Following leaked revelations at the end of March that Anthropic had developed a powerful new Claude model, the company formally announced Mythos Preview on Tuesday along with news of an industry consortium it has convened, known as Project Glasswing, to grapple with the cybersecurity implications of the new model and advancing capabilities more generally across the AI field.

The group includes Microsoft, Apple, and Google as well as Amazon Web Services, the Linux Foundation, Cisco, Nvidia, Broadcom, and more than 40 other tech, cybersecurity, critical infrastructure, and financial organizations that will have private access to the model, which is not yet being generally released. The idea, in part, is simply to give the developers of the world’s foundational tech platforms time to turn Mythos Preview on their own systems so they can mitigate vulnerabilities and exploit chains that the model develops in simulated attacks. More broadly, Anthropic emphasizes that the purpose of convening the effort is to kickstart urgent exploration of how AI capabilities across the industry are on the precipice, the company says, of upending current software security and digital defense practices around the world.

“The real message is that this is not about the model or Anthropic,” Logan Graham, the company’s frontier red team lead, tells WIRED. “We need to prepare now for a world where these capabilities are broadly available in 6, 12, 24 months. Many things would be different about security. Many of the assumptions that we’ve built the modern security paradigms on might break.”

Models developed and trained by multiple companies have increasingly been able to find vulnerabilities in code and propose mitigations—or strategies for exploitation. This creates a next generation of security’s classic cat-and-mouse game in which a tool can aid defenders but can also fuel bad actors and make it easier to carry out attacks that were once too expensive or complex to be practical.

“Claude Mythos preview is a particularly big jump,” Anthropic CEO Dario Amodei said on Tuesday in a Project Glasswing launch video. “We haven’t trained it specifically to be good at cyber. We trained it to be good at code, but as a side effect of being good at code, it’s also good at cyber.” He adds in the video that “more powerful models are going to come from us and from others. And so we do need a plan to respond to this.”

Anthropic’s Graham notes that in addition to vulnerability discovery—including producing potential attack chains and proofs of concept—Mythos Preview is capable of more advanced exploit development, penetration testing, endpoint security assessment, hunting for system misconfigurations, and evaluating software binaries without access to its source code.

In carrying out a staggered release of Mythos Preview, beginning with an industry collaboration phase, Graham says that Anthropic sought to draw on tenets of coordinated vulnerability disclosure, the process of giving developers time to patch a bug before it is publicly discussed.

“We’ve seen Mythos Preview accomplish things that a senior security researcher would be able to accomplish,” Graham says. “This has very big implications then for how capabilities like this should be released. Done not carefully, this could be a meaningfully accelerant for attackers.”

Project Glasswing partners, including some of Anthropic’s competitors, struck a collaborative tone in statements as part of the launch.

“Google is pleased to see this cross-industry cybersecurity initiative coming together,” Heather Adkins, Google’s vice president of security engineering, says in a statement. “We have long believed that AI poses new challenges and opens new opportunities in cyber defense.”

The Dome is big. It’s not portable, practical, or inexpensive. It accepts the romance of wood, or the brute power of propane or natural gas. Its height makes it versatile enough for steaks, fish, or other skillet meals. This pizza oven is designed to be a fixture in your life and backyard, bolstered by an ever-expanding accessory set. And it also more than earns its place there, once you buy a snap-on Neapolitan arch accessory ($60) to bolster its insulation.

The Gozney makes truly excellent high-temperature pizza. Most backyard ovens, even our other favorites on this list, tend to struggle to reach and maintain the 900-degree temps needed for proper Neapolitan crust. The Dome Gen 2 gets there in 20 minutes, it heats admirably evenly, and it’s responsible for the best pizzas that my colleague Kat Merck says she’s made in her entire life. This is worth noting, given that she was editor and recipe tester for pizzaiolo Ken Forkish’s iconic pizza book The Elements of Pizza. (For what it’s worth, Forkish also uses a Dome Gen 2 at home, while enjoying his retirement. He likes using dough at 67 percent hydration, while cooking at 900 degrees in the Dome.)

A couple caveats, however: Gozney often markets the Dome as being able to cook two pizzas at the same time. This is a silly thing to do at the temperatures you’re cooking at. Cook one pizza. If you use the Neapolitan Arch, it’ll make the oven’s aperture narrow enough that you’ll need to limit yourself to a 12-inch peel anyway. The price of a Gozney Dome also rises considerably once you start delving into the accessories. With the stand, cover, Neapolitan arch, wood fire control kit, turning peel, and 15 pounds of Gozney-brand kiln-dried hardwood, the final price for the Dome Gen 2 can rack up as high as $3,270.

Best Big Pizza Oven for Families: Ooni Koda Max

Ooni’s large oven is for everyone who is sick of feeding their families with multiple teeny-tiny 12-inch pies and just wants to make a massive 20-inch cheese pizza for all the kids at once. You can either attach a propane tank or hook it to your natural gas line. If this is a possibility for you, then I recommend the latter. Ooni has a new gas management technology that keeps the temperature consistent across the huge surface. But big, powerful ovens use a lot of fuel: Its 35,000 BTUs put this Koda Max nearly on par with a 3-burner Traeger griddle. That heat will also come pouring out the open front of the oven, which means the Max is not ideal for small patios.

If you haven’t heard of UpScrolled before, a brief primer: It’s a social media platform not too different from, say, Instagram or TikTok. You can share photos or short videos, follow accounts, comment on posts, and amass a following of your own. Nothing too earth-shattering, right?

UpScrolled founder Issam Hijazi would beg to differ. Indeed, his nascent company diverges from most Big Tech platforms in a few notable ways: UpScrolled offers an old-fashioned chronological feed, rather than one dictated by an algorithm ostensibly serving up content you’ll latch onto; the platform also promises not to share user data with marketing firms or other commercial enterprises. And Hijazi, who is of Palestinian descent, founded UpScrolled in response to widespread user allegations that some social media companies were censoring or shadow-banning their posts—particularly pro-Palestinian content. The platform explicitly vows “never” to covertly suppress content, provided it doesn’t violate UpScrolled’s community guidelines.

Aside from breaking with plenty of Big Tech norms, Hijazi’s stance is rare among Silicon Valley types for being uniquely, overtly ideological. (In our conversation, Hijazi told me that he “personally” ensured UpScrolled users couldn’t select Israel as a location when using the platform.) But the approach has resonated: When we first met in February, a mere eight months after Hijazi launched UpScrolled, the platform had rapidly amassed 2.5 million users following freakouts over TikTok’s deal with President Trump to form a US-based version of the company controlled by American investors. Hijazi was, at that time, UpScrolled’s only employee.

Today, as UpScrolled counts more than 5 million users, Hijazi has rushed to scale his team to meet the platform’s growing needs—particularly around content moderation. Recently, his company has found itself in the crosshairs of organizations like the Anti-Defamation League, which alleges it doesn’t do nearly enough to stomp out antisemitic and extremist content. During a wide-ranging conversation last week I asked Hijazi about those claims, and how UpScrolled is catching up with its own rapid growth.