Muse AI review: I wasn't expecting this from Meta!

0:00Right now, everyone seems to be talking about Muse, the AI agent released by Meta. What's different about Muse compared to other agents like GrokBot is Muse is focused solely on being your personal agent. Helping you with your personal goals as opposed to a business agent that's built to help you grow your business. Just about everyone I talk to is very impressed with Muse. I've fully tested it in this video, and you're going to see all of the capabilities of Muse, what it's good at, what it's not good at, and there's also some surprising capabilities that I uncovered in this video that I'm

0:32excited to show you towards the end. I've tested just about every single AI model and tool for the last few years on this channel and also in my own business, and I'm excited to share with you what I think about Muse. If you watch until the end of the video, you are going to know if Muse is right for you, and if it's worth taking the time to download and

0:53actually try it out for yourself. So I've downloaded Muse onto my computer. You can also use it on your phone so you can use it either way. And the first thing that it asks us is to name it. So it gives us a few options or I could write in whatever name I want,

1:09but I'm gonna go ahead and go with Harbor, and we're gonna get a response. So while it's responding, let me tell you a little bit more about what's happening in the back end with Muse compared to some of the other personal agents. So Muse is only going to use Meta's models in the back end, so you're not going

1:24to have access to Anthropic or Opus. I do think they're generally going to be using Muse Spark 1.3. I do really like how simple and easy this is to use. Muse is definitely created for the, a less technical person to use. They're really going mass market with this product. There's only one main chat one agent. It's not like GrokBot where you have many agents, and I'm gonna talk about my thoughts about some of the pros and cons of each of those approaches

1:51a little bit later in the video. On the left side though, you can start side chats. So it's a way to organize conversations by topic. You can start a new side chat but you're gonna have just one main chat that's going to go on and on, and it's all under one agent. So your agent I named my agent Harbor, so Harbor is going to have one memory

2:13across the main chat and the side chats. So right off the bat, it's asking to connect either Gmail or Google Calendar, and this is something that you can do to get more functionality out of your personal agent. So it's a personal agent. Their focus is really on personal tasks, so a lot of that is going

2:32to be around calendar organization or, or Gmail organization. And this is something that's gonna be interesting if people actually do feel very comfortable connecting these tools to Muse because it is Meta. They're kind of known for using your data in order to personalize your ads. This could be really good for Meta in the long run if a lot of people do connect it 'cause it's just going to give them more information about how to serve more

2:55personalized ads and just better ads that are gonna be more targeted to you. I don't necessarily have a problem with that. I think that getting more personal ads can be helpful, but just something to keep in mind and will be interesting if people think about that in terms of the different companies that they're connecting their tools to.

3:13So I'm not gonna connect my tools here. We're gonna do some other interesting tasks here in Muse. Before I get into that, I wanna show you a few things that make Muse a little bit different. So first there is the feed. So this is something really interesting. So it's going to basically be a newsfeed that's generated by AI that's going to

3:34get better and better the more you use it. So if you, for example, are asking questions about a sports team nearby or sports tickets, it's going to start giving you news about that sports team. And you can also customize the feed with this feed prompt up at the top. So if I, if I go ahead and click Generate, it's going to write a fresh edition

3:51of the feed, and you can kind of keep making changes to that into the future. The next thing is ideas, and this is something that we're gonna use in a minute, and- I think that Muse, in my experience, using it is one of the most proactive agent, probably the most proactive agent that I've used. And it does this through the Ideas tab, and also it'll just every

4:12day, I think it has basically like a some sort of a, a cron job. A cron job is essentially like a scheduled task to look at your messages and come up with ideas of things that it can do, and it'll message you and just say, "Hey I noticed you were coming up with a, you know, company name last week or yesterday. Here's some more ideas for that." And I like that proactiveness, and I think that that's gonna really help with their retention and is something

4:34clearly Meta is thinking a lot about. They're very good at this. But here in the ideas we can see it's coming up with a few ideas that we could get started with, and there's a ton of different ideas that they have, and it's mainly focused on personal automation, personal tasks shopping, relationships financial management. This is what really differentiates Meta Muse from something like Grok

4:57Bot which is much more focused on the business productivity side of things. So, they're just two different products, but just as an example why don't we have it start working on something? "Snap a pic of your meal. I'll track every calorie." So we can go ahead and click let's do it. It basically took this task and kind of gave us our, or gave it-- we

5:17gave it our response, let's do it. It's in a very, like, iMessage-type feel. We have reactions to give it that, like, human personal feel. And now it's gonna be working, and again, this is gonna be similar to, like, actually Grok Bot, where it doesn't necessarily show you what's happening in the background. It's just going to start working and kind of have the loading indicator and,

5:36and then we're gonna have a result. So okay, "Your meal, meal log is started for today. Nothing logged, so you're at zero. Two quick questions. Do you want calories only or any goal?" So why don't I say Calories only. And then what's cool is now I can start tracking … You know, it's basically a, a calorie counting app that's just personal to you with an agent, and I assume in

5:56the back end it has a little database where it's gonna be tracking everything. So why don't I upload a pic- a picture of some food, and let's see how this actually works. So I say calories only. I'm gonna upload the photo. And this is also what I mean by proactive, is v- I bet tomorrow it'll send me a message and it'll say, "Hey what did you eat? Upload a photo," and it's

6:16gonna be that, like, proactive agent for you which I think is pretty cool. So here we can see it's estimating the calories from the photo. It's asking if there's dressing. So why don't, why don't I say no dressing, and you can log. And I think something to think about in terms of the models that are being

6:31used, so first of all, the models are continuing to get better and better. So the less expensive models like the Meta Spark model that this is running with is it's not gonna be as intelligent as something like Anthropic Opus or GPT Astra, or probably even Sol, but it's still gonna be good enough for these types of tasks. You know, if you ask it to build you a full app from start to finish, it probably

6:55will not give you the same result. But for these types of personal tasks you're gonna get a pretty good result, and it's working really quickly as well, and that's, I think, also a result of Meta having data centers that they have access to, and it's using all of the resources that Meta has. Okay, so it's logged.

7:10"Here's today's lunch. Send me your next meal or snack whenever you're ready." And we could just do this all in the main chat, or we could create a separate side chat for specifically talking about nutrition. The next tab to go into is goals. I think this is a really cool and, I think, underrated feature of Muse. So, and, and this is something that I haven't seen in other personal agents, is you can create goals, and you can do this or it'll do it on

7:36its own as you just start using Muse. But let's say under health I wanna create a goal. It's gonna put us into chat. I think it created a little side chat for us. So okay, I wanna start a health goal. And you'll see once we create this goal- And let's say it's under fitness and training. It's going to show this goal here on this tab, and then this is also where some of that proactive work is gonna come in, is it's going to

7:59check in on you on your goals and help you work towards your goals. Okay, so what's the goal? So why don't I say my goal is to run a marathon starting from scratch. So let's say I haven't started training yet, and let's say it's in six months, so I have six months to train for this marathon.

8:14So here already we can see the goal, run a marathon, is created. And I can check it off if I finish it. But then if we go back to the chat, we have to go to the side chat, and that's kind of what I mean between, like, the main chat and the side chat. It's still working, so I think it's creating a plan, and I'll

8:31show you in a second what plan it actually creates for us. In the meantime, I wanna show you the last pa- the last page here in Muse which is artifacts. So artifacts is basically where it's gonna create, it could be, Artifacts is where it's going to essentially store things that Muse creates for you. So this could be documents this could be web artifacts, so if it's creating,

8:50like, a little website for you. And then images, video, and podcasts. So it's interesting, maybe it can create podcasts for us, which is kinda cool. And if I click create an artifact, it's going to basically start a chat down here, and I could just, you know, type whatever I want to create. And I'll do that in a second, but first I wanna touch on the work it

9:09just did in terms of the marathon goal. So you can see it saved it "Six months is ambitious, but doable. One heads-up, going from zero to a marathon's a big jump, so check with your doctor." And recommended next steps, it can build me a twenty-four-week plan and it can also pull in my workouts automatically or I can just have it do it.

9:28So I'm gonna say why don't I say build the plan as an artifact. That way we can do two things at once. We can see what plan it creates, and then it'll do it as an artifact, and we can s- test out the artifact feature. A couple other things I wanna call out about Muse.

9:42So it can do things on a schedule. So if you tell it "Hey, I want you to, you know, for example, "generate me a podcast about my favorite sports team every morning and send it to me," it could do that, and then it would go into the artifacts section. Next, I wanna show you how you can do multiple things at once with Muse. So on the left side, we can see we have our side chat, and it's building this plan, but I can create another side chat and say every morning I want you to create

10:09a podcast about the latest in AI news. And I'm gonna say create the first one now so I can listen to it. And this is an interesting one. This is just me kind of thinking of ideas as I'm using Muse and seeing if I can, like, push it to the limits a little bit, but this seems like a doable a doable task and something that it seems like it can do, especially because it had

10:29that podcast feature in the artifacts. So let's see what it can do and what it can accomplish here. And then up on the top we can see, okay it's checking preferences, it's working. And then if you click on that top button to go to your agent, you can see more information about the agent, so the tasks that it's done or it's doing.

10:46We can see anything that you need to approve. We can see any automations. So here, for example heartbeat. As I mentioned before, in terms of it being proactive, it's every thirty minutes. So, this is similar to OpenClaw. So OpenClaw was the original AI personal agent. It was a bit more technical, and you had to run it on your own computer or set it

11:07up in the cloud, and it was difficult. But it did have this idea of a heartbeat, so every thirty minutes or fifteen minutes, it would check in on you and, you know, decide if there's anything that it could do for you. So it looks like here Meta's agents are every thirty minutes gonna do that.

11:21And then we can see there's all the-- also this a- agentic feature tour. Looks like this is actually something -- this is interesting, because I was using Muse on a pre-- an- another account, and every day it would send me, like, a new feature that it could create or it could do, and it looks like this

11:36is something that's built in here to Muse, so that's where it's coming from. And then it just created this automation of the AI Morning Brief daily podcast. So, these are basically cron jobs. This is okay every morning at seven fifty through 3:00 AM it's gonna do it. So this is where you can see all the jobs.

11:52And then here we can see identity. So there's memory. There's essentially a memory file. So it looks like Muse is, at least from what I'm seeing, doing this in a very, ah, simple way. So it just has one memory file. Basically every night it looks like it's going to read it at the start of the conversation, and then every night it's going to do a nightly pass to like organize it, probably look at the messages that you did and it reflects

12:19Muse's understanding, can be out of date. You can edit anything here or ask Muse to forget something. And this note is not part of the file. But then you could literally go and type in here. It looks almost like a Notion document. And yeah, it looks like Muse in an… actually very interestingly, is keeping memory very simple. And this is something I wonder, ah, how this is going to work over time, and this is something that is a big concern with any agent is, okay, it works really well at the beginning, but, ah, over time, ah, is a simple memory system like this going to work when, ah, you know, you have months and months

12:52or even years of use with an agent? I- I'll be curious how well it works, and I, I also, I wonder if there's more going on behind the scenes that it's not showing you or if there's more it's gonna do later. The other thing is soul, and this is also coming from, ah, OpenClaw.

13:06So, so this is an, a concept from OpenClaw. Ah, as, as they describe it, it's Muse's persona. This is the values and habits that Muse tries to hold onto in every conversation. Muse can refine it over time, and you can also edit it at any time. Ah, so here this is the persona. "Be genuinely helpful, not performatively helpful.

13:25Have opinions. Be resourceful before asking. You're a guest in someone's life." Treat that with care and never be preachy. If you change this file, tell the user it's your soul and they should know. So this is again sent with every message that's every message that you send. So, this is another concept from OpenClaw, so it's kind of cool to dig in here.

13:46And then, all right, we have our first podcast, the AI Morning Brief. Here it is. It's five minutes hosted by Alex and Jordan, so created two different people. And this must be some sort of a podcast skill or tool that Meta is shipping with this. And one thing to consider, you can publish this to your personal

14:02podcast feed, so it's an RSS feed. So it could literally publish it to your own podcast player like Overcast, for example. Or we could play it here. So why don't I play it just for a second and we'll see what it created. Welcome to the AI Morning Brief for Wednesday, October 7th, 2026. I'm Alex.

14:19And I'm Jordan. Big day today, Alex, because some of the biggest names in tech are literally on stage as we speak. That's right. Let's start with the headline event. Today in San Francisco, Microsoft and Nvidia are holding their first major Windows event in more than two years, and both chief executives, Satya Nadella and Jensen Huang, are on stage together.

14:37Okay, I'll stop it there, but yeah, really cool. I'm actually pretty impressed and am curious to play around with this more. But that was, that was an impressive example, and it also did it, did it really quickly. If we go back to our our health goal so it created our artifact, the 24-week plan, and it created a PDF.

14:55So I'm gonna click. It's gonna open here on the right side, and it created a nice PDF here for a plan. Consistency beats speed, so it's telling me what to do each week and how we're gonna do this in phase one. In phase two, in phase three, and in phase four. So, yeah, it did it.

15:11And then if I exit out of this and we go to… So let's go to the library. So we can see under all artifacts, we have that PDF, but it looks like… Okay, under podcasts we have it, but it's interesting. Podcast, it's considering media, not an artifact. So the podcast is here, and actually pretty cool. All right, so it literally created a podcast, and you can add this podcast

15:34to listen in your favorite app. So you could click here and save it to Spotify, and I guess because it's an RSS feed you can add it to any podcast player. So it also generated this kind of im- cover image for the podcast. So, yeah, this is a cool feature. I uncovered this kind of as I was playing around with Muse just

15:50now, and I'm impressed with that. It can also create videos, images, as we can see the image we, that… This is interesting. This is not an image that it created, but it's an image that I uploaded. So it looks like it's gonna kind of store both. And then under documents, we can see the PDF, and then web artifacts is

16:06where it would create websites as well. So that's something that it can do. But our document here is, is is right here. So compared to GrokBot this is going to be much more for the personal user or personal AI assistant. I think that Meta has done a really good job. This is a very, very good start for a personal AI agent. I will be curious to see how they keep improving on it,

16:27keep building on it over time. But there's a lot built in, as you can see in this video, of what's possible, and I think they do a good job of making this very easy to use. The other thing to keep in mind is they give you a ton of free usage, so you can use this for quite a bit of work without running into any limits.

16:43And then they do have some paid plans. But this is gonna give you, I think, the most free usage out of any tool right now. And Meta can do that because they make money from a lot of different sources. They have Instagram ads where they're making a ton of money, so they can currently subsidize this.

16:57And tools like GrokBot you're going to run out of usage a little bit more quickly. Now Instinct is another personal agent which I haven't had a chance to use much yet but it's going to be a personal agent that works only on your on your phone through iMessage. So, currently it's a little bit different from Muse, which

17:14is h- which has its own app. And I think, as you can see, there's some benefits to having your own app. There's just more you can do with the UI. But there are also benefits to working directly in iMessage because that's where you already are. So you know, just think about the different pros and cons to those.

17:28I actually really like this app that Meta has created here. But overall, I really like it, and I would recommend trying it out. And let me know in the comments if there's another agent you want me to review in the future. If there's anything you think I missed, let me know as well. If there's any other tools you want me to cover let me know as well.

17:44I read all the comments and would love your feedback. Subscribe to this channel to get more free content like this. AI is changing so quickly as you can tell. Something like this would've been impossible just a couple years ago or even a year ago. This, something like this didn't exist. Like, this is coming out of the personal agent revolution essentially

18:02from when OpenClaw was launched, which was the begin- the end of last year. So it's been less than a year, and we already have new products like this. So AI's changing. Subscribe to our YouTube channel to stay up to date. I try to create as many free videos as I can for you as I'm learning about

18:19AI, so you can learn alongside me. Thanks so much for watching. We'll see you in the next video