Key Takeaways
- Anthropic built Claude Duet with clean UI polish and channel structures, but its mobile-only design locks it out of normal desktop work hours.
- Desktop competitors like Grok Bot and Muse win daily attention simply by staying visible on computer screens while users sit in meetings.
- High model latency and absent proactive notifications stop Claude Duet from functioning as an active workplace teammate.
- Single-player AI silos fail inside companies because knowledge work depends on shared context rather than private chats.
- AI tools built around channels mimic Slack's interface, but they lose their value when colleagues cannot collaborate inside the same workspace.
The Friction of the Second Screen
During a standard workday, knowledge workers sit at laptops and monitors. Expecting them to pick up a phone to assign work to an agent creates immediate friction. Claire Vo tested Anthropic's Claude Duet and hit that exact wall.
When an agent lives exclusively on a mobile device, it demands context switching. A founder tracking engineering sprints or reviewing copy already has eight browser tabs open. Reaching into a pocket to chat with an AI breaks focus. Competitors with dedicated desktop clients bypass this barrier entirely.
Vo noted that visibility drives adoption: “I think that's why like Grockbot does so well is there's a desktop agent. It's cute little icon. It's sitting there staring me in the face.”
A desktop client supports a natural split-screen routine. Users kick off a quick prompt on their phone while away from their desk, then return to their workstation to complete the project. Without a desktop client, Claude Duet leaves users stranded on a small screen with slow model latency and no native push notifications to alert them when complex tasks finish.
Single-Player Interfaces Limit Team Output
Work inside any company with more than two people is collaborative. Anthropic gave Claude Duet channel primitives that resemble modern chat tools, but restricted the experience to one user talking to one bot. That single-player architecture caps its utility.
When tools copy the visual structure of Slack without the multiplayer infrastructure, they miss why Slack worked in the first place. Slack succeeds because teams share context openly across rooms. Colleagues see queries, build on top of answers, and correct bad outputs in public threads.
“I'd say the final thing that's really missing is look, we don't work alone and Duet right now is clearly a agent for single player use,” Vo observed. “So while Claude Duet with these channels might feel like Slack, it doesn't have that multiplayer sharing aspect of Slack that has made me really effective even as a small business owner.”
If an agent cannot join a shared room, answer questions in front of three teammates, or allow someone else to fork an artifact, it remains a personal assistant. Personal assistants help individuals write drafts. Multiplayer agents change how entire teams execute.
What to Do With This
Audit your internal agent deployments this Friday. If your team relies on personal chat windows where prompts and outputs remain private to one employee, kill those seats and migrate the workflows into shared Slack channels or a desktop multiplayer tool. Test whether an agent allows a second team member to view, edit, and fork the generated artifacts without re-running the prompt from scratch.