Anthropic’s Mythos AI model, a powerful cybersecurity tool that the company said could be dangerous in the wrong hands, has been accessed by a “small group of unauthorized users,” Bloomberg reports. An unnamed member of the group, identified only as “a third-party contractor for Anthropic,” told the publication that members of a private online forum got into Mythos via a mix of tactics, utilizing the contractor’s access and “commonly used internet sleuthing tools.”

The Claude Mythos Preview is a new general-purpose model that’s capable of identifying and exploiting vulnerabilities “in every major operating system and every major web browser when directed by a user to do so,” according to Anthropic. Official access to the model is limited to a handful of companies through the Project Glasswing initiative, including Nvidia, Google, Amazon Web Services, Apple, and Microsoft. Governments are also eyeing the technology. Anthropic currently has no plans to release the model publicly due to concerns that it could be weaponized.

“We’re investigating a report claiming unauthorized access to Claude Mythos Preview through one of our third-party vendor environments,” an Anthropic spokesperson said in a statement to Bloomberg. Anthropic currently has no evidence that the unauthorized access is impacting the company’s systems or goes beyond the third-party vendor’s environment.

The model was reportedly accessed illicitly on April 7th, the same day that Anthropic announced it was releasing Mythos to a limited number of companies for testing. The group that gained the unauthorized access has not been publicly identified, though Bloomberg reports that its members are part of a Discord channel that seeks out information about unreleased AI models.

The group accessed Mythos by using knowledge of Anthropic’s other model formats obtained from a recent Mercor data breach to make “an educated guess” about its online location. Members have been using Mythos regularly since gaining access — providing screenshots and a live demonstration of the model as evidence to Bloomberg — though reportedly not for cybersecurity purposes in an attempt to avoid detection by Anthropic. Other unreleased Anthropic AI models have also been accessed by the group, according to Bloomberg.

Huawei has launched its passport-style foldable in China, ahead of similar devices said to be coming from Apple and Samsung. That makes the Pura X Max the first wide foldable phone that’s actually available to buy, with today’s launch now giving us the full specifications and price after its design was unveiled by Huawei last week. 

There are two versions of the Pura X Max, both of which are available to preorder starting today. The standard version starts at 10999 yuan (about $1,613) for 12GB of memory and 256GB of storage, or 11,999 yuan (about $1,800) for 512GB of storage. The Pura X Max Collector’s Edition starts at 12,999 yuan (about $1,906) for the 16GB RAM / 512GB storage configuration, or 13,999 yuan (about $2,053) for the 16GB/1TB variant.

The foldable’s wide display orientation is the biggest standout, opting for a book-like design compared to the foldables we’re familiar with. It features a 5.4-inch cover display with 3500 nits of peak brightness. Opening the Pura X Max reveals its 7.7-inch internal display, which tops out at 3000 nits, with both screens supporting a 1–120Hz LTPO adaptive refresh rate.

Apple and Samsung are both reportedly working on their own wide-style foldable phones. Various leaks have appeared that give us a potential early look at the rumored iPhone foldable and wider version of Samsung’s Z Fold, though both devices aren’t expected to arrive until the second half of 2026 by comparison — giving Huawei a significant head start on this emerging design trend.

On the hardware side, the Pura X Max includes a homegrown Kirin 9030 Pro chipset and 5,300mAh battery, with support for 66W wired charging and 50W wireless charging. Its triple camera setup includes a 50 megapixel main camera with variable aperture, a 50 megapixel telephoto lens, and a 12.5 megapixel ultra-wide shooter. The phone itself runs Huawei’s latest HarmonyOS 6.1, supports the Huawei M-Pen 3 Mini stylus, and provides a wide array of AI-powered features, including pose recommendations for photographers.

Huawei hasn’t confirmed whether the Pura X Max will be released outside of China, but that prospect should give Samsung and Apple cause for concern.

All empires eventually fall, and it seems the creative software industry has collectively decided that Adobe’s time has come. The Creative Cloud provider’s suite of design tools have been considered the industry standard for decades — despite unpopular decisions to fully embrace generative AI and abandon software licenses in favor of expensive, complicated subscriptions. 

Pricing in particular has given competitors an opening to attack. Some of the best alternatives aren’t just undercutting Adobe’s price — they’re available for free. People love free.

One example that was announced this week is Autograph, motion design software akin to Adobe After Effects. Autograph was acquired by Cinema 4D maker Maxon last year, and has now been relaunched with free access for individual users. It initially cost $1,795 for a permanent license (or $59 per month on subscription) when it launched in 2023, which was a hard sell compared to the $34.49 per month standalone After Effects subscription that Adobe demanded, and continues to charge today. And while Autograph isn’t directly comparable, it provides a similar suite of animation and VFX tools and doesn’t charge a dime.

Perhaps coincidently, Canva also dropped its own bomb on Adobe’s After Effects this week. Canva has made the full version of Cavalry available for free instead of locking the motion graphics software behind its own user subscriptions, after the design platform acquired it back in February. If that sounds familiar, it’s because Canva did a similar thing last year with Affinity — a trio of apps it acquired that provide similar features to Adobe’s Illustrator, Photoshop, and InDesign software. While Affinity Designer 2, Affinity Photo 2, and Affinity Publisher 2 were each a one-off $69.99 payment before (or $169.99 for all three), they’ve since been combined into a single, entirely free app.

Other Adobe apps also took a hit this week thanks to the latest DaVinci Resolve 21 update. The free multipurpose post-production software — which is already considered a rival to Premiere Pro — now includes photo editing features like color-correction, masking tools, and import support for Apple Photos and Lightroom Catalog files. The update also adds support for Affinity’s .af file format, making it easier to use another free app alongside DaVinci Resolve.

Even when the Adobe alternatives aren’t free, they’re becoming more attractively priced. Apple launched its Creator Studio suite in January, which includes access to a whole host of editing apps, including Final Cut Pro, Logic Pro, Pixelmator Pro, Motion, Compressor, and MainStage. The $12.99 monthly Creator Studio fee is more affordable than Adobe’s $69.99 monthly Creative Cloud Pro subscription by comparison, and Apple isn’t forcing users into a subscription plan. You can still buy one-time licenses for individual apps on Apple’s App Store. Take that Adobe.

When we covered that announcement, several themes appeared in our comment section. One was the collective shock at how low Apple’s pricing was compared to Adobe’s despite being, well, Apple. The other was that all the Creator Suite needed was a suitable Lightroom alternative to seal the deal. Apple may yet find a way to make it happen, but DaVinci has filled that gap in the meantime.

When you pair these recent announcements with creative software that was already free, or at least subscription free, then you have an industry movement that should give Adobe something to worry about. 

Procreate has made a name for itself for being staunchly anti-AI and releasing incredible digital illustration and animation software for iPads that you can buy once and keep forever. It’s also pledged to bring them to Mac desktop devices. Blender, the free open-source 3D computer graphics software suite, is continually gaining new features, and has proved capable enough to be used in Oscar-winning feature film releases. And Figma was so good that Adobe killed its own XD product design tool in favor of trying (and famously failing) to acquire the platform, which offers a free-to-use Figma tier.

Freedom from Adobe’s app ecosystem is actually starting to look plausible. And making that freedom increasingly free is the icing on the cake.

Adobe is fully embracing AI tools that enable creators to edit their work using descriptive prompts, instead of manually using specific Creative Cloud apps. The software giant’s new Firefly AI Assistant allows users to describe what they want to change by typing their own words into a conversational interface. 

Adobe says this marks a “fundamental shift in how creative work is done” by removing skill barriers and laborious tasks, while still giving creatives full control over their work. It’ll be “available soon” on the Firefly AI studio platform according to Adobe, though no specific launch date was provided in the announcement.

The unified AI interface, which builds on the Project Moonlight experiment that Adobe introduced at its Max conference last year, automatically performs “complex, multi-step workflows” to edit projects, utilizing specific tools and apps (including Firefly, Photoshop, Premiere, Lightroom, Express, Illustrator, and more) on the user’s behalf.

Users of the Firefly AI Assistant can instruct the chatbot to “retouch this image” or “resize this for social media,” for example, with Adobe’s AI agent then providing a selection of edits to choose from alongside surfacing specific tools or sliders that allow creators to fine-tune the results. For more detailed adjustments, creatives can also open the edited results in Creative Cloud apps to finish the project.

The Firefly AI assistant will learn the user’s preferences over time, such as preferred tools, workflows, and aesthetic choices, to help make the results feel more personalized and consistent. Adobe’s AI chief Alexandru Costin told The Verge that creatives will be able to choose whether to enable this feature, and can select specific projects for the AI assistant to learn from. Creatives can also create “Creative Skills” — tools that provide specific and consistent presets — that the AI assistant can execute, or select from a library of pre-made skills at launch.

This is Adobe’s latest push into the world of AI agents, having already launched specific AI assistants for apps like Adobe Acrobat, Express, and Photoshop. Adobe says it will also bring these agentic features to third-party AI apps like Anthropic’s Claude, allowing those users to access Adobe tools outside of its own Firefly and Creative Cloud platforms.

This announcement comes alongside some new image, video, and audio editing capabilities for Adobe’s Firefly platform, which are rolling out starting today. The Firefly Video Editor is now integrated with Adobe Stock for easy access to B-roll footage, and allows users to access new features for improving color adjustments and the clarity of spoken dialogue. New editing features are also available in the Firefly image editing tool — Precision Flow, which enables creators to make and compare a wider range of generated images without adjusting their prompts, and a new AI Markup tool that lets users control where edits should be made using brush and rectangle tools or reference images.

Google is launching a new Chrome workflow feature that allows you to reuse your favorite Gemini commands across multiple webpages. Any AI prompts can now be saved as “Skills” in the Chrome desktop browser, letting you instantly run them across any tabs you select.

“Until now, repeating an AI task — like asking for ingredient substitutions to make a recipe vegan — meant re-entering the same prompt as you visited different pages,” Chrome product manager Hafsah Ismail said in the announcement. “To make this easier, we’re launching Skills in Chrome, which lets you save and reuse your most helpful AI prompts and run them with a single click.”

Skills are rolling out to Chrome users with their language set to US English starting today. Once you have access to the feature, it can be managed by typing a forward slash ( / ) in Gemini and clicking the compass icon. AI prompts can be saved as Skills directly from your Gemini chat history on desktop, where they’ll then be available to reuse on any other desktop devices that are signed into the same Google account on Chrome.

The aim is to spare Chrome users from having to manually retype frequently used Gemini prompts or having to copy and paste them over from a saved list. Some of the Skills made by early testers include commands for calculating the nutritional information of online recipes and creating a side-by-side comparison of product specifications while shopping across multiple tabs, according to Google. 

The company is also launching a library of preset Skills that you can save and use instead of making your own. These ready-to-use Skills can also be customized to better suit your needs, providing a starting point without requiring you to create your own from scratch.