At a glance
WHAT IT’S REALLY ABOUT
Claude Duet impresses in polish, but falls short on speed
- Claire reviews Anthropic’s new mobile personal agent, Claude Duet, and tests whether it’s compelling enough to replace her current go-to agents like Grok Bot and Muse.
- Duet’s strongest differentiator is its exceptionally polished, trustworthy design, including clear onboarding, task lists, artifacts, and Slack-like channels for organizing workstreams.
- In practice, Duet feels insufficiently proactive, failing to re-engage her with meaningful prompts beyond standard briefings, which makes it easy to forget amid “personal agent fatigue.”
- Claire finds performance/latency to be a major drawback—slow task completion and weak loop-closing degrade the experience even when the underlying model quality is strong.
- Duet’s lack of a desktop app and limited multiplayer/team sharing make it less competitive for real work, despite being capable for basic SMB tasks like pipeline review and quick deck generation.
IDEAS WORTH REMEMBERING
5 ideasDuet’s standout strength is product polish and clarity of experience.
Claire sees Duet as a strong mobile companion for Claude with thoughtful onboarding and clear permissioning, aimed at SMB owners and busy professionals rather than power-user “agent swarms.” The overall UX feels calm, trusted, and cohesive from the first run.
Task lists and artifacts make agent work more visible and reusable.
Task lists help users track what they asked for versus what the agent is doing, while artifacts are promoted to first-class outputs that are easy to view and reuse (including previously created Claude artifacts). This makes Duet feel more like a lightweight work hub than a chat-only agent.
Channels are a compelling organizing primitive—but mainly for solo use today.
The “channels” concept (Slack-inspired subspaces) helps separate workstreams (e.g., a sales channel) and reduces the messiness of a single endless thread. It’s a promising primitive for organizing agent collaboration, even if the product isn’t yet truly collaborative across people.
Duet underperforms on proactive re-engagement, which hurts retention.
Despite being positioned as proactive, Duet didn’t re-engage her after initial setup and didn’t deliver enough high-value nudges beyond the familiar “morning briefing” pattern. In a crowded agent ecosystem, being calm and subtle can translate to being forgotten.
Latency is the biggest practical deal-breaker for synchronous personal-agent use.
Claire reports long waits (often 10+ minutes) for tasks/threads to complete or request permissions, which breaks the “instant colleague” expectation of personal agents. She’d often forget tasks were running, and the agent failed to close loops fast enough to keep her engaged.
WORDS WORTH SAVING
5 quotesThe bad news is I have personal agent fatigue. I am so tired of new personal agents.
— Claire Vo
To compete for my attention, to compete for being my number one personal agent, you have to be not good, not great. You have to be exceptional.
— Claire Vo
For a proactive agent, it just wasn't that proactive.
— Claire Vo
There's gotta be something better than the morning briefing to engage agent users on why they would come back to the agent over time.
— Claire Vo
The second thing that I don't love about it, and this has been my experience with Claude generally, is it's slow, man, and I get so disengaged waiting for the models to finish.
— Claire Vo
High quality AI-generated summary created from speaker-labeled transcript.
