AI Personal Assistants: Muse, Grok Bot, Cue Usage & Comparisons

A blogger visited Meta's Brooklyn event to try Muse, the company's new AI agent product, and shared a selfie from the venue.

After testing Grok Bots for a week, the author says marketing bots return to tasks on their own with context intact, and teases a livestream on handoff-ready marketing workflows.

An attendee shared a photo from Grok Bot Galaxy and said the event had incredible energy.

The author shares how his six agents divide work, with Grok Bot running Grok CLI to search X trends and analyze account weaknesses.

The author says dropping single local sessions for Cursor cloud agents gave 2-3x more output, and lists Grok Bot among the cloud agent camp.

A recap of Day 2 covering bot teams for sales engineering, sales, SDR and support roles, plus a Cupcake game prototype demo.

The organizer reports 130 sign-ups and 70 attendees at the first Grok Bot meetup in Salta, and demos using it to generate assets, update metrics and move YouTube episodes to Spotify.

A user compared two weeks of use: last week the weekly token limit ran out by day four, while this week, with double the usage and several extra bots, only 57% was used by the weekend.

The user says Grok, combined with request routing through their own device, is the strongest agent for heavy tasks, and reflects on having previously let weaker agents do deep research while they did repetitive work.

A user spent a full day using Meta's Muse as an executive assistant, with it pulling an itemized receipt from JetBlue into Brex and connecting to Granola, Gmail, Google Calendar and Notion to build a personal CRM.

Alexandr Wang says he only realized after seeing a post that Muse Spark 1.3 ranks second on the agent benchmark Agents Last Exam (ALE).

A user showed off their Poshy Grok Bot during a live session and said it sold a second pair of shoes on Poshmark the night before.