A solo creator running five YouTube channels, a trading community, and a real estate business walks through how Meta's free AI agent Muse actually handles his calendar, email, and content pipeline.
A solo creator runs five businesses by splitting AI into two jobs: coding assistants that build, and a free daily-ops agent that handles calendar, email, and content, with every output needing his approval.
This video is a working log of how one creator already uses Meta's free AI agent, Muse, to run five YouTube channels, a trading community, and a real estate business without hiring more help. Muse delivers a daily brief with calendar and deadlines, drafts and triages email for approval, turns Threads, Instagram, and Facebook signals into content ideas, and even follows up unprompted the next morning on a conversation from the night before. Everything funnels through a GitHub repo where Claude and Codex also review drafts before a human signs off. The creator draws a clear line: pricing, deal terms, and strategy stay human decisions, and he's upfront that AI buys back hours, not guaranteed income.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →
Cold open: before showing what Muse can do, the host warns that it requires real care about what access you grant it.
He frames the scale of what he's running solo, then introduces Muse as the daily-operations half of a two-tool AI stack alongside Claude and Codex.
He clarifies Muse's free tier covers his day-to-day business use, and explains why he still relies on Claude and Codex for anything resembling coding or app-building.
Muse opens each day with a brief covering calendar and deadlines, then handles email spam filtering and drafts replies he only has to approve.
Muse watches signals across Meta's apps and surfaces content ideas that feed his YouTube pipeline.
He describes Muse picking a conversation back up unprompted the next morning with an additional idea, a behavior he says beats Claude and Codex on this specific point.
Muse auto-drafts daily posts for Threads, Instagram, and Facebook that he reviews and approves rather than writing from scratch.
He draws a hard line: pricing, deal terms, and strategy stay human decisions, and AI's real value is bought-back time, not guaranteed income.
A walkthrough of the Muse interface on phone and desktop, including naming the assistant (his is named Rain) and customizing its avatar.
The Feed surfaces interests, reminders, and upcoming calendar tags in a scrollable format he prefers to doomscrolling Instagram.
The Ideas tab surfaces draftable next steps across productivity, financial management, health and fitness, and relationships.
The Goals and Library sections hold saved outcomes and artifacts, but he's candid that Muse inferring things he never explicitly told it made him uncomfortable enough to avoid installing it on his desktop.
His real production pipeline: Muse sends content ideas through a shared GitHub repo where Claude and Codex both weigh in, and he gets back a finished thumbnail, script, and title bundle to approve or send back.
A multimodal test: he sends Muse a photo of shoes from social media and it returns the exact product match with price and retailer links.
Asked to help a family member job hunt, Muse says it can handle 90% of the prep, resumes, cover letters, forms, but can't submit applications that require a verified human identity.
He commits to keeping Muse alongside Claude and Codex going forward, framing it as reducing friction rather than replacing judgment.
A rundown of how deeply AI threads through his schedule, meals, training, content, meetings, freeing attention for the work that actually moves revenue.
He draws the final distinction: Muse isn't as strong as Claude or Codex at coding or building an app, but it's right there on daily-task intelligence and responses.
Closing argument: constantly fielding offers to hire virtual assistants, he argues Muse now covers enough of that role that VAs are becoming obsolete for him.
Splitting AI tools by job and keeping every output in a review-approve loop is what lets one person run five channels, a trading community, and a real estate business without burning out.
“Warning, downloading Metamuse or using Meta AI or any Meta product can be very invasive.”
“No matter what you see on the internet, AI will not make you rich, but it can buy back a lot of the hours.”
“This thing knows things that I didn't even tell it, and that kind of made me uncomfortable.”
“Removing the friction of a lot of things makes your life easier.”
“I think virtual assistants are a thing of the past.”
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
Before showing what Meta's new AI agent can do, the host opens with a warning: Muse is genuinely invasive, so know what access you're granting before you install it.
The creator splits his AI tools by function rather than making one model do everything: coding assistants handle building and shipping, a daily-ops agent handles the repetitive business grind.
The three things he actually delegates to Muse day to day, deliberately narrow rather than handing over everything.
00:0000:09
00:1600:23
00:27
00:34
00:41
00:4900:56
01:02
01:09
01:16
01:2201:29
01:36
01:42
01:49
01:55
02:02
02:0902:15
02:22
02:29
02:35
02:4202:48
02:55
03:0203:08
03:15
03:2203:28
03:35
03:41
03:48
03:5504:01
04:08
04:1504:21
04:28
04:34
04:4104:48
04:54
05:01
05:08
05:1405:21
05:27
05:34
05:41
05:47
05:54
06:01
06:0706:14
06:20
06:27
06:3406:40
06:4706:54
07:00
07:07
07:1307:20
07:27
07:3307:40
07:47
07:53
08:00
08:06
08:1308:20
08:2608:33
08:40
08:46Add Modern Creator as a preferred source and Google shows you more of our breakdowns in Search, Top Stories, and AI Overviews. It only changes what you see, and you can undo it in your Google settings anytime.
20:43A $40M founder demos the three AI agents his team actually built — content, chief of staff, and product manager — then hands you the one mental model that makes all three work.
June 8th
17:05How a free AI agent builds persistent memory about your business and runs parts of it for you.
June 11th
15:25Every night at 1am, 3am, 4am, and 5am, a set of Claude Code agents build something, grade themselves against a rubric, and write their own rules before the next run.
October 6th
31:35A quarter-billion-dollar webinar seller grades AI task by task and finds one job it actually wins.
October 4th
26:19A creator reacts clip by clip to Meta's free AI agent, grading the calls, purchases, and file reads it actually completed on camera.
September 29th
40:24Two hosts react to eleven creators' real GrokBot setups, timing which automations are worth the tokens and which ones need a webhook instead of a timer.
October 3rd