search deals and only had to share some search data with its rivals. 

I say rivals, but in fact Google still dominates the search market. Indeed, its control has actually expanded a bit. By StatCounter’s numbers, in August 2025, Google search owned 89.89% of the market. By last month, almost a year after the company had been slapped on the wrist, its share had grown to 91.1%. 

Boy, wasn’t that a win?

The remedies phase of the government’s ad-tech antitrust case order is a similar Google win. No, it doesn’t overturn the April 2025 finding that Google unlawfully maintained monopolies in the publisher ad-server and ad-exchange markets, or that it illegally tied its ad server and exchange together. Instead, it determines how the company must change its conduct.

The DoJ wanted Google to divest itself of AdX, the company’s ad exchange for matching publisher inventory with advertiser demand. It also sought to require Google to open-source the final-auction logic in its DoubleClick for Publishers (DFP) ad server and, if necessary, divest its DFP business. 

Brinkema rejected all three requests. Instead, her short order accepts “most” of the parties’ proposed behavioral remedies, subject to modifications described in a memorandum opinion that remains sealed while the parties identify confidential information for redaction.

Specifically, the court directed Google and the plaintiffs to meet and submit a joint proposed final judgment within 30 days. If they cannot agree on the details, each side must submit its own proposed version.

Google — as it should — sees this as a win. As Lee-Anne Mulholland, Google’s vice president of regulatory affairs, said in a statement, “We’re very pleased the Court rejected the DOJ’s proposal to break apart tools that help small businesses reach new customers and grow.”

Far less convincing is the DOJ’s characterizing the decision as a win. “The Antitrust Division is pleased that the court ordered substantial relief in the Google ad-tech case. We are one step closer to restoring competition and bringing relief for the American people in online advertising markets.” 

Oh please! It does nothing of the sort.

To quote Laurel Kilgour, research manager at the American Economic Liberties Project, an anti-monopoly, nonprofit group: “Judges keep finding Google guilty, but Google keeps walking away with both its ill-gotten gains and its empire intact. Without actual structural remedies, antitrust rulings are just inconvenient speed bumps that allow Google to lock down search and ad tech markets today while using that same unchecked power to monopolize tomorrow’s AI frontier.”

Exactly so. 

The specific obligations Google will face are not yet public, because Brinkema’s explanatory opinion remains under seal. Earlier, Google had proposed making real-time bidding responses from AdX available to rival ad servers, removing Unified Pricing Rules, and accepting a monitoring trustee for a three-year supervision period. None of that will even dent Google’s ad business.  

When all’s said and done, Google is still in charge of both its exchange and publisher ad-server operations. I expect the minor changes required won’t amount to a hill of beans. Restoring competition in markets where the company was found to have used its vertically integrated position unlawfully? Give me a break. 

Public Knowledge, a free speech and internet advocacy group, described the decision as one that leaves Google’s monopoly power substantially intact. 

As Public Knowledge Legal Director John Bergmayer said, “The court found that Google illegally acquired and maintained monopolies in publisher ad servers and ad exchanges, and used its control of DFP and AdX to shut out rivals.” He concluded, “Telling a monopolist to do better is not the same as restoring competition.… A finding of liability means little if the remedy leaves the market fundamentally unchanged.”

Ya think? 

In Google’s last reported quarter, the company reported $81.6 billion in advertising revenue. Next year, when Google is being “punished,” I expect its revenue will be even higher. 

IT industry experts and analysts are still trying to piece together Nvidia’s surprise plan to pay $12.9 billion for open-source AI company Hugging Face.

Nvidia dominates AI with its GPUs, and the company generates billions of dollars in revenue through a proprietary approach to the fast-moving technology. Hugging Face, on the other hand, hosts open models and has been a neutral player between chip vendors and model labs.

“This is about Nvidia having more say in how the stack gets built,” said Stephanie Walter, analyst at Hyperframe Research.

Hugging Face is wildly popular with developers, and Nvidia is buying early influence with that crowd. “You have a better chance of being part of the production environment later,” Walter said, adding that she wasn’t sure how Nvidia reached a nearly $13 billion price tag for the acquisition.

“Hugging Face has near-uncontested market primacy over where developers go for open-weight model releases. Now Nvidia owns that,” said Mark Petty, senior director analyst at Gartner.

Nvidia’s chase for developers should force IT decision-makers to review how much of the AI stack they control, said Hector Liu, director of Institute of Foundation Models’ Silicon Valley Lab. IFM is part of the Abu Dhabi-based Mohamed bin Zayed University of Artificial Intelligence.

Liu said CIOs should ask themselves three questions: “Can you run the model on hardware you already have, without a dependency you didn’t choose? Can you see how it was built?” And, “is the license one you can build a business on?

“A model that passes all three is durable, no matter who buys whom next year,” Liu said.

IFM’s latest K2 Horizon model, which was introduced on the same day Nvidia’s deal was announced, is hosted on Hugging Face. The open model was built to answer all of those questions.

K2 Horizon runs on AI hardware from Nvidia, AMD, Cerebras, and major cloud providers, said Liu, who doesn’t expect that to change. “Nvidia has said Hugging Face stays an open platform for every builder and every accelerator, and we’ll take that at face value,” Liu said.

Nvidia pledged to maintain Hugging Face’s hardware and model independence, and said its compute won’t be required. 

Even so, Nvidia isn’t paying nearly $13 billion for a model repository, said Jake Newfield, CEO at Hermetiq — it’s buying the front door to open AI.

(Hermetiq develops AI build and code observability tools.)

AI-generated output is becoming abundant and the infrastructure that makes it testable, reproducible and deployable is becoming strategically valuable, Newfield said. “Nvidia can accelerate it with capital and compute, but the real test is operational neutrality,” he said. 

That involves assessing whether competing hardware remains equally supported across the tooling, benchmarks and deployment paths developers actually use, Newfield said.

Beyond distribution, Nvidia is also buying Hugging Face for data, Petty said. That data showed that agents overtook humans as its largest traffic source in July, the kind of insight that could prove valuable down the road.

“Every model pulled tells Nvidia what the market wants next,” Petty said.

Nvidia has every reason to grow demand for open models and to keep them cheap, Petty said. “That’s good for enterprises, as it prevents market consolidation around a small number of proprietary model owners,” Petty said.

Closed and open models will coexist and lead to a world “where you’re going to continue to train these models and run these models at scale,” Justin Boitano, Nvidia’s vice president for Enterprise AI, said in a press conference after the deal’s announcement.

Nvidia will benefit “through the training that’s done and the inference that’s done on our hardware as models get diffused into the ecosystem at scale,” Boitano said.

That’s one motivation to keep the ecosystem open and neutral, so “developers can work wherever they want to work,” Boitano said.

Open source makes AI more accessible by lowering the barriers to experimentation and adoption, said Jon Carvill, senior vice president of marketing at AI chip maker Nuvacore. “Bringing Nvidia and Hugging Face closer together should help accelerate that choice, access and innovation,” Carvill said.

But there are still unanswered questions around whether Hugging Face will remain open and how much proprietary control Nvidia might exert, said Jack Gold, principal analyst at J. Gold Associates.

Microsoft traveled a similar path with its acquisition of open-source repository GitHub, which “did not really pan out as well as Microsoft hoped,” Gold said. “With Nvidia’s acquisition, will it still be as open to competitive hardware-software access, or will there be some barriers employed?”

I surprised myself this morning when I came across an interesting Apple-commissioned report — Rethinking critical AI infrastructure — I’d not seen before. It looks at the shifting expectations for AI infrastructure and recognizes that enterprise users want (and need) secure, on-device AI solutions for critical parts of their business.

That’s why tens of thousands of companies are already investing in Macs, because they recognize that Macs do indeed do AI. The study, published earlier this year and put together by Omdia, reflects insights gathered across 1,500 conversations with enterprise tech leaders and practitioners, noting that for many in business the current cloud-based approach to AI fails to deliver on three key metrics:

• Costs: Current pricing models seem unsustainable. Particularly when it comes to agentic AI, costs climb fast and business users need to get those costs under control. 
• Security: Even the most secure cloud services include some degree of data risk. When it comes to using AI for regulated data in industries such as healthcare, business users need much more security than the cloud inherently provides. After all, data that is not transmitted will not leak in transmission.
• >Capacity>: Workload requirements change and capacity needs to scale. That can boost the cost of accessing additional cloud capacity, or impose limitations in the event it can’t be found. It’s also true that while frontier models can provide all the bells and whistles of AI for advanced tasks, the vast majority of the AI work does not require anything near as much power. As Omdia explains: “57% of enterprise models are under 10 billion parameters, well within the capabilities of modern devices like MacBook Air or the entry-level MacBook Pro.”

What they think

As you might expect, the researchers believe on-premises AI set-ups respond to all three needs; not only that, but once you’ve coughed up cash for the necessary computational infrastructure, you don’t have to pay much more. “On-device infrastructure has near-zero marginal cost after initial investment, enabling unlimited experimentation without budget constraints,” the report said. 

Basically, once you’ve invested in on-premises capacity, you can divert mundane AI tasks to those machines for processing — limiting costs, boosting security and releasing capacity, turning to cloud-based models only when higher end AI solutions are required. While that’s good news for Apple, that’s bad news for many AI companies’ revenue models. (Perhaps they should have recognized that even the most advanced LLM’s will run on a standard iPhone eventually.)

The other advantage is that if AI is not used as widely as expected across a company, the same hardware can be used for other company tasks.

What’s actually happening

Enterprises already using AI are learning these lessons, which is why we see more of them buying Macs for these tasks. They do so because Apple’s computers deliver the computational power and performance to run AI effectively, from chip design to power consumption to the OS itself. Apple has intentionally built its platforms to be the best in class for running AI on device, and the Unified Memory architecture Apple has created in Apple Silicon scales really well, meaning you can run ever larger LLMs on Macs. 

It’s not just Macs, either. An iPad can run up to 14 billion parameter models quite happily; a Mac Studio reaches 480 billion; and a cluster of four Mac Studios will take you all the way to 1.6 trillion parameters using off-the-shelf cables.

To put that into context, Omdia found that 57% of the AI models typically used by the enterprise come in at under 10 billion parameters, which implies that enterprises could run a huge chunk of their AI tasks on an iPad, an iPhone, and certainly on a Mac. The ability of Apple’s ecosystem to scale is precisely why most AI developers at frontier model companies already use Macs. “Organizations that build AI solutions in-house adopt Mac for AI workloads at nearly double the rate of organizations buying commercial solutions,” the report explained.

The takeaway

Apple is emerging as an important component of an overall ecosystem for applied AI in the enterprise — or anywhere else — challenging frontier models with a scalable, controllable, economical, and secure approach to deployed AI that delivers most of the bang expected for the enterprise buck. 

While Apple paid for the report, that doesn’t necessarily invalidate its conclusions, which are not myopic around the Apple platform. Apple does not replace everything else, it just becomes one of the pillars to build success with AI. Companies can use other AI services and solutions, but they’ll want Macs along for at least some of the ride. And as the models themselves evolve and become slimmer and more refined, the platforms that run them best will deliver the advantage business users need.

Now, we need Apple to develop tools for the management, deployment, and governance of these solutions.

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