Muse Six Months Out: The Questions Meta Hasn't Answered

Muse launched on September 8, 2026. By the time this series closes, it's a little over three weeks old, and in that short window it reportedly topped app charts in the US and Canada, expanded to small businesses with more than a dozen named connectors, and announced a companion "Meta Enterprise Platform" led by newly hired executive Chirantan Desai. That's a genuinely fast launch trajectory. It's also a trajectory built almost entirely on announcements, not on six months of usage data — and the honest version of a "what's next" article for Muse is a list of what Meta hasn't yet said, not a prediction dressed up as analysis.

The Confidential VM deadline is the biggest one

Meta has committed to shipping Muse Confidential VM — full end-to-end encryption of a user's VM, data, and conversations with a key only the user holds — by year-end 2026, a claim Meta frames as meaning "not even Meta can access it." That's roughly three months out from today. It's the single most consequential unshipped commitment in the entire product, because it's the feature that would move Meta's privacy promise from "architecturally separated, per Meta's own description" (the current Sentinel-gated Secure VM, covered in Ad Data and Agents: What Muse Promises to Keep Separate) to "cryptographically enforced, independent of Meta's own good faith." Whether it ships on time, and whether outside researchers get to verify what it actually encrypts, is the single most watchable fact in this entire series over the next quarter.

Pricing beyond "free for most features" is still undefined

Zuckerberg has said Meta expects to profit "by taking a small fee from transactions" rather than through a paywall, and Muse is described as "free for a huge number of tokens" — but no source reviewed for this series specifies what happens at the edges: what a paid tier actually includes, what usage threshold triggers it, or what percentage fee applies to a Link-processed transaction. For a product actively expanding into small business — where transaction fees on Shopify and Stripe-routed purchases could add up meaningfully for a business with real margins to protect — this is a live commercial question with no public answer yet, not just a consumer curiosity.

The small-business/ad-account boundary needs a real answer

We flagged this directly in Small Business Data: What Plugging In Shopify Actually Exposes: Meta's ad-data-separation promise is written for personal conversational data, and it's unclear whether it extends with the same force to a small business's Muse-connected sales data and ad account access. As more businesses connect, this stops being a hypothetical and becomes a live question a regulator or journalist can actually test by comparing a business's Muse behavior against its subsequent ad targeting. That's the kind of test that six months of real usage would make possible in a way launch-week announcements can't.

1Password, Shop Pay, and the credential story are still "coming"

Both were listed as forthcoming at launch, and neither had a public ship date as of this writing. Until 1Password support actually lands, Muse's "no visibility into passwords" architecture, discussed in One Password and Muse: Credential Management Without the Copy-Paste, only covers the connectors Meta has already built — meaning the long tail of tools without a native integration remains a gap the product hasn't closed yet.

Glasses integration is a "coming months" promise with real stakes

Muse's arrival on Meta's AI glasses line — hands-free activation by name, and the ability to act on whatever the user is looking at — was announced at Connect 2026 for "coming months," with no firm date. This is worth watching not just as a feature but as a trust escalation: an agent that can see what you're looking at through a wearable camera is a materially different privacy proposition than one reading your connected apps, and none of the safety framework described for Muse's software side (Sentinel, the audit trail, per-connector permissions) has been explained in the context of a device that's always pointed at the world around you.

The audit trail's actual interface hasn't been shown

As covered in Audit Trails: How Much Visibility Does Muse Actually Give You?, Meta has described the existence of a complete audit trail without showing what it looks like, how far forward "plans to do" actually projects, or how long records are retained. That's a reasonable thing to withhold at launch. It's not a reasonable thing to still be withholding in six months, if Muse is going to keep expanding into small-business and eventually enterprise use cases where a real compliance-grade audit log matters.

The honest framing for where this goes

None of the above is a prediction that Muse will fail to deliver on any of these fronts — Meta has, so far, shipped roughly on the cadence it announced, and the underlying architecture described (Secure VM, Sentinel, scoped connectors) is more substantive than a typical launch-week feature list. But a fair account of "Muse six months out" has to be built from what's genuinely unresolved rather than confident extrapolation from a three-week-old app-chart placement. The three products that launched together this September — Muse, OpenAI's Dots, and xAI's Grok Bot — are all making the same underlying bet, that people will trust an agent with a persistent computer and standing access to their digital life. The next two quarters, not the launch week, are where each company's specific version of that bet either holds up under real use or doesn't. This series has tried to lay out exactly where to look when that verdict starts to arrive.

Part 30 of 30 in the Meta Muse series. Previously: Will Muse Need Claude Code Skills, or Make Its Own Ecosystem?. Browse more builder insights or explore AI skills at aiskill.market.