Stratechery on Meta enterprise: separate contexts or one work-and-life agent?
Stratechery on Meta enterprise: separate contexts or one work-and-life agent?
Source. Ben Thompson, “One More Note on Agents, Meta Connect, Meta Enterprise Platform,” September 29, 2026, subscriber text delivered by Stratechery subscription. Meta’s September 28 announcement confirms the business initiative and named Muse/Muse Code products; it does not prove enterprise customer adoption.
Argument in order. Thompson first argues an agent may reduce the difficulty of smartphone/web interactions for nontechnical users, then treats Muse as the new organizing principle behind Meta’s device lineup. He reads Meta’s enterprise platform announcement as a strategic error: consumer distribution and entertaining, intimate use are different from enterprise selling, tenant controls and work identity. He invokes his earlier disagreement with Zuckerberg’s one-identity view and argues a separate work and private context is desirable. He considers the alternatives—that work is part of life and that enterprise revenue helps underwrite costly compute—but still favors a small-business entry point rather than a full enterprise pivot.
Area consequence and challenge. If design, support, sales and engineering roles share a ‘personal’ agent, what data may follow it across customer/team/private boundaries, and who can approve action? A single conversational surface is not a single permission domain. Connect to roles, coordination and interface question. Thompson’s clean consumer/enterprise incompatibility claim is his strategic judgment, not tested against an actual cross-context deployment. Meta explicitly says enterprise security/privacy will be built in; neither that assertion nor Thompson’s skepticism supplies error rates, retention, revenue or integration workload.