The 10 use cases
1 · Negotiating bills — 00:44
Muse logged into AT&T for him (credentials from a secure store), found that paying from a bank account instead of a credit card saves ~$10/line/month, and walked him through enrolment — Peter typed the bank details himself for privacy. **Bill: $200 → $160/month, ~$480/year**, from a single 5–10 minute conversation. He’s now using a “finances” thread to work the same trick on Tesla insurance, comparing Geico / Progressive / State Farm.
2 · Custom animated avatar — 02:48
Replaced the default avatar with a red panda entirely by chat: four generated images → “make it cuter” → four more → picked one, and Muse added animations by default (looking around, typing), then generated more clips (waving, eating bamboo). Dead end the video notes: he couldn’t push extra animations onto his profile.
3 · App connections + message triage — 04:11
Muse connects to a long list of everyday apps, and is distinctive for connecting natively to Meta’s own (Instagram, Threads, Facebook, Messenger). He set up a messages thread to monitor IMs and flag anything needing a follow-up; a scheduled task can check unread messages daily. His finding: Instagram/Threads DMs were 100% spam — but that’s exactly the value. Limit: extending beyond the list fails (Linear wouldn’t connect), and he likes that Muse stays personal-use focused instead of demanding API keys.
4 · Facebook Marketplace — 05:49
Looks up items to buy (“single golf clubs near me” → three options), lists his own (iPhone 13 Pro — photos, price, the rest), and — because Muse also has Messenger — messages sellers and negotiates price on his behalf.
5 · Personalized feed — 06:49
Asked for important AI news on topics he cares about; now skims it every morning and can click through to the browser. He can blend in personal material (his newsletter). His line: Facebook’s own feed is AI slop and ads; Muse’s is more valuable. Caveat he predicts: Meta will add ads once it’s popular — it’s ad-free today.
6 · Goals + habit tracking — 08:01
Muse builds a to-do list from past conversations, plus a goals system. Demo: breaking his habit of bringing his phone to the bedroom — it interviewed him first (what pulls the phone in? what would you rather do?), made a plan, and built an artifact that reminds him at bedtime and checks in each morning, charting streaks. Honest reporting: he’d failed two nights running.
7 · Ideas tab / proactive suggestions — 10:15
Because most people don’t know what to type into a chat box, Muse proactively suggests things it can do based on what it knows (stage his video shoot packet, lock in a lower bill, triage his inbox, snap a meal and track calories). Demoed: a side-by-side electric razor comparison — it browsed Best Buy and recommended Panasonic. Also includes offbeat ideas like messaging the government division to fix a pothole or busted street light.
8 · Local events, shows and dinner — 12:14
Looked up reviewed ticketing in the Bay Area → the Magic Theater for Sept 30 (his anniversary), found a nearby dinner spot, and can book (“two general admission tickets”) via its browser.
<br />
9 · Weekend scout — 12:47 (1/2)
A standing job to surface interesting things to do nearby — it found him a comedy day in Golden Gate Park.
<br />
10 · Phone calls — the killer feature — 13:01
He asked Muse to negotiate his Xf (2/2)inity/Comcast bill. It asked for a voice (male/female), his full name and number, then called Xfinity while he ate breakfast, found a $60/month promotional rate (+$10 off for autopay), and returned a transcript of the call — “Hi, this is Haley calling on Peter’s behalf… I’m calling to see if Peter can keep the same internet plan but lower the $84 monthly bill.”