Remember when Meta launched Threads and it hit 100 million signups in under a week? The tech press declared Twitter dead by Thursday. Then the daily active user count fell off a cliff, and Meta spent the next year and a half quietly rebuilding the thing into something people actually opened on purpose. I reviewed Threads twice — once during the hype, once six months later — and they were basically two different apps with two different verdicts.
I bring that up because Meta’s Muse is now posting launch numbers that beat ChatGPT’s early mobile run, and the framing around it feels familiar.
What the numbers actually say
Here’s what’s been reported. Appfigures estimates Muse pulled 2.8 million global installs in its first 12 days, with 359,000 daily active users on iOS. Separately, Apptopia ran a matched U.S.-and-Canada iOS comparison and found Muse outpacing ChatGPT’s early mobile download pace. Other coverage has cited a 730K download figure, which likely reflects a narrower window or region — worth keeping in mind when you see the number quoted flatly.
So: strong initial adoption. That much is real, and the fact that two independent app-intelligence firms landed in the same direction makes it harder to dismiss as a single bad data pull.
Why the comparison is softer than it looks
As someone who spends most of his week installing AI tools and deleting most of them, I want to flag what this comparison is and isn’t measuring.
ChatGPT’s mobile app arrived months after the web product had already become a phenomenon. There was no Meta-scale distribution machine behind it, no install prompts inside apps that billions of people already have open. Muse launches into a completely different set of conditions. Comparing day-12 download curves across those two situations tells you something about distribution muscle. It tells you very little about whether the product is good.
The metric I care about isn’t installs. It’s what the DAU number looks like at day 90 relative to day 12. 359,000 daily actives on 2.8 million installs is roughly 13% — and early-window DAU-to-install ratios are notoriously flattering, because the people who installed yesterday are mechanically counted as active today. That ratio is the thing to watch, not the raw total.
The reviewer’s actual question
When I evaluate a tool for this site, download counts sit near the bottom of my list. What I’m asking is simpler:
- Does it do something I can’t already do in a tool I have open?
- Does it stay useful after the novelty wears off, or does it become another icon I swipe past?
- Is the distribution doing the work, or is the product doing the work?
That third one is where Muse’s numbers get genuinely interesting, and also where the verified data runs out. Nothing in the reported figures separates organic pull from Meta’s ability to put an app in front of an enormous installed base. Both produce the same install chart. They produce very different six-month charts.
What I’d tell you if you asked whether to install it
Try it, but hold your conclusions. That’s not fence-sitting, it’s the pattern I’ve watched repeat with every major AI app launch in the past two years. Early numbers reward reach. Staying power rewards usefulness. The two correlate less than you’d expect.
The honest read on Meta’s position is that this is a real and somewhat unusual win. Meta has spent a lot of money and public credibility on AI products that landed with a shrug. Beating ChatGPT’s early mobile pace, even on a favorable comparison, is a better headline than Meta has had in this category for a while. It gives them something to build on.
It also puts them in a spot where the next data release matters more than this one. If Muse’s daily actives hold or grow through the next quarter, the comparison to ChatGPT starts to mean something about product quality rather than reach. If the DAU line sags while installs keep climbing, that’s the Threads pattern again, and everybody involved already knows how that story reads.
Where this sits on my list
Muse goes in the “watching, not recommending” bucket for now. Not a knock — that’s where most tools sit for their first few months here. I don’t have enough usage data to tell you whether it earns a slot on your home screen, and neither does anyone else writing about it this week.
Check back when there’s a day-90 number. That’s the review worth reading.
🕒 Published: