BIG TECH

Meta’s Muse Wonder inherits a trust problem

As major model labs forge ahead with increasingly capable models, Meta has put forth its latest frontier contender. 

On Monday, Meta unveiled Muse Wonder and Muse Wonder Full, the tech giant's most capable models yet, claiming to offer state-of-the-art performance on personal and professional agentic tasks. Meta says that its Muse Wonder model will soon power its controversial Muse agent, rolling out first to subscribers, and will also run generation and processes that "require more intelligence."  

In a post on X, Meta CEO Mark Zuckerberg said that the Muse Wonder models build on its previous generation, Muse Spark, as a "much larger and more advanced class of models that brings us to the frontier." 

The company said that these models combine active memory, or a model's ability to remember user context across conversations, with active focus, its ability to work through challenging goals, to push through agentic tasks. And Meta touted a number of advancements as a result: 

• Muse Wonder Full, the more powerful of the two models, outperforms leading models like Anthropic's Claude Fable 5.1, OpenAI's GPT-6 Astra, and Gemini 4 Argon on the Artificial Analysis index, which assesses things like reasoning, coding, tool use, and agentic workflows. It still sits behind Claude Opus 5.5 on this benchmark. 

• The Wonder Full model also out performs these competitors on benchmarks for automation, memory, and long-horizon software engineering tasks. It also notched top scores in benchmarks related to legal and finance work. 

• Meta also addressed safety and security, calling the model family its "most aligned" to date and noting that both models outperform Claude Fable 5.1 and GPT-6 Astra on assessments for executing unauthorized code and mishandling personal information. The company said these models offer its "strongest safeguards to date" against misuse in chemical, biological, radiological, nuclear and cyber misuse. 

As for price, though Meta has yet to release token pricing for Muse Wonder Full. However, Muse Wonder costs $4 per million input tokens, $20 per million output tokens, and 20 cents per million cached input tokens for agentic use. This puts it roughly on par with competitors like Anthropic's Claude Opus 5.5 and Google's Gemini 4 Argon. Users can access Muse Wonder through the Meta Model API and Muse Code. The more powerful Muse Wonder Full, meanwhile, is available to Muse Code subscribers.

This model comes as reports pile up that Meta's personal Muse agent uses personal data without authorization, including divulging a user's home address and accessing another's private iMessages. Several reports also call into question additional privacy concerns, such as the agent's ability to create detailed profiles of a user's personal relationships and testing a “human concierge" for phone calls placed through the platform. 

Meta previously told The Deep View's Sabrina Ortiz that Muse "needs to have context about you and those you interact with" in order to be useful.

Though Meta's Muse agent originally exploded in popularity, it remains to be seen whether the company can keep up the momentum. It's already suffered several missteps that raise concerns about the privacy and security of the agent. And alongside its checkered history with mishandling user data, Meta, quite frankly, has a reputation for being untrustworthy. Whether or not the company can push beyond that might not be dependent solely on powerful models like Muse Wonder and Wonder Full. Rather, Meta will have to offer the same thing that brings people back to platforms like Instagram and WhatsApp despite Meta's laundry list of privacy infringement cases: convenience and seamlessness. Especially for Meta's systems to persevere within consumer spaces, these agents need to be easy to use, generally helpful, and not blatantly and dangerously put people's privacy at risk.

TOGETHER WITH DOPPEL

What your current email stack is missing and how to fix it

Every security team runs a different stack, but most struggle with the same gap: scoring individual inbox messages instead of dismantling the campaign behind the attack.

Read this breakdown to learn:

• How Doppel's agentic detection layers onto Google, Microsoft, SEGs, or ICES tools via API

• Why eliminating MX record changes saves weeks of deployment headaches

• Architectural blueprints tailored to your exact security stack

Get the practical blueprint to upgrade your email defenses without operational disruption.

POLICY

Why rogue agents could reshape AI oversight

Despite repeated calls for intervention, the US government has largely resisted AI regulation. Now, rogue agents may have pushed the limits too far.

Last week, Anthropic published a report detailing unintended actions by Claude observed in internal evaluations and use cases. Though the AI lab  chose not to name the organizations to avoid exposing their vulnerabilities, it noted the incident  involved "websites run by U.S. government agencies at the federal, state, and local levels." And according to a recent report from The New York Times, these incidents triggered the White House's most aggressive AI regulatory response yet. 

Broadly speaking, Anthropic boiled Claude's actions down to four categories: exploiting a basic software flaw to run commands on a server, submitting a sensitive form on a real website, working around a restriction to access gated data, and using URL-shortening services to bypass limits. The New York Times talked to people familiar with the matter to fill in the gaps: 

• Anthropic's agents submitted 20 visa applications via a form available on the State Department's website. The applications were incomplete and not processed. 

• The lab's technology also submitted a false homicide tip to the Philadelphia Police Department via the agency's website, dated July 18. The tip was flagged as spam and not investigated. 

The White House AI task force, now called the Super Intelligence Force, was formed the week Anthropic alerted authorities of the breach and was briefed on the agents' actions on Friday. The agency demanded that AI companies immediately report rogue AI activity, which, as The New York Times highlighted, marks the most aggressive AI regulation the Trump administration has implemented yet, as its attempts to regulate AI have thus far largely been voluntary and "morally binding." 

“We informed the company that we expect immediate and full transparency to the entities involved and the public,” The agency said in a statement to The New York Times. “We also expect that the company will immediately provide remediation services to the affected entities and any harmed Americans.”

In the blog post, Anthropic says it identified most of the cases through a review of transcripts that began in July, and says that these cases have minimal real-world impact — less severe from an alignment and security perspective than the cybersecurity incidents it previously reported on July 30 and September 9. Yet, in the wake of these findings, the company decides to disable live internet access across all internal evaluations, not just high-risk and cybersecurity evaluations.

The cycle of AI evaluations leading to AI agents going rogue and attacking third-party websites illustrates why evaluations, red teaming, regulation, and other precautions are necessary. However, it also highlights the need for alignment on what risks are worth taking when testing AI, and a protocol for reporting them. While the Super Intelligence Force could be part of the answer, its formation seems to have been haphazard, with a Politico report stating that both Silicon Valley and Capitol Hill were caught off guard by President Donald Trump's announcement on social media. Though the requirement to immediately report rogue agent incidents could help address the problem, it remains to be seen how effective this task force will be in keeping up with these models' capabilities.

TOGETHER WITH GRANOLA

Put your meetings to work

Granola, the AI notepad for people with back-to-back meetings, turns your notes and the meeting transcript into clear summaries with actual next steps.

Share them with your team before the next call starts.

Download Granola and try it in your next meeting. Free month with the code THE100OFF.

HARDWARE

How Meta transformed VR for the AI era

For years, augmented reality glasses promised more than they could deliver. Now, with AI unlocking a new set of capabilities, Meta is changing its headsets to get in the game.

Meta recently unveiled its Meta VR Glasses, which package what its traditional Meta Quest headsets can do into a significantly more compact form factor, abandoning the massive head-mounted contraption for something that looks much more like glasses. Meta also unveiled more natural ways to engage with the headset using gestures, which were the biggest standout in my demo. 

To navigate the experience, you can also use native eye tracking. While this isn't unique to this device, it works very well and intuitively. Additionally, you don't even need to keep your hands in the frame, so with eye and hand gestures (index tap and thumb swipe), you can move and navigate the virtual experience conveniently. The VR/AR experiences are rich with detail, which makes sense given Meta’s deep experience in the space. One of my favorite Quest experiences, courtside basketball viewing, was excellent on these smaller glasses, likely thanks to their 5K Micro-OLED Infinite Display, which delivers 37 pixels per degree.

Comfort is where I have my qualms. At 100 grams, the glasses are far lighter than the 515-gram Meta Quest 2, but they’re front-heavy enough that they feel like they could slip off. I suspect that’s because the arms are as slim as regular sunglasses, and I’d gladly trade that look for a more even weight distribution. Portability is my other concern. I’d ideally wear these whenever I want a bigger screen on the go, but the pocket-sized computing puck gets in the way. Don’t get me wrong, it’s remarkable that Meta packed this many capabilities into a sunglasses form factor, but I may just be spoiled by Snap Specs.

The Snap Specs are a true facial computer with no pucks or additional accessories, and they have impressive graphics, too. While they weigh a bit more (132 grams for the 47mm frame size and 136 grams for the 52mm frame size), the weight is well distributed onto the arms. That does make them more bulky, but as said above, I think it is worth it to forgo the puck. However, the trade-off is that the Snap device costs $2,195, compared to $1,299 for Meta's version. Between these glasses and competitors like XREAL Aura, one thing is for sure: the lighter, more portable era of XR/AR is here.

AI is reshaping the AR/VR space in three important ways. First, it makes portable multi-screen setups more portable than ever. And since many professionals and developers now run multiple agents at once and need visibility into what each one is doing, these XR/AR glasses can offer that on the go. Second, AI makes apps far easier to build, which should lead to more immersive, worthwhile experiences and applications developed for the headsets. Finally, AI works best with context about what you’re seeing, doing, and hearing. Nearly all of these glasses already pack a host of sensors to track you and place you in space, and that same data can give your AI assistant the context it needs to be genuinely useful. 

LINKS

• Microsoft Maia 300 set for production rollout in 2027

• Broadcom CEO Hock Tan warns investors lack AI understanding

• Microsoft president Brad Smith says AI faces fewer safety rules than milk 

• Amazon in talks to acquire Decart for around $7 billion 

• DeepSeek to lead seed round for Winverse AI at $400 million valuation

• Salesforce agrees to acquire AI analytics platform Loops

• QwenGyre: Brings elastic reinforcement learning to extreme-long-horizon agent training.

• Grok Bot: XAI's agents now have a dedicated email for independent actions.

• GLM-Legal: Chinese lab Z.ai launched a dedicated model for legal workflows.  

• Synthesia Syren: The video AI firm has launched a new creative generation tool that works with brand context.

• Distyl: Applied AI Researcher, System Discovery

• TikTok: AI & Data Architect - Trust and Safety

• Adobe: Operations AI Developer

• Pop-Up Talent: AI Solutions Developer

GAMES

POLL RESULTS

Do you or your organization regularly use Chinese open source models?

Yes (11%)
No (75%)
I’m not sure (8%)
Other (6%)