Valve is continuing to broaden SteamOS support beyond its own hardware, with the latest beta adding "initial gamepad support" for a few recent gaming handhelds and improving compatibility for a few others. Ars Technica reports: Those newly supported handhelds include MSI's Claw 8 EX AI+, the extremely expensive, extremely powerful handheld built around Intel's Arc G3 processor that launched this summer. While SteamOS has long included beta support for "other AMD powered handhelds," Valve has only recently been working on offering compatibility with Intel-based handhelds like the MSI Claw line. This weekend's update suggests the company hasn't abandoned those efforts and that Intel chips aren't being left out of the SteamOS conversation.
SteamOS 3.8.25 also newly supports the controllers built into Ayaneo's Android-focused Pocket S2 and the more recently released Windows-based Konkr Fit. Valve also says the new OS offers "improved support" for older MSI Claw devices and the OneXPlayer line of portable gaming PCs. This weekend's update follows June's SteamOS 3.8 update, which stressed "improved compatibility with recent Intel and AMD platforms." That update specifically called out improved controller support on handhelds from the likes of GPD Win, Anbernic, and OrangePi, as well as improvements for various Lenovo Legion Go, MSI Claw, and OneXPlayer devices. The report notes that the new SteamOS update also updates the Linux driver powering the new Steam Controller, "allowing it to work with 'native controller applications' even when Steam is not running."
PUBLIC DEFENDER By Brian Livingston Realistic-looking videos that are the most-often recommended by the popular TikTok social media app regarding health care contain no real people 40 percent of the time, a new analysis of the site’s mini-movies has found. Many of TikTok’s top-rated videos on every topic are AI-generated but look exactly like actual […]
UI changes, lost productivity, untransferrable knowledge... Long-time Slashdot reader theodp notes the folly of switching to an unfamiliar new environment to be "more productive"...
Champions of the switch often insist workers will be able to hit the ground running and get the same or even more work done from the get-go in the new unfamiliar environment because it's 'more productive', a conclusion often shaped by conventional wisdom or their 'gut' feeling. Decision makers may also mistake their own 'light-user experience' for universal experience and believe all platforms are fungible.
And while it's one thing for a business to delay their delivery dates and increase costs of by introducing a switch in computing environments that's of unquantified and possibly dubious value, how about when a switch in computing environments is made that affects schoolchildren's ability to learn for a period of time?
While a MacBook Neo is indeed cute, what might be the possible consequences of a decision to move 25,000 students off Windows and onto MacBook Neos...? Or even the impact on a single kid who moves from a MacBook school to a Windows or Chrome school (or whose home computing environment isn't aligned with their school). Or students whose teachers decide learning in their classrooms should now additionally require an understanding of the Minecraft computing platform/UI (used in over 40,000 school systems.). Learning Management Systems and now various AI Edtech tools used by schoolkids are also subject to change.
"Even if students and schools eventually benefit from a new platform, one wonders how much learning is lost during the switch," theodp asks, "and what the financial and educational ROI and payback period look like." Are there bigger problems than we realize? Share your own thoughts and experiences in the comments.
What are the costs of switching between Windows, MacOS, and Chrome?
"I've been a Linux IT administrator for 30 years now," says crhylove (Slashdot reader #205,956), in a new video on his YouTube channel. "I used to be a data center architect, at one point I was the IT director for 13 school districts... I just haven't been happy overall with all the other operating systems out there... "
I've been working really diligently for the last few weeks on a new distro called Spaced Linux that is exactly what I want to use. It's heavily focused on the [zoom-friendly 3D] Compiz desktop, very efficient using the rolling Devuan Ceres backend, and avoids dependency hell by promoting mostly Flatpak usage.
It's still early, but it's already working great for me, it's been my daily driver for a few weeks now.
There's also user-friendly themes — everything from Linux, Android, GeoWorks, and Mac OS X to Windows 11, Windows XP, and even Windows 3.11 (dark or light theme). And you can also run Windows apps under Wine. "And let me tell you, Nvidia runs perfectly on here now..." he adds in the video. "It just works really, really well out of the box."
He's now working on voice-command features, bur argues that Spaced Linux "is already extremely usable, extremely efficient, extremely fast, and extremely reliable.... And it's only going to get better and better."
Senators from both parties Thursday probed Trump administration officials about whether federal agencies and foreign governments are coordinated enough in the battle against scammers, something witnesses told the Foreign Relations Committee they were working to remedy.
At least 13 federal agencies have authorities to counter scams, raising questions about whether someone needs to be in charge of all those efforts. And while there was some bipartisan sentiment at Thursday’s hearing that the Trump administration has taken good actions to battle scammers, both lawmakers and administration officials said that scam operations have demonstrated that cracking down on them in one place often just leads to them going elsewhere.
Sen. Pete Ricketts, R-Neb., compared the situation to an international initiative that gained prominence in the 1990s to counter drug trafficking, Joint Interagency Task Force South.
“Given that today’s scam centers are similarly transnational, combining cybercrime, human trafficking, money laundering and cryptocurrency, has the threat reached the point that we should establish a comparable multinational coordination mechanism?” he asked.
Sen. Jeanne Shaheen, D-N.H., focused on federal coordination: She paraphrased a former federal official who said, “there is nobody that is heading that effort up across agencies. We need to treat this like combat, and so we need somebody in charge.”
Shaheen, the top Democrat on the panel, is a co-sponsor of the bipartisan Scam Compound Accountability and Mobilization (SCAM) Act, which seeks to unify federal efforts on the subject.
A State Department official told Shaeen scammers were a national security priority for President Donald Trump, and that his executive order on the topic sought to tackle coordination.
“I do understand that this is a whole-of-government approach, and many agencies are focused on this,” said David Bedard, deputy assistant secretary at State’s Bureau of International Narcotics and Law Enforcement Affairs “The Action plan that was directed by the president is currently in the interagency review process to deconflict some of the concerns that you have raised. We certainly think the task force that will be implanted through the executive order will solve the problems you might be referencing.”
There’s also an international plan under the task force, he said. Currently, the administration shares intelligence on scammers with foreign allies, and Interpol has “productive” channels to work through there and is setting up its own task force, Bedard said, but there are concerns about other countries taking similar, duplicative action.
There have been signs of progress on the international front, Bedard and another State Department witness told the panel.
Michael DeSombre, assistant secretary at the Bureau of East Asian and Pacific Affairs, said Trump has raised the subject with Chinese President Xi Jinping, and that China has used its influence in Asia as its own citizens have become scam victims. Still, there’s been more progress in countries where the United States has stronger relations, such as Cambodia, than in those where ties aren’t as close, like Burma and Laos.
In Cambodia, one key has been pursuing scam center bosses first and foremost, Bedard said.
The UK’s AI Security Institute said that their AI research system took “unsanctioned” actions over the internet to engage in “sustained, potentially harmful activity directed at real people.”
According to a blog post and technical report released Tuesday, the organization noticed “unusual data transfers” leaving their research systems through the pseudo-anonymous Tor network on July 28. The models, which included Anthropic’s Mythos 5 and OpenAI’s GPT-5.6-Sol models, were being tested for their cybersecurity capabilities.
Like the OpenAI and Anthropic incidents before it, the offending agents were attempting to solve a challenge. On 10 of 122 runs taken by two models, they took a combined 19 malicious actions, including attempting to insert malicious code into a real open-source software project and creating fake online identities to contact human maintainers and press them to approve it.
It also inserted malicious prompt injection instructions in places “where it reasoned that other automated AI systems might pick them up and execute them.” Sometimes multiple agents would collaborate on a task, with one model leaving public messages or notes for the other on GitHub.
However, AISI said their incident differs in one critical way.
“Importantly, this was not a case of a model escaping its secure test environment, or ‘sandbox,’” the institute wrote in an unsigned blog. “As was standard in our cyber testing, we had intentionally permitted internet access, and model-provider cyber classifiers were deliberately disabled – conditions that do not reflect how frontier models are made available to the public.”
The institute said that while their own design and evaluation choices “to some degree” enabled the malicious behavior, the models still displayed “novel, potentially deceptive behaviors” and were executed “to an extent and severity that we did not anticipate.”
The same day, OpenAI acknowledged AISI and another third-party cybersecurity tester, private firm Irregular, had recently reported incidents where their models exceeded “beyond their intended testing boundaries.”
That included GPT-5.6-Sol reusing GitHub tokens left by another agent to attempt to recover accounts and to access a DNS server that contained malicious payloads meant to exploit a software vulnerability.
The setup did not work, there is no evidence any real resolver queried it, and the infrastructure was removed when the evaluation ended,” the company wrote in an unsigned blog post Tuesday.
OpenAI said it would review its own third-party testing procedures to focus on higher risk evaluations, assess requests by third-parties to enable internet access, stop conditions and other features for their models.
The incident with Irregular occurred on July 29 during a Capture-the-Flag cybersecurity evaluation of OpenAI’s models. Due to a “misconfiguration” the models were allowed to access the public internet, where they encountered and exploited a real domain, mistakenly believing it was still in a test environment.
OpenAI said an investigation by Irregular is ongoing, but also found that the models had found and used credentials for the site at one point. The blog also references other additional potential cybersecurity incidents.
“Irregular has informed us that all of the issues identified pertaining to the incident are no longer active and relevant safeguards were added to the testing environment,” the blog said. “Irregular has also communicated about related incidents involving other labs from the same testing environment.”
CyberScoop has reached out to Irregular for comment.
The incidents were made public the same day that the White House met with Anthropic, Open AI and other frontier AI companies to preview a new framework for evaluating models before they’re released publicly. Some media outlets have reported that after an executive order, export controls and other actions, the administration does not plan to make the new framework public.
The Trump administration’s haphazard and opaque interventions into artificial intelligence security matters could catapult Chinese alternatives into broader acceptance, posing new security risks altogether, a group of Democratic senators wrote to top administration officials Monday.
The five senators said that the administration’s handling has alternated between too passive, such as when OpenAI models escaped testing in the Hugging Face hack last month, and overstepping, such as when the Commerce Department suspended access for any foreign national to Anthropic’s Fable 5 and Mythos 5 in June.
“The Administration’s ad hoc and unpredictable approach undermines U.S. competitiveness, heightening market incentives to adopt open weight models from vendors based in the People’s Republic of China (PRC),” wrote Sens. Kristen Gillibrand of New York, Adam Schiff of California, Mark Warner of Virginia, Chris Coons of Delaware and Mark Kelly of Arizona.
In the Hugging Face hack, the senators wrote that “the Federal Government cannot be passive as these capabilities emerge.”
In the case of the Fable 5 and Mythos 5 suspensions, the senators said that the administration “utilized an infrequently used authority to direct Anthropic to suspend all access to its Fable 5 and Mythos 5 models for foreign nationals (including foreign national employees inside the United States) citing an undisclosed national security concern later described as a narrow jailbreak finding.”
Because Anthropic couldn’t immediately assess users’ nationality, the firm had to disable both models for everyone. The administration and Anthropic negotiated for 18 days behind closed doors before reaching an agreement, the lawmakers complained.
“While the Administration may have been responding to real security concerns to protect the United States, even justifiable interventions can create broader harm if the standards and decision-making processes are opaque, ad hoc, or unpredictable,” they said in their letter to leaders in the White House, Office of the National Cyber Director and departments of State, Treasury and Commerce. “Moreover, when the Executive Branch exercises authority delegated from Congress, such as in the conduct of export control administration, it is essential that it keep Congress fully apprised of its actions and procedures.”
During the time Anthropic was under export controls, the stock price of “an entity-listed Chinese lab” nearly doubled, the senators said. And while Hugging Face was breached, the company “had to” rely on a Chinese open-weight model due to guardrails on U.S. frontier models.
“If American models are perceived as subject to sudden access disruptions based on a black-box U.S. Government process, or as unreliable because U.S. AI labs are overcorrecting in the face of this black-box process, companies and governments in the United States and abroad may hedge by adopting Chinese or other foreign models instead,” the senators contended. “That outcome would undermine U.S. technological leadership while increasing exposure to systems that may carry risks of PRC or otherwise directed censorship, espionage, IP theft, and other supply chain security risks.”
Their letter asked for answers to questions about the standards the administration uses to determine the national security risks a frontier model presents, what legal authorities it will use to invoke restrictions, which agencies are responsible for which decisions and more.
None of the offices or departments the letter was addressed to immediately responded to a request for comment.
The letter follows inquiries at the state level, where 15 attorneys general asked OpenAI for more information regarding the security incident at Hugging Face.
Amazon’s security researchers say a hacking group tied to North Korea targeted small, little-noticed software packages more than a year before it struck one of the internet’s most widely used programming tools.
The company’s threat intelligence team said Wednesday at a media roundtable at its Arlington, Va., offices that the same group linked to the recent compromise of the open-source axios software library also planted malicious code in a package called typo-crypto in March 2025, a full year before the axios breach. Researchers found the connection while tracing domain records tied to the axios attack back to earlier activity.
“We believe the March 2025 typo-crypto campaign was a rehearsal,” said CJ Moses, Amazon’s chief information security officer, adding that the target’s small scale let the group test its methods “without putting that on the big stage.”
Amazon said the group also compromised two other packages, debug and chalk, in September 2025. Until now, those three incidents had not been publicly linked to the same actor. Security researchers track the group under several names, including UNC1069, Sapphire Sleet and Stardust Chollima.
Axios, debug and chalk are code libraries used by software developers around the world to build applications. Axios alone is downloaded more than 100 million times a week. “That number represents real organizations putting real code into production systems every single week,” Moses said.
In the typo-crypto case, the malicious file was named “core.js” and was made to look like a legitimate, unrelated package called core-js. Amazon said the file activated only when it received a specific numeric input, then reached out to a server controlled by the attackers to download a second piece of code. That second stage was written differently depending on whether the infected computer ran Windows, macOS or Linux. The code combined encoded text with a cipher, a method Moses said was meant to slow down analysis, including by AI-based review tools, without relying on heavy encryption.
Amazon said the typo-crypto package had few downloads compared with axios, debug or chalk. Researchers believe that initial target served as practice, letting the group refine its approach before turning to more widely used software. “They did what a lot of people do: crawl, walk, run,” Moses said.
In each of the four cases, Amazon said, the attackers built a relationship with a maintainer who already had access to a package, then used that access to publish an update containing hidden code. “They didn’t break through a window,” Moses said. “They basically earned the trust of an employee to hand them the keys.”
Cybersecurity firm Wiz separately found that about 1 in 10 cloud computing environments were affected by the debug and chalk incident within a two-hour span, a finding Moses cited to illustrate how fast the impact spread. “Going from there not being a vulnerability, to there being a vulnerability, to there being an exploited vulnerability … used to be days to weeks. Now it’s hours to minutes,” he said.
Rick Anthony, senior engineering manager at Amazon Web Services, said the research further shows how attackers face two basic problems in these types of incidents: getting malicious code into a package that will eventually run inside an organization, and keeping that code hidden from developers or security tools. He said groups are increasingly building reputations as legitimate contributors over time.
“Let me get my package deployed in as many places as possible so that I can spring the trap later,” said Anthony, describing the mindset behind the approach.
Researchers said generative AI has made it easier for attackers to produce code, documentation and contribution histories that look authentic. Anthony also described a technique in which attackers register package names that AI coding tools sometimes generate by mistake, so a developer following an AI suggestion could install malicious software without making any typing error of their own.
The findings come two years after a separate incident involving a program called xz-utils, in which an attacker spent time gaining the trust of the software’s maintainers before inserting a backdoor. Moses pointed to that case as an early example of a pattern now appearing “at scale” and tied to a nation-state.
Since that incident, separate groups have been running roughshod over open-source software. Another group known as TeamPCP has compromised and injected malicious code into more than 1,000 software packages over a four-month span this year.
ISSUE 23.30.1 • 2026-07-28 By Susan Bradley • Comment about this alert Today, I want to tell you that I’m less worried about the side effects of this month’s security updates from Windows, Apple, and Linux distros. I’m more concerned about security in the AI industry in general. July saw the first side effects of […]
Apple’s macOS 27 release — dubbed ‘Golden Gate’ by the company — is packed with new features, despite Apple saying that this year was more of an evolution than revolution. There are plenty of additions that have got tongues wagging and pens scribbling across the internet.
Yet for every Siri app and Liquid Glass tweak, there are many more upgrades that people are barely talking about. That means that if you don’t dig below the surface, you might miss some of the interesting new ways you can use your Mac once you’ve downloaded and installed macOS 27.
Here, we’ll run through five of our favorite little-known inclusions in the macOS Golden Gate beta. Get to grips with all of them and you’ll have fresh ways to use your Mac on a daily basis.
Visual Intelligence
(Image credit: Apple)
Previously, Visual Intelligence was only available on iOS. You’d point your iPhone at an object, press the Camera Control, and Visual Intelligence would identify the object or pull out details from an event flyer and add them to your calendar. Now, similar powers can be harnessed on your Mac.
But you might be asking how this will work on a Mac, seeing as you can’t just go up to something and take a picture of it with your laptop. Fortunately, Apple has adapted Visual Intelligence to macOS 27 in a handy new way.
Much like when you take a screenshot, you can drag a box over part of your screen using your mouse. When you let go, you can ask Visual Intelligence to help you with what you have selected. You can ask Siri AI a question about the image and it will respond. You can then go back and forth with the chatbot if you need more information.
If you highlight an event, you can add it to your calendar. Even better, if you use Visual Intelligence on a festival schedule with multiple acts playing at different times, for instance, you can add as many or as few to your calendar as you want, all saved as separate entries with relevant info like the band that’s playing and the stage that you’ll find them on.
This is one of my favorite features in macOS 27 as it’s not only useful, but highly flexible and intelligent. Visual Intelligence is able to understand many different inputs and the contexts surrounding them, turning Siri into much more of the powerful assistant that it was always meant to be.
Automatic password strengthening in Passwords
(Image credit: Future)
The best password managers often alert you if your logins stored in the app are weak or compromised, but that usually means updating them to stronger alternatives yourself.
If that requires manually visiting affected websites and changing your login details by hand dozens or even hundreds of times, the task can feel incredibly daunting. Faced with that kind of situation, many people likely just give up, preferring to stick with their untrustworthy passwords rather than go through the hassle of changing them all one by one.
That’s an obvious security risk, and it’s one that macOS 27 has attempted to put right. Although the following change hasn’t had much airtime since the feature was announced at Apple’s Worldwide Developers Conference (WWDC) in June, I feel like it could be one of the most important features in the new operating system.
In macOS 27, all you need to do is ask Apple’s Passwords app to fix your dodgy passwords for you. Just one click and it’ll get to work visiting affected websites and updating your logins without you needing to lift a finger.
It’s a great example of how useful AI agents can be, all backed up by Apple’s well-known security and privacy credentials.
Natural language in Calendar
(Image credit: Future)
How many times have you opened the Calendar app on your Mac and tried to create a new event with a phrase like “Movie night with James tomorrow at 8pm at home,” only to find none of the details — such as the time or the place — are properly added to the event?
It’s been a recurring problem for years, and a frustrating one, too, as Apple fans have watched other apps race ahead with similar features while Calendar lags behind.
Now, that’s been improved. Changes in macOS 27 mean that you just need to enter an event using the type of natural language that you’d use in an everyday conversation. Any details that Calendar can pick up on — like a person, place, time or date — will turn gray. Click them and select the suggested detail and they’ll be properly added. This feature understands your context, too, such as where your home is located.
Sure, similar features have been present in rival calendar apps for years, and that could explain why Apple’s Calendar update hasn’t exactly lit the world on fire in the weeks and months since it was announced.
But having this feature is nonetheless an excellent addition for users of the Calendar app. The more naturally you can use your apps, the more effortless they become, allowing you to quickly be done with them so you can get on with what matters in your day. This natural language upgrade for the Calendar app lets you do exactly that.
App resizing in iPhone Mirroring
(Image credit: Apple)
Of all the clever new features announced last year as part of macOS Sequoia in 2024, iPhone Mirroring was one of the more intriguing. As the name suggests, it mirrors your iPhone onto your Mac’s screen, bringing up a software view of your device where you can open apps, type messages, and basically use everything on your iPhone through the Mac sitting in front of you.
But there was one annoying quirk with this system: you couldn’t resize the iPhone window to use different aspect ratios. That meant any apps you opened on the virtual iPhone were limited in size and failed to make use of your Mac display’s larger proportions.
Thankfully, Apple has fixed that up in macOS 27. Instead of being stuck with the unchanging iPhone window of the past, macOS 27 now lets you drag the on-screen iPhone and change its dimensions as required.
You can now make the iPhone window wider or taller and the content inside the app will automatically adapt. It’s far better than what we had before.
So, if you’d given iPhone Mirroring a try before but were put off by the tiny window you had to work with, macOS 27 could make using your iPhone on your Mac a little more comfortable.
Improved support for ultra-wide displays
(Image credit: HP)
Before macOS 27, using an ultra-wide display with your Mac could be a downright pain. Many ultra-wide resolutions were not supported natively in macOS, meaning you had to settle for something that wasn’t quite right for your display, which could in turn negatively affect your macOS experience.
And when competing operating systems like Windowscould support ultra-wide monitors in a more comprehensive way, the frustration for Mac users was understandable.
Worse, if you ever unplugged your monitor and then plugged it in again, macOS would forget your previous display settings, forcing you to set up your ultra-wide monitor as needed every single time. If you relied on one of the best MacBooks as your daily driver and you frequently disconnected it from your desk display, this was a regular irritation.
In macOS 27, though, that’s finally a thing of the past. You’ll now be able to run an ultra-wide monitor up to 5K resolution and up to 120Hz refresh rate, meaning far more high-end displays will now work with macOS 27 without a hitch.
And Apple has fixed the disconnection issue so that hooking up your monitor no longer requires you to dial in your settings every time. Instead, macOS remembers your preferences so that no manual adjustments are needed.
That makes this a seemingly small change on the surface, but one that will be appreciated by anyone who owns an ultra-wide monitor and has long struggled with the limitations of macOS.
Code found in the iOS 27 beta suggests Apple is developing a system that could restrict leased iPhones when customers fall behind on payments. The discovery follows a recent Bloomberg report that Apple may soon launch a new "Apple Upgrade" leasing program, allowing customers to pay for hardware through monthly installments. 9to5Mac reports: The code describes a system called App Managed Features, which allows an authorized financing or provider app to enroll an iPhone and perform ongoing status checks. If the contract is no longer in good standing, Apple's system services can place the iPhone in "Restricted Mode," which blocks access to most apps until the payment or contract issue is resolved, while keeping a small set of apps available. The fixed allowlist currently found in the iOS 27 beta includes: Accessibility Reader, App Store, Health, Magnifier, Phone, Clock, Settings, Wallet, Passwords, and the Restricted Mode interface itself.
Apps that can send critical alerts, such as Messages, Home, and certain medication or safety apps, may also remain accessible. However, the provider appears to have some control over those exceptions. The code does not appear to cancel, suspend, or otherwise modify App Store subscriptions associated with blocked apps. As a result, a subscription could continue billing even while access to its app is restricted.
Additionally, there isn't a fixed number of missed payments that automatically triggers the restrictions. The financing provider's app decides when to lock the device based on its own policies. Finally, the new framework also introduces a new type of activation lock called "Partner Finance Lock," which is meant to prevent users from erasing, reselling, or stripping a restricted device for parts.
Frontier AI companies often refer to their models as “helpful assistants” or try to compare them to entry-level employees.
But new research from the UK’s AI Security Institute reinforces how large language models suffer from a common flaw that would land many human employees in hot water with their employers: they cheat.
In other words, these models are so committed to completing their tasks that they will break the rules, cut corners and deceive their own users to accomplish them.
“Every model we have tested for this behavior attempted to cheat,” the AISI report said. “Models did not reliably report this behavior when asked, and often did not reason about it in their chain-of-thought, suggesting that detecting cheating will likely require robust monitoring methods.”
The research tested OpenAI’s ChatGPT 5.4, 5.5 and 5.6 models, along with Anthropic’s Claude Opus 4.7 and Mythos Preview.
AI models from Anthropic and OpenAI were tested for cheating. All demonstrated some level of deception in their tests. (Image Source: AISI)
The AISI ran models through a series of “Capture-the-Flag” cyber evaluations, where they obtained points by performing offensive cybersecurity related tasks, like exploiting a particular vulnerability or reverse engineering compiled code.
The research defines “cheating” as “taking an action that is out of scope for the task or explicitly disallowed by the rules, in order to achieve a goal through a shortcut, workaround, or unintended solution that the task was not meant to, or should not, permit.”
It captures a range of observed behaviors from models, like searching the internet for solutions, attacking or escalating privileges on unrelated systems, probing evaluation software to gain access to the task solution.
AI models fail to identify when they’re cheating, and often justify it as acceptable when challenged by users (Image Source: AISI)
Models also failed to acknowledge when they were cheating to complete a task, and less than 50 percent said the rule-breaking was “wrong” when challenged on it by a user.
A model’s propensity for cheating was not related to its capability, meaning newer or advanced models aren’t more likely to cheat. Rather, researchers think it stems from the techniques used during a model’s training and alignment that are to blame.
But even if rates of cheating remain constant, the problem could still worsen over time. As newer models in the future could become more proficient and learn more effective cheating techniques.
This deception also makes it difficult for labs like AISI to verify their own work, which relies on evaluating trustworthy outputs from AI systems.
Models like Claude Mythos Preview and GPT-5.6 Sol justifying their cheating to users. (Image Source: AISI)
The research underscores how AI systems can go to drastic lengths to complete their task, including blowing through or circumventing a company’s IT and cybersecurity protections.
In one instance, AISI researchers said a model was inadvertently given a cyber capability evaluation that was misconfigured and impossible to solve.
“The model tested was so persistent in attempting to cheat that it wrote and ran code on an external service, hosted on the open internet outside of AISI’s systems, in an attempt to access our evaluation infrastructure, triggering a security alert in AISI’s systems,” the report said.
While AISI said there were no data leaks or damage from the incident, the model could have successfully accessed their evaluation system had they not had monitoring in place. The institute said it implemented further controls on internal systems in response to the test.
The researchers said there are “significant consequences” to a status quo where we can’t trust models not to cheat. The behaviors are especially problematic in areas like AI safety and security research, as well as cyber operations and military decision-making, where trust outputs from the AI systems are critical.
Today, AISI said it can detect LLM cheating through a mix of manual review and LLM monitoring, but that may not always be true, and future models may be better at hiding their actions from human overseers.
“A more fundamental fix would be to train the models not to cheat in the first place – but given this kind of behavior was reported in frontier models more than a year ago, robustly aligning it away may not be easy,” researchers wrote.
After a year and a half spent downplaying calls for AI safety regulations, the Trump administration has sharply reversed course, embracing a level of government scrutiny of frontier AI systems before public release–a far stricter stance than the Biden administration took.
An executive order designed to be friendly to the AI industry was meant to let the federal government briefly review some new models on a voluntary basis.
When the Trump administration, suddenly and without much warning, slapped export controls on Anthropic’s Fable 5 and Mythos 5 in response to private sector threat intelligence reporting, the U.S. AI industry officially entered its regulatory era.
But key questions and gaps remain. It’s not clear why the administration drew the line where it did, or whether they will move it again in the future.
While newer models like Mythos and OpenAI’s Daybreak do have stronger cybersecurity capabilities, the private sector reports the administration relied on describe capabilities already available in older commercial, open-source and Chinese models that nearly anyone can access.
CyberScoop spoke with current users of the latest frontier models, including OpenAI’s ChatGPT 5.5 and Fable 5, to learn more about what these models are currently capable of in offensive and defensive cybersecurity.
Cybersecurity experts and former government officials say the administration may be playing catch up on threats that have been building for years as it has more fully realized the national security implications of the technology.
Are the models breaking new ground or just breaking things?
Users of Chat GPT 5.5, introduced this past April, and Fable 5 tell CyberScoop those models have been largely helpful to their work, even as they complained about high token usage and safety guardrails that hinder, but don’t meaningfully prevent, defensive cyber tasks.
Eyal Webber Zvik, chief strategy officer at Cato Networks, a cloud and cybersecurity network provider in OpenAI’s Trusted Access in Cyber program, said they use GPT 5.5 and later OpenAI models to scan and triage internal codebases for vulnerabilities, test new safeguards and provide “highly autonomized service” to their customers.
Zvik wouldn’t disclose how many bugs 5.5 has found but said the company’s view is that it helps both find bugs that humans missed and rank which ones to patch based on factors like each bug’s exploitability.
“It is now a native part of our development environment and cycles, and we use those models to scale our entire codebase and make sure what we release into the service that our customers use to run their networks and network security has the least likelihood of having any vulnerabilities that can be exploited,” said Zvik.
John Hopper, vice president of engineering at SpecterOps, an identity security company, said newer models like GPT 5.5 are sharper and more persistent in pursuing their tasks.
“That can be a good or bad thing,” he noted.
One metric that SpecterOps tracks is how long it can keep a particular agent working before it moves off task or fails. That metric “matters a lot” because the longer an agent works without human help , the more agents a single operator can run at once.
Hopper said this provides defenders with immense value, and pushed back on the idea that the offensive capabilities the models offer are automatically more beneficial to malicious hackers. There is “a modicum of grounding that the industry needs when we talk about these models.”
“Yes, AI frontier tools will lower the barrier of entry, but these problems have always existed,” he said. “I don’t actually believe that AI is going to remove the needle in the haystack problem, but by howdy, using my two hands to find that damn needle, compared to using a backhoe, I can tell you which one I’d rather be driving.”
Eran Kinsbruner, vice president of product marketing at software security firm Checkmarx, told CyberScoop that later models like OpenAI’s Codex Security and GPT 5.5 are noticeably easier to set up and run with local systems, even for less technical users. That alone gives them an edge over many cybersecurity tools where interoperability is a constant concern.
However, GPT 5.5 burns through tokens at a much faster rate. He recalled one instance of using it to scan a medium-sized repository in three different programming languages.
“After 26 minutes I almost ran out of tokens, and it didn’t provide anything, just created a threat model for me and told me you want to buy more tokens?” he said.
In other instances, some of the scan results he received were not comprehensive.
Further, he expressed frustration with some of the guardrails designed to prevent risk – like only allowing users to scan local files but not code repositories like GitHub – “makes not too much sense” given how often developers must work with remote code.
Those kinds of guardrails – which can prevent models or developers from injecting malicious code or prompting into their models – sit at the heart of the debate in Washington D.C. and around the world. Some users feel differently about their utility.
Kinsbruner said that doesn’t make sense for organizations like his, which work with thousands of different enterprise organizations with thousands of different code repositories spread across the internet.
“I cannot imagine how large-scale developers could just jump into this solution and make it an enterprise-grade, enterprise-level, de facto cybersecurity solution” out of it, said Kinsbruner.
OpenAI did not respond to a request from CyberScoop for an interview on GPT 5.5. The company has since released another model, GPT 5.6, that they said is more efficient at token use.
The White House’s crash course in AI cyber risk
The White House keeps changing its line on whether and how the U.S. government should limit the release of commercial frontier models. The shift comes from lessons learned since coming into office in Jan. 2025. Trump threw out Biden-era regulations meant to steer the industry toward safer models. Top officials like Vice President JD Vance argued against restricting industry progress.
Less than two years later, administration officials worry about the impact of speed and scale – two things AI excels at – in cyberspace.
According to Will Loucks, senior director of intelligence at the Office of the National Cyber Director, over the past two years the number of exposed and known vulnerabilities has shot up. Threat actors exploit those flaws faster before defenders can fix them. Once inside, the time from initial access to full network control shrinks.
“So in other words, every stage of the cyber operations lifecycle that a threat actor has to move through to get to a victim network and achieve an outcome, they’re just moving through more quickly faster,” said Loucks at a July 16 event in Washington D.C.
Speaking about AI in particular, Loucks said one of the defining characteristics of the technology is its ability to lower barriers for threat actors.
“Sometimes speed and volume have a threatening aspect alone, even if sophistication isn’t quite increasing in the same way, and the reason for that is because it places pressure on defenders…to triage alerts more quickly,” he said.
Jordan Rae Kelly, former director for cyber and incident response on the White House’s National Security Council during Trump’s first term, told CyberScoop that the changes over the past two years reflect the lessons the White House has learned on the issue since returning to office.
In the early days of this administration, Kelly said, “there is a sense and a spirit that the Biden administration was limiting AI and there was a kind of a rip-it-all-off [attitude], everybody go and do whatever, we will be the biggest and boldest and brightest.”
“I love that talking point, but I think what you’ve seen is probably an education over the last 19 months, where people [in the White House] have said that’s a challenging premise to put into place, knowing about the potential downsides and capabilities,” she added.
Michael Daniel, former White House cyber coordinator under President Barack Obama, thinks the horse may already be out of the barn.
Daniel, now head of the Cyber Threat Alliance, a membership nonprofit group focused on cyber threat information sharing between industry and government, said his members report that AI is being used to do things “faster and at a slightly bigger scale” but aren’t yet seeing the flood of exploitation that analysts have warned about. Not yet.
“I think what we’re seeing right now [and] talking about is ‘okay, where are the step changes [in the cyber threat landscape] actually going to occur?” said Daniel. “Are we and when will we see the explosion in vulnerability reporting from these Mythos-like capabilities? That’s what’s really got their attention right now.”
But Mythos and OpenAI’s Daybreak models are restricted to select organizations, and neither has publicly released its most powerful cybersecurity models to the public. That dynamic won’t last.
The UK’s AI Security Institute estimates that open source and foreign LLM models are between 4-7 months behind frontier U.S. models. In that setting, it’s hard to stop the development of AI models worldwide through export controls or other limits.
“It’s not like we’re buying ourselves five to ten years on this,” he said. “We’re not, and so I’m not sure the impact on the defenders who are trying to obey the law is worth whatever small hiccup we cause for our adversaries.”
Kelly said there’s merit to the administration’s current position, even if it took time to get there. Many federal cybersecurity procedures that operated even a decade ago – such as a Vulnerabilities Equities Process that could take days or weeks to consider the pros and cons of keeping an exploit – are no longer practical.
“All of that work to some degree, is out the window, because you can’t meet with the regularity you would need to meet to adjudicate vulnerabilities that are being found in seconds and exploited in minutes,” said Kelly.
But Kelly and others say that’s also because AI capabilities in cybersecurity are developing faster than policymakers can react, even in the best of times.
Key questions remain and the administration’s balance between national security and backing domestic industry will likely shift in response to new events. The administration wants a framework that can predict and manage the risks of AI models today and tomorrow. That may be harder than it sounds.
“Do I think they’ve been clear? No,” said Kelly. “But I think it’s a place where clarity is really hard to achieve.”
The home appliance giant LG Electronics USA said this week it plans to suspend any apps built for its smart TVs that turn one’s television into an always-on residential proxy node. The move comes less than a month after researchers found that more than 42 percent of games and other apps available for download on LG’s webOS store allow unknown third-parties to route their Internet traffic through a user’s TV.
Proxy SDK prevalence among smart TV apps for LG (webOS) and Samsung (Tizen OS) televisions. Image: Spur.us.
On July 2, we featured research by the security firm Spur that examined the prevalence of residential proxy software development kits (SDKs) in smart TV apps. Spur found more than 42 percent of apps available for download on LG smart TVs include SDKs that turn one’s television in a proxy node indefinitely, and that more than a quarter of the apps made for Samsung’s Tizen operating system had similar residential proxy components.
Responding to questions about Spur’s research, LG Senior Vice President John Taylor told KrebsOnSecurity the company was working with app developers to remove the residential proxy option from their apps on the webOS platform. Developers that fail to comply, he said, will find their apps suspended.
“A residential proxy network is not an intended use for LG smart TVs, and LG Electronics is working with developers to remove the residential proxy option from their apps on the webOS platform,” Taylor said. “If this option is not removed, these apps will be suspended.”
Taylor said LG is committed to keeping residential proxy networks out of its smart TV apps going forward, and that the company’s review of those apps is “well underway now.”
“As part of our ongoing efforts to enhance platform quality and the user experience, LG will continue to strengthen our evaluation process for developer-submitted apps, including those that incorporate residential proxy SDKs,” Taylor wrote in an emailed statement.
App makers looking for ways to monetize their creations can turn to residential proxy providers, which pay developers to include SDKs that turn the user’s device into a residential proxy node that is rented to paying customers. In the case of LG and Samsung smart TVs, Spur found residential proxy SDKs bundled with everything from simple games like Pac-Man to screensavers and file utilities.
A Pac-Man smart TV app from Bright Data offers users the choice between viewing ads in the game or agreeing to allow their TV to serve as a residential proxy node. Image: Spur.us.
Spur’s report found the residential proxy network Bright Data accounted for a majority of proxy SDKs across both Samsung and LG smart TVs. In a statement shared with KrebsOnSecurity, Bright Data said its network is built on consent and responsibility and operates by LG and Samsung terms.
“Every peer opts in through a dedicated screen and receives value in return; every customer is vetted, and our practices have now undergone a second independent audit by PwC,” the statement reads. “We remain committed to an open, transparent internet where legitimate businesses, researchers, and institutions can responsibly access data that lives in the public domain.”
Bright Data and other proxy providers named in Spur’s report all say they follow rigorous know-your-customer processes to validate legitimate uses of their services, which is often heavily tied to content-scraping activities by said customers. The proxy companies also say they incorporate technological countermeasures to prevent proxy service customers from being able to interact with and control other devices on the proxy user’s local network.
Spur argues the problem is not that residential proxy networks exist, but rather that they are being embedded at scale in devices that most consumers do not think of as computers and are not equipped to audit.
“A one-time consent prompt buried in a TV app is not a substitute for meaningful transparency, ongoing control, and platform oversight,” Spur’s Trevor Sutter wrote. “The risk is amplified when consent comes from individuals within the household who use the device but shouldn’t give consent, such as minors.”
LG’s announcement that it is culling residential proxy SDKs from its app store is welcome news, but the company recently came under fire for another questionable partnership: Pimping McAfee security products via software drivers included in its high-end LCD monitors.
Earlier this week, the Youtube channel Gamers Nexus showed that certain LG LCD monitors will automatically install an app that promotes paid McAfee antivirus subscriptions, and that the app arrives through Windows Update without an approval prompt.
Update, July 22, 1:06 p.m. ET: Added statement from Bright Data.
Apple's MacBooks seemingly have some big changes on the horizon. One of the biggest is the rumored MacBook Ultra packing that OLED screen, and this is something to be excited about, don't get me wrong. Well, worries about the price aside, and wider issues around Mac pricing overall, but there's something I'm more concerned about now.
Namely the direction in which Apple is heading with its M-series silicon, following the latest rumor dump from prolific leaker Mark Gurman. In last weekend's Power On newsletter for Bloomberg, Gurman dropped a whole lot of fresh info about Apple's CPU roadmap and the incoming M6, M7 and M8 chips.
The leaker tells us — and bear in mind, this is all just rumors, but eye-opening speculation nonetheless — that Apple is doing something very different with the next range of M silicon. The M6 will consist of just a vanilla chip, with no more powerful variants, which isn't something Apple has ever done before — there's always been a Pro and Max version (if not an Ultra, too).
This is because Apple apparently wants to usher in the M7 series more swiftly, with the M6 debuting late this year and the M7 following in the first half of 2027, just six months (or so) later. The M7 will have a full complement of Pro, Max and Ultra chips, the former two following at the end of 2027, with the Ultra in 2028. Gurman believes this is Apple's current plan and the reason for rushing to get to the M7 is AI.
Cue multiple groans, no doubt. Gurman theorizes: "The reason for this break with tradition: AI. Apple had been planning major neural-processing upgrades for the M7 family and ultimately decided those improvements were important enough to justify accelerating the next generation rather than completing the M6 lineup."
We're further told that beyond this, Apple is "already developing M8 chips with even greater AI capabilities" to land in 2028.
Gurman concludes: "The takeaway is that AI is no longer just another feature Apple's chips need to support. It is now shaping how those products are designed and when they are shipped."
A wise path to follow?
(Image credit: Lance Ulanoff / Future)
So, what am I worried about here exactly? Obviously there's a strong suggestion that AI is taking much more of a center stage role with Mac chips, and given what's currently going on with MacBooks, I've got to wonder about the wisdom of this — and how it might affect Apple's PC sales.
Currently Mac sales are strong and considerably buoyed by the launch of the MacBook Neo in March 2026, a budget laptop that has undoubtedly flown off the shelves. However, Apple's recent price hikes have poured cold water on those sales prospects to an extent, even hitting the MacBook Neo, which went up by $100 in the US (£100 in the UK, and AU$150 in Australia).
Granted, while it's odd for a device to get a price rise like this so soon after launch, given the demand around the Neo, it's not going to be sidelined with this increase (although the upper-tier model with 512GB storage now looks shakier and out of budget territory, frankly). The harder pill to swallow is the harsher MacBook Air price hikes, and as we pointed out a few weeks back, some wallet-friendly Windows 11 laptops now look a lot more tempting after Apple's price increases.
Looking on Reddit, it's not hard to find a lot of unhappiness about the MacBook price hikes. There's considerable angst from some people who were thinking of pulling the trigger on a new Apple laptop, and then watched those price increases swoop in, and are now feeling the opposite of buyer's remorse (holdout's regret?). Some of those folks are acting with their wallets, closing them and vowing to wait for a better buying opportunity now that they've missed the boat, as it were.
That could be a long wait, though. Of late, we're hearing a lot of gloomy news about the RAM crisis getting worse later this year, and in 2027 — maybe a lot worse. The memory supply situation is what caused Apple to jack up Mac prices, of course, and it's difficult to see how the MacBook maker will continue to weather this storm if it worsens next year, given that its resources in terms of built-up inventory of RAM (or SSDs) are clearly running dry (or have run dry).
Rumors have continued to persist about Apple potentially tapping Chinese memory chip makers for its RAM needs, as an alternative to the big three mainstream suppliers (Micron, Samsung, SK Hynix). That's a worrying development in itself, and it's doubtful whether even Tim Cook can pull off navigating that kind of negotiation before he leaves his Apple CEO role (there are various possible hurdles to clear and governmental complications, put it that way).
Diving into AI headfirst
(Image credit: Apple)
I don't think any of this bodes particularly well for next year's Mac sales given the prospect of pressure around more price hikes (although yes, other PC makers face the exact same unfavorable conditions). And then we come back to this apparent renewed focus on AI, which hasn't gone down well on Reddit and other social media.
Here's one Redditor who sums all this up neatly: "I was about to buy a new MacBook to replace my M1 Pro with an ailing battery and small storage. But then Apple jacked up the prices by almost 20%. And now they are diving into AI headfirst. It might be time for me to look at Framework laptops."
There's a strong backlash against AI right now, whether that's about sprawling data centers and them chugging huge amounts of power, or Copilot in Windows 11 and indeed Apple Intelligence. So, it's no surprise that the idea of "diving into AI headfirst" has been greeted by a lot of skepticism.
What exactly is Apple Intelligence going to serve up in terms of compelling features in the next year or two? The trouble is that this feels very much up in the air, yet Apple is supposedly building its Mac strategy around this now. It seems like a bit of a leap into the unknown to say the least, and I'm not surprised that it's making Mac fans nervous.
Okay, so this AI focus is still just a rumor, although Gurman is one of the more reliable Apple sources out there, and it appears to be a well enough fleshed-out piece of speculation. Even if it's true, Apple may yet switch away from this course, but for now, I'm struggling to find the logic of shifting to a heavier AI focus so soon.
Maybe the reasons for this will become clearer as we see how Apple Intelligence develops in the near future. However, in the meantime, I think there are reasons to be concerned about Mac sales potentially flagging next year, between disgruntled MacBook holdouts, the continued RAM crisis, and cynicism around Apple's apparent new AI-led approach.
The Mission 1 Pro is as rugged as you’d expect any GoPro action camera to be, with a highly water-resistant design — even without an additional casing — and a shock-resistant body that can survive tumble after tumble, but with 8K recording quality it enables you to capture footage with a real cinematic feel.
Thanks to impressive subject tracking and GoPro’s HyperSmooth stabilization tech, the Mission 1 Pro can capture the action smoothly whether you’re wearing it yourself, or following your subject along a trail or down a piste.
Plus, thanks to the new larger 1-inch sensor, low-light video quality is better than ever for a GoPro — letting you take advantage of the roughly two hours of 4K 60p recording time available to you on a single charge even as the sun starts to set.
(Image credit: Future)
Meanwhile, HLG HDR format and GP-Log2 support, control over ISO, EV, white balance and shutter speed, and the freedom to record in 8K/30p open gate — the full height and width of the versatile 4:3 aspect ratio sensor — will give you more artistic and post-processing options that make this action camera feel more like a professional piece of kit.
The downside of all this is that the Mission 1 Pro doesn’t come cheap. It’ll set you back $699 / £599.99 / AU$1,099.95 if you aren’t already a GoPro member, while members can take advantage of a healthy discount, with prices starting at $599 / £509.99 / AU$949.95. This sure is pricey even for one of the best action cameras when DJI rivals cost substantially less, but if you’re going to take full advantage of its cinematic capabilities, it’s certainly one to consider.
The only real disappointment is that its photography capabilities left me wanting. With its optional point-and-shoot camera-style grip I had hoped that the GoPro Mission 1 Pro might perform well as a stills snapper, but it did struggle at times. Hopefully, the ILS model with its Micro Four Thirds lens mount will deliver a more all-round performance when it debuts later this year.
The grip does at least help make the Mission 1 Pro much easier to handle when shooting video, and it feels like a useful upgrade for anyone who plans to use this GoPro as their new vacation camera, and take advantage of its improved video and audio quality, and durable design.
GoPro Mission 1 Pro: Price and release date
$699 / £599.99 / AU$1,099.95 (non-member price)
$599 / £509.99 / AU$949.95 (member price)
Other models in the series are the Mission 1 and Mission 1 Pro ILS
The GoPro Mission 1 Pro is available to buy now for $699 / £599.99 / AU$1,099.95 (non-member price, or $599 / £509.99 / AU$949.95 for members), while the more basic Mission 1 will cost you $599 / £529 /AU$949.95 (non-member price, or $499 / £439.99 / AU$799.95). The Mission 1 Pro ILS costs the same as the Pro, but it won’t launch until later in 2026.
Other bundles include one with a point-and-shoot camera-style grip, which I had for my review. This bundle feels quite pricey at $779.99 / £679.99 / AU$1,229.95, seemingly with no member discount at the time of writing.
The Mission 1 Pro and Mission 1 share a lot of the same features and hardware, but the Pro version unlocks 8K 60fps footage and 8K Open Gate 4:3 filming for greater editing flexibility. You’ll also unlock slow-mo 4K video at 240fps, and 1080p up to 960fps for more extreme slow-mo shots.
(Image credit: Future)
Meanwhile the ILS model is interesting as it takes away some features — such as built-in autofocus and the ultra-wide lens of the regular Pro — to give you more control, with manual focus control and a Micro Four Thirds (MFT) mount, enabling you to attach a wide variety of lenses for different cinematic effects. If you’re more of a cinematographer than a casual user, the ILS model could be what you’re after.
That said, even the base-model Mission 1 is a fair bit pricier than the GoPro Hero 13 Black — our current pick for the best overall action camera. If you care about low-light shots the Mission 1 could be worth the price premium, but for a lot of people who are just looking to capture fun clips of their action sports escapades even the basic model may be overkill.
Value score: 3/5
GoPro Mission 1 Pro: Specs
Type:
Action camera
Sensor:
50MP 1-inch type CMOS
Displays:
Front: 1.4-inch LCD
Rear: 2.59-inch OLED touchscreen
Memory:
None; supports V30 and Class A2 rated microSD cards
Video:
8K up to 30fps (no crop) / 8K/60p (cropped)
ISO range:
Video: 25-6400
Photo: 100-3200
Burst shooting
60fps
Processor:
GP3 SoC
Connectivity:
BlueTooth, USB-C, 5GHz Wi-Fi 6
Dimensions:
80.1 x 52.1 x 44.3mm (including buttons and lens without hood)
Weight:
207g
GoPro Mission 1 Pro: Design
The ideal action camera is one that can handle just about anything you throw at it, and the GoPro Mission 1 Pro GoPro is a camera that’s designed to be pushed to its limits.
If you’re a fan of underwater video, the Mission 1 Pro boasts a design that’s waterproof to 66ft / 20m without the need for a housing — plus the hydrophobic lens cover should help keep water splashes and droplets from obstructing your view massively when you’re using the camera above the waves.
It’s also rated to work in temperatures between -20C and 45C, with its power and shoot buttons designed to make them easier to press while wearing the kind of thick gloves that are necessary in frigid climates.
(Image credit: Future)
The Mission 1 Pro’s shock-resistant body can also survive tumbles and drops, giving it an almost indestructible feel. While I never pushed the Mission 1 Pro too far, I never felt like I needed to be careful with it.
Moving on from the action-focused aspects, the 2.59-inch OLED rear touchscreen display offers enough real estate for you to easily manage and track your settings while also leaving plenty of room to clearly see your compositions.
Meanwhile, the 1.4-inch front LCD is a good size for vloggers to track what is/isn’t in frame as they talk to camera, with a convenient and concise readout of your shot settings letting you know that you haven’t, for example, accidentally swiped to a different mode while turning the camera 180 degrees.
Included in the box are a lens hood to help reduce glare — just note that it can fall off easily, so keep an eye on it, or be prepared to pick up replacements — a 2150mAh Enduro 2 battery, an adhesive mount, mounting buckle and a thumb screw.
You’ll need to provide your own microSD card, as there’s no internal storage. The Mission 1 Pro relies on V30 and Class A2-rated microSDs for the best performance; A1 cards will work, but will encode more slowly.
You can pick up a series of attachments for this camera, including ND filters; however, the bigger sensor size of this camera means the Hero 13 Black’s suite of lens attachments can’t be used here, and sadly there aren’t, as yet, equivalent accessories for the Mission Pro 1.
Other GoPro attachments, such as harnesses or helmet/bike mounts, should be compatible, as the Mission 1 Pro relies on the same bracket as models that came before.
(Image credit: Future)
Lastly I’ll mention the point-and-shoot camera-style grip attachment, which was supplied to me for testing alongside the Mission 1 Pro and can be purchased separately ($99.99 / £89.99 / AU$154.95) or as part of a bundle. The grip is incredibly simple, yet effective at transforming the Mission 1 into something that feels more like a regular point-and-shoot compact.
When it comes to capturing quick videos, such as clips at the beach or at a festival, the grip makes it easy to hold the camera without the risk that your hand will get in the way of the lens, although as mentioned, the Mission 1 Pro doesn’t perform as well as a stills camera — more on that below.
Design score: 5/5
GoPro Mission 1 Pro: Performance
To make video capture easier than ever the Mission 1 Pro features a slew of preset capture modes that adjust settings based on your needs, with modes for Vlogging, SportPOV, TimeLapse, Open Gate, SlowMo, Low Light, Dive and plenty more.
You can tweak the settings for each mode to fine-tune your image quality needs, and also manage modes by hiding ones you never use, or adding custom settings that are your go-tos.
For most scenarios, I found the 4K open gate option to be the most versatile. The image quality is solid, and you have the option to reframe the footage to 16:9 or 9:16 later based on your needs — plus it’s less of a drain on the battery than shooting in 8K.
Vlog-style recording is made easy thanks to the front screen, and automatic subject tracking that centers you in the frame — ideal for casual vloggers who constantly find they steadily drift out of shot unless they pay close attention to the camera they’re holding.
The audio quality is decent too, picking up voices and essential sounds, but cutting out wind noise.
Slow mo is also a lot of fun, though as you’ll see in my example video above, raw untuned footage at resolutions lower than 4K can start to look grainier than is ideal — upscaling and/or sharpening the image with editing software is essential to get things to look better.
The larger sensor also helps in low-light conditions, especially with the new Low Light shooting mode. Unsurprisingly, in proper darkness the camera still struggles, but with some illumination remaining the camera performs well in this mode.
(Image credit: Future)
Where I’m less convinced of the Mission 1 Pro’s abilities is when it comes to stills photography, which is a massive shame considering you have the option to add the point-and-shoot grip.
Whenever I tried to take photos I found myself fighting against its lack of autofocus — or even manual focus — and the less-than-ideal zoom quality. The camera has a fixed focus with a minimum focus distance around 60cm — which counts out close-up photography — and an optimal focal distance of about 1.5m. This also means it isn’t focused to infinity, so landscapes will appear a little soft in your shots, especially since the larger sensor has a shallower depth of field than previous GoPros.
In addition, as soon as you zoom in you’ll find the image quality drops a fair amount, although this is to be expected from the digital zoom this camera uses.
The overall quality isn’t terrible, though, and the point-and-shoot grip does offer some useful versatility for video recording beyond the more traditional action camera use-cases.
Performance score:4.5/5
FutureFutureFutureFuture
Testing Scorecard
Attribute
Notes
Score
Design
With its rugged build and larger display you couldn’t realistically want more from the design of the Mission 1 Pro. My only gripe would be that longtime GoPro fans may be frustrated that they can’t, for now at least, bring all of their lens attachments over to this new camera.
5/5
Performance
The Mission 1 Pro delivers some great-looking video with great audio even without external accessories, and even in low light. I’m skeptical of the camera’s photography credentials, though.
4.5/5
Value
The GoPro Mission 1 boasts a slew of pro-level enhancements, although these do come with a higher price tag.
3/5
Should I buy the GoPro Mission 1 Pro?
Buy it if…
You want a professional-quality action camera With 8K recording and a full suite of settings that you can adjust, the Mission 1 Pro delivers as professional-grade camera more so than its predecessors.
You don’t want to think too much Like its predecessors, the Mission 1 Pro boasts a selection of preset recording modes that you can fine-tune to suit your video needs. If you want good shots without putting in a lot of thought and effort, the Mission 1 Pro will still help you capture great shots.
You want a small camera that’ll last With a hardy exterior that can survive high heat, freezing cold, and excursions deep below the water, and roughly two-hour battery life when shooting 4K/30p, the GoPro Mission 1 Pro will be ready when you need it.
Don’t buy it if…
You want to capture high-quality stills As an action camera the Mission 1 Pro is great, but as a stills camera I found it wanting, even if there is a dedicated point-and-shoot grip accessory.
You’re on a budget The GoPro Mission 1 Pro is a lot pricier than other action cameras out there, and if you’re a more casual and budget-conscious user the upgrades might not be worth the outlay.View Deal
You want lens options The Mission 1 Pro doesn’t support the lenses GoPro has previously offered with its action cameras like the Hero 13 Black. For lens variety you’ll want to check out the Mission 1 Pro ILS model when that launches later this year.View Deal
Also consider
GoPro Hero 13 Black A more budget-friendly option, and, even years after launch, still a solid action camera performer, the GoPro Hero 13 Black doesn’t boast the 8K quality of the Mission 1. However if you’re after something smaller and lighter for more casual use it’s an excellent alternative.
DJI Osmo Action 6 If you’re already part of the DJI ecosystem, want tools like wireless Hi-Res audio recording when using a DJI Mic 3, or are just after some frankly excellent battery life, decent waterproofing, and a generous 47GB of built-in storage, this DJI camera could be for you. Just note that its video is capped at 4K resolution.
Insta360 Ace Pro 2 If 8K is a must, the best alternative for you is the Insta360 Ace Pro 2. Admittedly it isn’t as capable as the GoPro Mission 1 (the 8K quality doesn’t support the best image stabilization) and overheating is an issue, but the Insta360 Ace Pro 2 is a superb all-rounder.
I spent just over a month with the Mission 1 Pro, taking it out on excursions to snap photos and videos to test its capabilities not just as an action camera, but as a point-and-shoot video capture tool — with that latter case aided by the point-and-shoot grip.
If you’re planning to use the GoPro Mission 1 Pro as your vacation camera — aided by its durable design and waterproofness — the grip is super helpful at helping to keep you hands out of frame, and as a pure action camera the GoPro fits the same mounts and accessories as previous models like the Hero 13 Black (save for the lens attachments).
Inspired by shows like Hunted and Jet Lag The Game, a lot of my testing focused on running and vlogging with the action camera to capture my fitness progress and to record personal cuts of games of hide-and-seek across London — passing the camera between my friends so they could record their stints too. They found the GoPro super easy to use, with the smoothing tech and subject tracking ensuring thaty vlogs and action shots look great without requiring constant focus on what the camera is doing — you just press go and it’ll come with you.
BrianFagioli writes: Valve has released Proton 11.0-1, a major update to its Windows compatibility layer for Linux that makes more games playable while fixing a long list of bugs affecting existing titles. The release restores compatibility for many EA games after a recent EA App update, moves classics like Resident Evil (1996), Resident Evil 2 (1998), Dino Crisis, and SHOGUN: Total War from Proton Experimental into the stable release, and adds support for games including Gothic 1 Classic, X-Plane 12, Breath of Fire IV, and Deadly Premonition. Valve also fixed crashes in HELLDIVERS 2, restored No Man's Sky VR support, improved Steam Overlay compatibility with EA games, addressed KDE and GNOME desktop issues, and rebased Proton on Wine 11.0 with updated graphics components. The full list of changes can be found here.
Login
