Category: Alex Finn

Anthropic just released their Grockbot in Chad GBT dots killer. It’s called Claw Duet and it might just be the best personal AI agent out there. In this video, I’m covering everything Claw Duet. What it is, how it works, how to get set up, how to get access to it, because not everyone get access to it, and which AI agents you should be using. This one is an exclusive banger. Let’s lock in and get into it. So this is Claude Duet and the reason why it’s so small on my screen here is currently it is mobile only. It is only on the mobile Claude app. It is in super early access. Only a few people are getting into it. Highly recommend signing up for that weight list. I put the link down below for that. But let’s get into what it is and what makes it different than the other AI agents out there. And real quick before we do that, instantly becomes the number one name for an AI agent out there. Here’s your power rankings for best AI agent names. Duet Hermes at number two. Great, great name. Muse at number three. I like Muse a lot. Rockbot at number four. Solid name. They just got a little lazy. Just put bot at the end of the Grock. That’s fine. Five. Open Claw. Claudebot was significantly significantly better. Six dots. Terrible name. Awful name. That’s the power rankings and names. But now, let’s get into the actual app and see which agent is better. So, Claw Duet, much like the other AI agents out there, uses all your plugins. So, it instantly connects to your email, your calendar, uh all the other services you use out there that you currently use with Claude. So, it knows everything about you. It has all your data and can take care of anything you need. Here’s what’s really interesting about this and why I actually think this is the best UI for an AI agent ever. You basically have two sides of what an AI agent can be. You have Grockbot, right? Grockbot is on one side, which is just a long list of bots and agents, and you choose which agent you want to talk to. They all have their own context and skills and everything, and you choose one, and you talk to it. This is like the anti-god thread, which I’ll show you in a second. It’s a long list of agents. On the other hand, you have like a dots for instance, which is more of a god thread. One thread. You put everything into the one thread. It’s a true personal agent, just one agent bot. Muse is like this. Hermes is kind of like this, but then they added like the bots modes. Now there’s like 20 different modes. Open claws like this. One god thread. You just put everything into it. There’s upsides and downsides to each upside of the kind of long list of bots like Rockbot is is you can have many different bots with many different responsibilities and the context don’t get confusing or overlap or anything like that. You have bots for very specific tasks which makes those bots very good at those specific tasks. On the other hand, you have the God thread, which it’s not quite as good at very specific tasks because it doesn’t just have the context of that one type of task; but it makes it very easy to use because you just go to one thread and you put everything in that one thread. So, it’s less of like a cognitive load because you don’t have to think about which bot to use, but it’s not quite as good at like those individual tasks. Just to give an example of like why a Grockbot with a long list of bots is like a really good setup is because I have Barry. Barry’s like my chief content officer when I can just go into Barry, give an idea, and it knows it knows I’m trying to give it ideas about content. I can say, “Hey, her newsletter.” And it just knows I’m trying to make a herder newsletter. It can help me give an outline, schedule it for me, get the right audience in my email newsletter tool, all that. You go to the right bot, it has the right context automatically, and it does really good work. God thread like dots. I can’t really do that because I do all my work in that one thread. So, I have to be a lot more clear about what I’m trying to do. So, what is Claude Duet? Is Claude Duet more of a god thread type agent or is it more of a long list of bots type agent? Well, it’s actually the best of both worlds. So duet, you have one god thread, but as you give it tasks, as you give it specific tasks to do, it actually takes those tasks and breaks it down by category for you. So I’m giving it some tasks for an app I’m building called Henry. So I brain dump a bunch of things I want to do and it figures out, okay, they want this task done for Henry. Puts a hashtag Henry. Okay, Alex wants this task done for his video game, so does #nol district. So it’s one god thread, right? this one chat right here. But when you give it tasks, it breaks it down by what it calls channels. They’re actually called channels or categories. Creates tasks for each one of them. Puts it in the right channel, the right category, and keeps it neatly organized with the right context. So, it’s actually the best of both worlds where it has the positive of the god thread, right? Where you only need to put things into one thread. You don’t need to have this cognitive load. Which bot do I talk to? let me go through my long list of 20 different bots. You put it all in one thread, but then that one god thread decides which context that task should go to and breaks it down as these tasks for you. I don’t manage this list of tasks. This la list of task is all managed by the god thread chat. So I don’t need that cognitive load of like which contact should I be using, which bot should be doing this. No, I just say hey do this for me. And it goes okay, I’ll give this over here. We’ll start doing this for Henry. start doing this for your game. We’ll start doing this for your content. And it gives it the right skills, plugins, and context. So, it’s really the best of both worlds. The simplicity of the God thread of like a dots, but also the power of the separated out contexts like a Grockbot. What I really like about, and there’s many other things I really like that we’ll go through, and I’ll go through in a second which agent you should be using for which task. I really like this plus button down here. So if you click this plus button on the task screen, it’s basically like a brain dump mode. So if I go here, click on the keyboard, I can just brain dump everything that’s on my mind and it will take all of that, break it into individual tasks and just start work for me. It really seems that Claw Duet more than any other AI agent out there was built to lighten your cognitive load as much as humanly possible. You just have ideas in your head. It takes those ideas, breaks it down to actionable tasks, and does those tasks really well. So, let me just show you what an idea of like a brain dump would be right here that it breaks down into tasks for me. So, I just put in everything that’s on my mind into this task creator. I said, I want to write a newsletter on Herder. I also want to start planning out multiplayer in null district and how that would work. I also just got a second DGX Spark. I want to see what the best model are for that cluster. And I’m just going to hit send. And what now Duet’s going to do is it’s going to break those down into tasks. I either save it for later. I got to say start, start start. And then now let’s look at my task list in my active. And you can see all the tasks that are waiting for my review. And if I scroll down, you can see which ones uh need my input and which ones are doing work. And now it’s doing the multitasking of a Grockbot, but it has the simplicity of a chat GPT dots god thread, right? with like a Chad GBT dots or like a Hermes if you’re just using a god thread, you can’t just put a bunch of things in because it’s not able to multitask quite as well. You can in dots if you say like, “Hey, make a thread for this, make a thread for this, make a thread for this, make a thread for this, but you’re kind of manually managing all these threads.” With Claw Duet, again, you don’t have to think about that. You don’t have to think about the multitasking. You just say whatever is on your mind and it breaks it down into these tasks in the task list and starts going. My favorite invention of all time is full self-driving on my Cyber Truck. The reason why it’s my favorite invention of all time is I can go in my car, plug in my destination, and I can just sit there and my car drives me an hour away to wherever it is I’m going. And my entire cognitive loads lightened up because I don’t have to think about driving. I don’t have to be engaged in the driving. So I can use my brain to focus on other things, to focus on other problems to solve. It makes driving so much nicer. Claude duet is kind of like the full self-driving of AI agents where your cognitive load’s lighting up so much because you don’t have to think about the multitask. You don’t have to think about which bot to go to. You just say what’s on your mind. It breaks in task and figures out how to do it and it just does it. So I can lighten my load. I was at the gym earlier today and I had some ideas of what I wanted to do that day and I just sent it and it just broke up a task and started going. But this goes further than just breaking up into tasks. This isn’t just a good bot at making tasks. It also is incredibly good at being proactive with ideas. Here’s what I mean by that. So, I’m at the gym today. I really want to work on Henry Move it forward. Here’s my average conversation with Claw Duet. I’ll show you how simple this is to get a lot of work done. Claude Duet is constantly making recommendations on what you should build next without me even asking. It just says, “Hey, by the way, I think we should work on these five things next. Do you give me the go-ahad?” Right? So, I wake up today, I say, “Hey, let’s work on Henry.” He goes, “Okay, here’s what we should build out next.” It gives me really good recommendations of what to build. A lot of times, I’ll go to AI agents, I’ll say, “Hey, what should we work on next?” And it gives me like total slop. These are actually really good recommendations on what I should be implementing next. Then you can’t see it here, but it has like buttons it automatically creates on with each one of these features in them. I just hit the button and it starts working on it. One of the buttons I hit was start the pipeline fixes. I click the button. Boom. It goes, it starts working. It creates the PR. Then it comes back a couple minutes later. It says, “Should I open the PR for you to review?” Then has a yes button or a no button. I hit the yes button. By the way, it’s using Claude code for all this. Right now, Claude Code with Opus 55 is the best coder out there, hands down. No question about it. So, instantly, Claude Duet becomes the best coding personal agent out there. So, I say yes, open the PR. It creates the PR. Boom. It’s good to go. It’s done. I Okay, what’s next? It gives me five more really, really good ideas on what to work on. Gives me the buttons on which one I should do. I click the button which I want. Creating the signin. Boom. It’s good to go. Makes the PR says, “Okay, you ready to go? Want me to merge it?” I say yes. What’s next? Gives me more ideas. I click okay. Go. This is like the way you work with Claw Duet, which I love. Which is like, hear me out. Now, this is going to sound weird, but hear me out. I believe the future of working with AI is you sit there and you have a button and you just sit there and you press the button over and over and over again like this. And as you’re hitting the button, more things get done. You get more productive, you create more value, you make more money. I think that is the future of AI. Not having to have this huge cognitive load on what to do and get work done. You sit there, your agent knows you really well. It’s able to make incredible recommendations. You just sit there, do it, do it, do it, do it, do it, do it, do it. And boom, at the end of the day, your business is moving forward and you’re making more money. Duet is the closest I’ve seen to that vision of where I think AI agents are going out of any agent I’ve worked with. I’ve literally spent the entire morning just saying, “Yes, do it. Yes, do it. Yes, do it. Yes, do it. And my app Henry has moved forward so much, which is incredible. So, it is super proactive, but it’s proactive in this different way where it’s not like like for instance, dots. Dots is very proactive, but it’s proactive in this way where it’s like I get an email, it reads the email automatically and then recommends like a response that email. That’s great. I like that. But Duet’s proactive in like a different kind of way where it’s finding unknown unknowns, uncovering those unknown unknowns and recommending them to you to get done. I didn’t think of any of these features to build out. It’s like, hey, we should build these features next. So, I like this type of proactivity a ton. It’s really, really cool. I’ve gotten a lot more work done with Duet uh over the last day because of this proactivity. Speaking of unknown unknowns, if you look down here, there’s also a make section. If we click on the make section, it has a bunch of templates for recommended tasks we can do next. So again, this is where Duet really focused on which was finding the unknown unknowns, finding things to do that you haven’t thought of, lightening your cognitive load, and all you need to do to get work done is clicking a single button. It wants to make an ads plan for my Vibe Coding Academy, by the way. Vibe Code Academy, number one community AI. Link for that down below. It wants to make a LinkedIn ads plan. All right, sounds good to me. Let’s click that. Let’s get it popping. Boom. It already made it for me. I actually I actually think it proactively made it. That’s incredible. Uh it didn’t even generate when I clicked it. It automatically opened. That’s uh pretty amazing. So that’s done already. It proactively made all of these documents for me, which is incredible. So again, it’s just thinking of things to do and proactively doing them. But the important part here is those things it’s thinking of doing are actually great. They’re actually helpful. They’re not just AI slop. There’s a few other things here. I’m going to go through those few other things and I’m going to go through weaknesses where it’s behind all of the other agents. Then I’m going to give you a breakdown of power rankings of which AI agents you should be using. But a couple other things to note, if you click the hamburger menu icon up here, you have what’s called channels. So you do have a way if you want of breaking this down into kind of dedicated side chats. But even when you do that, you don’t really have to worry about talking to the right one because it will automatically assign tasks to the right channel if you just go to your God thread and give it a bunch of things to do. It’ll automatically assign to the right channel. So the channels are more for just letting your agent know what are the different areas of your life. So super proactive, lightens your cognitive load more than any other AI agent. But is this the best one ever? Well, there are still some weaknesses to Claw Duet. Right now, it’s just on mobile. There’s no desktop component. There’s no desktop app. I assume this is cuz it’s in super super early access and I assume a desktop app will come soon. So, that kind of stinks. I I I I wish I could kind of lock in on my computer and talk to my agent. It is what it is. Two, and this might be the biggest weakness. This might be a reason why I still use Grockbot a ton at the moment, is that it doesn’t really have a computer of its own. It can do claude code cloud agents where it spins up cloud agents to write code for you but it doesn’t have its own box. It doesn’t have its own computer. So to give you an example of why that is a weakness, I started a new GitHub organization for Henry. So it’s separate from my normal GitHub account. So to commit code for Henry, my agents now need to connect themselves to a new GitHub account. For Grotbot, that’s super easy. I say,”Hey, go and connect yourself to the new GitHub account.”It goes on its own computer, which is already logged into my GitHub and then just switches organizations and adds itself by itself to that organization. Because it has its own computer, it can do anything I can do. Claude Duet on the other hand though doesn’t have its own computer. It just has access to all your plugins as well as Claude Code cloud agents. So, it can’t open its computer, connect to a different service or anything like that. So, I had to personally and manually go and log into my GitHub and choose the new account and connect it manually and like that was a little bit extra work which kind of stunk. Grockbot, it doesn’t matter what you need to do. You just say no Grock, you do it and it will just figure out how to do it. Like I have my fantasy football lineup I need to set this weekend. I was on the go. There was an injury. I had to switch out the injury. Saquon Barkley is now back in the lineups. I needed to switch him in. Grockbot, I said,”Hey, go. Can you put Squan back in my lineup? It opened up its computer. It’s logged into my fantasy football account. It switched the lineup. Claude can’t do that. Doesn’t have its own computer. There’s no Yahoo Fantasy Sports plugin. So, it’s just something I had to do manually. So, not having a cloud computer is a weakness for some specific types of tasks. I’m actually not 100% confident they’re going to give it its own computer. Out of all the companies out there, Anthropic is definitely the most cognizant of safety, cognizant of making sure their agents don’t do anything crazy that can like hurt your security or privacy or anything like that. So, I think giving these agents their own computer opens up a lot of security concerns. And so, I think Anthropic’s probably going to be super careful with that. And I don’t know if that you get your own computer. We’ll see. I hope they do. I hope they to the side of like, hey, let the user do whatever the hell they want. But I get it. You don’t want like people who don’t know what they’re doing going on these cloud computers and just putting all their credentials in and maybe doing something not so secure. But if you’re someone who relies on your agent do absolutely everything, it’s a little bit of a weakness the fact that it doesn’t have its own cloud computer. So which agent should you be using? Here’s where I’m at currently and what I’m doing for pushing my coding projects forward. I like Duet. I think the access to Claude code and Opus 55 is unbeatable. It’s incredible. The proactive thinking, the figuring out what to do next, gives Duet a very distinct edge for knowledge work: for going online, creating documents, doing different things, opening things up in a browser, submitting the form here, then downloading that form, opening it up on your third computer, taking it off the third computer, moving it over here. I like Grockbot. Grockbot A has its own cloud computer. Beacon can control any of your computers. That’s another thing Duet can’t do at the moment is actually control your computer, download files, things like that. Grockbot can do all that. And because of that freedom and that flexibility of being able to do things on a cloud computer, move it to your on-site computer, things like that. I lean Grockbot for those tasks. Personal agent, I’m trying to figure out how to get an iPhone Duo tomorrow. It comes out tomorrow. My pre-order is for like the first week in November. I want it tomorrow on launch day. I’m having Muse go out and figure out how to do that. Muse is super good at going on the internet, like figuring out how to buy things, reserve things, things like that. kind of personal agent. I like Muse for that. And then open source. You got to have your open source agent, right? So, I have a second DGX Spark I just got. I’m I’m running like local models on there. Uh Hermes is powered by those local models because it’s open source. You can do whatever the hell you want. So, I’m using Hermes for more of my open-source free-for-all, doing experimental things. Listen, all of these agents are spectacular. All of these agents are so good. I’m just like an AI crazy person, so I use all of the agents. But if you only had to choose one, you can get away with using any of these to do all your tasks, right? You can get away with it. So like, you don’t need to be like me where you use every AI agent, you can choose just Muse and do everything or just duet and do everything. That’s totally fine. But if you’re trying to hyper optimize like I am, this is what I’m using. If you learned anything at all, leave a like down below, subscribe, turn notifications. Which agent are you using? Which agent would you like me to do more deep dives on? Let me know in the chat down below. I also got the number one newsletter on planet Earth. Ship it weekly, completely free. Link for that down below. Hope this was helpful. Turn on notifications. We got some banger videos coming up. Turn on that notification. Some amazing videos coming. Hope this is helpful. We’ll see you in the next video.

Hermes Agent has released some monster updates lately that have completely changed how you use the AI agent. If you use these new, sometimes hidden features the right way, you’re going to be a productivity machine. In this video, I cover all the new features and show you my brand new Hermes agent setup. I also show you what I’m using. Am I using Grockbot, Hermes, Chat GBT, and which tasks I use them for. If you stick with me until the end of the video, you’ll be a master of the best open-source AI agent out there. Now, let’s lock in and get into it. So, one of the biggest updates from Hermes lately is the bot mode. Here is how you need to be using the bot mode. For those who don’t know, Grockbot comes out a couple months ago, takes the world by storm. Hermes quickly comes out with their bot mode. What is the bot mode? Bot mode allows you to create multiple Hermes bots very very easily and manage them and talk to them all at the same time here. Sort of like iMessage. Previously, Hermes was very much a single agent setup where you just talk to one agent. It’s just kind of the god mode agent that does everything for you. Now with Hermes bot mode, you have multiple bots, different profiles, different memories, personalities, tools, all of that. But here’s the mistake I see a lot of people use with this new feature. They treat it like Grock where they make a whole bunch of bots that all are like different human roles, the developer role, the marketer role. I don’t think that’s the right way to do this. I think that’s fine for Grockbot. For for Hermes, you want to make profiles and agents that kind of match the strength of what Hermes is. And the strength of what Hermes is is you can use any model, any provider you want with each one of these agents. So, I’m not making agents that have different roles. I’m making agents with Hermes, which is GPT6 Ash, Claudy, which is Opus 5. I’m making agents for all the different models so I can take advantage of the strengths of each one of those models. If I need to build something, I go to Hermes, my GPT6 Astra model. If I need some front-end work done, I go to Claud. If I have simple tasks I want done, I go to Quen, which is actually powered by a local model running on my DJX Spark. So, it gets it done completely for free. So, my recommendation to you is this. Go to your Hermes agent, say, “Hey, set me up a new bot powered by Claude or whatever other models you want to use.” Hit enter. It will set up the new bots for you in the bot mode. By the way, this is only in Hermes desktop. And all you have to do is click the bot mode here, and it will create the new bots for you. Do that, then you can use the strengths of each model for whatever task you’re doing. If you want to do cheap coding, make sure you make a Grock bot in there. If you want to do super fast coding, I actually make a Meta bot in there. Meta’s coding models are actually good now. So, make sure you set those up. We’re actually going to be covering 10 awesome new features here. So, let’s get to the second feature, which is their advanced local model support. Now, stay with me here. It doesn’t matter which computer you have, this local model support is going to be relevant for you. So, I really, really like this. If you go to settings and then you go to providers and then you click on run models locally, click that. What’s going to happen is it’s actually going to look at your hardware and then determine the best local model you should be running on it. Doesn’t matter what you have. If you have a Mac Mini 16 GB, you have a super powerful Mac Studio like me, does not matter. It will figure out the best model for your device and then basically manage it for you. All you need to do is click set up for me. It’ll almost be like LM Studio where it’ll download it for you, load it, test it, run it, and it’ll put it into a Hermes bot for you. So, you will have your own Hermes bot powered by a local model with unlimited intelligence. Now, I’m going to be honest. If you have a bad computer, if you have a smaller computer like a Mac Mini, will it be the best intelligence in the world? No. But you’ll still get free intelligence. So, even if you have a smaller computer, you can use it for really simple tasks to let it run and do it for free. So, everyone should be doing this now. Go to providers, local models, and see which local model it wants to load for you, and then load up. Just click set up for me. You have nothing to lose by doing this. Hermes agent is insanely powerful, but it’s just one piece of the agent puzzle. Luckily, my friends over at HubSpot put together a free AI agent cheat sheet that helps you figure out which AI agents should be used for which jobs without having you need to test a bunch of tools yourself. Inside the guide, it covers seven of the biggest AI agents right now. Chad GBT agent mode manis claude coowork openclaw zapier nadn clawed code for each one it breaks down what it’s best at how hard it is to set up what it costs and when you should actually use it you also get copy paste starter prompts these prompts cover workflows like competitor research lead routing and daily briefings so you can go from comparing tools to actually using them right away is the skip if you need section which basically tells you hey these agents you don’t need to use them if you do these There’s so many agents out there, it’s good that they tell you which ones you don’t need to use. If you enjoy Hermes, you want to know what other agents you should be using, too. This is the resource for you. Grab the free cheat sheet in the link down below, and thanks to HubSpot for sponsoring the video. The next big update I’ve been using on is Hermes HUD. This is a great new feature, especially if you play video games. This is like a gamer’s dream. I recently built like an open- source version of this for all my other agents, but it’s great that Hermes built this in. So basically it allows you to make this like kind of mini version of Hermes. This little HUD, you can move it anywhere on your computer you want. It always sits on top of whatever is on your screen. So if you open up a game, you open up War Dogs or World of Warcraft, whatever you’re playing, this will sit on top of it. So anytime you can go in and you can talk to it. You can say, “Hey, what’s the stock price of MU Micron?” But here’s the cool part. It has access to whatever is on your screen. So if you’re playing a game, you’re playing World of Warcraft and you’re on a quest and you’re stuck, you don’t know where to go, maybe it has you killing like a specific like highlevel elite bore, you can go in on Hermes Hut and say, “Hey, where is this boar located?” And it’ll show you, it’ll tell you exactly where to go cuz it can see your screen. So it’s like the best like cheat code for video games ever. If you had a Game Shark when you were a kid like me, I know my millennial is showing right now. This is basically like that or just an easy way to access your agent. No matter what’s going on your screen, you don’t have to keep minimizing and opening and minimizing. It’s just always a HUD that’s there working for you. So, I love it. If you’re on Mac, all you need to do is do commandshiftth. If you’re on Windows, it’s control shift H. Puts it in HUD mode and you’re good to go. You can move this around anywhere on your screen you want and you always have Hermes agent ready to go. It’s amazing. Which brings us to feature four, which is it now has chat GBT voice mode. If you’ve watched my previous videos, you know I am absolutely in love with chat GBT voice. I think I think it’s such a powerful tool. Well, guess what? You now can hit this button right here. Open it. Hey, can you tell me what the status of my other agents are? So, it has the strengths of Chad GBT voice, which is it’s real time, interacts with you very, very quickly, but it also has access to all its tools. So, it can look what your other agents are doing. It can orchestrate your other agents. It can delegate tasks to different agents. It’s basically chat GBT voice but in this nice multi- aent setup. All you need to do is put in your OpenAI API key and it’s good to go. I think the costs are like really small. It’s like 5 cents a minute or something like that. So, it is really, really cheap and it’s really awesome. If you haven’t used Chad GPT voice yet, it is such a sick way to use AI. It is clearly the the way of the future, the way we’ll be interacting computers in the future, which is through our voice. I recommend when you’re doing work next time and you’re thinking about typing your Hermes agent, just use voice instead. Sit there, talk to your microphone and just delegate tasks out. Just stream a conscious as you come with tasks. Give it to your Hermes agent and now it’ll be amazing at delegating it to your different bots you set up, which is much related to the next feature I really love that they just released, which is the ability to use the hey Hermes keyword to talk to your agent. So, as you can see here, I just said it out loud. Hey Hermes, can you tell Claud to build a 3D firstperson shooter using 3JS? And as you can see, it gave the command to the other agent to build the 3D firstperson shooter. So now basically anytime you’re sitting at your computer and you think, “Oh, I wish I could do this. I wish I could do that.” You just say the Hermes keyword and it starts listening to whatever you have to say. This is awesome. You’re multitasking. You’re sitting at your computer. You think of things. You say the keyword and you just start shouting out tasks to do. It’s been really helpful for me. I’ve been able to get way more done. I’m telling you, voice makes you so much more productive. The next feature I’m really enjoying is its sub agent functionality. It now has really nice user interfaces for sub agents. So, you can track what each sub agent is doing. A lot of the AI agents out there and the tools, they don’t give you visibility into all the agents. I really like this. So, check this use case out. Please use sub agents to design three concept poster images for dystopian shooting video game called Null District. It’s a stylistic cyberpunk dystopian looter shooter. What’s going to happen when you give a task that has multiple threads and strands to do? It’s going to spin up a bunch of sub aents, but now it’s going to give you a really nice interface for them. As you can see here, here are the three different sub aents. You can see which models they’re using. You can click in and you get an entire window with every single thing the agent is doing, which is really cool. If you want to change each sub agent, which model they’re running. So maybe you want to do like Claude for one sub sub aent to see what kind of design you get, Muse for another, you can go in here and do that. You can change which skills each sub agent uses, things like that. That’s like the strength of Hermes, right? Is it’s so configurable. You can do anything you want, any model you want, any provider you want. So, I highly recommend doing that when it uses sub agents. Have it use different sub aents. Have it use Opus, have it use Chad GBT, have it use Grock, and see how the different sub aents perform. Really good if you want to do something like design related where you can see different options after it’s done. The next feature that’s great is the cron jobs are greatly improved. They now have memory of past cron jobs. So, if you’re anything like me and you have your agents do regular research, daily research, weekly research, whatever, it will now remember its past research and not repeat its research again. So, so as an example, I have it research AI stocks every day. I like to invest. I like to invest in AI companies. Before it would go and it basically research a lot of the same companies over and over and over again cuz it had no memory of its past crown jobs and what it did during those crown jobs. Now the memories of those crown jobs have improved. So now you can see here here’s one of the research reports it did recently where it went through found a bunch of stocks Micron Micron AVGO. But if we look at the next report re it researches an entirely different set of companies. As you can see it even says previously researched. It shows it right there, which is amazing. So now your cron jobs are a lot higher quality. I’d recommend doing all your research cron jobs inside Hermes cuz now it has that memory, which is really, really nice. So I think the number one question I get on all my videos now is like, Hermes, you’re not talking about Hermes. I still use Hermes a ton. Hermes is still a fantastic tool. It is the only usable open-source AI agent, I think, right now. It is excellent. So, you want to use it for the strengths that you get out of an open-source AI agent, which is one, you can use any number of models, and two, you can customize it any way you want. So, if I’m doing network administrator stuff, which is loading local models, my different computers, managing my tech, managing all my different computers, I like to use Hermes for that because Hermes is a lot more willing to jump between my computers. There’s a lot of guard rails on the other AI agents when it comes to that. Also for personal work, for the crown jobs, for doing things that might require or be better with multiple models, I like Hermes. For the deep technical work, I like Chad GBT work only because that harness, that desktop harness is excellent. The computer use in that is just the best, bar none, the absolute best. So, I like Chad GBT work for that. In more general knowledge work, I like Grockbot cuz I think the Grockbot interface is excellent. I think it’s super easy to use and also it has the mobile app. Hermes still doesn’t have a mobile app, which is that’s a big weakness. Grockbot has the mobile app. So, for quick general knowledge work, I like to go to my Grobot app on my phone, shoot off some commands real quick, and I get a lot of work done. That’s my stack. I’m still using Hermes. It still manages all my computers for me. I also love the iMessage uh integration, so I can go to iMessage and quickly shoot off prompts as well. That’s what it’s great for, too. But, you want to mix and match. You want to use the different agents and use it for each of their strengths. I hope this was helpful. If it was, leave a like down below. Let me know in the comments which AI agent you use the most. Hermes, Crockbot, Work, Open Claw, whatever you’re using the most. Let me know down below. I’m curious. I’ll start making more videos about whichever agents you guys like. Subscribe and turn on notifications. All I do is cover all the AI agents, how to use them, how to get the most out of them. I hope this is helpful. I’m so appreciative if you’d watch this video and I’ll see you in the next one.

Your Grockbot just got a hundred times smarter. Five major updates to Grockbot just dropped. I’m about to go through them all. How to take advantage of them, how to be way more productive. I will then show you which AI agents you should be using for which use cases. There’s like a 100 AI agents out there right now. Is Grockbot the best? I’ll show you which ones you should be using. Then I’m going to give you two use cases you can steal from this video that will instantly make you way more productive with Grockbot. This is going to be an absolute banger video. Let’s lock in and get straight into the first update. Here it is. Your Grockbot now has its own email address. So, here’s what you need to. This is rolling out slowly. By the way, some people reported they have access to it. Some people don’t. I just checked. I have access to it. You can now ask your Grockbot to get its own email address. So, instead of having to give Grockbot access to your own email or anything like that, it can have its own email address. You can forward emails to it. It can read those emails. It can send emails. And most importantly, it can sign up for its own accounts with its own email address. All you need to do to set this up is go to your Grobot, go to your mainbot. We’ll talk about main bots in a second, and then say, “Hey, claim me an email address.” It’ll check for you if it can set it up. Again, some people have access, some people don’t. I think it’s a slow roll out. And then you can just say, “Hey, uh, can you make it, alexgrockbot.com?” Alex is free. I got alexrockbot.com. I feel like that’s a major win, by the way. Anyway, and it claims it for you. It’s good to go. Now, my Grockbot has its own email address. Please do not email my Grockbot. Again, this gives you several major advantages. Now, if you get an email from somebody and you want your Grockbot to take a look at it, forward it to your Grockbot’s email address. If you’re having a conversation with someone and you want your Grockbot to schedule you a meeting, add your Grockbot as like a CC on the email and say, “Hey, by the way, Grockbot, can you schedule the meeting here?” But as I said earlier, I think maybe the most powerful use case here is you can say, “Hey, Grockbot, I need you to sign up for an account here. I need you to download this app over here. Please use your own email address to sign up for it.” Now, instead of having to get you to put in your password and sign in with your account or whatever it is, it can just sign up on its own, create its own account and keep your workflows moving much quicker. Quick, easy win to Grockbot. Just go and say, “Hey, can you claim me an email address?” and then get that done. Big easy win. Love it. Very cool feature from Grockbot. The second major update they just dropped is called MainBot. What is MainBot? Well, if you haven’t been paying attention lately, MetaMuse has been taking over the space. It’s what everyone’s talking about. It’s been getting a lot of attention and for good reason. It is one bot that does basically everything. You have one Godthread chat that you go to and it goes and can handle your entire life for you. I think the SpaceX AI team saw that Metamuse is getting a lot of attention like, “Okay, I think there’s some things we can do here to capture that audience that Metamuse has gotten.” And what they decided to do was give you the ability to have a main bot. As you can see, my main bot here is Slate’s got that star next to it showing you it’s the main bot. And now Slate has much more visibility into what other agents are doing by default. So by default, it knows what my long list of agents here, what they’re all doing. And Slate knows it’s your main bot. So it’s going to become a lot more proactive for you. That’s really the strength of Metamus is because one bot has access to everything and has kind of the god thread. It can be very proactive across your entire life and suggest things to do. Now that Grockbot, you can choose a main bot. It can do the same thing. To give you an example, here’s a proactive message I got this morning. It went through, checked my calendar, checked my email, said, “Hey, you need to lock in the date for the interview. Hey, you got a doctor’s appointment today. Hey, you got to respond to California because they want $10 trillion a month for health insurance. Do you want me to handle any of this for you? Right? It was super proactive. It looked at my entire life, everything I got plugged in, said, “Hey, I can do all these five things for you.” So now Grockbot is much more proactive. You need to set your main bot. If you’re logged in, you update for the first time, it’s going to have a popup that asks you to choose who your main bot is. Otherwise, you can go in and just message your bot, say, “Hey, I want to make you the main bot in Grockbot.” and it’ll do it for you. And then what you’ll notice is one, again, it’ll be way more proactive, message you, ping you, hey, this broke, hey, you got to do this. Hey, you got this coming up. Do you want me to do this? And two, it’ll have better visibility and delegation powers. So, you can just go to your main bot for everything. And it will delegate tasks very, very easily and nicely to your other bots. Mainbot, big update for someone who I personally I think I lean more towards the god thread model. Just going to one thread, handles everything for you. This is a major win. I really like this. This next update might be honestly the biggest one of them all. Elon Musk, the trillionaire himself, announced this morning that Grockbot moving forward is not just going to use Grock models. It’s going to also use Claude Opus 55, Midjourney, Sunno, and any of the other leading AI models out there depending on the task you’re giving it. So, if you give it a super advanced coding task, it’ll lean on Claude Opus 55 since right now it’s the best coding model. If you give it a super advanced kind of photo editing task, it’s going to use MidJourney. This is massive. It’s huge. And it shows you the position of strength that SpaceX is moving from. One, it just makes Grockbot much better, right? Grock 47, great model, quick model, cheap model, but it’s not the best model. And so that was kind of holding Grockbot back a bit is that I just didn’t have the best intelligence. It can handle most of my tasks for me, but if I had a super complex task, I’m going to Claude, right? I’m going to Claude Opus 55. I had to use other tools. Now SpaceX AI is saying, “No, we’re going to put it all in Grockbot. You just do everything in Grockbot.” Now, you might be going, “Hey, that’s strange. Isn’t like Claude a competitor to Grockbot? How can they do that? Why are they how would they implement their competitors? That doesn’t make much sense.” Well, no, no, no, no. Elon, there’s a reason why he’s a trillionaire. Anthropic signed a multi-billion dollar deal with SpaceX AI earlier this year to use their massive Colossus data center. So, when you go to Claude right now and you send it prompts, there’s a very good chance those prompts are being sent to SpaceX AI data centers and that’s where the Claude models are running. So yes, when it uses clawed model, that money is going to anthropic, but at the same time, that money is then going from anthropic back to SpaceX AI because all of their stuff is running on SpaceX AI infrastructure. That is why SpaceX AI is like an unbelievable company and Elon Musk is an unbelievable entrepreneur because he is really vertically integrated in the entire AI space. He has the models. The Grock models are very strong. He has the applications. Grockbot is probably top two uh agent harnesses out there at the moment. And then on top of they also have the infrastructure. So they have the data centers, they have the GPUs, they have the chips. They’re the only completely vertically integrated AI company. Open AAI is trying to build data centers right now. Anthropic is not even trying to build data centers. They’re just a model company. But SpaceX AI, they’re already there. They have the biggest data centers. They have the most GPUs. They have the most compute. So, they can make money off the entire stack. And now they’re building terra. They’re going to be creating the GPUs as well. I mean, SpaceX AI stock. I mean, it’s if you believe in the AI movement, if you believe AI is going to get only bigger and bigger. I I don’t know how you don’t invest in SpaceX AI stock. So, anyway, that’s the big update. There’s no kind of change needed for you other than you just keep using Grockbot, I guess, because uh now it’s going to be better. It’s just going to be smarter for you and you don’t have to do anything about it. It’s just going to start using smarter models for you, which is amazing. This is like it’s almost like the cursor of AI agent harnesses because it can use any model. And believe it or not, the cursor team built it. So, I guess that kind of makes sense. Next up is a more advanced cursor integration. So, this is sick. Grockbots’s become a lot better at coding now. can do a lot more interesting things. So, cursor cloud agents are now more deeply integrated. So, it has some neat interesting tricks. Now, first of all, you can track the cloud agent live as it’s working. So, I had my Grockbot build a new feature into Henry intelligent machines. Uh it sent it sent me this link to the cursor cloud agent. Now, I can see everything it’s doing, everything it’s building. I can change the model if I want to. So, if I want to switch to Opus 55 to make sure it’s building with that model, I can do that. But the really cool part is what it shows you next. Check this out. It now also records videos of all the changes it makes and sends you the recordings and screenshots. Look at this. It made a change on my landing page. It now has this video. It recorded itself. It recorded itself of the change it made. It changed some colors of buttons here. And you can see it clicking it showing the different animations. Changes it makes now get recorded on its own computer and sent to you and even uh sends you the screenshots too. I guess this is the before and then the after. Really sick. So, it’s a lot more intelligence, a lot more interactive. It’s a lot more multimedia. Really, really cool. It’s making coding now with Grockbot pretty incredible between it being able to use Opus and now being able to send you videos and screenshots of everything it’s doing automatically. You can even see in the library over here, everything it’s doing. You can check on it later. Really incredible. So, make sure you’re using the cursor agents inside of Grockpot as well. And here is another update I think is really cool. If you live on X like I do, you can now tag bot in any of your posts or replies, give it a command, right? So Christa from the SpaceX AAI team met her at a Grockbot event. Very nice. Uh it says you can now tag bot to start working on something. I said, “Hey bot, give me all the use cases of what everyone’s doing with this new feature.” Grockbot replied on it. Sending this to your Grockbot now. And then inside Grockbot, you can see it messaged me directly uh from my main bot, every use case people are doing with this new tag feature in Grockbot. It even sent me an entire markdown file with the use cases in it. Uh this is pretty wild. Look how much that that’s it’s amazing. So now when you’re on X, if you’re not using X, I think you should be. I think you should be building a social media presence. I think distribution is one of the last modes in the AI age. If you’re using X, you should be using the atbot. you see a cool tip, maybe you see a tip I put on X, you can just reply, “Hey bot, send this to myself and execute on it for me and it’ll do it for you.” So, I want to give you two use cases, two tips for Grobot. You can really use this with any AI agent at the moment uh that I’ve been doing lately that I really like. It is a morning prompt and a night prompt. In the morning, I go in and I say, “Which approvals do you need for me right now?” I kick my day off with this every single morning. This very simple prompt does something unique though. What it does is it causes your Grockbot to go look through all your emails, look through everything it knows about you, your linear, your notion, everything you have it connected and find tasks it can do for you, right? It finds things you might have forgotten about and do those tasks for you, which is really, really cool. And then you just say approve, approve, approve, approve, and it goes and it does those tasks for you. On the flip side, you have the night prompt. So, right before I go to bed every single night, I go to my Grockot or go whatever. Last night it was with Claude. I was building a game and Claude and I go, “Hey, I’m about to go to sleep. What’s something you can work on for the next eight hours that pushes us closer to our goals?” And it comes up with a whole bunch of really interesting, unique things it can build or it can do for you. Then you just hit approve and it goes. These two prompts are really powerful because they unlock unknown unknowns, right? They find things you haven’t thought of that you can do that the agent has awareness of because it has access to all your context. It can figure those things out. You say, “Hey, go do that for me.” And it does it. It’s really good at finding things and figuring out things you didn’t know your AI agent can do. So, morning prompt, night prompt. Do those two things. Morning, hey, what approvals do you need for me? At night, hey, I’m about to go to bed. What can you build for me? So, there’s a lot of different AI agents. I cover them all on this channel. I cover every single AI agent. Sometimes, I guess that pisses people off that I talk about every AI agent. Guys, reminder, this isn’t a football team, right? You don’t have to be loyal to any of these agents. You can use them all. I am constantly changing which ones I’m using based on their strengths. You don’t have to. Any of these Grockbot, DOT, Muse, HARC, whatever. You can just choose one and use them for everything. But if you want to hyper optimize like I do, here’s kind of how I’m optimizing right now. For all my knowledge work, I’m using Grockbot. I think these improvements are amazing. I think for just general day-to-day knowledge work, it is excellent for building. I am now using Grockots. I’ve been using Dots mostly for building lately. The challenge with Dots is there’s just a lot of kind of rough edges with it at the moment. It can only work on one computer. You can’t put it across all your computers and Claude Opus is the strongest model right now. Now that Claude Opus is inside Grockbot and now they have these deep cursor cloud agent integrations. I’m liking Grockbot for building. Although I will say this, if Dots drops an update tomorrow where it can go across all your computers, it’s a little bit faster and maybe they drop Astra 61, I’m going to switch back to Dots for building. And that’s that’s just the way it is. Like you don’t have to be like me, right? I’m trying to hyper optimize. I’m trying to get like the best agent for the best use cases at all times. For me, this is just my setup now. That’s what’s best at the moment. Tomorrow, Dots might come out. I’ll put out a video titled Dots is the greatest AI agent ever and I’ll tell you why you got to use it and you just got to deal with it because that’s how I roll. And then on the personal assistant side, I misspelled assistant. I should have had my agent build this for me. I’m using Muse and Hark. Hark is brand new. It just dropped yesterday. I’m liking it a lot. So, I’m playing around with that a lot. Uh Muse, it just knows everything about me. So, I’m doing some things there with it as well. I’m playing around with those two. These two are really your like personal assistant tasks. you’re like, “Book me the reservation, book me your Uber, respond to this email, kind of the more personal assistance stuff.” That’s what I’ve been using there. If you learned anything at all, make sure you leave a like down below, subscribe, turn on notifications, sign up for my newsletter, my AI newsletter. It’s number one AI newsletter I’m playing with. 40,000 subscribers. It’s completely free. I will send guides on all of this to my newsletter subscribers. Link down below. Hope that was helpful. See you in the next video.

Back in January, I told the world that the Mac Mini was going to be the king of local AI. I also said Apple was going to win the local AI race. People called me insane. They destroyed me on social media for saying that. Well, I just got the brand new Mac Mini M6 and we’re going to prove them all wrong. I’m going to attempt to push this thing to its absolute limits. I’m going to show you how to load up local AI models on a Mac Mini the absolute easiest way possible. No technical skills required. Then we’re going to see what kind of workloads this thing can handle. Can you cancel your cloud subscriptions? Can you cancel chat GPT with just a Mac Mini? We’re going to find out. Or will there be egg all over my face cuz this thing won’t be able to handle anything. Let’s put it to the test. I’m going to plug this sucker in and we’re going to see what we can do with local AI on a Mac Mini M6. This one is going to surprise you. Let’s lock in and get into it. All right, so I just plugged the Mac Mini in. Here you can see the stats. It is an M6 uh with 32 GB of RAM. Let’s get this going. Let me show you first how we’re going to load local Mazda. This is the absolute easiest way to do it. What we’re going to do is we are going to load up Hermes desktop. So, for those who don’t know, Hermes Agent is a fantastic open-source agent, runs on your computer, is your own personal assistant, your own personal AI employee, and can be used with any model out there, both cloud and local. So, I’m going to download this now. If you want to follow along with me, do it with me here. It’s completely free. Everything I’m about to show you is completely free. Easy. No technical skills required. I will move the Hermes application into applications cuz that’s something you still do on Mac. Then we’re going to load up Hermes here and we’re gonna open it up. All right, installation super easy. It’s going to download everything, get it prepared. Again, no tech needed here, and then I’m going to show you how we’re going to use Hermes to figure out which AI models we can run locally on our Mac Mini and then actually load them up and start using it. Okay, so I’m installing Hermes Desktop. That’s going to get set up. For those who don’t know, Hermes, amazing local open-source AI agent makes it really easy to load in any model you want. It’ll even load local models for you. It’ll find the best local model for you and load it super easy. We’re going to have this set up. Straight up beginners can do this. Even if you’ve never loaded local model before, you can do what I’m doing. So, follow along me. I’ll put a link down for Hermes Desktop, too. Completely free. All right. All installed. It took like 45 seconds. Let’s launch this up. And boom. We are in Hermes agent. Now, let’s load up a local model. They have super nice functionality here. They don’t there no one’s sponsoring this video. This is completely unsponsored video. Everything I’m telling you here is because I think it’s the absolute stone cold best. Let’s go into settings in the top right of Hermes agent here. Next, we’re going to go to providers and then we’re going to go run models locally. What will automatically happen is your Hermes agent will decide what is the best model to run locally on your computer. For me, it was Quen 3635B. So, if you’re like me, you have a Mac Mini that’s 32 gigs of memory. It’s not a terrible amount. That’s not a huge amount. That’s a decent amount. We can now run Quen 3635B. This is a great model. A lot of people compare this to like Opus 45 somewhere in that area. For those who don’t remember Opus 45 back in the day, that was sick. That was mindblowing when it came out. Now, it’s running locally on your desk on a Mac Mini. So, let’s set this up. All you got to do is click set up for me. Bada bing, bada boom. It’s going to start downloading it. It’s about 21 gigs. So, you need 21 gigs on your hard drive uh to download it. Not too much. And in 4 minutes, we are going to have a Opus from last year level model running completely on a Mac Mini. No internet required, no cloud subscription, don’t have to pay for anything. It’s all going to be running locally. But we’ll see how is it going to be any good? We’re about to find out. By the way, right after we load this, I’ll go over kind of all the strengths and weaknesses of local models. So, you’re thinking, why the hell would I want to load a local model? We’ll go through that in a second. I’ll go through all the advantages why you’d want to do something like we’re doing here. For now, let’s get this loading and let’s test it out. And real quick, as the model is loading, let me just take you through what are the advantages of doing local AI on Apple. Why I’m bullish is because number one, you can use the full memory of your Apple device for AI. What does that mean? Apple devices have what’s called integrated memory. Meaning the memory you have on your computer. This Mac Mini for instance, 32 gigabytes of memory that can be used for all graphical processing as well. That means all 32 GB can be used to load up local models. That’s a lot because even top-of-the-line Nvidia chips, which don’t get me wrong, Nvidia incredible, worldchanging chips, tons of strengths. I have a lot of videos on DGX Sparks and all that. Definitely check that out as well. But the advantage of Apple chips is you can load bigger models because you can use all of your memory for those models. So even though this is a cheaper sub,000 computer, we can load big models onto it because all the memory in it can be used for local models. On top of that, there’s a lot of software advantages. It’s a great UX. I think Mac OS is a great UX. A lot of people use it as their daily driver to do work. So now the thing you’re using for your daily driver to do work, you also have your local AI running on it as well. So if you’re doing creative work, now you can generate images on the same device you’re using to do your work. And so these are the big advantages of going with an Apple device. And again, Nvidia computers, they have their own advantages. I have a 5090. I’m able to run local AI models in there lightning fast like crazy. It will be a little bit less performant on Apple chips. Although the new M5 Ultra, by the way, I will be showing off the M5 Ultra, the 512 GB Mac Studio in just a couple weeks. Now, make sure to subscribe and turn on notifications immediately for that cuz when that video drops, that’s going to be an absolute stone cold banger. That’s going to be worldchanging technology. Turn on notifications for that. But at the moment, Apple’s still catching up a little bit on the speed side. We’ll see what the M5 Ultra brings us in a little bit, though. So, these are the advantage of going with an Apple computer. Also, not listed here is just straight up price, right? Mac Mini. You can get it for like $800, $900 and run a local model. That’s pretty mind-blowing. All right, let’s get back to Hermes. All right, it looks like it must have been loaded cuz we’re back at the Hermes screen. Let’s click here to choose a model. I scroll down. Local Quen 3635B. Let’s click on that. All right, so we got it loaded. The model’s right there. For those who are brand new to the channel, this is Hermes Asian. It’s your own AI employee. You can ask it, do whatever you want. Has control over your computer, can build things, can talk to you, can answer questions. It is now powered by our local model. So, first of all, to answer the first question, can you load local models onto a Mac Mini? Well, the answer is yes. We now have Quen 36 in here. But let’s see how it perform is. If it stinks and is slow and sucks, then well, it’s going to be kind of pointless. Let’s see how good this is. Hey, are you there? And I’m going to hit enter. Uh, let’s see what happens. Let’s see what type of performance. So for those who don’t know also by the way the first time you send a prompt to a local model it’s going to take a little bit longer. They have to load it into memory. So just as you saw there load into memory now moving forward it’s going to be much faster. Also by the way if you’re following along this is your first time loading up local models and you just did it. Congratulations. You should give yourself a pat on the back. You should feel good about that. You just loaded the most important technology in the history of our species onto your own computer and now it’s running on your desk. That is pretty amazing. Hey, I’m here. What can I do for you? Since this is our first conversation, why don’t have a quick profile so I can be more useful? I’m Alex Finn. I’m a content creator and entrepreneur. I create content about AI. say hi to my audience. And in just in a second, we’ll do a kind of a a performance check here. See how fast this is. But look, look at this. You can see how it’s thinking. So, I see all the thinking here. Let me save that to memory. This is pretty fast. I’ll be honest. I’m doing this live with you. I did not do any sort of pre-ests or anything. Like, literally, this is my first time doing this with this Mac Mini. Hey, Alex. Here’s a message you can share. Hey everyone, I’m Alex Finn. Well, I’m Alex Finn, a content creator, entrepreneur diving into the world of AI. Stay tuned. Feel free to tweet it however you like. Where do you usually post your content? YouTubex, uh, all of those places. I I did not edit this. What you just saw was completely real time. It’s responding very, very quickly. I mean, that feels like it’s as fast as cloud AI. It’s creating memories for itself. So, it’s remembering things. This is really good intelligence. What we’re seeing here, like a year or two ago, would have been mindblowing frontier. Now, I’m running it locally on a Mac Mini. This the one we’re demoing on right now is on my desk over here. By the way, I have like 20 Mac minis, by the way. I’m like a Mac Mini freak. Anyway, so let’s do this. Let’s test. Let’s see how good this can be. We tested the speed. We verified the speed is pretty dang good. Now, let’s do this. And like I’ll even show you like the timestamps here. See this? 12:34 is when we sent the prompt. We got the answer 12:34. So, less than a minute, it’s responding. Less than a minute. We’re getting responses here. All right, let’s do let’s have a build something interesting. Can you build me a snake game using single file HTML? All right, let’s see how fast it can actually build us apps. Is this usable for apps? We’ll start with something small like building a snake game, then we’ll get to something a little bit more advanced out, maybe like a 3D firsterson shooter. So, here we go. The user wants a snake game in a single HTML file. Let me build a complete polish one. I love how you can see the thinking. That’s another one of the advantages of local LMS. And again, advantages of local LMS right after we do these tests and everything, I’ll go through that. I’ll educate you. I’ll get your brain bigger. We’ll get your Abdullah obligata bigger. But what’s amazing about that, you can see all the thinking, right? A lot of the times when you use cloud models, the thinking is hidden. So, you don’t really know how the models work. But because you’re running the models locally on your own computer, everything happens in here, right? Nothing goes to the internet, nothing goes to the cloud, everything happens locally. So, you’re able to see the way it thinks, the way it operates. You have full control over how it works. And that’s why I believe in a lot of ways local LLMs are the future. They are the future of what most people using cuz it’s going to be private. It’s all going to happen locally. None of your private details and data go to the cloud. And that’s pretty amazing to think about that all this incredible computation is happening just right here. And as I was speaking, again, this is all real time. Uh the game is being built. Looks like we’re almost uh done here. It looks like I wrote a good amount. 78 lines of code just like that. All right, so I’m gonna open this my browser. Let’s Let’s test this out. And it’s going. It’s going. Okay, so I’m getting the the cookies. Uh the snake’s getting bigger. It works. It works. The snake title still on the screen. That’s fine. We can prompt in a second to fix Oh, come on. You’re going to put in the corner. Are you serious? Oh, look at this. Look at skills. That’s skills, baby. No one watching this can do that. I’m true born gamer right here. The fact that you get that corner cookie. All right, let’s see what happens when we die here. Boom. Game over. We got our score that was built completely locally. Let’s take it to another level. Let’s say this. Can you build me a single file HTML 3D firsterson shooter? Again, completely local. All on a Mac Mini that you can get for under $1,000. All right. Boom. It’s done. Here’s your 3D shooter. Has an entire 3D engine. T tons of enemies be able to play the game. Shoot, reload, muzzle flash. All right, let’s play this. Let me load this up in a browser. Here we go. Let’s start. We have It said it was like Wolfenstein. That’s for sure. It definitely looks like it. We’re playing it. Oh, I’m shooting the enemy. I’m shooting the enemy. Got experience points. Map in the top right. Is this the prettiest best 3D firsterson shooter of all time? No. And that’s not the point. This isn’t a Frontier cloud model, but this is a model running completely locally on like the cheapest computer you possibly can buy. You have the most advanced technology of all time, artificial intelligence running 100% on your desk on a computer this small. I will say this, if I showed you that demo 2 years ago from a cloud model, you would have been blown away. You would have been absolutely blown away. Is it great compared to like Opus 55 today? No. No, it’s not. But the point is this. It is running locally, completely private, secure, running on your own computer. So, let’s go into conclusions, then we’ll talk about all the advantages of local AI and why you want to do what I just showed you. Conclusion, can you run local AI on a Mac Mini? Yes, I think you can. Not only can you run really smart intelligence like Quen 36, which again is like Frontier from a year ago, but it’s also pretty good speed. So that whole thing, oh, you can’t do it on Mac Mini. Alex Finn is wrong. That’s dead. I’m right. I just proved it. I have the receipts I just showed you. But okay, but Alex, uh, I can use Chad GBT. I can use Opus. Why would I do this and buy a computer? Well, a few things. One, this is free. Yes, you have to pay for the computer itself, but I this I think was like an $800 or $900 computer. Now, if you’re paying $200 a month for Chad GBT and this level of intelligence is good for you, you have it paid off in like 4 months. So, long-term, you’re saving money. It’s private. When you send prompts to your AI model right now, it goes to the cloud. It goes to data centers and servers. It does not stay on your computer. So the things you’re saying are not totally yours. It’s not totally private. When you do a local model on a Mac Mini or any computer you have on Nvidia, whatever, it is secure. It stays on your computer. It’s completely private. I mean, you can be disconnected from the internet and it’s still going to work. You have total control. If you want to customize this model, you can. If you want to train what’s called a Laura and make it your own and give it your own voice, you can do that. If you want to learn how to do that, let me know down below. Do you want like advanced local AI videos like how to train the model Lauras things like that? Let me know down below in the comments. I’ll make like an advanced local AI video for you. And I think this is the most important part of it all. You know, every time I say try this AI, try that. I think the number one comment, well, what’s the ROI? What’s the ROI with that? When will I make back my money? I don’t think everything you have to do in your life has to be like an exercise of ROI. I don’t think you need to map out every penny you spend on like a spreadsheet like, oh, if I don’t get my ROI by week six, I’m not going to do. I think you’re allowed to do things for the fun of it. And while I do believe there are tons of other advantages to local AI, I think you’re allowed to do things because it’s fun. And this is extremely fun. It’s extremely fun getting a computer and then running local AI on it and having like your own personal private intelligence. I think that’s extremely fun. And if you’re watching this, if you watch uh the Alex Finn channel, you’re a member of FinnFam, I give you permission to have fun. You’re allowed to have fun. So, if you just want to do this and you don’t care about the ROI, then do it. Have fun and do it. It’s also extremely educational. You’re learning about the most important technology of our lifetimes, AI. you’re learning about it in the most basically hands-on way you possibly can by loading it locally on your own computer. And so it’s just straight up educational. And I think that’s important. I think AI is going to be around for a long time. I think it’s only get 10,000 times bigger. And I think it’s going to basically control the world economy. So you should be learning about it. You should be hands-on. You should be an AI expert. And this is a great way to do that. Did you learn anything at all? Leave a like down below. I have the number one AI newsletter on planet earth. Link for that down below. It’s called ship it weekly. Ship it link down below. Join that. I hope this was helpful. Let me know what you want me to cover in the next video. You want me to cover more local AI stuff? You want me to cover more kind of clawed code stuff? You want to see more building stuff? Let me know down below. I hope this was helpful. I hope you enjoyed it. Thank you so much for watching channel. I am so deeply appreciative you choose to learn from me. Thank you so much. See you in the next video.

Hermes Agent has released some monster updates lately that have completely changed how you use the AI agent. If you use these new, sometimes hidden features the right way, you’re going to be a productivity machine. In this video, I cover all the new features and show you my brand new Hermes agent setup. I also show you what I’m using. I using Grockbot, Hermes, Chad GBT, and which tasks I use them for. If you stick with me until the end of the video, you’ll be a master of the best open-source AI agent out there. Now, let’s lock in and get into it. So, one of the biggest updates from Hermes lately is the bot mode. Here is how you need to be using the bot mode. For those who don’t know, Grockbot comes out a couple months ago, takes the world by storm. Hermes quickly comes out with their bot mode. What is the bot mode? Bot mode allows you to create multiple Hermes bots very very easily and manage them and talk to them all at the same time here. Sort of like iMessage. Previously, Hermes was very much a single agent setup where you just talk to one agent. It’s just kind of the god mode agent that does everything for you. Now with Hermes bot mode, you have multiple bots, different profiles, different memories, personalities, tools, all of that. But here’s the mistake I see a lot of people use with this new feature. They treat it like Grock where they make a whole bunch of bots that all are like different human roles, the developer role, the marketer role. I don’t think that’s the right way to do this. I think that’s fine for Grockbot. For for Hermes, you want to make profiles and agents that kind of match the strength of what Hermes is. And the strength of what Hermes is is you can use any model, any provider you want with each one of these agents. So, I’m not making agents that have different roles. I’m making agents with Hermes, which is GPT6 Ash, Claudy, which is Opus 5. I’m making agents for all the different models so I can take advantage of the strengths of each one of those models. If I need to build something, I go to Hermes, my GPT6 Astra model. If I need some front-end work done, I go to Claud. If I have simple tasks I want done, I go to Quen, which is actually powered by a local model running on my DJX Spark. So, it gets it done completely for free. So, my recommendation to use this, go to your Hermes agent, say, “Hey, set me up a new bot powered by Claude or whatever other models you want to use.” Hit enter. It will set up the new bots for you in the bot mode. By the way, this is only in Hermes Desktop. And all you have to do is click the bot mode here, and it will create the new bots for you. Do that, then you can use the strengths of each model for whatever task you’re doing. If you want to do cheap coding, make sure you make a Grock bot in there. If you want to do super fast coding, I actually make a Meta bot in there. Meta’s coding models are actually good now, so make sure you set those up. We’re actually going to be covering 10 awesome new features here. So, let’s get to the second feature, which is their advanced local model support. Now, stay with me here. It doesn’t matter which computer you have. This local model support is going to be relevant for you. So, I really, really like this. If you go to settings and then you go to providers and then you click on run models locally, click that. What’s going to happen is it’s actually going to look at your hardware and then determine the best local model you should be running on. It doesn’t matter what you have. You have a Mac Mini 16 GB, you have a super powerful Mac Studio like me, does not matter. It will figure out the best model for your device and then basically manage it for you. All you need to do is click set up for me. It’ll almost be like LM Studio where it’ll download it for you, load it, test it, run it, and it’ll put it into a Hermes bot for you. So, you will have your own Hermes bot powered by a local model with unlimited intelligence. Now, I’m going to be honest, if you have a bad computer, if you have a smaller computer like a Mac Mini, will it be the best intelligence in the world? No. But you’ll still get free intelligence. So, even if you have a smaller computer, you can use it for really simple tasks to let it run and do it for free. So, everyone should be doing this now. Go to providers, local models, and see which local model it wants to load for you and then load up. Just click set up for me. You have nothing to lose by doing this. Hermes agent is insanely powerful, but it’s just one piece of the agent puzzle. Luckily, my friends over at HubSpot put together a free AI agent cheat sheet that helps you figure out which AI agents should be used for which jobs without having you need to test a bunch of tools yourself. Inside the guide, it covers seven of the biggest AI agents right now. Chad GBT agent mode, Manis, Claude, Co-work, OpenClaw, Zapier, NAND, Claude Code. For each one, it breaks down what it’s best at, how hard it is to set up, what it costs, and when you should actually use it. You also get copy paste starter prompts. These prompts cover workflows like competitor research, lead routing, and daily briefings. So, you can go from comparing tools to actually using them right away. My favorite is the skip if you need section, which basically tells you, hey, these agents, you don’t need to use them if you do these. There’s so many agents out there. It’s good that they tell you which ones you don’t need to use. If you enjoy Hermes, you want to know what other agents you should be using, too. This is the resource for you. Grab the free cheat sheet in the link down below. And thanks to HubSpot for sponsoring the video. The next big update I’ve been using on is Hermes HUD. This is a great new feature, especially if you play video games. This is like a gamer’s dream. I recently built like an open source version of this for all my other agents. But it’s great that Hermes built this in. So basically, it allows you to make this like kind of mini version of Hermes. This little HUD, you can move it anywhere on your computer you want. It always sits on top of whatever’s on your screen. So if you open up a game, you open up War Dogs or World of Warcraft, whatever you’re playing, this will sit on top of it. So anytime you can go in and you can talk to it. You could say, “Hey, what’s the stock price of MU Micron?” But here’s the cool part. It has access to whatever’s on your screen. So, if you’re playing a game, you’re playing World of Warcraft and you’re on a quest, you and you’re stuck, you don’t know where to go, maybe it has you killing like a specific like highlevel elite boar, you can go in on Hermes Hut and say, “Hey, where is this boar located?” And it’ll show you. It’ll tell you exactly where to go cuz it can see your screen. So, it’s like the best like cheat code for video games ever. If you had a Game Shark when you were a kid like me, I know my millennial is showing right now. This is basically like that or just an easy way to access your agent no matter what’s going on your screen. You don’t have to keep minimizing and opening and minimizing. It’s just always a HUD that’s there working for you. So, I love it. If you’re on Mac, all you need to do is do commandsh. If you’re on Windows, it’s control shift H. Puts it in HUD mode and you’re good to go. You can move this around anywhere on your screen you want and you always have Hermes Agent ready to go. It’s amazing. Which brings us to feature four, which is it now has chat GBT voice mode. If you’ve watched my previous videos, you know I am absolutely in love with chat GBT voice mode. I think I think it’s such a powerful tool. Well, guess what? You now can hit this button right here, open it. Hey, can you tell me what the status of my other agents are? So it has the strengths of Chad GBT voice which is it’s real time interacts with you very very quickly but it also has access to all its tools. So it can look what your other agents are doing. It can orchestrate your other agents. It can delegate tasks to different agents. It’s basically chat GBT voice but in this nice multi- aent setup. All you need to do is put in your OpenAI API key and it’s good to go. I think the costs are like really small. It’s like 5 cents a minute or something like that. So it is really really cheap and it’s really awesome. If you haven’t used Chad GPT voice yet, it is such a sick way to use AI. It is clearly the the way of the future, the way we’ll be interacting computers in the future, which is through our voice. I recommend when you’re doing work next time and you’re thinking about typing your Hermes agent, just use voice instead. Sit there, talk to your microphone, and just delegate tasks out. Just stream a conscious as you come with tasks. give it to your Hermes agent and now it’ll be amazing at delegating it to your different bots you set up which is much related to the next feature I really love that they just released which is the ability to use the hey Hermes keyword to talk to your agent. So as you can see here I just said it out loud hey Hermes can you tell Claudia to build a 3D firstperson shooter using 3JS and as you can see it gave the command to the other agent to build the 3D firsterson shooter. So now basically anytime you’re sitting at your computer and you think, “Oh, I wish I could do this. I wish I could do that.” You just say the Hermes keyword and it starts listening to whatever you have to say. This is awesome. You’re multitasking. You’re sitt your computer. You think of things. You say the keyword and you just start shouting out tasks to do. It’s been really helpful for me. I’ve been able to get way more done. I’m telling you, voice makes you so much more productive. The next feature I’m really enjoying is its sub agent functionality. It now has really nice user interfaces for sub aents. So you can track what each sub agent is doing. A lot of the AI agents out there and the tools, they don’t give you visibility into all the agents. I really like this. So check this use case out. Please use sub agents to design three concept poster images for a dystopian shooting video game called Null District. It’s a stylistic cyberpunk dystopian looter shooter. What’s going to happen when you give a task that has multiple threads and strands to do? It’s going to spin up a bunch of sub aents. But now it’s going to give you a really nice interface for them. As you can see here, here are the three different sub aents. You can see which models they’re using. You can click in and you get an entire window with every single thing the agent is doing, which is really cool. If you want to change each sub agent, which model they’re running. So maybe you want to do like Claude for one sub sub agent to see what kind of design you get. Muse for another, you can go in here and do that. You can change which skills each sub aent uses, things like that. That’s like the strength of Hermes, right? Is it’s so configurable. You can do anything you want, any model you want, any provider you want. So I highly recommend doing that when he uses sub agents. Have it use different sub aents. Have it use Opus. Have it use Chad GBT. Have it use Grock and see how the different sub aents perform. Really good if you want to do something like design related where you can see different options after it’s done. The next feature that’s great is the cron jobs are greatly improved. They now have memory of past cron jobs. So, if you’re anything like me and you have your agents do regular research, daily research, weekly research, whatever, it will now remember its past research and not repeat its research again. So, so as an example, I have it research AI stocks every day. I like to invest I like to invest in AI companies. Before it would go and it basically research a lot of the same companies over and over and over again because it had no memory of its past cron jobs and what it did during those crown jobs. Now the memories of those cron jobs have improved. So now you can see here here’s one of the research reports it did recently where it went through found a bunch of stocks Micron Micron AVGO. But if we look at the next report, it researches an entirely different set of companies. As you can see, it even says previously researched. It shows it right there, which is amazing. So now your cron jobs are a lot higher quality. I’d recommend doing all your research crown jobs inside Hermes because now it has that memory which is really really nice. So I think the number one question I get on all my videos now is like Hermes, you’re not talking about Hermes. I still use Hermes a ton. Hermes is still a fantastic tool. It is the only usable open- source AI agent I think right now. It is excellent. So, you want to use it for the strengths that you get out of an open-source AI agent, which is one, you can use any number of models, and two, you can customize it any way you want. So, if I’m doing network administrator stuff, which is loading local models, my different computers, managing my tech, managing all my different computers, I like to use Hermes for that because Hermes is a lot more willing to jump between my computers. There’s a lot of guard rails on the other AI agents when it comes to that. Also for personal work, for the cron jobs, for doing things that might require or be better with multiple models, I like Hermes. For the deep technical work, I like Chad GBT work only because that harness, that desktop harness is excellent. The computer use in that is just the best, bar none, the absolute best. So, I like Chad GBT work for that. In more general knowledge work, I like Grockbot because I think the Grockbot interface is excellent. I think it’s super easy to use and also it has the mobile app. Hermes still doesn’t have a mobile app, which is that’s a big weakness. Grockbot has the mobile app. So, for quick general knowledge work, I like to go to my Grobot app on my phone, shoot off some commands real quick, and I get a lot of work done. That’s my stack. I’m still using Hermes. It still manages all my computers for me. I also love the iMessage uh integration, so I can go to iMessage and quickly shoot off prompts as well. That’s what it’s great for, too. But, you want to mix and match. You want to use the different agents and use it for each of their strengths. I hope this was helpful. If it was, leave a like down below. Let me know in the comments which AI agent you use the most. Hermes, Grockbot, Work, Open Claw, whatever you’re using the most. Let me know down below. I’m curious. I’ll start making more videos about whichever agents you guys like. Subscribe and turn on notifications. All I do is cover all the AI agents, how to use them, how to get the most out of them. I hope this is helpful. I’m so appreciative if you’d watch this video and I’ll see you in the next one.

All right. So, I want to make one thing clear here. Claude Opus 55 is the best AI model ever released and is the one you should be using for pretty much everything right now. So, I was very lucky. Enthropic gave me early access to this model. I have been using it non-stop. And when they took it away a couple of days ago, I legitimately felt sadness. Just to start off with here, and I’ll go through demos, take you through it, all that, show you what it’s built, it is smarter than Fable, and it’s not even really close. And it is smarter than Chad GBT6 Astra. That’s right. It is the smartest AI model out there and yet at the same time it is significantly faster and it is significantly cheaper. So when Anthropic gave me early access to this model, they didn’t tell me what the name was. They said, “Hey, here’s a model. Test it out. Let us know what you think.” And I could have sworn on everything this was going to be like Fable 5.5 or Fable 6 or whatever. Like it was just such a massive leap over the results I was getting from Fable. Then this morning they messaged me and they go, “By the way, that was Opus 55.” I could not believe it. It’s the first thing I said. I’m like, “No, there’s no way. This is Opus.” I mean, it was smarter than Fable. I was taking code that Astra wrote. I was giving it to Opus and it was going, “Oh, this code’s disgusting. Let me clean this up for you.” Like, it was just finding issues left and right with basically all the code any other model has written for me. The speed is unbelievable. I mean, I can’t remember Opus ever being this fast. I mean, let’s not even compare it to Fable or Astra because it is just light years faster than both of those, but even comparing it to previous Opus models, it just feels zippier. Like, I was testing it early and I thought, “Oh, they must have had some sort of breakthrough here cuz the speed is incredible.” And so, they release it today. I finally get my hands on the benchmarks when they release it and I’m just like, I just can’t believe it. They actually reduced the price. So, not only is it cheaper per task than any other model out there, they also just reduced the straightup price. You literally never see this. They decided, hey, we’re just going to make less money on this model. I guess the idea is, and this is a great idea, if this model is so good, people are just going to use it more and we’ll get more usage out of it and make more money off it. And it makes sense. It’s both consumer friendly and friendly for the company. the price goes down, the cost per task goes down, and the intelligence goes way up. As you can see here, there’s nothing even close. You can see Astra, it’s smarter than Astra. The high setting on Opus 55 is better than the max setting on Astra 6. All while being like a fraction of the cost. I mean, you very rarely see this. This is like the trifecta of great speed, intelligence, and cost. all better, all improved, all like best-in-class. That is unbelievable. And if you’re going to be using Claude Opus 55, you want to make sure you get the absolute best out of it. And the way you get the best out of it is with Claude Skills. Many people don’t know how to set up Claude Skills. That’s where my friends at HubSpot for Startups come in. They just put out this free resource that gives you five clawed skills that help you crush founder-led content. And once you have these five skills installed, you stop prompting Claude from scratch every time and start working with a system that already knows your workflows. Inside, you’re going to get five installable skills that turn your raw work into voice matched channelready content in under 2 hours. You got a voice builder that actually sounds like you. Look at this output. A weekly capture skill that takes your notes, Slack messages, whatever, and turns into actual real content. My favorite is the hook and angle generator. This has made creating content so easy for me. I just give it kind of rough ideas and it gives me hooks I can use, angles, drafts. It just makes creating content so nice. Grab all five skills from the link down below. And thanks to HubSpot for sponsoring the video. And now, despite all of that, that’s not even the best part of this model. I’ve been talking to it non-stop for the last week and let me tell you this, it is the best model to talk to ever. There’s been like this crazy issue going on with AI the last few months and it’s been pissing me off so much. Basically, every model to come out since like Opus 48 has been unbearable to talk to. And this is across all of the different companies. This is Chad GBT, this is Grock, this is all of them. Basically, every model to come out from any company since Opus 48, which if my memory serves me correct, was like May, has been unbearable to talk to. They’ve all developed like this weird language that they talk in that’s like not even human. They use all these strange terms. They use smoke test a thousand times. They just like use this weird jargon that doesn’t really make sense. And you read like paragraphs of text from all these models. and you read it to the end, you go, “What did I just read? That is dead. That is over. They fixed it with Opus 55.” What was amazing about Anthropic models and why I like ride or died with Anthropic like a year ago was it was just straight up the best models to talk to. They felt warm. They felt human. They talked concisely. You It was just like pleasant to talk to the models more on a level than any other company out there. They lost it. It pissed me off so much. And now it is finally back. Opus 55 is just as good, if not better to talk to, than Opus 45, than Sonnet 35, the OG model I love talking to. It’s better that it talks human again. So, I’m going to go through so many of the things it does better right now with this live demo I’ve been doing. So, I’ve been basically been building uh this really complex game for the last week with it as as well as many other things, too. But, I think this makes for the best demo. But the way it talks is so easy to read. So nice. The full build and four tests are running in the background. About 30 minutes. I’ll be notified when they finish and there are no test results yet. The build replaces the running binary. So I had to close your game. Like it just talks like a human being. If this was Opus 5 or Astra 6 or Gro 46 or whatever, while those are great great models, like the way they talk is like the full smoke test has passed. the smoke test has passed the functionality and the binary has reset the B boundaries and it is now up to you to be smoke test to make sure it is available like it’s just like all these extra words and jargon and this is just straight up hey we built it looks good here’s everything we improved I’ll look at the screenshots check the numbers and relaunch the game when it finishes every test passed here’s the things we did it has like nice charts and tables very easy to read it’s like they trained this just to be the easiest to read model ever bullet points make it super easy to read as well. I say, “Hey, there’s a lot of issues with the game. What can we prove?” Hey, I tested the game. Here’s all the issues I’m seeing. You took 601 damage. Like, it just talks like a human being. It’s also spectacular at debugging. I handed, as I said earlier, a bunch of code that I wrote with Astra 6, which again, Astra 6 incredible. It was my go-to model up until this point, but it took all the code. It just found so many issues. Like, I I I’m really really imp. And I didn’t even ask it to debug. I just had it and said, “Hey, can you improve these things?” Like, “Oh, by the way, I found all these things. Do you want me to fix it?” It’s just so good. It’s so smart. It’s so intelligent and feels human. So, I want to show you an example of like the depth and thinking it gives you in just a couple prompts. This is a full top-down 3D extraction shooter. It builds in a few prompts. There’s so much depth to this game, even more depth than I was getting out of Astra. And it does it so nicely with so few bugs. I think that’s like kind of the underrated part here is like it writes bug-free code. I don’t know how they train this model, but it’s like shipping me the most code that’s bug free that I’ve ever done before. So, let’s just go and just look at this. So, you go in, it has this like entire weapon inventory system. You can go right in here and there’s so much depth to this game. It might not like look at it at first, but you go just first of all look at all the items in the game. Look at the map and everything in it. It has different locations. Now, we got enemies. I can kill the enemy. There’s like a full like loot system, the loot wheels. You can put it in your, as you can see, there’s like full equipment and all that. It’s really amazing what it was able to build for me. The 3D modeling also basically just as good as Astra 6. Has a full extraction system. I can hold E to extract. I get out with all the loot I came in with. By the way, would you play this game? Let me know down below. One other thing is its ability to come up with novel ideas. If you’re anything like me, you do tons of brainstorming with your models. Opus 55 has been unbelievable at coming up with novel ideas. So, I just bought the new Apple Watch Ultra 4. I go to Opus 55. I say, “Hey, can you help me come up with ideas for how we can hack this? Apps we can put on it, different AI things we can do.” He comes up with all these ideas, builds it out for me, and what we end up doing is I can now talk to my agents on my wrist. So, all my agents, I built this custom app for the Apple Watch where I can see all my agents working as they’re building code or whatever, and I can hit a button and talk to them and text them while I’m on the go. No phone required. It just like came up with all these really interesting ideas. I said, “Yeah, that sounds good.” And it went ahead, built it. Figured out a way to like put developer apps on my phone and on my watch. I have no idea how to put apps like on my watch. I don’t know how Apple iOS watches work. It just is like, “Hey, by the way, I put the app on your watch.” I’m like, “No, you didn’t.” And I look down and there’s just the app on my watch. I have no idea. I didn’t plug my watch into anything. It just figured it out. So, it’s super intelligent, super fast, super cheap. Comes up with these novel ideas. I said, “Hey, I have like 17 computers here. What are some cool things we can do with local models?” It set up this entire local model system for me. If you use AI to bounce ideas, which you definitely should be doing, just coming up, hey, how would you use this? What would you create here? What would you build here? This is like the best model for that ever. Now, the one downside is is like, do I believe the Clawude Code harness is as good as the Chad GPT harness, for instance? No. The Chad GPT harness has like voice mode, which is just like mindblowingly good. Claude does have a voice mode, but it is not quite as good as Chad GPT’s voice mode, although I’d expect Anthropic to come out with improvements for that soon. So, the harness isn’t quite there yet, but this model is so good that I do believe you need to be using it. It solves basically every problem and challenge Anthropic has had for a while now. From usage, it’s way more efficient to the way it talks, it’s way easier to talk to, to just being able to get things done. They finally are back. Anthropic is back. They’ve been out of the game for like 5 months now, which has been really unfortunate, but they are back. And when companies come back and they compete, us, the consumer, always ends up winning. We are again in the two most exciting weeks in AI history this week and next. Make sure to subscribe and turn notifications to stay uptodate on all the latest news on what is going on with AI. I’ll give you all the demos, everything, all the exclusives. You’ll find them here on this channel. I hope this was helpful. I am so grateful you’d watch my content and I’ll see you in the next

Grockbot just released for Tesla. That’s right. You can command an army of AI agents from your car. Today, I’m going to show you what the future looks like. SpaceX AI reached out to me, gave me early access to this incredible new feature, and I’m going to show you what it looks like to command an army of agents from a robot that drives you around. That’s right. Today, we’re going to get in this incredible machine. We’re going to sit in there. I’m gonna get all my work done while the robot drives me around. No steer, no nothing. Lean back, get tons of work done. I am lucky enough to have it preloaded into my own personal Cyber Truck. By the way, check this out. Just check out the wrap job if I don’t say so myself. How beautiful it is. This is going to be the car we’re driving around. Say, highly recommend you get one of these. The best car I ever drew in my life. Let’s hop in. Let’s get in and see what the future of work looks like. Rockbot inside of self-driving Teslas. Let’s go. All right, here we go. Let’s get inside here. Let’s go. We are going to set this up. All you have to do to start using your Grobot is talk to Grock. Let’s not get DCAD here. Here we go. Put this in. Here’s what we’re going to do. Let’s start right now. Hey Grock, can you see my Grock bots?

You’ve got 17 Grock bots all idle right now. Your exec team is Slate as chief of staff, build as CTO, Barry for content, Dusty for community, and Kelly for HIM. On the shelf side, you’ve got mirror running product. As you can see, can see all of my Grocks. Let’s put this pause for a second. It has awareness of all your Grock bots you have, and you can just start talking to any of them. So, here’s what we’re going to do. I’m going to plug it in and out. We’re going to start heading and get a burger and we’re going to get more work done on the way to a burger than we ever got done in our life. So, here we go. I’m going to navigate in and out. Boom. There we go. We have full self-driving going. I’m We’re in the middle of this parking lot. I’m just going to hit the start self-driving button and we’re off. I’m going to sit back here and relax. Hey Grock, can you do me a favor and can you talk to the Henry intelligent machine team and have them start building an alternative landing page for the app we’re building?

Done. I sent the ask to Kelly, Hank, and build on the Henry intelligent machine site. They’ll coordinate on direction and scope, and I’ll have them report back with a plan and any decisions you need to make. So, we started our work here. I just put Grock on mute so he doesn’t listen to me while I talk to you. We started our work. Anyway, what were you going to say? We’re off now. We’re off to burgers already on it. He’s checking what Kelly and Hank have so he doesn’t widen the scope. I’ll let you know as soon as they have something concrete. So, as you see, I’m getting I’m getting up to he’s syncing with Kelly on scope. Then he’ll propose a page structure and start the draft in the repo. Sounds like they’re moving fast on it. As you can see, I’m getting status updates as we go. I’m just sitting back getting status updates to you and she’ll have a short plan back to you shortly. Our car is driving. I haven’t touched the wheel once. Let’s get off mute here. Let’s give our next commands. Can you go to our content team and can you have them prepare five drafts for me for a tweet that talks about how amazing Grockbot in the Tesla is and how I can get so much more work done while I’m in full self-driving mode and Grock bots are just working for me. Can you get five drafts? I can compare.

Send to Barry. He’ll have five drafts for you to compare shortly. Barry being my content team writing tweets for me. Hank on the landing page until you pick an angle. No rush on that. Hank got a head start anyway. Kelly’s pulling his draft so nothing gets built. Amazing. Can you do me a favor? Can you go uh maybe you go to Slate, my chief of staff, and can you have them? Here they are. Draft one story. Oh, here’s our drafts. Can you can continue telling me the drafts? Draft one story. Yesterday, I did a 40minute errand in the Tesla FSD drug. Brock bots ran my content drafts, academy replies, and a research pass in the background. I got out of the car with more shift than most people get done at a desk. Draft 2 contrarian. People think FSD is about not driving. Wrong. It’s about getting a second work block every day that used to be dead time. Car drives, rock bots work. Let’s move on to the next task here. Can you go to my chief of staff slate and can you have him put on my calendar for October 1st uh that there’s a new reveal for the Roadster? I want to make sure I can watch that. Slate’s

on it. He’ll get the Roadster reveal on your calendar for October 1st. He’ll confirm the time with you once it’s set. Can you go to build my CTO and can you ask him to go onto my Omachi laptop uh and install Grock Build so I can bunch build a bunch of new apps on that laptop tonight. Slate found your Google calendar needs reconnecting first. He’ll slot the Roadster reveal for October 1st at 7:00 p.m. Pacific, unless you’d rather a different time. Just reconnect when the sign-in card pops up and he’ll finish it. That works. Can you go to build my CTO and can you have him install Grock Build on my Omachi laptop so that I can build a whole bunch of cool new things when I get home later?

Builds on it. He’ll work with Marchie Dev to get Grock build installed on your Omari laptop before you get here. That sounds amazing. On to the next task. We’re getting so much done here. Grock, thank you so much. Grock, can you do me a favor? When I get home, I’d love to be able to play a video game. Can you go on my Mac Studio and build me a 3D firsterson shooter on there?

Bill Junior is building it now. A 3D firsterson shooter on your Mac Studio, ready to play when you get home. He’ll ping you with a link to launch it once it’s playable. We’re almost at In and Out now. We already have multiple apps building for us. Laptop version 1.030. Grock build is installed for us. We have a new landing page uh being built for my app right now. We have multiple tweet drafts being written for us. Starting a new repo for a browser based FPS. He’ll have it ready to open in your browser when you get home. I haven’t touched a steering wheel once. Super plus control plus G opens Grock build straight from the keyboard. That’s amazing. Uh, let’s see what we want to do next here. All set on that front. Anything else you want lined up before you head home? Can you just give me a status update on uh the landing page that we kicked off earlier?

By the way, beautiful Mountain View, California. If you want to give them a shot of Mountain View, look how amazing this street is. The road, couple trees over there. Absolutely gorgeous. Not a replacement for the live homepage. Back to the star of the show. Finished the design and Hank’s actively building against it right now. Kelly and Build will review once the PR lands, but there’s no URL yet. Hank’s still coding. By the way, if you haven’t been in a Cybert truck yet, you have to you have to get in one. It is the best car ever made. Check it out. Look at all the space we got here. Look at all this. Like unbelievable amount of space in here. The screen is massive. I can lean back. I can relax while I’m going anywhere. A huge screen. Chargers here. Get yourself a Cyber Truck. Unbelievable car. You know what? I’m feeling kind of tired. Let’s do this. Hey, Grock. Can you take us to the closest Phils? I’m feeling a little tired. I can use a little energy.

There are a couple of Phil spots nearby. The closest is in Palo Alto on Forest Avenue. Just take us to that one.

We’re headed to Phil’s on Forest Avenue in Palo Alto. About 5 1/2 miles, roughly 10 minutes. So, you should be sipping something energizing by 5:43. And as you can see, all I need to do is talk to my uh my car here. Takes me wherever I want. Haven’t had a touch of the steering wheel once. Uh it’s absolutely amazing.

And uh hey, Grock, can you just give us a uh a couple status updates on where we stand with our work? Did we get the 3D firstperson shooter done? How’s that alternative landing page looking?

I sent the question to Bill Jr. and build. They’ll dig into the shooter and the landing page and report back. And as you can see up here, look at this robo taxi. Look at this robo taxi approaching. This is the future here. Soon you won’t even see this steering wheel. Absolutely gorgeous. Look at that car. Can’t wait till that releases. Bring it to Bring it to Palo Alto, Elon. Bring it to Palo Alto. What are we waiting for? Us here in California, I don’t pay 70% taxes. not to be able to use your robo taxi. Bring it to California. Elon, what are we waiting for? You’re teasing me left and right here. I see every other mile. I see one of your robo taxis. Bring it to California. Uh anyway, as you can see, this is what the future of work looks like. You’re looking at right now. This is the future productivity. Sitting in your car, sitting in your robot, it drives you everywhere. Your Grock on your car can talk to all of your Grock bots, can pass it off to all your bots. Uh, it it really is amazing. The Grockbot, the the Gro built into the car. After you sign in, it connects to your Grockbot account and it can just see all your bots and talk to them. Uh, you can get status updates, you can check your email, you can check your calendar, you can shoot off emails. Imagine waking up, you’re heading off, maybe you got a 45minut drive to work. You sit back, you put in your destination, you start talking to Grock, you get your email shut off, and by the time you get to your office, you got more done than all your co-workers combined, you didn’t touch your steering steering wheel once. Life is good. Life is beautiful. I believe this is the future, right? I This is what’s beautiful about AI is you can just It allows you to do more of what you love and less of what you don’t love, right? I’m doing what I love. I’m here to I don’t have to sit in an office. I’m not chained to my desk. I can go out and about. Maybe I want to go for a surf. I plug in the beach. I get a whole bunch of work done in my Tesla Cyber Truck. And by the time I get to the beach, all my work’s done. I can go surf. I can enjoy the things I love in life. That’s the beauty of AI. Don’t Don’t listen to these people telling you AI sucks. They suck. AI is beautiful. That is Grockbot in the Tesla. Get it now. Uh, I have a referral code for Teslas, but I don’t know what it is. So, uh, if you want to buy one off this video, DM me so I get like $200 off my next car. Uh, subscribe, like, turn on notifications, sign up for the newsletter, and click on as many things as possible my profile that boosts all the algorithmic stuff. Thank you so much for watching. See you in the next video. There you go.

Gro 47 just released and it’s not only an excellent model, but is the first model trained to be used in an AI agent harness. That’s right. This model is literally built for Grockbot. In this video, I’ll cover why this model is so good, show you how to use it, and most importantly, show you my Grobbot setup and workflow that will skyrocket your activity and makes the absolute most out of Gro 4.7. This one’s a banger. So, let’s lock in and get into it. So, Elon did it. He dropped the long awaited Gro 47 today. It’s the one we’ve been waiting for. Let’s talk about what is different with this model. The number one news story here for me, and after this, we’ll go straight into Grobbot. I’ll show you my setup. I’ll show you which bots you need. I’ll show you the workflows for using this model. All of that. So, stick around. But the number one story here is that this model was trained to be used natively in Grockbot. That means it was built to be your personal assistant. It was built to do things. It was built to be proactive. It was built to execute on tasks. It was built to be just as good, if not better, than human beings at doing work on your computer. And this is huge news because I personally believe Grockbot was the best AI agent harness. And that was on Gro 46. That was before this upgrade. Now, there are a lot of other agents out there. There’s now Muse. There’s a few others. I think Grockbot has been my daily driver for weeks now since it released. So, this upgrade is awesome and I’ll show you a lot of the awesome things it can do. Some other amazing things, same price as Gro 46. Gro has been the most costefficient model out there for months and months and months. That only continues. They did not raise the price. They improved performance, same price. From a benchmark perspective, I’ll show you those in a second. It’s better than Opus 5 and it’s half the price of Opus 5. That is pretty mind-blowing. At this point, I don’t know really what the point of using Opus 5 is, but I am sure we will hear more from Anthropic soon. This is also SpaceX AI’s first entry into what I am calling the great agent race. What do I mean? I believe all of these AI companies now care less about you using their models and more about you using their AI agent harnesses. There is so much more to be won by SpaceX having you use Grockbot than it is just by using their chatbot because now they can see everything you’re purchasing, emails, all of that. There’s so much more data to be had from these AI agent harnesses. So, I believe this massive race for the AI agent harness has begun and this is their first entry into and I believe everything we see moving forward from these companies are are going to be all about the harness, Grockbot, Muse, all that rather than just the model. So this is big from a performance perspective. I like this chart because you can see price to performance here. As you can see, Opus 5, Grock 47, it is above Opus 5 and basically medium and above and it is significantly cheaper. So there’s pretty much no reason to use Opus at this point. Fable 51. They never claimed this was going to be a Fable or Astrolevel model. Although Elon saying the next one might be that Fable or Astro level model, but this is going to be your daily driver, your main workhorse. When you need tougher tasks, you go with Fable and Astra. When you’re looking to kind of do your basic everyday tasks, Gro 47 at this moment in time, we got a very exciting two weeks ahead of us is the winner for those kind of daily tasks. Now, let’s get into the meat and potatoes here. This is what you’ve been waiting for. This is the actual helpful part and that is Grockbot using this brand new model in Grockbot. I get a lot of questions around AI agents. Should I be using Hermes, Grockbot, Muse, all these? We talk about a lot of different AI agents on this channel. I personally over the last few weeks, Grockbot has been my daily driver. I think Muse is stronger in a lot of different ways. I think Chad GBT work is stronger in a lot of different ways. But for getting just basic tasks done, basic knowledge work tasks done, firing off code really quick, understanding your emails, doing a little bit of research. I think Grockbot is that fast ball down the middle pitch you rely on that just gets it done better than all the others. That’s why I’ve been relying on Grockbot. So, what I’m going to show you in a second is my workflow with the brand new Gro 47 inside Grockbot. what I’m doing to get way more work done, get way more code out there, get much much more done. But just from a highlevel perspective, Muse is great. I just think Rockbot is built more for the power user. I’m a power user. Odds are if you’re watching this channel, you’re a power user. I think it’s just better for power users because you can have all of these different agent personas, right? With their own personalities, their own tools, their own memories. for the power user, the non-make me a reservation user. I think this is the best setup you can have, right? I think news is great, but I think it’s built for the like the make me a reservation crowd. The crowd that like wants to save like 8 seconds and have their AI agent go and make a reservation. I’m not one of them. I make my own reservations, but when I want to get hard work done, when I want to pump out tons of PRs and tons of code, I go to Grockbot. I’m not going to Muse for that. So that’s why I still recommend using Grockbot. Now let’s talk about using Gro 47 in Grockbot. First I’ll show you my setup. Then I’ll show you my workflow using Gro 47 using the cloud cursor agents which I think are excellent. So my setup is this. As you can see it over here on the left. basically for every important area of my life. Whether it’s for Henry intelligent machines, the startup I’m building, whether it’s for my content team, whether it’s for product research, whether it’s for my community, for all the different important areas of my life. Another project I’m working on right now, I basically have three main bots for every area. The project manager, the developer, and the designer. I think for every important area of your life, you should make these three bots inside a Grockbot. The reason I do that, I think this is a perfect way to separate your team, separate their duties. I find that they are really good at determining which task to give to which bot really well when you have these three roles. So, what you want to do if your setup is not perfect just yet, go in, think of the four to five important areas of your life, then build those sections on your left here and then put the three bots in each. The project manager, the developer, the designer. Then your chief of staff can easily message each one. The project manager is going to be your executive on each of those teams. So you can create a group chat called the exec team which has all your project managers in that one chat. So if you ever need status updates or things like that, you can just go to your exec team chat and get the status updates from the project managers. This is not a setup you can do in Muse. This is a setup you can technically do in like Chad GBT and Claude, but you would need to like spin up threads and then pin the threads and then rename them with names. So, it’s kind of like you’re taping and gluing together the fix here. Grockbot just made for this use case and that’s why it’s still my daily driver. By the way, if you are a Grock fan, you need to immediately subscribe and turn on notifications cuz there is way more exciting Grock content coming. I filmed a video over the weekend which is my most anticipated video I’m ever putting out that’s coming out soon. So, make sure to turn on notifications for that. I’m really excited about this one. It’s going to be a lot of fun. Subscribe, turn notifications, and leave a like if you learned anything at all. Also, one last plug. Number one AI newsletter on planet Earth, completely free. Link for that is down below. Click to sign up for that. You won’t regret it. So, let’s go into the workflow. I personally am building my app Henry intelligent machines almost completely with Grockbot at the moment because I have a really good workflow going. What is this workflow? Well, here’s what makes Grockbot so good. It is really good at being autonomous and proactive, right? Chad GBT6 Astra, best model on planet Earth now, without a doubt. The best at building things, all that. But it’s not autonomous. It doesn’t just do work, right? You have to tell it exactly what to do. It does what you tell it to do really well, but it doesn’t autonomously pick up tasks and do things while you’re not looking. Grockbot does that really, really well. When you connect your Grockbot to the cursor cloud agents, to the team I showed you earlier, developer, project manager, designer, and then to some sort of knowledge manager, whether it’s notion or linear that has tasks in it. You can basically just go to your team, say, “Hey, start building.” and they’ll go and just build for hours and hours if not days and days like they’ve doing for me here. So, here’s the Gro 47 inside Grockbot workflow that’s absolutely blowing my mind. First of all, make sure you’re on Grock 47 in your cursor cloud agents. All you have to do that is say, “Hey, change our cursor cloud agents to Grock 47.” You hit enter on that. Whoever you’re talking to, whatever Grockbot you’re talking to is going to go change it to Grock 47. It’s not that by default for some reason. So, make sure you do that. Once you do that, we can start setting up this workflow. It’s been amazing for me. That’s helped me pump out PRs. So, you go to your agent and you give it the idea of what you want to build. I want to build my own agent harness, right? Whatever you’re building at the moment. For me, I’m building an AI agent. So, I give it to my agent here and I say this is what I want to build. You give it details. Then you say create as issues inside linear/notion for this project. So here’s what we’re doing. You’re giving details about whatever you want to build to your grobbot. Your grobbot is taking all those details and breaking your project down into tasks. These are called issues in linear. I don’t know what they’re called in notion. They’re just tasks in notion. And they’re breaking it down into little building blocks for your project. And then they’re going to go and take that and put that in that project management tool. You can use linear notion for this. Notion’s like completely free. Linear I think I’m paying like $20 a month for, but has like a really nice UI. So it’s completely up to you. But as you can see here, I have a ton of different tasks and issues in here that my Grockbot made. So once you have that, your Grock bots can now just work off this board. They can go autonomous. So now you can say and go to your Glock one and say start taking taking tasks off the board and working on them. Review all your work in verify before you create a PR. Then you hit enter on that. What’s going to happen is your Grockbot, which it’s been doing it for me for literal days now. This is happening over days here. It’s just I’m, as you can see, there’s zero messages from me here. It is just working autonomously literally for days taking issues and building them out. And what’s amazing here and it’s what makes Grockbot so good is they just autonomously talk to each other. The different bots I have been sent messages to this bot for weeks. They’re just talking to each other, the project manager, the researcher and developer taking different tasks, asking each other for permission, asking each other for second opinions and building out my project. And because I have hundreds of tasks on my linear that I had them write out and develop in detail, they can just go and go and go and go for days on end. And they do it completely autonomously. Now, I know what the first push back you’ll give me is, “Alex, what about usage? What about usage?” Yes, Grockbot usage isn’t the best in the world. It’s not the best in the world. Uh, out of all the AI agent harnesses, that’s probably Muse. Muse just gives you a crap ton of work for free. But here’s what you do. As long as you make sure they’re using cursor cloud agents to code, then you save on usage because it’s not Grock doing the work. It’s your cursor credits that are doing the work. And if you’re paying for a cursor plan or a Grock plan, you get cursor credits with it, right? So, you’re saving usage because the cursor cloud agents are doing it. I’ve been doing this workflow and I’m getting so much work done autonomously. It’s like actually ridiculous. This is the only AI Asian harness I use that’s so good at autonomous work. Like if I’m doing handson work, like hands-on, I want to make sure it’s done well. Yeah, I’m using Chad GBT because I do it step by step by step. It’s not autonomous, but it’s super smart. But if I have kind of basic again fast ball down the middle task, build this out, change this, put this, then I have Grockbot doing it. I can just do it auton.

It’s an amazing workflow. If you aren’t using Grockbot like this yet, you need to be cuz now with Gro 47, it’s doing it so much better. It’s so much more accurate. The code is so much better. I am really enjoying Grockbot. I just think it’s an excellent user experience. And now the brains just got a nice free upgrade, right? It’s the same price as Grock 46. You just wake up today and your Grock bots are smarter than yesterday. That’s pretty good. What do you do in Grockbot? What’s your setup? Let me know down below. I’m very curious. And what AI agents do you want to hear more? If you want more Muse tutorials, do you want more Grockbot tutorials? Do you want more Hermes tutorials? Let me know down below in the comments. I make all my videos based on your feedback. So, I want to know what you hear. I hope this was helpful. We live in the most exciting time ever. These next two weeks are going to be crazy, by the way. These next two weeks are going to be insane. We have so much coming out. Dev Day is next week for OpenAI. I guarantee we get a crap ton of releases. I’m hearing birdies about different things. I think a lot of other companies will be shipping things. Next week and this week are going to be probably the two most exciting weeks in the history of AI. You need to lock in. You need to turn on notifications. I will have videos about all of them as they come out. This is the number one destination in the world for AI news and most importantly practical AI advice for getting the most out of these magical tools. Hope this is helpful. So grateful you’d watch my content.

Meta’s Muse is possibly the best AI agent out there. The issue is is I don’t think it’s a right fit for a lot of people. In this video, I cover what Meta’s Muse is, what it does so well, how it compares to Hermes, Open Claw, Grobbot, Chat, GPT were Claude Co, and if you are a part of the group, I think should actually switch to it. By the end of this video, you’ll be a master of Meta’s Muse and will know if you should be taking your talents to meta. Now, let’s lock in and get into it. So, this is MetaMuse, the most talked about AI agent at the moment. And there are some things in here that really, really impress me. The first thing I’m going to do is take you through everything it does, all its capabilities, all its features, how to use it, what I’ve been doing with it. After that, I will tell you who it’s for. So, who is the actual group of people I think should switch to it. I don’t think everyone should switch to it, but I will let you know right after if you are one of those people that should be switching to it. I’ll go through all of the strengths and weaknesses, and then I will rank it. I’ll stack rank it, how it compares to other AI agents, and I’ll show you what my current stack is, which AI agents I actually am using. But let’s go through it. Here is Metamuse. It is what I call a godthread agent. This is a term I’m now coining called the Godthread agent. What does that mean? It is not like Grockbot where it is kind of multi- aent. You have a bunch of different named agents. It’s not like chat GBT work where you have a whole bunch of different threads for one agent. It is a god thread agent which basically means you have one chat. You can make side chats. Those are kind of like exceptions but it’s built around just having one mega chat you do everything in. Now I’ve gone back and forth on do I like the god thread way or do I like the grobbot multi- aent way or do I like the one agent with multiple threads way. I think it really depends on what kind of work you do. I think for the more hardcore people who use AI agents for a lot of different types of tasks. I actually do think the kind of multi- aent multi-thread way is good to go. But for the people, and this is a little spoiler on who I think this is built for, the people who are like the reserve me a movie or reserve me a restaurant or get this bill negotiated down use case. I think this is actually a better form factor for them. I call that the make me a reservation AI agent crowd. The make me a restaurant reservation crowd doesn’t need 10 different named agents to make a reservation at a restaurant. They just need one agent they can quickly go on their phone to open up and say, “Hey, find me a reservation at a Greek restaurant near me.” So, it is one God thread, but there are so many little details in here that I actually really love. So, over on the right hand side, I instantly fell in love with this feature, and I wish this was built into every other agent. They have an activities feed so you can see everything you’ve done with your agent that day by day by time. This is brilliant. I wish I had this for everything I do on my computer. I wish I just had I don’t know if you’re like me but sometimes at the end of the day you feel a little down cuz you feel like you could have got way more done. And I feel like this is like the antidote to that where at the end of the day you can go see everything you got done and actually like wait I was way more productive than I thought. I don’t know if it’s a part of the human condition where you just always think you got less done than you really did. But this is great. I come back to CNA and go, “Oh, wow. I actually got more than I thought I actually enjoy.” It makes me feel a lot better about my day. So, I love this. They also organize a whole bunch of other things here. I really like different approvals. So, when it needs approvals for you, upcoming tasks, it has. So, like this is your crown jobs. And then identity. This is another thing I really enjoy that they did. You kind of have your two sides of the customization spectrum when it comes to AI agents. You have Hermes, which is customize literally everything. Everything is tinkerable. Everything’s a drop down. Change everything you want. Then you have Grockbot where you really don’t customize anything. You don’t choose models. You don’t edit personalities. You don’t see memories, anything like that. Then Metamuse is kind of in the middle where you can tinker and edit anything you want, but it’s kind of hidden away. So for the power users, you have that option, which is great. I do wish I could change the personalities on Grockbot. I do feel like kind of the way Grockbot talks is a little weird sometimes. I don’t quite understand what it’s always saying. It kind of has like Claude Opus syndrome where it says things sometimes but doesn’t really mean anything. Metam uses kind of that sweet spot where it’s hidden but you can find it for the power users. That’s sick. There’s also voice functionality coming soon. They’re talking about how you can use voice to talk to your agent. Well, speaking of voice, that brings us to today’s sponsor who’s paying my bills. If you’re building an AI agent with voice, which every AI agent out there nowadays has to have voice, you need to make sure your AI agent is able to properly listen to you and talk to you. And that’s where Assembly AI comes in. If you build the smartest AI agent in the world, but it can’t understand what you’re saying in chat or it gets what you’re saying wrong. It can’t say your email address or things like that, then the whole AI agent stinks. That’s where today’s sponsor, Assembly AI, comes in. They have the real time API that literally in real time, I’ll show you this in a second, takes your voice and turns it into text in the most accurate way you’ve ever seen an AI agent do. And what’s amazing about is it takes the context around what your agent is asking and use it to understand your voice better. So, for instance, if it knows your agent is asking you about email addresses, it’ll be able to understand what you’re saying is an email address. So, let’s do this. Let’s see how awesome this is. Hey there, Assembly AI. My email address is alex1% better.io. This is a realtime demonstration of your voice AI API, your real time API. Look how fast this is coming in based on what I’m saying. And boom, there you go. Look at how good that it’s literally real time as I’m speaking. You can use this for AI receptionists, personal assistants, and even like customer support agents. Assembly AI handles listening. you connect it to whatever AI model you want. Get your API key for free at the link down below. And thanks to Assembly AI for supporting the channel. They also have many different other features here. I really like some of these features. So, one of the main use cases a lot of people do are like morning briefs, afternoon briefs, content briefs, different things like that. They built a feed so you can just take all the brief stuff out of your agent and put it into your feed. So, you can customize this feed any way you want. One of the big things I do with all my agents is like I have it update me and give me a brief on tech and AI releases every single day so I can spend less time on social media here. Brilliant. You can just customize your feed so you get that information right in your feed. You don’t have to kind of spam your AI agent with like crown jobs and things to get that information. It just lives in its own feed. Next is ideas. This is just like custom recommended ideas that Meta came up with that you can use with your MetaMuse. I talk about this a lot and that’s blank canvas syndrome. I think a lot of people have which is they get a blank canvas. They get an AI agent you can do anything in the world with but they because it’s a blank canvas they don’t know what to do with it. Right? They get less done because it is a blank canvas. MetaMuse fights this by giving you a whole bunch of uh different recommendations for what your AI agent can do. So if you come in here and you don’t know what to do with it, here’s a whole bunch of recommendations. Goals. I actually sneaky love this. This might actually be my favorite feature of them all in Metamuse, which is goals, which is basically the brain dump to reverse prompt I talk about all the time, but built into the AI agent. So, you can click a goal, it asks you a bunch of questions, right? So, I clicked on the goal to grow my creator business. And what did it do? It went in and started asking me a bunch of multi-choice questions on that goal, kind of coaxing me to brain dump in it. And then from there it reverse prompted. So it basically said here’s everything I can do based on that goal. One of my goals was I want to grow on LinkedIn. It came in and came up with a bunch of ideas on how it could help me grow on LinkedIn. It actually ended up building me a carousel for LinkedIn. It’s not something I ever thought of doing with this. It helped me come up with it. It helped me find the unknown unknowns because of this goal functionality. So I actually am a really big fan of this goal functionality. Anything that helps you come up and uncover unknown unknowns I think is going to be massive when it comes to AI because I think that really is people’s biggest challenge when it comes to AI is just not knowing what the hell to do with it. This goals functionality is fantastic for that. It also ensures you get productive things done with your AI agent which is awesome. Then we have library and you know I’ll get into this a little bit later too which is some of the strengths of Metamuse over the other agents. One of the huge strengths of Metamuse is that it is way more proactive than other AI agents. And one of those things it’s super proactive about is creating artifacts. For instance, I said, “Hey, research me so you can find out who I am and get to know, you know, what I do and my career and things like that. It automatically created uh this website with everything it knows about me.” I didn’t ask it to create this website. Now it has an artifact with all its information on me, which is sick. The first thing I did with Metamus is askked like, “Hey, can you do research on like what people are doing with Metuse? What are the different use cases people are doing online?” Boom. Without me asking, it created this artifact, which is basically this HTML site that has the top use cases people are doing with Muse. That’s sick. That now lives in my library. I didn’t ask for any of this. It is super super proactive with just like creating artifacts, creating helpful visuals. I don’t ask for any of this and I think it’s really really powerful. I think the next objection people are going to give me over this is, okay, it’s building all these sites, it’s making images, videos, podcasts. Alex, won’t that like run out of usage? When I’m in Grockbot and it makes me a video, there goes my usage for the week. Won’t this run out my usage? Well, here’s another strength of Metamuse. You get a spectacular amount of usage. They give you a tremendous amount of tokens completely for free in Metamuse. That is amazing. There are different meta paid plans built in like Instagram and Facebook that give you even more usage on top. It’s kind of confusing, but what you need to know is you get a lot of usage and you get a lot of that for free. It’s one of the benefits of this being made by a company that is getting revenue many many other ways other than AI, right? Like Meta, they get most of their revenue from Instagram ads, from selling like cheap Chinese t-shirts on Instagram in the ads, right? That’s how they get their money. Now they can take those profits and funnel it into their AI stuff like Metamuse. So you get tons of usage. So that’s great as well. One of the biggest strengths of Grockbot is its deep integration with X, right? Like that’s its moat is no other AI agent can get you content from X post things like that and that’s really helpful. Well, Meta their moat is they can get a lot of content from Instagram and Facebook. So, for instance, I can say things like, “Can you find me an Instagram reel that talks about Muse? I want to see what other people are saying about Muse.” So, we get a Casey Neistat reel it links to. Does this super quick, by the way. Boom. Takes me to Instagram. And here’s a Casey Neistat uh Instagram video on Muse. Boom. Deep integration with Instagram. Just something mildly ironic here. I feel like if you split up where the AI doomers are and where the AI positive people are, I feel like there’s more AI positive people on X and like all the AI doomers are on Instagram. So, it’s kind of funny that this is linked to Instagram and uh Grockbots’s linked to X as you can see. Fair warning from the comments. Reactions are pretty polarized. Lots of meta privacy skepticis. I think that’s pretty funny. Anyway, but this if you value it is one of the modes of Metamuse is that it can get all this data from Instagram and Facebook. Now, I’d argue, and I don’t want to get on my soap box here, I’d argue for most people, consuming more Instagram content is probably a negative, but there is a group of people where getting more Instagram content is hugely positive. I’m going to get into this a little bit more later, but the ecommerce crowd, right? Like, if you are in e-commerce, the best place on planet Earth to be on is Instagram. You can make like multi-million dollars with e-commerce using Instagram ads. That’s amazing. And so for those types of people, this is going to be an unreal integration. Now, I think when it comes to Facebook, most people just think, “Oh, isn’t that where my aunt goes? Isn’t that where all my grandparents and ants go?” Yes, that is true. Facebook is just people’s aunts and grandmothers. But Facebook does have something that’s still pretty interesting, and that is Facebook Marketplace. So, there’s a deep Facebook marketplace integration with Menuse. If you’re trying to pick up more Mac minis like me, you can just go in here and say, “Hey, can you find me Mac minis on Facebook Marketplace?” And it’ll find you a bunch of Facebook Mac minis. That is really sick as well. I mean, I feel like there’s like business opportunities you can do here, arbitrageing different things. I don’t know. I feel like there’s something here when it comes to automating Facebook Marketplace. I don’t know what it is just yet, but if you’re one of those people that like live in Facebook Marketplace, I actually know there’s a lot of people like that, this is going to be great for you because now you can pretty much automate Facebook Marketplace. MetaMuse has fantastic computer use as well, but here’s the exception to that. MetaMuse is completely cloudbased. There is no local component of this. It can’t control your computer. There’s no local computer use whatsoever. Everything is in the cloud. It has its own VM and all the work it does there. Grockpot on the other hand has both cloud and local. It’s up to you what you want to do. While Hermes is basically completely local and chat GBT is both cloud and local. So this in my eyes is kind of a weakness as a power user. I want it to control my computer. I wanted to be able to go through my tail scale network and go to my DGX Spark, go to my RTX5090 computer, go to my Omachi computer and change things around and load local models and do things like that. That’s a big use case for me. So for me, that is a massive weakness, the fact that it can’t do anything locally. But for most people, the make me a reservation at a restaurant crowd, that doesn’t matter, right? You just want them to go online, their virtual computer, and make you a reservation. you want it to go and get you a ticket to Resident Evil tomorrow night in PaloAlto, it can do that for you. It does that really, really well. As you can see here, it got me a ticket for Resident Evil tomorrow night. I’m super excited about that. It did it in like 10 seconds. Here’s another strength. It is unbelievably fast. It is like lightning fast. So, you’re getting tons of usage lightning fast. The Meta AI team uh clearly turned things around and have things going in their favor lately, which is great to see. It’s I I’d really like them to catch up to OpenAI and Anthropic and SpaceX. So, what we’ve seen so far is Metam Muse is extremely fast. Its computer use is incredible online in VM. You get tons and tons of usage. It is extremely proactive. I’d say even more proactive than the most proactive agent I’ve ever used before this, which was Instinct. I’ve been testing Instinct on the side. I’d say it’s big strength is it proactivity, thinking of things to do and just doing them. MetaMuse actually beats that in my eyes. Its deep connection with Instagram and Facebook is great. Especially if you are in the ecommerce world. Like if you’re in the e-commerce world and you’re doing meta ads and all that to to grow your e-commerce business, then that is fantastic and it is absolutely brilliant at the make me a reservation use cases. It’s just great at that. So, real quick, we’ll go over strengths and weaknesses and I’ll tell you who this is for and if I can recommend you to actually use it. It strengths dirt cheap. You get tons of usage for absolutely free, which is incredible. Basically unheard of in the AI age. It’s lightning fast. Anything you ask to do, it does it very, very, very fast. It’s super proactive. It’s building artifacts, websites, podcasts, videos, just out of the blue. You’re not even asking for it. The Instagram Facebook integration is perfect if that’s what you’re into. I’m going to be honest. The Instagram Facebook integration is unbelievable. It’s great. It pulls reels, posts, whatever you need. It’s excellent. Fully customizable. So unlike Rockbot, you can customize the memories, personality, all that. And the mobile app is very, very strong. Now for weaknesses, there are three pretty major weaknesses for me. One is the weak ecosystem. Metamus’s ecosystem is Instagram and Facebook. That is it. If you use Grockbot, the ecosystem is cursor. So you get cursor cloud agents in it, which are excellent. You get X, so you get all the X social media posts and you get some other things that may or may not be coming soon that I may or may not have heard of uh from a birdie, right? They have a very strong ecosystem. Chat GBT work, they have a strong ecosystem. You get codec, so you get all its coding abilities built in. You get chat GBT voice, which is great. Built into chat GBT, you get their best-in-class image gen. You get other things like their little microcontroller. I have it creator micro keyboard, which is great. They have a very strong ecosystem. So if you’re using the Chad GBT work agent, you get to use the entire ecosystem of Chad GBT products. For Metamuse, the ecosystem is Instagram and Facebook. I haven’t been on Facebook in 15 years. And for me, Instagram, I it’s just not for me. Instagram is not for me. I will just leave it at that. But I I do not want to spend more time my life on Instagram or pull in Instagram reels. So that ecosystem doesn’t do it for me. Now if that ecosystem does it for you then this might be a great option right like if you are super as I said before e-commerce the Instagram ecosystem is actually fantastic but right but that that’s all their ecosystem is at Metamuse is Instagram and Facebook there’s no frontier models right so you’re using the meta models which the metas-park model is very good you know it’s like it gets the job done it’s not weak you’re not going to really complain about the intelligence of this uh agent at all but it’s not Astra right it’s not fable so you’re not getting that like next level give it a prompt to build something and it oneshots this absolutely incredible product and it’s cloud only. So if you’re a power user like someone like me who you expect a lot out of your AI agent you expect it to jump around your tail scale network control all your different computers load local models things like that it’s not going to be able to do that it just has its cloud VM so who is this for who would I recommend Metamuse for? So, if you’re one of those book me a reservation AI agent users, like if you’re one of those people that like you don’t want to Google search the restaurant you go to, then click reserve a table, like you need an AI agent to do that, this is perfect for you, right? It’s perfect for those people who just need a VM cloud and an agent goes on it and books your reservations. Great for you. If you’re super into Instagram, right? If you’re in the Instagram ecosystem, if you do e-commerce and run Instagram ads, or you just want to get reals or see what the latest thirst traps on Instagram are, great for you. That’s not for me. I spend most of my time on social media on X trying to get like news and like learn about AI and stuff like that. But if that other Instagram stuffs for you, you go do that. If cost is a factor, this is a great great option, probably the best option, right? It is basically free. Anyone can use it. I haven’t paid a penny yet and I’ve used it a ton. If you want the absolute cutting edge, as I said, there are a lot of features here that are the cutting edge of AI agents. It’s so proactive. It builds you HTML sites and artifacts for everything you do. That’s really, really helpful. It’s also blazingly fast. It just feels really, really nice to use. The speed is incredible. And if you just need an assistant, if you just need something that checks your email, sees when your next calendar appointment is, which to be honest with you is probably a majority of people. Probably a majority of people in this world, that is perfect for them. That’s exactly what they need. If you watch this channel, that might not be what you need. I cover a lot of like deep tech use cases. I cover building things. I cover loading local models and using Linux distros and doing things like that. This is not really for that crowd. This is for the general, hey, I need someone to get me a ticket to a movie. Hey, I need someone to get me a reservation at the Greek restaurant down the street. This is for you. For me personally, I’m not one of these users. I don’t use Instagram. I don’t make restaurant reservations. Call me crazy. I’m fine spending the 10 seconds it takes to type in the name of my favorite restaurant, then click reserve. Like, that’s I that’s always like the kind of the first use case people show. AI agents, I don’t care. I like booking my restaurants. Call me crazy. Uh cost. I’m already paying for every service on planet Earth, so it is what it is. The proactivity is nice. I do like the proactivity and the speed. I actually think that is a big strength and I really like that. But I don’t need an assistant. I don’t need a personal assistant. I need an AI agent when the moment a new open- source model drops, it goes looks at my computer, the hardware, downloads the perfect version, it loads it up, tests it, builds a harness for me, and then texts me and says, “Hey, the model’s good to go. You can start talking to it.” That’s what I need. Metamuse isn’t that. If you are a more casual AI agent user, I think this is great for you. I just think for, you know, if you’re someone who enjoys the videos I make, a lot of the videos I make that are around building amazing things with AI agents, starting businesses, doing really intense things, local models, building your own harnesses. Metamuse doesn’t really address many of those use cases really at all. So, I personally will not be switching. So, what will I be using? How does this compare to the other agents? This is my current stack, right? For deep technical work, I’m doing Chad GBT work. Fact of the matter is Chad GBT6 Astra is the best AI model on planet Earth and it isn’t remotely close. So if I’m doing anything with any sort of complexity, I want the best intelligence in the world. Chad GBT work with six Astra provides that to me. Grockbot for general knowledge work. I have basically an AI agent bot built out for every responsibility in my life. And so they each have very specific context on the very specific parts of my life. And so for my general knowledge work that needs good context on what I do, my business, things I’m interested in, I use Grockbot for that. So and for network administrator personal work, I use Hermes. This is a big workflow for me as well. I’m constantly using different AI models, running experiments, doing things like that. Hermes, which is an open-source harness that basically does whatever you need locally on your computer, I think is the best way to do that at the moment. Metamuse does not fall into this for me. I think is a very cool tool. I’m interested to see what they build out. But like ecosystem matters. I think that’s the most important factor here because anything MetaMuse does that’s like stronger than everyone else on the software side, they can just it could just be stolen, right? Like Grockbot can add a lot of that proactive stuff basically overnight. Chad GBT work can add like the little customization features basically overnight. Really the only moes that these agents and companies have at the moment is the ecosystem. That’s the moat. None of these companies will be able to copy the Instagram moat. It’s impossible. Metamuse won’t be able to copy the X moat, right? Or other moes that they’re potentially getting soon. So the platform, the ecosystem is the moat here. And for me, the ecosystem around MetaMuse just doesn’t it just doesn’t do it for me. I don’t I don’t need any sort I don’t need to see more reals in my life to be quite honest with you. I imagine if you’re a power viewer of this channel, which I recommend you should be, hit subscribe, turn on notifications, leave a like. You’re only getting the best information in AI on planet Earth by subscribing to this. You are probably something similar to me where you’re doing a lot of building, you’re doing a lot of deep knowledge work. you’re probably using these other tools in my stack a little bit more. So, I showed you the strengths, I showed you the weaknesses, I showed you who this is for, I showed you which tools I’m currently using. Let me know down below which of these tools are you using, which are your daily drivers. Talk about your stack down below in the chat. I want to see what you guys are using. I hope this was helpful. Leave a like if it was. I am so grateful you’d watch my videos and learn from me. I really, really am very, very appreciative and grateful of you guys. Thank you so much. I’ll see you in the next

It actually happens. Chat GPT 5.6 is out and I am totally blown away. But is it better than Fable 5? I have a definitive answer for you. In this video, I’ll cover what Chad GBT 5.6 is, the improvements it made, do an indepth breakdown against Fable 5, show you who the clear winner is, and then show you an amazing workflow that will change how you use AI forever. This is the most important video you’ll watch in a very long time. Make sure you lock in and let’s get into it. Days like today are the absolute best. If you’re into technology, congratulations. You are going to have a lot of fun right after this video. This is Chad GBT 5.6. I’ve had early access. I’ve been using it a ton. So, all the people that go in comments go, “Oh, how’d you get a video out so quick? What the hell? What’s going on?” Shut up. I’ve been using it. Now, let’s get into it. Chad GPT 5.6 Six three models, Luna, Soul, Terra. The one we’re going to talk about today is Soul. That’s their biggest model. That’s the one everyone goes, “Oh, how’s it compared to Fable? Is it better than Fable?” That’s the one we’re going to be talking about is Soul. So Soul’s the smartest. Terra is kind the medium. Luna is the efficient one. The chop GPT Mini. So Soul is the one we care. I care about brains. So let’s talk about that. They also have many different thinking levels now. They have like seven different thinking levels. Ultra is the one I’m using. The reason why we can use the smartest Chad GBT model with the highest level thinking is because Chad GBT is so kind when it comes to usage. It gives you a ton of usage. So you don’t have to be scared like you are with Anthropic of using the highest thinking levels. Then two prompts later you have no more usage. Chad GBT you can use the highest thinking levels and still get a ton of usage out of it. They have a fast mode which is in my opinion revolutionary. I’ll tell you, you’re getting Fable level thinking on a really, really fast speed. One challenge I’ve had with Fable over the last couple weeks is it’s just slow. It’s a slow model. But Chad GBT56 Soul Ultra on fast mode, lightning fast. It is the best model in some ways. We’ll get into the fable comparison in just a second. Stick around with me here. We go in depth. Who’s the winner? I hand a trophy out. It is the best model in some ways, but it is without a doubt the best harness. Codeex, which we’ll also talk about that in a second. It’s gone, is the best harness at the best price at the best speed. Some also also some changes announced today. We’ll go into in-depth breakdowns on subsequent videos. Let me know what you want me to do. Most Chai GBT work, some of these other things. Codeex app no more. It’s gone. They combined it with the chat GBT app, which I actually believe is a very smart move. Codeex was the best harness. They’re bringing all those nice features over to the Chat GBT app. It’s allin-one. I didn’t like the separate apps. I’m a fan of this move. Codeex is no more. It is changed on mobile as well. So, if you use codecs on mobile, if you use codecs on iPad, which I highly, highly recommend you do, it’s now called remote sessions. They really want you using this in a very specific way on mobile, which I’ll talk about a little bit in a second as well. I actually think the iPad experience now is like the best experience. I I use the iPad a ton. I’ll talk about that, too. But the iPad experience for chat, I don’t even think they’re like trying to make it good. It actually is very good. I’ll talk about why I think it’s awesome, too. But let’s do the comparison everyone and their mothers is waiting for Chad GBT56 with Fable 5. Which is the winner? Well, I’m going to show you the comparison between the two, what the differences are, and then I’ll tell you the winner once we go through this. We’ll make this quick. I have been using the absolute booty load out of Fable 5 the last week. I don’t even know what booty load means. I just dropped that. As you can see here, I hit extra usage yesterday. I spent $500 of credits in the last 24 hours of Fable 5. I’ve been using the absolute hell out of it. So, I am in depth in the weeds on Fable 5. I know how it works inside and out. How does it compare to 5.6? Well, let’s take a look at this here. Fable 5 is still better for planning and highlevel thinking and understanding of large code bases. It is the best when it comes specifically to that. I’ll give you an example. I have like an entire software factory I have built out inside of codeex and claude code that allows me to kind of autonomously build a lot of software. It’s very complex, in-depth. If you want a video on that, let me know down below. I’ll do a whole video on loops and software factory if you’re interested. But here’s the thing. When I gave that to Fable 5, I said, “Hey, look at my entire software factory and understand how it works.” It understood it pretty much instantly. It was able to pick it up and start doing work in the factory like it was been there. Like it’s been there for years. When I handed the software factory over to Chat GBT 5.6 initially when I got early access, I’m going to be honest, it did not pick it up immediately. It took a while. I had to kind of nudge it in different directions until it finally grasped how it worked. So when it comes to understanding large code bases, being a high-level wise thinker, understanding from like a CEO level what’s going on, I would say Fable 5 is better. But now let’s talk about actual execution, actual building. 5.6 the winner. 5.6 is the winner for many, many different reasons. The code quality is just as good as Fable 5. The code is putting out is just as good, but it is more eager. So, the tables have kind of flipped here where Chad GBT 5.6 is more eager than Fable. So, it will get tasks done. You can give it major major tasks and it will grab it. It will get it all done. It will add a whole bunch of extra features to it. It’ll be fantastic at that. This was claude strength for a long time. That kind of eagerness of getting a ton done was Claude strength for a very long time. Chad GBT has overtaken that. So when it comes to execution, when it comes to actual building, 56 is the winner. Now after this comparison we’re going through, I’m going to actually show you the workflow. Everyone in their mothers is going on Twitter saying, “Oh, use Chad GBT as a sub agent inside of Claude Code.” No, that’s not the best way to do it. I’ll show you the best way to do it in a second. You’ll learn the best workflow once we go through this. Once I declare the winner between Chad GBT 5.6 ICS and Fable 5 workflow coming right after this. Let’s talk about price and this is going to be a dealbreaker for most people. Fable 5 unreasonably expensive. So expensive I actually think this model will die. I think Fable 5 is the smartest AI model, potentially the greatest technology ever made. I think it’s dead in the water. I was shocked when I started going into API pricing and I within a few hours used hundreds of dollars of credits. That’s absolutely insane. No normal person is going to want to pay that price. And you even look at the enterprise right now. Companies are cutting out frontier models and using things like GLM 5.2 cuz costs are getting so high. I don’t know who Fable 5 is for. It’s absolutely genius level. I wasn’t prepared to tell everyone, hey, you need to be paying API pricing for it. And the moment I started paying API pricing for it, I’m like, wait a second. No, no, no, no, no. Done with this. Done with Fable. uh unreasonably expensive. It is so unreasonable in fact that I believe they’re going to push up Opus 5. I think in the next week Opus 5 comes out. I believe I don’t think it’s Fable level. I believe it’s very close to Fable level, but at way more reasonable prices. I just don’t see how Fable fits into the picture with what I’m about to show you with Chad GPT. No one in their right mind, unless you’re just outrageously rich, is going to be spending $500 a day on credits. And this was even like a full day of coding. This was like an afternoon like 1:00 p.m. to 8:00 p.m. of coding. $500. That is crazy. Chad GBT 5.6. It’s more affordable, right? It’s way more affordable than Fable 5. It is still costly. I have never I’m going to be honest with you and I do this all day. I’ve never hit my limits with Chad GBT before. That’s never happened for me. 55 everything before that. I’ve never hit my limits. I hit my limits for the first time yesterday using Chad GBT 5.6. 6. That’s never happened before. So, it is more expensive than the other Chad GBT models. I did hit my 5hour window. I didn’t come close to the weekly window. 5 hour window though, I hit the limit. I started having to use API pricing. I was using Soul at ultra level. So, like if you go super hardcore on their most expensive model, you could hit limits. But at the end of the day, it is way more affordable and reasonable than Fable 5. UI was always Chad GBT’s weakness. It’s now good. It’s good enough at UI now. They they’ve certainly improved it. It’s good enough. It’s not a concern now where you need to jump to Claude every single time. Claude is better at UI still. Here’s where I actually think the winner comes in. And this is why I am going to declare Chad GBT 5.6 the winner. The harness around Chad GBT is so much better. It’s so much better. The computer use is significantly better. The browser use is significantly better. Claude has made jumps here lately, but Chad GBT, the harness, doing things for you, signing up for websites, getting API keys. I’ve never been more confident when giving a model a task that it will do literally whatever it takes to get it done. It will pop open the browser, will control your computer, it will do whatever it needs in order to get that task done. And that is a feeling and emotion that I’ve never had with Claude inside their harness. Which for me leads me to crown Chad GBT 5.6 the winner. Because that feeling of knowing when you give a task to a model, knowing it will get done is an invaluable feeling. Even though there’s some weaknesses when it comes to planning, highlevel thinking, figuring very large code bases out that it doesn’t really match up with Fabelon, the fact that you can give it a task and you know for certainty at the end it will be done for a reasonable price leads me to make Chad GBT 5.6 the winner. The question becomes also little things here. I’ll do a video. Let me know down below. Do you want a full video on Hermes using Chad GBT 5.6 as well? I’ll do that. But this leads me to the next section of this video, which is the workflow. How do you actually use it? We talked about what is better, how it compares to Fable 5, what the new changes are. How do you actually use it? Everyone in their mother’s been telling you if you pay attention to Twitter, if you’re in the Twitter bubble, oh, set up Chad GBT 5.6 as a sub agent inside Claude Code. No, do not do that. I will show you the best way to do this. And before we do that real quick, if you learned anything so far, leave a like down below, subscribe, turn on notifications. All I do is make the absolute stone cold best content about AI on the entire internet. And if you want to work with me live, I do live boot camps every single Friday in the Vibe Coding Academy. It’s the number one community for people building with AI on planet Earth. We’re also doing a $1,000 hackathon right now. It ends in two weeks. if you can still join it. Get in, build something cool, submit it, you can win $1,000. We’re doing it every single month. Join the Vibe Coding Academy link down below for that. Let’s get into the workflow now. So, this is the Codeex app. I’m using Codex app for now cuz I am filming this right before the model comes out. So, I’m still using the Codex app before it’s combined with Chad GBT. But, here’s how you want to use this. Here’s how you want to use Chad GBT 5.6. First, we’ll talk about from the harness level. You want to choose one computer or device you own to be kind of the head development machine. The reason why you want to do this is the remote control capabilities of Chad GPT are incredible. You can control Chad GPT from any device you have, iPad, iPhone, any of your computers, Mac Mini. So what you want to do is choose your kind of main developer machine. This should be a desktop device. I’m using one of my Mac Studios, my Mac Studio 1. I have three of them. Not to flex on you, but kind of a flex. I’m using my first one as my main developer machine. Now, every other device you have will be a node to that. Here’s how you want to think about it. So, you have your main development machine where you’re using Chad GPT or Codeex and then every other device you own will control that. So if you have a desktop and you have a laptop, your laptop controls it, your iPad, your iPhone, everything controls this one place. And the reason why you want to do it this way is so that you have all your code in one place instead of having some features on your laptop, your desktop, some features on your iPhone, your iPad, it all goes to one place. And because the remote control functionality is so good on the Chat GBT and Codeex apps, you can easily control from all your different devices. So, choose a device. Make sure it’s a desktop if you have one. The reason being is you can leave your desktop on 24/7, 365. Laptops, you got to close it. Puts it in sleep mode, different things like that. So, make sure you have kind of a desktop device, a Mac studio, a gaming computer, whatever you got, make that your main computer. Even if you just have like an old dusty laptop in your closet that you can put on 24/7 mode and plug it in, use that. Something that can be up 24/7, 365. So from there, you have your projects and you have your chats inside of each of the projects. I’m going over kind of highlevel tips in here. We can do a deep dive building out a full app end to end on a subsequent video. If you want, let me know down below. But what we’re going to do is you want to kind of have your factory admin chat in each project. This is the highlevel chat where you can be like okay merge the code, create a PR, do this where it’s kind of managing your entire factory and then every other chat from there is going to be building out specific features for that project. So you have your main admin chat where you can go in from any of your device say hey what’s the status what’s uncommitted code we have what PRs do we have to merge things like that it’s kind of your highle admin and then every other chat is for specific features you’re building out now we said earlier Fable 5 is a better planning model and I I stand by that so what you want to do is you want to use Fable 5 for planning. So, if you have a project, you open it up inside Claude Code, you go into Fable 5 mode. I’d go on high. If you go Ultra Code or higher, you spend hundreds of dollars for a single prompt. Just use high. When you’re doing any sort of planning, you’ll come into Clawude Code. You will ask for the plan. Hey, this is what I’m thinking about building. Please build out a very large spec. If you’ve been paying attention to all my videos, you know I talk about linear and how I use linear in my entire process. So if you’re using linear, have it build the spec out in linear. If not, totally fine to say, “Hey, build a markdown file.” Then what you will do is you will take that markdown file, go into your projects, drop it in, and say, “Here’s the plan. Build this out for me.” Now, a lot of people are telling you to use a Chad GBT sub agent inside of Claude Code to do execution. I don’t like that at all because when you do that, you’re missing out on all the amazing features of the Chad GBT harness. The Chad GBT harness does so many things so well for you. So, for instance, I’m building out my app, Henry Intelligent Machines, link down below if you want to get an early preview of that. And I’m having it test some of the action generation inside the app. In Henry, you launch a venture and then it generates a bunch of actions that you need to do to push your business forward. I wanted to test that. Typically, when testing that, you would need to open up a browser, go through the flow yourself, click on things yourself. But all I do is I go in and I say, “Hey, test this out for me. Load it up into a browser, walk through it for me, do it five times, and then build a report on that.” And that is what Chad GBT does. It opens up its own browser and runs all the servers. It gets whatever API keys it needs, it plugs them in, and it runs through everything. So, it said what I tested, actually opened it up, started a venture, did a bunch of different actions, tested everything out. It did all the testing for me, something that would have typically taken me an hour before, took 24 minutes for chat GBT while I did other things. So your workflow is get the planning from claude code, move it over to chat GPT and then you give the actions to chat GPT. Here is the mindset I have. Anything you would manually need to do before when vibe coding, I just tell Chad GBT to do it. Anything it is, it doesn’t matter what it is. If they say, “Hey, sign up for this account.” I go, “No, you do it.” If they say, “Hey, can you grab this API key?” I say, “No, you do it.” If they say, “Hey, can you connect these two things together or move this file over here?” I say, “No, you do it.” Chad GBT literally 100% of the time I have said, “Hey, no, you do it.” It’s figured out how to do it. That’s why this harness is so good. When it comes to computer use, when it comes to browser use, it is absolutely next level. It is bar none not comparable to Claude Code whatsoever. That’s why the workflow is get your plans from Claude Code, move them over to Chad GPT codeex and then you say, “No, you do this. You handle this and it will get done and it will save you tons of time.” Now, while you’re in here, I do sole ultra for everything. It’s the stone cold best. If you’re on the $200 a month plan, you can get away with using fast for a lot of things. Again, I hit my limits using fast here for the first time in my entire life in Chad GBT. So maybe you don’t use fast quite as much, but if you’re on soul ultra on normal speeds, you can use up your $200 plan, like you you’re not going to hit your limits and you’re going to have plenty of usage, which I cannot say for Fable because I used the plan up in like a day and then when I went into API pricing, it was outrageous, out of control. 5.6 is the winner. It is my new daily driver. It is what you need to be using for pretty much everything outside of highle understanding, planning, and some kind of UI tweaks if need be. It’s still pretty good at UI. By the way, for the hardcore Alex Finn fans who know all about the world famous Alex Finn benchmark. Yes, it blew away the benchmarks, this is the first person shooter test. It is absolutely incredible. It has it has crits, dynamic enemies. It has great explosions. It has great waves. The graphics are fantastic. The sounds are amazing. I don’t know if you can hear the sounds or not. The power-ups are incredible. Like, there’s like 10 different dynamic enemies. Uh, it’s incredible. Chad GBT 5.6 6 won the benchmark test. I am really really impressed. And what’s incredible is over the next few weeks, we’re actually expecting GPT6. We’re expecting the next level. The the rumors I am hearing are that it’s coming very soon. They’re already done training it. They’re just doing testing now. So, expect more videos like this very, very soon. If you learn anything at all, likes, subscribes, turn notifications. Again, live boot camp tomorrow. $1,000 hackathon next couple weeks. Join the Vibe Coding Academy link down below. I love you all. It’s an honor to teach you these things. It’s an honor to work with you.