The Wire · Meta
Llama and Meta's open-weight models and AI research — the open end of the frontier, tracked for builders.
OpenAI cut GPT-5.6 Luna API prices by 80% to $0.20 per million input tokens and $1.20 per million output, while Terra fell 20% to $2 and $12. Fast mode gives Sol up to 2.5 times Standard speed at twice the price. OpenAI says Sol-assisted kernel work lowered serving cost by 20% and improved token-generation efficiency by more than 15%, while Luna delivers year-old frontier performance at roughly six cents per task-dollar and nearly nine times the speed. The edition connects cheaper models to Amazon's reported $1.8m, 860%-over-budget coding task, Gemini Robotics 2 whole-body control, Nscale's Anyscale acquisition and Okta's roughly $200m Permiso deal.
If you thought Windows Recall was creepy, buckle up. Microsoft is about to shoehorn even more AI into its desktop operating system, giving Copilot access to every file on your PC. But don't worry, Redmond insists it's totally secure and safe since the agents powering the functionality will run entirely on your PC — some of the time. Microsoft describes the new functionality, announced alongside a new Nvidia N1X powered Surface Laptop, as "hybrid intelligence." Built atop Microsoft's HydraFusion model router, Windows apps and Copilot features will soon be able to use local models when it makes sense and cloud-based models when it doesn't. For example, Windows might use a local model for simple or privacy-sensitive tasks, or decide complex or performance-sensitive requests need OpenAI's GPT-6 or Anthropic's Claude. Local agents will run in Microsoft's Execution Containers (MXCs), agent sandboxes with fine-grained permissions applied at runtime. According to Microsoft EVP of Windows and Devices Pavan Davuluri, MXCs mean Windows is now "the most secure platform for agents." (That claim will surely age well because models never escape their sandboxes and routers never misbehave.) To support this hybrid intelligence functionality, Microsoft is extending support for Windows ML, its inference framework built on the ONNX runtime, to Llama.cpp – one of the highest-performing inference engines for local AI inference. By adding support for Windows ML, Llama.cpp should be able to take advantage of a broader range of accelerators. Microsoft is rolling out its hybrid intelligence functionality to GitHub Copilot — its AI code assistant — next week and its Copilot app over the next few months. The new Copilot features look and sound suspiciously like OpenClaw in a business suit. Copilot's agentic functionality will not have access to your entire file system, opening the door to advanced search functionality and productivity tasks through integrations with Microsoft Office. In a demo
Read full story →Good Monday morning, I'm on my second coffee of the day already so please excuse any jitters that come through today's newsletter. We've got a big constitutionality ruling on Flock, a scoop about a major fire Meta needed to put out pre-Muse launch, and a lot more from trawling the World Wide Web. Let's get into it. THE BIG STORY Meta Rushed to Fix Muse ‘VM Escape' Vulnerability Immediately Before Launch In the immediate weeks before Muse’s launch, Meta engineers found several security vulnerabilities in the company’s viral AI agent product, at least one of which could have allowed malicious users to break outside of Muse’s intended environment and access Meta’s own sensitive databases and services, 404 Media has learned. The issues were so severe that they reached Mark Zuckerberg and staff worked overtime to fix them. These specific vulnerabilities were discovered before the launch of the product but required a multi-team “mad dash” to fix “a sudden spike in reported KVM escapes,” according to an internal post by Meta executives to its core infrastructure team seen by 404 Media. In order for Muse to work, a user gives the AI agent access to various important services and accounts that they own. On Meta’s end, each individual Muse instance runs on a kernel-based virtual machine, which connects to, but is supposed to be isolated from, Meta’s own critical infrastructure. A “KVM escape” is when, through a security vulnerability, a Muse instance is able to escape from that virtual machine and interact with the system that runs it, or with other users’ virtual machines. According to a Meta source, as well as internal security documentation and internal posts viewed by 404 Media, at least one of the vulnerabilities could have allowed an outside attacker — that is, a normal Muse user — to access data in sensitive internal Meta databases. At least one of the vulnerabilities was related to an exploit found in Linux kernel-based virtual machine code in July. 404 Media granted
Read full story →Elon Musk says SpaceXAI, the company formed by merging SpaceX and xAI, will be renamed SpaceXSI after President Donald Trump ordered US federal agencies to replace "Artificial Intelligence" with "Super Intelligence" in all comms. Trump's September 29 executive order directs federal agencies to switch from AI to SI in official communications and other non-statutory documents. Musk endorsed the change, posting "No more AI. SI. It's better," and confirmed plans to rename SpaceXAI: "Yes, we will make that change." He gave no date for the rebrand. At the time of writing, the spacexsi.com domain was listed for sale. SpaceXSI sounds somewhat like a performance car of the 1990s, or perhaps a family wagon with sporting aspirations. This writer remembers putting his driving license at risk in a Peugeot 106 XSi, occasionally finding that ambition exceeded ability. We asked OpenAI whether it plans to rebrand as OpenSI. The company has yet to respond. Trump's order does not require private companies to change their names. Tech leaders including Mark Zuckerberg already use "superintelligence" to describe their ambitions. Trump's order goes further by applying the label to technologies covered by the existing federal definition of AI. Musk has form when it comes to rebranding. Twitter became X in July 2023, nine months after he bought it. "SpaceX is a super intelligence company," Musk posted. Bad news for NASA, which probably thought it had something to do with rockets. ®
Read full story →In the immediate weeks before Muse’s launch, Meta engineers found several security vulnerabilities in the company’s viral AI agent product, at least one of which could have allowed malicious users to break outside of Muse’s intended environment and access Meta’s own sensitive databases and services, 404 Media has learned. The issues were so severe that they reached Mark Zuckerberg and staff worked overtime to fix them. These specific vulnerabilities were discovered before the launch of the product but required a multi-team “mad dash” to fix “a sudden spike in reported KVM escapes,” according to an internal post by Meta executives to its core infrastructure team seen by 404 Media. In order for Muse to work, a user gives the AI agent access to various important services and accounts that they own. On Meta’s end, each individual Muse instance runs on a kernel-based virtual machine , which connects to, but is supposed to be isolated from, Meta’s own critical infrastructure. A “KVM escape,” then, is when, through a security vulnerability, a Muse instance is able to escape from that virtual machine and interact with the system that runs it, or with other users’ virtual machines. Do you know anything else about Muse's security? I would love to hear from you. Using a non-work device, you can message me securely on Signal at jason.404. Otherwise, send me an email at [email protected]. According to a Meta source, as well as internal security documentation and internal posts viewed by 404 Media, at least one of the vulnerabilities could have allowed an outside attacker — that is, a normal Muse user — to access data in sensitive internal Meta databases. . At least one of the vulnerabilities was related to an exploit found in Linux kernel-based virtual machine code in July . 404 Media granted the Meta source anonymity to speak about sensitive security matters. Several of the vulnerabilities were in the underlying Linux virtualization software that Meta uses for Muse. The secur
This week, the White House got nearly every major tech CEO in one room — Zuckerberg, Bezos, Musk, and Anthropic’s Dario Amodei among them — to sign an AI safety pledge that President Donald Trump called “morally binding.” Trump also signed an executive order officially rebranding AI as “super intelligence,” and meanwhile, Meta and OpenAI are putting friendlier faces on their AI products, even as the biggest money […]
Prior to joining Airbnb as CTO in January, Ahmad Al-Dahle was head of generative AI at Meta and led the launch of its open source Llama models over 2023-2025. Now Al-Dahle is in charge of turning Airbnb into an “AI-native company”. What that means in practice is using AI internally to speed up product development, then using those same capabilities to transform the customer experience . We’re calling this an “inside-out AI” approach and in this article we’ll dig into the details. An example of inside-out AI: Airbnb used a custom internal tool called Everest to help accelerate the launch of a new external service. (Diagram by Latent Space) Al-Dahle spoke with Latent Space about Airbnb’s AI transformation. We began by asking him why he made the shift from building frontier models at Meta to deploying them at Airbnb . “Fundamentally, I like to chase where I believe the hard frontiers live ,” he replied, pun perhaps intended. “We kind of knew [at Meta] what the flywheel looks like, we understood how to improve the model capabilities, generation on generation.” Al-Dahle’s tweet in January on joining Airbnb from Meta The next challenge, Al-Dahle thinks, is deploying models at scale. At Airbnb, his goals are to “ push people to work differently , because these tools are transforming how you work,” and to “ take these systems and deploy them in production , in creative ways that actually add value to a core user experience.” Jumping to prototypes, code is the artifact From its launch in 2008, Airbnb has been considered a technology company operating in the hospitality industry. The company went public in 2020 and now has a market capitalization of approximately $93 billion (at time of writing). So it’s a big operation to try and make “AI-native.” Al-Dahle reeled off a few statistics to show that Airbnb is AI-pilled now: 60% of its code is now AI-authored , it has shipped nearly 80% more features and improvements year over year, and pull-request throughput for the average en
An independent briefing for builders: the whole field read continuously, every story scored for relevance, and the noise left off the page.
300+ curated sources. Every story scored 1–10 for builder relevance by Claude's frontier model. The filler never makes it to the page.
GPUs, datacenters, power deals, and inference economics: the infrastructure layer that decides what every builder pays. Our signature coverage.
Every story is sourced. Every score is computed. We show our work and link to originals.
Every briefing closes with The Call: one falsifiable claim with a date on it. When we're wrong, we say so in print. Opinions are cheap; ours get scored.