What we know so far: As the server battle between AMD and Nvidia enters a new phase, the two companies have begun trading jabs through performance estimates and early benchmarks. While Nvidia-approved results suggest its Vera processors outperform most AMD Epyc chips, Team Red believes its upcoming Venice lineup can leave Vera in the dust.

AMD recently published performance projections claiming its upcoming server CPU platform will dramatically outpace Nvidia’s latest showing. AMD’s estimates directly reference earlier results from controlled benchmarks that had favored Nvidia’s processor.

Team Red’s next data center CPU platform recently entered production and is on track to launch later this year. Built on AMD’s Zen 6 architecture, Epyc Venice chips will offer up to 256 cores and 512 threads. The lineup also marks AMD’s transition to TSMC’s 2nm process, a jump directly from the 4nm Epyc Turin that skips the 3nm node entirely.

AMD is projecting a 70% overall improvement in performance and efficiency over Turin, along with a 30% increase in thread density.

Nvidia, for its part, formally launched its Vera server CPU at GTC in March. The Arm-based SoC packs 88 cores and 176 threads. In recent benchmarks, Phoronix described Vera as the most capable Arm processor it has ever tested, outclassing Intel Xeon and AMD Epyc across most workloads. However, the tests were conducted at Nvidia’s headquarters and came with several restrictions to ensure Nvidia’s sign-off.

AMD drew on Phoronix’s figures when building the methodology for its Venice projections.

Comparing core counts per CPU, node power, nodes per rack, and a 100kW rack power budget, the company estimates Venice will deliver 3.3 times Vera’s per-rack performance. AMD also projects its 192-core Epyc 9965 Turin and the 128-core Intel Xeon 6980P GNR-AP can reach 2.37x and 1.46x of Vera’s output, respectively.

AMD is also challenging Nvidia on per-core performance, claiming a 64-core Venice chip can beat Vera by 27%, with the 96-core variant edging it by 11%.

As both processors target AI workloads, AMD argues that Venice’s higher core counts will translate into a meaningful advantage for agentic AI deployments. Even so, the true performance gap will remain uncertain until independent benchmarks arrive.

While promoting Venice’s theoretical performance, AMD is already hinting at what comes next. “Verano” will be AMD’s first CPU designed specifically for AI infrastructure. That chip is expected to introduce the Zen 7 architecture. Supply chain reports suggest Zen 7 will target TSMC’s A14 node, a 1.4nm-class process that would mark AMD’s entry into the angstrom era and deliver further gains in performance and efficiency beyond 2nm. AMD has not confirmed those details.

The iOS 26 update has the second-worst adoption rate of all iOS releases since 2015, falling behind both iOS 18 and iOS 8.

Though all eyes might be on iOS 27 and its AI-infused Siri, which debuted at WWDC, the software has only entered beta testing.

Most iPhones, or 79% of all devices to be more precise, are currently running iOS 26. This is according to Apple’s App Store data for June 2026, which also revealed that 86% of all devices introduced in the last four years have iOS 26 installed.

While these figures might seem impressive when taken at face value, the iOS 26 adoption rate is actually worse than that of the preceding iOS 18 update. In June 2025, 82% of all iPhones ran iOS 18, more than the 79% currently running iOS 26.

When taking into account Apple’s App Store data from all iOS releases from 2015 through 2026, the iOS 26 update actually has the second-lowest adoption rate, with only iOS 17 seeing fewer user upgrades.

The exact adoption rates for iOS 8 through iOS 26 for all compatible iPhones are the following:

• 2015 iOS 8: 84%
• 2016 iOS 9: 84%
• 2017 iOS 10: 86%
• 2018 iOS 11: 81%
• 2019 iOS 12: 88%
• 2020 iOS 13: 81%
• 2021 iOS 14: 85%
• 2022 iOS 15: 82%
• 2023 iOS 16: 81%
• 2024 iOS 17: 77%
• 2025 iOS 18: 82%
• 2025 iOS 26 79%

Note that all of these adoption rates were taken in June of the corresponding year, with the exception of iOS 12 usage data, which was published in August 2019.

At 79%, the iOS 26 adoption rate is below the 82.3% average from 2015 through 2026. Still, it’s not all bad news.

In February 2026, only 66% of all iPhones were running iOS 26, meaning more users have upgraded in the last couple of months. This is to be expected, though.

Even so, not everyone has embraced the controversial iOS 26 upgrade. According to Apple’s June 2026 App Store data, 14% of devices are still running iOS 18, while 7% are still on even older releases.

Apple’s data on newer devices running iOS 26

Since 2020, in June of every year, Apple has published iOS adoption rates for devices “released in the last four years.” At 86%, the adoption rate for iOS 26 is identical to that of iOS 17 in 2024. Meanwhile, iOS 18 had a higher score, at 88%.

The adoption rate of iOS 26 is also lower than the 87.6% average when Apple’s data from 2019 through 2026 is taken into account. Previous iOS updates, like iOS 12 in August 2019, saw lower adoption rates among newer devices, so iOS 26 is not an outlier in any sense.

Apple’s exact numbers for “all devices introduced in the last four years,” since the company released such data, are:

• 2019 iOS 12: 85%
• 2020 iOS 13: 92%
• 2021 iOS 14: 90%
• 2022 iOS 15: 89%
• 2023 iOS 16: 90%
• 2024 iOS 17: 86%
• 2025 iOS 18: 88%
• 2026 iOS 26: 86%

While iOS 26 has fallen behind iOS 18, possibly due to its Liquid Glass design language, the difference is likely not enough to cause alarm for Apple. Looking ahead, iOS 27 could boost iOS adoption rates.

Rather than controversial design changes, Apple says the iOS 27 update will deliver improved performance on older iPhones, with app opening speeds now being 30% faster, relative to previous releases.

Additionally, the same iPhone models that support iOS 26 can be updated to iOS 27, including the iPhone 11. As such, iOS 27 seems to have little to deter potential upgraders. Time will tell if the iOS 27 adoption rate will actually be higher than that of iOS 26, though.

Microsoft shared a 30-minute gameplay video for Fable right after the latest Xbox showcase. The footage offers the most detailed look yet at how Playground Games plans to handle the return to Albion. Players step into the role of a hero who begins as a child discovering unusual powers in the village of Briar Hill. A time jump then moves the story forward to adult life, where decisions start to shape both the character and the surrounding world.

Playground Games based the game on three key ideas from previous entries. A fairytale tone is undoubtedly evident, but it is balanced by a particular British sense of humor and a heavy emphasis on making decisions that will have long-term consequences. The new story stands on its own, but it has a very familiar approach to fantasy.

Xbox Series X – All Digital Gaming Console – 1TB SSD – Includes Wireless Controller – 4K Gaming – 120FPS…

• FASTEST, MOST POWERFUL XBOX — Experience next‑generation performance with 12 teraflops of processing power, delivering smoother gameplay, richer…
• TRUE 4K GAMING UP TO 120 FPS — Enjoy breathtaking visuals with true 4K resolution, HDR, and frame rates up to 120 FPS, plus support for advanced…
• 1TB CUSTOM SSD & VELOCITY ARCHITECTURE — Reduce load times dramatically with the custom 1TB SSD and XBOX Velocity Architecture, enabling…

The combat system functions efficiently because it is divided into three fundamental categories: Strength for close-up and melee combat, Skill for ranged attacks, and Will for magic. Some of the videos show some very innovative uses for that magic, such as transforming an enemy into a chicken and then shooting them with a fireball, or sneaking up behind someone and teleporting past for a quick strike before the green orbs light up to indicate you’re accumulating experience.

Reputation is based on what others see you do, not on an abstract number. So, assisting someone in need may earn you a virtuous or merciful reputation, which may result in a more welcoming reception from the community. If you cause disturbance in public, you may gain a reputation as a troublemaker, which could make things more difficult in the future. The system is clever since it allows you to build a reputation through a combination of good deeds as well as selfish goals, and you can even try to shape people’s impressions of you by spreading rumors or paying off local gossipmongers.

Is it more than just fighting and main quests? You can truly establish your existence in the world. You can purchase a home, open a bar, a blacksmith business, and so on. You can even develop relationships with others, some of which may lead to your own family or a shared home. Instead of always rushing to complete the main story, the game lets you to take a break and enjoy the journey.

Albion, the land in which you play, is full of little side paths and detours. There is no one “correct” way to play, so go explore, do some local jobs, and start building a life in any community you find yourself in. Towns look to be populated by real people, rather than just waiting for the hero to show, because the residents have their own schedules rather than simply hanging out till you arrive.

The visuals are bright and stunning, with a timeless quality about them. The villages are full of flowers and thatched roofs, and the locals look to be going about their everyday lives rather than standing around waiting for you, which is one of the most impressive elements of this demo, given that the entire game will be released on February 23rd, 2027.
[Source]

A former engineer at Elon Musk’s xAI has filed suit against the company and its parent SpaceX claiming he was fired for raising concerns about AI safety.

Devin Kim, who left xAI in September 2025, filed the suit in a California state court on Tuesday. The complaint comes days before SpaceX is set to join the public markets in what’s shaping up to be the largest IPO in history.

According to the lawsuit, which TechCrunch has viewed, Kim became a prominent voice for AI safety while working on Grok, xAI’s AI chatbot. He allegedly complained repeatedly about xAI’s failure to prioritize safety in Grok’s development, a product that has since come under fire for a range of safety and behavioral issues. In particular, Kim was concerned with the possibility that Grok could foment discrimination and help spread information about weapons of mass destruction.

“Grok, of course, proved Mr. Kim right by engaging in spectacular displays of online hatred and vitriol, with the model likening itself to Hitler (‘MechaHitler’),” the lawsuit reads. “Following the Hitler debacle, Mr. Kim worked to re-evaluate Grok’s political bias and discriminatory tendencies.”

A few months after Kim departed xAI, Grok made headlines again when the chatbot was used to flood X — Musk’s social media platform that also falls under the xAI umbrella — with nonconsensual sexual imagery.

The lawsuit also positions Kim as a whistleblower who was concerned about xAI’s alleged disregard for AI safety as “unlawful” in areas such as internet regulation, consumer protection and unfair business practices, and arms and explosives regulation, among others. 

xAI and SpaceX did not immediately respond to requests for comment. 

Kim’s focus on AI safety predates his time at xAI. While working at Scale AI, Kim worked on early safety AI initiatives, like leading a project that produced training data for AI to train systems to detect harmful content and comply with governance policies. Last week, the nonprofit Center for AI Safety, which focuses on AI risks, named Kim as its president.

Interestingly, the lawsuit doesn’t implicate Musk himself as a reason for a lack of safety. Rather, Kim’s lawyers describe Musk as having directed xAI to follow the law and implement appropriate safety and testing processes. Instead the claim targets Kim’s supervisor, xAI co-founder Jimmy Ba — who left the company earlier this year — saying that Ba ignored Musk’s directives and retaliated against Kim for pushing for safeguards, in an effort to “silence his repeated complaints about AI safety and biases.”

The lawsuit portrays Ba as someone who vehemently opposed AI safety measures, allegedly telling Kim at one point “AI will kill us all anyway,” and who was instead driven by a mission to make xAI the first to reach superintelligence. 

“In one instance in or around August 2025, Mr. Ba attempted to thwart EU safety regulations during the release of Grok Code 1, misrepresenting aspects of the model in order to avoid legally required testing,” the complaint says. “Mr. Ba indicated that he would rather release an unsafe model than a poor-performing one. Mr. Musk ultimately had to intervene.”

According to the lawsuit, Kim intended to give a presentation of his findings the week of September 15, 2025, but Ba called him into a meeting and told him they should “go [their] separate ways” without providing a satisfactory reason. 

TechCrunch has reached out to Ba for comment. 

Kim is seeking compensatory and punitive damages, as well as a declaratory judgment that xAI and SpaceX’s conduct was unlawful.

from the calls-are-coming-from-inside-the-house dept

The more things change, the more they remain the same. That could be said of anywhere in this country, now that the Trump administration is trying to turn the clock back to 1940, if not 1840.

But it’s especially true in Los Angeles, where law enforcement agencies have apparently learned nothing, despite being the ignition source of two riots. The 1965 Watts riot was provoked by racist, abusive actions of the LAPD. The 1992 riots were similarly provoked by the racist, abusive actions of the LAPD.

Before, between, and after, Los Angeles law enforcement agencies haven’t done much to improve. When not actively thwarting federal investigations and running illegal jailhouse informant programs, the Los Angeles Sheriff’s Department has hosted any number of “gangs” composed of officers who are more willing than others to engage in violence and rights violations.

The LASD’s gangs have made headlines for most of the last decade, including stuff that would otherwise seem to be the broadest of satires:

Los Angeles Sheriff’s Deputy Allegedly Removed ‘Unauthorized” Sheriff’s Gang Tattoo With A Bullet

It’s admittedly hilarious, but only in the darkest sense. While absolutely absurd, it also indicates that LASD officers (especially those who are in LASD gangs) feel the solution to every problem — including tattoo removal — is to start blasting.

A handful of people who’ve run on “reformer” platforms have either failed to be elected, or have been elected only to renege on their reformation promises.

The LAPD covers less area and has fewer officers than the Sheriff’s Department. But it still has nearly 9,000 officers, which is only about a grand short of the LASD total (10,000 officers). If nothing else, basic mathematics would strongly suggest the LAPD would be just as receptive to internal gangs as the Sheriff’s Department.

The LAPD internal investigation leveled a troubling allegation: Officers in a specialized unit tasked with combating street gangs had themselves behaved like a gang.

In 2023, officers in the San Fernando Valley were accused of making dozens of improper traffic stops and attempting to hide their actions from their supervisors by switching off their body cameras.

Advertisement

When confronted by Internal Affairs detectives, according to the findings of a months-long probe, officers in the Valley’s “gang enforcement detail” said they were engaged in a “gun hunting competition,” with each firearm-related arrest tracked on a whiteboard in their office. Cops with the most seizures would pose for pictures with pro-wrestling-style championship belt that had “Mission GED Pistoleros” emblazoned on the buckle.

And so it is. While this opening salvo of paragraphs merely suggest some members of the LAPD were more prone to doing bad stuff than others, the Internal Affairs report makes it more explicit.

The report said the Valley unit was a “law enforcement gang.”

That report was buried by the LAPD for almost three years. But that burial proved temporary. The report — which had previously only been seen by LAPD officials and some city lawmakers — prompted further inquiries. And those further inquiries generated answers that raised even more questions:

LAPD leaders said at the time that the problems were confined to that one division. But a new case involving similar allegations against anti-gang officers operating out of South L.A.’s 77th Street patrol area has reignited questions about whether there are deeper issues across the department.

Oh, the fucking irony. An anti-gang squad that behaves like a gang. Wow, imagine if we’d ever seen this anywhere else multiple times. I mean, say the first thing that comes to mind when I say “rampart.”

It’s tempting to simply say that no one cares. But I don’t think that’s true. I do think a lot of people care, including LA lawmakers who want to see real reform. The problem is that the people with the most power don’t care. That not only includes law enforcement unions, law enforcement officials, elected officials (including sheriffs), but also the handful of lawmakers who actually think law enforcement officers should be allowed to violate rights while performing their duties.

That’s the headwind reform efforts face. While thousands (or millions, in this case) may recognize the problem and want reform, it only takes a handful of powerful people to prevent their voices from being heard. And while it’s easy to tell people to vote their way back into power, we only need to look to the White House to see how facile and futile the “vote the bastards out” suggestion is. It’s something that should have been addressed years ago, because if you give the bastards an inch, they’ll entrench a mile. If Los Angeles is going to fix this, it will require the concerted efforts of people who are more motivated to protect their paychecks than serve the public. I wouldn’t hold my breath.

Filed Under: gangs, lapd, lasd, los angeles, los angeles county sheriff’s department, los angeles police department, police gangs, police misconduct, police violence

With new generations of AI models fueling both rapid software vulnerability discovery and the potential for faster exploitation by malicious hackers, the United States Cybersecurity and Infrastructure Security Agency released a new directive on Wednesday that requires more rapid and efficient software patching by federal civilian agencies. The “binding operational directive” (BOD) lays out a rubric for how quickly bugs must be fixed based on four assessments of urgency, with a turnaround time in critical cases of just three days.

Chris Butera, CISA’s acting executive assistant director for cybersecurity, told reporters on Wednesday that the goal of the directive is to help agencies prioritize, so they can address the most problematic vulnerabilities first while taking more time to remediate bugs that pose a less-pressing risk. The directive comes as private companies and governments have been scrambling to assess the extent of the cybersecurity reckoning that AI vulnerability and exploit development capabilities could unleash.

“Prioritizing IT and security operations attention on the most at-risk assets is particularly important now given advancements in artificial intelligence, which allow threat actors to find and exploit vulnerabilities in [federal] assets,” Butera said on Wednesday. “Defenders cannot afford to take weeks to patch systems that can be autonomously exploited en masse.”

The CISA directive’s criteria for evaluating patch urgency includes looking at whether a vulnerability is in a system that is publicly exposed, whether the bug is listed in CISA’s Known Exploited Vulnerabilities Catalog, whether an attacker could automate all of the steps to exploit the vulnerability, and how much access an attacker would get to the target if the bug were exploited. A vulnerability where all four points apply must be fixed within three days, according to the new directive, and the agency must also execute a “forensic triage” process to determine whether systems have already been compromised.

The directive supersedes two previous CISA orders related to patching timelines for urgent vulnerabilities—one from 2019 and one from 2021. Those established a framework in which the most critical bugs had to be patched within 15 days of detection and another class of high-urgency vulnerability had to be remediated within 30 days. And both encouraged faster patching for severe flaws when possible. Even before the AI era, in 2021, CISA wrote that “threat actors are extremely fast to exploit their vulnerabilities of choice: of those 4% of known exploited [vulnerabilities], 42% are being used on day 0 of disclosure; 50% within 2 days; and 75% within 28 days.”

US federal cybersecurity has improved significantly over the past decade, but it still often lags, thanks to funding shortfalls and competing priorities. CISA’s Butera said that the agency developed the new assessment rubric and the directive more broadly with these limitations in mind. He noted, for example, that the three-day deadline for the most urgent vulnerabilities isn’t, say, 24 hours, because such a short timeframe would not be feasible for most agencies.

New AI capabilities are already changing the landscape of vulnerability detection and bug hunting. And as this spurs new urgency in patching, many researchers have started to conclude, essentially, that no amount of patching will be enough—and that the software development community globally must work to adopt new, architectural or systemic approaches to invalidating whole classes of vulnerabilities at a time.

“CISA’s directive has its heart in the right place, but it only tackles half the challenge,” says Emily Long, CEO of the cloud security firm Edera. “If your architecture doesn’t limit what an attacker can reach after a breach, you’re just running faster on the same treadmill. Patching will always be important, but we should be talking more about containment by design.”

CISA’s Butera seemed to acknowledge this evolution on Wednesday. The new directive “is an initial step to counter the increased capabilities of emerging AI models,” he says. “Yet there is still more work to do.”

A man raises his phone as police move into a crowd. The video is shaky, loud, immediate. Within minutes, it is online. Within hours, it is everywhere. This is how accountability works now. Something happens, someone records it, and that footage can show what really happened, sometimes contradicting official accounts. It can empower citizens and create consequences for officials.

But the footage’s life cycle does not end there.

In recent months, civil liberties groups have warned that adding facial recognition to consumer smart glasses could turn everyday recording into something more troubling: real-time facial identification. It reflects a broader shift already underway, where images and videos captured for one purpose can later be searched, matched, and used for another.

An ouroboros is an ancient Egyptian symbol, a snake or dragon eating its own tail. As I began to see patterns in my broader research on surveillance corporatism and governance lag, I began using the term “surveillance ouroboros” to describe this recursive pattern of observations intended to hold power accountable becoming new input for the same surveillance infrastructure.

Facial recognition changes accountability

During the George Floyd protests in 2020, people filmed police in real time. Phones were pointed at officers, not at each other. The goal was simple: to show what the state was doing. That footage spread quickly and became part of a much larger pool of public data.

At the same time, reporting from outlets including The New York Times and BuzzFeed News showed that law enforcement agencies were using facial recognition tools, including systems built by Clearview AI. Those systems were built from billions of images scraped from across the internet, including publicly available photos and videos.

The basic approach is now routine: People record the state, or anything else—as in the January 6 attack on the U.S. Capitol—and the state compiles that footage and data into a searchable environment, which may later be used to identify some of the same people who made the footage.

Facial-recognition systems used by law enforcement are increasingly outpacing the legal safeguards.

A 2024 Government Accountability Office review found that federal law enforcement agencies continued to expand their use of facial-recognition systems for criminal investigations despite ongoing concerns around training, privacy protections, civil-liberties safeguards, and oversight. Earlier GAO findings showed that agencies had conducted roughly 60,000 facial-recognition searches before formal training requirements were put in place for personnel using the systems.

The American Civil Liberties Union and other groups have warned that these tools could be used to identify people from images shared online, including protest-related footage. Concerns about facial recognition led some U.S. states and cities, including San Francisco and Boston, to restrict or ban government use of the technology, while federal agencies have continued to face scrutiny over how such systems are tested, deployed, and audited. A 2024 analysis published in Internet Policy Review warned that facial-recognition systems used by law enforcement are increasingly outpacing the legal safeguards meant to govern them, creating growing tensions around data protection, oversight, and proportional use.

The spy network that built itself

Surveillance used to require infrastructure. Cameras had to be installed and data had to be collected deliberately. That is no longer the case. People carry cameras everywhere. They record constantly and upload in real time. Events are documented from multiple angles without planning or coordination. The cumulative result is a continuous stream of usable data: faces, locations, timestamps, and interactions. The Internet of Things also waits all around us, gathering information and releasing it when people least expect it, as Andrew Guthrie Ferguson describes in a recent excerpt of his book Your Data Will Be Used Against You.

RELATED: “Sensorveillance” Turns Ordinary Life Into Evidence

Similar dynamics are emerging globally. A recent analysis in the International Journal of Law and Information Technology examined how facial-recognition systems in China and Japan are expanding faster than the legal frameworks governing them. Reporting by The Guardian described the limited legal protections around the rapid deployment of AI-assisted surveillance infrastructure across parts of Africa.

There used to be a clear distinction between surveillance and accountability. Surveillance meant the powerful watching the people; authorities tended not to share their imagery except under duress or a court order and usually after a long delay. Accountability meant the people watching the powerful, and often publishing imagery immediately to head off or counteract official mischief. That distinction no longer holds. The same footage can serve both roles. A recording meant to expose misconduct can later be used to identify someone else entirely.

Surveillance ouroboros is not a future risk. It is already here.

This dynamic persists because people still need to record. In many places, it is one of the only tools available when formal accountability breaks down. When oversight institutions weaken or fail, public documentation becomes a substitute. In that environment, people turn to visibility. But that visibility comes with a cost. The more people that document, the more data that exists. The more data that exists, the easier it is to search, match, and store. Every video feeds the ouroboros. People are not feeding the system because they trust it. They are feeding it because the alternative is silence.

Most of the people in these videos are not the focus. They are in the background, passing by or standing nearby. But that distinction does not matter once the footage enters a system. Today’s facial recognition can identify even a face that passed through the corner of a frame. Someone who did nothing can still become part of a dataset without ever knowing it. As recognition systems improve, older footage becomes more useful, and invasive.

No single decision created this outcome. It emerged gradually through more cameras, better recognition, larger datasets, and easier integration. Each step made sense on its own. Together, they changed what recording means.

Public recording is still necessary. Without it, many forms of abuse would remain hidden. But recording is no longer just exposure. It is also contribution. If you published imagery or video last year, you may already have contributed to a system you have never seen, but the ouroboros has.

Surveillance ouroboros is not a future risk. It is already here. Every time someone presses publish, they are doing two things at once. They are exposing power, and they are helping build the system that the powerful will later use to track the less powerful.

• After four years, we finally get a second Kingdom Hearts 4 trailer
• Disney characters were few and far between, but two worlds were teased
• We also didn’t get a release date

It’s tough being a Kingdom Hearts fan. The series is an all-time great, but if its platform-hopping origins weren’t frustrating enough — with entries exclusive to different consoles and even different regions for a while — we must now face the fact that Square Enix loves taking its time between entries, and between a new game’s announcement and release.

Kingdom Hearts 3 was first showcased around five and a half years before its release, and it seemed Kingdom Hearts 4 might take just as long between merely its first and second trailers. That is, until we got a brand new look at the game at Nintendo Direct, saving this year’s Summer Game Fest event for me just as I had conceded the Disney meets Final Fantasy crossover would be absent yet again.