Why Meta's Muse AI agent is soaring in popularity

Vídeo em Inglês, publicado pelo canal Yahoo Finance.

Video transcript

Transcript from available captions; it may contain errors.

Let's take the biggest catalyst I think for all of this which is uh the really good news for Meta around its Muse product which is a new product launched uh I think two weeks agoish >> uh a new kind of agentic service that you can offer. I downloaded it this morning. >> Oh, you did? >> I want to just read out. >> I did not. But it gives you examples of what it can do. And I do want to just read some of these off because it even being kind of steeped in this stuff, it's still startling for me to see. Um, I can score you tickets to events this weekend. Tell me what you're buying. I'll give you a bunch of side by sides. Snap a pick of that old bill. I'll pay it, file it, and then follow up with you once you've gotten confirmation. Let me triage your inbox. Let me get back money that's yours. I'll check every state database to find old uh, you know, owenses for you. I can get you back money when the price drops on an item. All of this is so tied. >> Do we trust it to do all this stuff? So, this was my thought right away. And there was a tweet that went viral from um another one of our hosts on the Yahoo Finance Network um A Jazz last night about he got his flight got delayed and he put up a tweet saying that he filed it with Muse and within two minutes he had the refund from Delta and a new flight booked. >> That's >> that's one example. That's one example. That's one example. >> Uh how many downloads does this thing have within 5 days of its launch? 750,000 >> in the first 13 days. 2 and a half million. >> One example. Okay. I >> But if that many people are downloading it, you have to think like I don't maybe you don't maybe you don't have to think it's working yet, but there is certainly excitement and that excitement's been enough to really carry the stock forward. >> I I hope it's ready for prime time because one of the whole concerns like forget about the hacking, forget about that it's going to kill us. There is also a whole concern about verification. That's right. >> Who is responsible if it screws up the refund? If you're asking it to pay your bill and it something goes wrong and it doesn't pay it on time, >> who owes the electric company? >> Well, there's been who's responsible. >> There's been examples of agents saying stuff like you're like, "Hey, can you change my my seat for my my my flight?" And it goes on and says, "Uh, I accidentally cancelled your flight or >> yeah, whatever." >> Right. And then you get the AI, "Oh, that was my bad. I'll make sure to not do that again." Well, now I don't have a flight. It's just you taking it back to like where we are with with Facebook I mean or sorry meta stock doing okay right uh >> we saw some analysts like you squalally talking about this is like the first true kind of way to get non you know social media type revenue non-advertising revenue from a product right because a lot of the features you talked about have to be have be paid part of >> that's right a lot of this you do have to pay the subscription premium it's a premium model that's right >> and then the big question is like Julia you're saying do I trust Meta they're the most like let's just say aggressive with user data on and exploitationally historically so I want to have it triage my my my my Gmail or my email I mean I I don't know if I want to do that >> I don't know I mean you know Mark Mahaney at Evercore is also a big Metabol and he put out a note reiterating it's out before $860 price target on Meta um and one of the things he pointed out that I thought was interesting is like all of the attention has been on the enterprise there really hasn't been um a monetization of consumer AI yet the real use case like we've been waiting for the thing and he's excited that maybe this is the thing but >> that this could be the the personal use case >> right but again to come back to enterprise versus consumer the enterprise has scaffolding that's right >> the enterprise has verification the enterprise hopefully has redundancies and like you by yourself you don't necessarily have that so again listen I understand a lot of people excited maybe it's awesome but that would be my concern >> well I'll tell you who's excited about it is Amazon, >> right? >> We saw Amazon um their web retail side of the business come out and basically kind of block the Muse software and say you cannot use this to go shopping on our platform. Now Shopify took the other side of that and immediately partnered with Muse to get the Muse product onto its Shopify platform. But for Amazon, this immediately raises questions of you've built a model around selling products to people and advertising while you're doing that. The point of the ad is for the human to see something and buy it and then you get paid based on how much you know your ad gets seen if this is all robots doing the shopping. Does the ad model on that side drop out? And then also how does this control product diversification, product spread, consumer purchase behaviors, all of that gets impacted by this? >> Well, at some point I wondered like do agents respond to ads? >> Do we make new ads that agents that are designed for? >> There's a whole discussion. Uh Corey uh Johnson talked about how this is a whole market. They're not even talking about the the millions and millions of agents that are seeing stuff like that. Right. That's right. Just real quick about you're talking about downloads. Uh someone did a search for Instinct AI, which is the other agent. When you do that, it's all meta ads. All all in um sorry all news ads.

Assistir ao vídeo original no YouTube

Assuntos: Agentes de IA, Muse. Publicado no YouTube em 2026-09-22.

Seleção automática por assunto e idioma. O vídeo pertence ao canal de origem.

About Muse Wonder

Muse Wonder is a family of Meta AI models introduced on October 12, 2026, made up of Muse Wonder and Muse Wonder Full. According to Meta, they are its most capable models for personal and professional agent tasks, trained at Prometheus, Meta's gigawatt-scale data center. Meta highlights two capabilities: Active Memory, which keeps the user's context across conversations, and Active Focus, which keeps the model working on a hard goal, trying alternatives and checking in before important actions.

Meta says Muse Wonder is available through the Meta Model API and Muse Code, that Muse Wonder Full is available to Muse Code subscribers, and that Muse Wonder will later power experiences in Muse, starting with subscribers. The evaluations cited by Meta include Harvey's Legal Agent Benchmark (LAB), FocusBench and SWE-Marathon, with comparisons against Muse Spark 1.3.

Fact sheet

• Models: Muse Wonder and Muse Wonder Full (Meta)
• Introduced: October 12, 2026
• Capabilities: Active Memory, Active Focus
• Training: Prometheus data center (Meta)
• Access: Meta Model API and Muse Code; Muse Wonder Full for Muse Code subscribers
• In Muse: Planned, starting with subscribers
• Official page: Meta – Muse Wonder

Read also

• I Tested Meta’s NEW Muse AI — Is It Actually Good?
• Master Meta's NEW Muse AI in 10 Minutes (What can it do?)
• Meta’s Muse AI Agent Cost $0. What’s the Catch?

---

Sources and references: Meta – Muse Wonder · YouTube description of the video by the Yahoo Finance channel.