I Tested OpenAI's Dots vs. Meta's Muse: What You Need to Know
A hands-on, day-in-the-life comparison of two new persistent AI agents, covering memory, pricing, integrations, and which one actually gets used.
Posted
today
Duration
Format
Review
educational
Views
6K
266 likes
57 · 43
Big Idea
The argument in one line.
OpenAI's Dots launched feeling rushed and unfinished next to Meta's Muse, which wins on interface, free access, and editable memory, even though Dots runs on a smarter model and plugs directly into an existing Codex and ChatGPT workflow.
Who This Is For
Read if. Skip if.
READ IF YOU ARE…
You're already testing persistent AI agents like Dots, Muse, or GrokBot and want a side-by-side on memory, pricing, and integrations before picking one.
You run most of your work through ChatGPT, Codex, or a coding assistant and want to know whether a dedicated agent layer on top is worth adding.
You're choosing a tool to recommend to a non-technical friend or parent and want to know which interface is actually easier to use.
SKIP IF…
You want a tutorial on how to build or code an agent yourself. This is a user-facing product comparison, not a build guide.
You have no interest in OpenAI or Meta's ecosystem and only use open-source or self-hosted AI tools.
TL;DR
The full version, fast.
OpenAI's Dots and Meta's Muse are both persistent cloud AI agents you talk to, schedule tasks with, and connect to messaging apps. Dots runs on OpenAI's Astra model and plugs directly into an existing Codex and ChatGPT setup, including Slack. Muse runs on Meta's Spark 1.3, has editable soul and memory files, a native WhatsApp connector, and a free tier with up to 100 million tokens a week. Muse's interface is more polished: a today view, an idea feed, and one-click podcast generation. Dots feels rushed, mismanages its own project threads, and even misdescribes its own behavior. Pick Dots if you already live inside Codex and ChatGPT; pick Muse if you want the easier, cheaper on-ramp.
Free for members
Chat with this breakdown — free.
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.
Opens comparing Dots and Muse side by side, including the clip of Dots failing live on stage at OpenAI's Dev Day and getting muted on the livestream.
01:40 – 04:32
02 · Features and Memory
Covers background work, model access, and workspace setup, then how each agent's memory works: Dots inherits ChatGPT memory while Muse exposes an editable soul file and memory file.
04:32 – 06:58
03 · Slack, WhatsApp, and Ecosystems
Shows Dots integrating into Slack (DMs and channels) versus Muse's native WhatsApp connector and side-chat threads, plus how each leans into its parent company's ecosystem.
06:58 – 08:18
04 · Pricing and Model Performance
Compares Astra (Dots) against Spark 1.3 (Muse) on benchmark scores and speed, and lays out Muse's free 100-million-token weekly tier against Dots' paid-subscription requirement.
08:18 – 10:57
05 · Muse Interface Walkthrough
Tours Muse's today view, approvals, schedule tab, heartbeat, goals, feed, and one-click podcast and artifact creation.
10:57 – 13:02
06 · Dots Interface and Limitations
Tours Dots' voice calling, terminal access, and scheduled tasks, and catches it scattering new Codex threads outside the user's existing project and misdescribing its own capabilities.
13:02 – 16:30
07 · Verdict and Final Thoughts
Lands on a four-point verdict: Muse's interface advantage is real, Astra is the smarter model, GrokBot's multi-bot structure may scale differently, and Dots still feels like a rushed catch-up product.
Atomic Insights
Lines worth screenshotting.
OpenAI's Dots visibly failed on stage during its own Dev Day demo, and the team muted the live stream rather than let the audience watch it recover.
Dots and Muse are both single-thread persistent agents, not multi-bot systems like GrokBot, so you build a relationship with one assistant instead of spinning up many.
Muse's memory system is split into a soul file for persona and a memory file for facts, preferences, and commitments, and both are human-editable text.
Dots inherits its starting memory from existing ChatGPT history, which can surface stale context a user has to manually correct before the agent understands current priorities.
Muse connects natively to WhatsApp for two-way chat, while Dots integrates directly into Slack, including group channels a whole team can message.
On paid benchmarks, Astra (Dots) scores higher than Spark 1.3 (Muse) on overall intelligence and coding, but Spark answers roughly 3.6x faster in tokens per second.
Muse is free to start with up to 100 million tokens a week, while Dots requires an existing paid ChatGPT subscription just to unlock the agent.
Muse's interface groups work into a today view, an approvals queue, and a schedule tab with a recurring heartbeat that proactively flags things from email and documents.
Dots kicks off new Codex threads outside a user's existing project structure, scattering work into a general recents list instead of the project it was meant for.
When asked how it manages scheduled tasks, Dots gave directions that didn't match its own actual interface, undermining trust in its self-described capabilities.
Muse's free-form goals and feed features are built to give non-technical users ideas for what to automate, rather than assuming they already know what to ask for.
Meta's ecosystem tilts Muse toward Instagram, Facebook, and Marketplace connections, while OpenAI's ecosystem tilts Dots toward coding and existing ChatGPT work.
Takeaway
How to actually choose between Dots and Muse
WHAT TO LEARN
The interface and free access decide adoption faster than raw model intelligence, so the smarter model under the hood doesn't automatically win the tool.
01First Impressions
Watch how a product launches under pressure, not just what it launches with: Dots visibly failed during its own live announcement, and muting the stream is itself a signal about how finished it was.
Both agents let you name and customize your assistant, but customization is surface-level. The real differentiator is what's underneath it, not the avatar.
02Features and Memory
A new AI agent inheriting your old chat history isn't automatically a shortcut. Stale memory that doesn't reflect your current priorities can be worse than starting fresh and re-explaining.
An editable memory file you can read and correct directly builds more trust than a black-box memory system, even if the underlying model is less capable.
Persistent cloud agents increasingly come with their own always-on computer, so scripts and API connections still work even without a native integration.
03Slack, WhatsApp, and Ecosystems
Which messaging app an agent plugs into isn't a minor feature. It decides whether the tool fits into a team's existing habits or asks everyone to open one more app.
An agent that can be added to a team Slack channel in three clicks changes who uses it: it stops being a personal tool and becomes something coworkers interact with directly.
The ecosystem a company already owns pulls its agent's roadmap in that direction: expect coding-first integrations from an OpenAI product and social and commerce integrations from a Meta product.
04Pricing and Model Performance
A faster, cheaper, slightly less intelligent model can still be the better choice for day-to-day knowledge work, where responsiveness matters more than winning every benchmark.
Free access with a generous token cap lowers the bar to actually trying a tool, which matters more for adoption than whether the paid tier is the better deal.
When evaluating two models, separate 'better at coding and automation' from 'better at everyday assistant work.' The same model rarely wins both by the same margin.
05Muse Interface Walkthrough
Organizing work into a clear today view, an approvals queue, and a schedule tab makes automation legible, so a user always knows what the agent already did versus what it's about to do.
A recurring automated check-in that surfaces things unprompted turns a reactive chatbot into something closer to a proactive assistant.
Giving users pre-built ideas solves the 'I don't know what to even ask for' problem that stops a lot of non-technical adoption before it starts.
06Dots Interface and Limitations
An agent that can't accurately describe its own features breaks trust fast. If it sends you somewhere a setting doesn't exist, you start doubting everything else it tells you.
Losing track of which project a new task belongs to defeats the entire point of an organized workspace.
Voice as a feature only matters if the underlying interface it's attached to already works. A good voice mode wrapped around a buggy core doesn't fix the core experience.
07Verdict and Final Thoughts
A tool that's inexpensive to try, has editable memory, and connects to what your family already uses can outweigh 'the underlying model is smarter' as the deciding factor for who you recommend it to.
The strongest verdict isn't 'X is better,' it's 'X is better for this workflow, Y is better for that one.' State the workflow, not just the product, when you make a recommendation.
Expect converging products: once one competitor ships an interface feature people love, the other visibly re-implements it within weeks. A gap between two competing agents is a snapshot, not a fixed ranking.
Glossary
Terms worth knowing.
Dots
OpenAI's persistent cloud AI agent, launched with a customizable avatar, that runs on the Astra model and connects to ChatGPT, Codex, and Slack.
Muse
Meta's persistent cloud AI agent, the top app in the Apple App Store at launch, that runs on the Spark 1.3 model and connects to WhatsApp and Meta's apps.
Astra
The model currently powering OpenAI's Dots, benchmarked as more intelligent and better at coding than Muse's underlying model.
Spark 1.3
Meta's in-house model powering Muse. Scores lower than Astra on raw intelligence benchmarks but generates output roughly 3.6 times faster.
Soul file
A Muse document that defines the agent's persona and personality, editable by the user like a plain text file.
Memory file
A Muse document listing facts, preferences, and commitments the agent has learned about the user, editable directly instead of buried in opaque memory.
Heartbeat
A recurring automated check that Muse runs roughly every 30 minutes, proactively scanning email and documents and surfacing alerts or reminders.
GrokBot
xAI's competing agent platform, referenced here as a third organizing model that lets a user create multiple distinct bots instead of one persistent thread.
the entire video's verdict lands in five words→ TikTok hook↗ Tweet quote
06:30
“If I wanted to have my parents or people that I know that aren't technical at all, I would probably recommend Muse.”
gives a concrete, relatable litmus test for who each tool is for→ IG reel cold open↗ Tweet quote
12:36
“Dots doesn't even know how it works.”
damning one-liner that undercuts trust in the product→ newsletter pull-quote↗ Tweet quote
15:19
“I said, Dots should have been called Bugs.”
quotable pun already tested as a tweet→ TikTok hook↗ Tweet quote
The Script
Word for word.
Read-along
Don't just watch it. Burn it in.
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.
17px
Today, I'm gonna be comparing OpenAI's Dots and Meta's Muse on things like their features, their functionality, and their feel. Also talk about pricing and where I think these apps are headed so you can use your time wisely. And of course, I don't wanna waste any more of your time, so let's hop right in.
Okay, so we have Meta's Muse, which quickly became the number one app in the Apple App Store. And then of course, we have ChatGPT's Dots. Now, my immediate reaction when I got in here and started playing around with Dots is that it felt like a bit of a rushed.
project to try to catch up because they saw how Muse was taking over and they realized we have a better model and we already have a lot of people inside of our ChatGPT ecosystem. Let's give them essentially what Muse is, but in our ecosystem. And it just feels super rushed.
You can see even during Dev Day yesterday, when they announced this new product, it literally failed. She was trying to call it here and it just wasn't responding and it was taking forever. And she ended up saying, oh, I think Dottie or whatever the name is, is having a slow day.
And they ended up like, just kind of pushing past it but they like muted it on the live stream so everyone that was there live in person saw the live demo they both have the ability to name your little muse or your dots and you can customize it a little bit as you can see same exact thing here inside of the chat gbt or the codex desktop app where i'm able to access my dot right here now the first thing that surprised me was i was expecting dots to be more of a grok bot feeling app where we would have a bunch of different bots that we could create and we can have them talk to each other but this is way more like muse where it's kind of one thread or one agent that you're talking to that kind of helps with everything on a personal level.
So we will look at the UI and we'll compare the differences and we'll talk about which one I like more. But first, let's just talk about some of the foundational stuff. So background work.
These are both persistent cloud agents that can have schedules and you can have conversations with and they both have their own cloud computer. The model powering dots right now is GBD6 Astra and the model powering Muse is Muse Spark 1 .3. I'll talk a little bit more about the differences between those models.
later in the video. As far as the existing work setup, you already have a direct connection in Dots to ChatGPT, Work, and Codex if you have the desktop app and if you just allow access. As far as Muse, it has its own computer and tools, and they both have a really good connector ecosystem.
So kind of like the super simple single sign -on, just connect with a button rather than having to do a bunch of API keys and things like that. And because they both have their own computers, they still can run scripts and you still could connect to APIs that might not have a native connector quite yet. They can both create documents, presentations, interactive dashboards.
They can create all of that and keep them in the ecosystem. As you can see here, if I go to my library inside of Muse, I can see these different artifacts, whether that be a document or whether it be more of a live dashboard web artifact. And then inside of my dots, you can see that there's a bunch of ways that we can see the outputs here, whether those be markdown files or dashboards and other things, of course, especially because with dots, your memory will start with your ChatGPT memory, and then it will start to develop its own context and learn over time.
So right away, my dot was connected to all my ChatGPT memory and it was saying, oh, you know, you run this business and you use NNN every day and you do this. And I was like, okay, that's really old memory.
That's ChatGPT memory. Let me plug you into my actual local files and my codex memory so that you have more real -time data of what I'm actually doing and what are my priorities. But what I really like about Muse is in here.
For the average person, let's say you're coming in here and you click on your Muse. Mine is named Charlie in this case. There's a little identity tab over here where we can see the soul file and the memory file.
So the memory file, we can actually like edit things. This is like facts and preferences and commitments like that. And then if we go back over here and we go to the soul, this is like Muse's persona.
So, you know, if you have built a Hermes agent in the past or OpenClaw and you've kind of worked with these memory and soul files, you will be familiar. But even if you're not, it just makes it super easy and visual because you can just sort of see it from this interface. Whereas with dots, it's a little bit more abstract as far as how it's understanding memory.
And you don't have as much to do with sort of like the profile. I don't know why that looks super pixelated right now. But overall, I'd say like the Muse interface is just a better experience, especially if you're thinking of someone who's non -technical.
By the way, guys, I've got this completely free SOP for you about getting your first AI automation client. It's going to go over the exact steps that has been proven for hundreds of our AIS Plus members to get their first paid gigs. It goes over the one sentence service pitch that can get you started today, why your first client should cost you money, the five minute video that answers can this person actually deliver before you've actually...
received any money. What to do when you have zero case studies. There's so many good things in here that are going to help you out.
Even if you already do have clients, I would recommend grabbing this because like I said, it's yours completely free. So if you want to grab this, there's a link for it down in the description. Let's get back to the video.
Now here's another difference. With communication, you can obviously do this stuff from your phone. You can use the Muse app or the ChatGPT app to talk to your dot or to your Muse.
With the dots, you can instantly integrate them into Slack, which is pretty cool. So this was literally like a three -click setup where I added this app into my Slack. So now I could chat with my actual dot right here and it would go ahead and respond.
If I open up the desktop app real quick, you can see in here that you can see that it's thinking. So you see the top, it says thinking, there's a little light bulb. And then in my Slack, that's how it's gonna respond.
But then also inside of a channel, I could have it talk to me in here by inviting it to a channel. So your team could actually message your dot, which I think is pretty cool. And once again, you can respond in here.
and it will show in the app that it has received this message, that it's thinking, and that it's trying to reply inside of that thread. And as you can see now, it is thinking, and then it's going to shoot a message back into the thread right there. It said, hey, I'm here.
So that is pretty cool. But with Muse, what you can do is you have a bunch of different chats you can manage. So you've got your main chat, and then you can set up different side chats for different topics.
And you also have a native WhatsApp connector, which is pretty cool. So on the app in Muse, this is view only. So what I would do is inside of my WhatsApp, whether I'm on my desktop or I'm on my phone, I can just say, hey, how are you today?
And this obviously is Charlie, my Muse responding in real time. And then in the Muse app, I would be able to see this interaction, but not respond in here. So it's a separate thread as far as my main chat, my WhatsApp channel, and then these side chats.
And then what else is very interesting is the ecosystem. So because Dots is already integrated into Codex and ChatGPT, If you are already living heavily in Codex and ChatGPT, then Dots probably makes more sense at least right away when you want to figure out what this feels like.
But with Muse, it has some of these, you know, like the meta ecosystem. So Instagram, Facebook, Facebook Marketplace, and it just feels a little bit more, you know, detached from like this coding stuff. Anyways, that's just a very quick rundown.
We're going to dive into some more stuff here. But overall, after playing with these each for about a day, this is how I feel if I could just sum it up. Dots fits your current work environment.
If you're using Codex and ChatGPT, it can use local skills. It doesn't feel like you're kind of starting over explaining a lot of things like priorities and things like that. And it offers a direct route into Codex and Slack.
Whereas Muse is much easier to evaluate for free because you can get started on this for free. It's inexpensive paid entry, editable memory, WhatsApp access, strong meta integration. Overall, like if I wanted to have my...
parents or people that i know that aren't technical at all i would probably recommend muse it just has a better experience and when you talk about this actual pricing on the muse free you can get up to 100 million tokens a week and you can get started like i haven't yet paid for muse and then you can get on subscriptions and you can increase your tokens whereas in chat gpt with dots you have to be on a paid subscription in order to actually get access to building your first Now there's something out about right now, they're not actually attributing your dot usage to your weekly subscription, but I think that's just to sort of build some, I guess, adoption inside, and then they're definitely gonna have that ding against your subscription.
Now also right now, we talked about how Muse is powered by Spark, and Dots is powered by Astra. So let's talk a little bit about these models. Here's some rough benchmarks.
Astra is a little bit more intelligent, but when we think about like professional work and knowledge work spark doesn't do bad at all so creating some of these documents and helping you with some of these routine tasks and figuring out how to unblock yourself things like that now when it comes to building apps and software and automations obviously ash is going to be a bunch better with coding it's going to be a lot better but muse is also going to be a little bit faster so i guess what i'm trying to say here is if i was building a bunch of little automations and scripts that I wanted to run and things like that and dashboards with one of these tools, I'd probably choose Dots.
But like I said, for general knowledge work, and if I'm just using this kind of as an assistant to help me keep on track, manage emails, remind me of things, manage the team, I would probably be pretty comfortable using Spark. So let me just show you, I've said quite a few times that I think the interface of... Muse and the experience of Muse is just better, let me show you why.
Inside of Muse, you obviously come in here and you create one thread. So with Grokbot, it can be a little intimidating to people because there's the ability to keep creating more bots, but here you just create one bot. Now, what you see right here is you see today.
You see things that have been worked on by date. You also click on this next tab, which is approvals, if there's anything for you to approve. So super nice, easy interface to approve things if you've got a bunch of automations running.
You have a schedule tab. So you can see here are your reminders. You can see here are things that are running.
You can even check in on the heartbeat that runs every 30 minutes. And the heartbeat will proactively look at your Google Docs and your email, and it will give you alerts about things. And it will even give you reminders like, hey.
Use me, set this up. You know, like it just feels like it's more user -friendly there. And then of course you have the identity of looking at things like your soul and memory.
Now also you'll notice on the left -hand side, we can search. So we can search through a bunch of different chats to figure out, oh, I was talking about this with Muse, where'd that go? You can also look at your different chats like I talked about.
You can manage them here, which I really liked because it's not as overwhelming as having to build a bunch of different bots. It's more like, oh, you can have a couple of side chats if you have one very specific project that you want my help with. Otherwise use the main chat.
And here's the channel that you've connected me to. And this is plural. So what I'm assuming is they're going to start to add more, probably Slack, probably Teams, Discord, things like that.
A lot of times I hear from people, okay, if I was to use AI, what would I even use it for? Well. Muse does a good job of like giving you ideas.
You can set goals, which I think is really cool. Optimize my sleep, health goals, relationship goals, career goals, interest goals. You can come in here and look for ideas.
It'll tell you what it should be thinking about and how you should be kicking off like these different automations and things like that, building artifacts. It's really, really cool. And then if you go to the feed, you can see that you can actually customize what you want to see.
You can see things about sleep because I've been talking about sleep. You can look at open AI. You can look at all these different things.
And if you like this, you can discuss it and you can just shoot off a task to your muse to say, hey, can we dig into this a little bit more? Build me an artifact for this.
Let's chat about this, whatever. You can also like things, which is gonna change the algorithm to give you more things that are actually relevant to you in your feed. I think this is a really cool experience.
And then if I go to my artifacts, you can see that I have all of these here, documents or web. We can also create images, we can create videos, and then you can create podcasts, which I think is really cool. So maybe you want to turn your feed, Daily Tech News, into a morning podcast, a three -minute morning podcast, or you want a weekly podcast on what your team accomplished this week or how your workouts were this week.
This is very, very cool. And something that I think like average people getting into this whole AI stuff, they'll think that this stuff is much cooler. Then if we go into Dots, the interface just is not good.
It feels rushed. It feels very simple. The one thing I think that it has nice, which Muse doesn't have, is that you can call it.
Dot which is actually the thing that I showed earlier which failed during the actual dev day launch But you can call it you can talk to it. It will respond in real time. It's pretty good voice It's it's pretty like the latency isn't bad at all They both have their own computers as you can see and it this saves logins So very similar to grok bot where you can save logins.
You can look at the actual sort of like terminal And you can also create and save files in here as well. This is how you set up that Slack connection.
And then you can also review recent activity. It's not very easy to manage your scheduled tasks inside of Dots. As you can see, here's one, right?
Plan tomorrow. And this is a repeat daily task at 8 p .m. Central.
We have some advanced settings as well. But I was asking Dots, I was like, how do I actually manage this stuff? And it was like, oh, click on my profile and go to scheduled.
There's no scheduled in here. I guess the way you do it is you go to your scheduled tasks as if you were looking at like your codex scheduled tasks or your ChatGPT work scheduled tasks, and you can see them in here. But why would you not be able to manage your scheduled tasks inside of the dots interface?
I thought that was just a little bit bizarre. And it also was telling me that it was really good at managing my codex threads, which was cool, but it really doesn't do it. Here you can see, I was like, oh, okay, so you're kicking off new threads and it's just appearing in my recents.
It's not actually putting it inside my HERC 2. So I obviously like to manage my threads right in here in my HERC 2 project, but it would kick them off down here. So right here, you can see here's a thread that it kicked off.
Actually, no, not that one. This one, you can see that it was sent by HERC. So my dot is basically kicking off these tasks, but it's not putting them inside of my HERC 2 project.
So I'm like losing the organization. So that's leading me to this point, which is like, Dots doesn't even know how it works. I love when I can talk to Cloud Code or Codex or Grokbot about how it works so I can get a better understanding of how I should utilize it.
But Dots will just tell me things that's like, okay, that's not actually how you work. So how do I, you know, I start to lose trust over what kind of knowledge it actually has. But hey, at least this little guy is cute and customizable because ultimately people seem to care way more about that.
Okay, so ultimately, where do we sit now? Well, I think that I'm going to continue to play around with both of them. But if I had to choose one, for me, I would be choosing Codex.
I would be choosing Dots simply because I live inside of Codex and Cloud Code. I don't care as much about this super polished and super nice interface because I already have ideas and I already have goals set and I already have a lot of the stuff set up already. But if I didn't, I would 100 % want this.
Because Muse's interface is so much better. Their advantage there is meaningful. But what I think is interesting is that Dots or OpenAI is probably getting so much feedback around Dots and they've obviously been looking at Muse and taking inspiration from it, which is why I'm so confused.
I was expecting Dots to come out and have a very similar style interface, but it's just like lacking so much. But what they're going to do probably is they're going to go to Muse and be like, okay, cool. I like this, you know, ideas tab.
Let's put that inside of dots. I like the way that these artifacts are set up. And, you know, OpenAI also did announce their spaces thing, which is going to be cool.
But they're just going to take features of what people really like about Muse or what people really like about Grokbot and just put it in there as well. So what I think is going to happen is like Muse, Grokbot, dots, they're probably going to all sort of blend into one sort of tool, meaning they're all going to look. functionally very very similar the differences will be some of the native connections and the capabilities of integrating with what you already have and obviously the underlying model that powers the thing and then pricing you know like the cheaper ones might be perceived as better to certain audiences and that was kind of my next point which is i think astra is just obviously a much better model than muse spark so that is one advantage that it has right now i also think that grokbot right now is a different organizing model because it has a bunch of Crock bots that you use together rather than just kind of working with one thread.
And I think that there's pros and cons to each. I think that the. average user who is not deep into the world of AI will probably like the interface of Muse and Dots more because it just seems more simple.
Oh, I have one personal assistant that I can talk to and delegate work to rather than trying to figure out how many bots I should make inside of Crock -Bot. And then finally, just a little bit of disappointment here that Dots feels so rushed. I mean, I was really excited when I heard that OpenAI was working on their own sort of always -on agents.
I thought it was going to be like Crock -Bot. I thought that it was going to be really, really crisp and slick, but it just feels very buggy.
I actually tweeted this last night and I said, dots should have been called bugs because I was having so many issues. I was going to make a video yesterday, but I was just getting so frustrated because nothing was working. And I was like, I couldn't get it on my phone and my desktop app wasn't working.
And I was just a little frustrated. It felt very buggy and it still does, but it does feel better today. But it's just interesting how limited dots feels at the moment, especially when you had, uh, very clear example like Muse to pull from.
So anyways, like I talked about earlier, I think it's really about understanding what are the things that matter most to you? Are you already kind of ingrained in the ChatGPT Codex ecosystem or are you more ingrained in like Instagram and Facebook and that kind of stuff might be more useful for you right now? Would you rather be able to talk to this thing on WhatsApp or would you rather be able to have this in Slack with your team?
I mean, I think the ability for my team to in Slack be able to talk to my, essentially my Codex, I think that that's a pretty cool like functionality or for me to be able to assign my codex things from slack and have it be sent right back into slack i think that is super cool and i'd like to see that come to other apps as well but that is going to do it for today so if you guys enjoyed you learned something new please give it a like it helps me out a ton and as always i appreciate you guys making it to the end of the video i'll see you on the next one thanks everyone
The Hook
The bait, then the rug-pull.
Two persistent AI agents launched within weeks of each other. Meta's Muse rocketed to the top of the App Store, and OpenAI answered with Dots, which visibly broke on stage during its own announcement. Nate Herk spends a day living inside both to find out which one is actually worth the desk space.
Frameworks
Named ideas worth stealing.
14:44list
The four-point verdict
Muse's interface advantage is meaningful
Astra is the better model
GrokBot has a different organizing model and may scale better
Dots feels like a rushed catch-up play
A closing summary slide ranking where each competing tool wins, used to land the review's verdict in one glance.
Steal forany head-to-head product teardown that needs to end on a scannable verdict instead of a paragraph
CTA Breakdown
How they asked for the click.
VERBAL ASK
04:00product
“Yours Free: An SOP for Getting Your First AI Automation Client”
Mid-roll pivot away from the app comparison into a dedicated pitch slide for a free downloadable SOP, funneling into the AI Automation Society community, before returning to the Dots walkthrough.
Add 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.
Add to Preferred SourcesOpens your Google source preferences with us pre-loaded. Tick the box and you're done.
A hands-on tour of a synced, phone-controllable AI agent team — agent computers, teachable skills, scheduled routines, event triggers, and where it stops making sense versus Claude Code or Codex.
A creator turns Dario Amodei's three filters for a solo billion-dollar company into a real product: AI software that quality-checks other companies' AI agents.
Anthropic's agent-skills team explains why they stopped building a new agent for every job, and the four habits that make one general-purpose agent actually reliable.
One creator ran two frontier AI agents through the same 15 real work tasks and tracked the winner, the time, and the exact dollar cost for every single one.
A two-hour gap after Claude Fable 5.1 shipped, a site-wide AI outage, and a blog post that got pulled mid-cycle — the strange week OpenAI introduced GPT-6 Astra.
00:13
00:30
00:43
01:08
01:20
01:32
01:45
01:57
02:09
02:22
02:39
02:47
02:59
03:11
03:36
03:48
04:00
04:13
04:26
04:32
04:50
05:03
05:15
05:40
05:52
06:05
06:17
06:29
06:42
06:54
07:06
07:19
07:31
07:56
08:08
08:21
08:33
08:45
08:58
09:10
09:23
09:35
09:47
10:12
10:24
10:37
10:49
11:02
11:11
11:31
11:39
11:51
12:03
12:16
12:41
12:53
13:05
13:18
13:30
13:42
13:55
14:07
14:20
14:32
14:57
15:09
15:21
15:34
15:46
15:59
16:11
20:32
11:40
09:47
Video of the Day38:35
Video of the Day14:45
05:42