Claude Managed Agents Will Make Millionaires (do this now)

So, these uh Claude code managed agents are a pretty big deal. And uh in this video, I'm going to explain why uh some of the ways that I'm thinking about it, how I've been playing around with them, and why I think it's really important for you guys to sort of linked in with the video that I posted recently around blocking the noise out. Claude Code managed agents are an incredible opportunity to block the noise out and figure out how to use this thing to create a business, whether an offer as an agency, build a SAS around it, whatever you want to do. But we have an incredibly powerful tool here. And I think figuring out just how to use this and apply it to a given problem in the business context is going to make a lot of millionaires. So if you guys don't know who I am, my name is Le Mley. I've been building and telling AI solutions to businesses for the past 3 years. I've taught thousands of people how to start a business doing the same thing. The new Claude Code managed agents that have come out recently are a pretty big deal. And firstly because they are basically if you've been watching some of my content recently where I'm talking about AI operating systems and how claude code can be used to basically run a business or be used as an a centralized AI workspace that can basically be a form of AGI for a business that is very closely tied into the claude managed agents which is what we're talking about here and that the claude managed agents allow us to take the incredible power that we were able to achieve with an AI operating system or workspace within claude code but put it into the cloud and they have managed infrastructure for us to use and we're only having to pay for the actual runtime. So, it can be sitting up there. You create this really powerful agent. You connect it in with a bunch of tools and knowledge and and basically the same thing you'd have on your on your laptop locally, but be able to run that as a part of your application. As I've been playing around with this, it's become obvious to me the connection between what a lot of people are talking about right now as continual learning and the issue around these LLMs not being able to learn continually. This can be solved and with these massive 1 million context windows which we have with things like cloud managed agents. Now you can create these incredibly powerful essentially AGI style systems by taking the claw managed agents uh infrastructure and giving it the right materials and and resources and infrastructure around it. It needs to act like a a ever learning AGI. And this is through things like having databases that is connected to. So of course it can it can find kind of a system of record. It can store data and it can pull it from the databases. You can have all sorts of like markdown files or text files in there for it to read as well to as context to be able to get up to speed on things. You can build custom tools. But once you have all of this set up, then there's this interesting piece about continue learning in memory. And so when you have this Claude code manage infrastructure that can uh sort of go on in a massive long conversation, maybe you could plug it into some kind of super assistant app for a given vertical or a specific problem that you can solve for. Maybe you can wrap it up and package it and give it to your clients if you're an AI agency and say, "Hey, just talk to this thing." It's going to be hooked up into everything like a AI operating system, but you're going to talk to it. And over time, I I've been going quite deep on what's called cognitive architecture and how we can given the limits of these LLMs that they have a million tokens and they don't seem to be able to learn continually. How can we create the equivalent of a cognitive or a replica of the human brain? some kind of cognitive architecture that allows LLMs to learn it. And so, for example, if you if you created an application that was acting as a personal assistant for some kind of use case or it was where a a given employee was working uh for the whole day and just chatting back and forth and using it as their clawed workspace or clawed uh environment for doing their work. At the end of the day, you could have it go to sleep and just like our brains and it's honestly the going through this AI revolution that we're going through and getting really nerdy and into it has made me become such a like basically being much more awe of just how incredible our own biology and I was already pretty big into biology. I knew how cool all of our stuff was, but this is like the particular the brain and like we're trying to find all these ways for it to be this continual learning and we're just looking like man the the brain's kind of already got it figured out. This is the only intelligent organ that we know that is able to do this. And so after a given day of chatting to one of these AI assistants that you might give and deploy to a business at the end of the day, it has certain databases. It has certain documents. It has kind of long-term and short-term memory. And when it goes to sleep, it will be able to go back and analyze the full conversation. It'll be able to kind of update all of the information across the databases. It'll be able to update the text documents as well. And then so when you start a new session the next day, it can kind of jump back in and be like, "Okay, wow. Now I know all of this stuff. I've got this log of all the stuff that I've done over the past few days. I'm going to scan the last like 5 to 10 days worth of stuff that was done." And you have this very, very intelligent long-term assistant. And I think this is the big unlock we're going to see here. You guys might have played around with Open Claw and this and that. This is the clawed competitor to open claw. If you guys have been following my channel for a while, you know that I've kind of leaned on the claw code side quite heavily and being like, "This is much more deterministic." I think some of the open core stuff is is interesting I think for people who are really really non-technical getting that wow aha moment of playing around with it but I've definitely leaned more heavily onto the claw code side and now we have two competing and multiple competing actually uh different products or or systems we can use for these kinds of solutions and claude code managed agents in particular is what I've been diving into on how could you turn this incredible platform I think you could almost block out all of the other noise that's going on in the AI space right now and say how could I solve a problem with this how could I take this whether it's in a single agent or a multi- agent set up, find some kind of business problem, find just just turn it into some kind of packaged AIOS offer for a business and say, "Look, you're just going to talk to this thing. I'm going to work out some kind of cognitive architecture around it. It's going to update itself overnight and and sort of refresh its memory and and update all of the irrelevant uh bits of data across the databases and the in the documents as well. Uh, and be able to have this really intelligent super assistant that can do stuff for their business." And this kind of system, every business owner I show it to, they just light up. I've done two different masterminds now in person working with founders to get these things set up. And these kinds of systems really are like the the big jump ahead that they need to be able to have some kind of workspace that allows them to work faster, have all of the data in one place, and to be able to build automations and do things on top as well. Now, the building automations on top part is probably not what you want to be selling uh at this point. Point of this video is just to tell you to get in there and start playing around with it and listen to what I've said about that cognitive architecture piece because when you have a million tokens, you can just keep rambling forever. But having an intelligent system for being able to go to sleep and like update its memory over time, uh, but it can feel like a continuous conversation, I think that's really the the key product unlock when it comes to either building products for your clients or building products as like a SAS. The user experience now can be so much smoother if you can get that cognitive architecture right. So what does this mean for openclaw? For all of you open claw users out there or would be open core users, this is claude's competing product. This is the answer to openclaw. Um now I think there's definitely still use cases for it. Um and by far the easiest way I think mainly the setup piece has been the most difficult part for a lot of people where I've ended up with open core is just within my claude code AI operating system that I use to run my business. I've created a / new- employee command and I will use the hosting a deployment for that. So I can just hook into the hosting API and I'm able to very very easily use their specific setup for creating these kinds of hosted openclaw agents. They just make it like so easy. And so what I've been able to do is to hook that into my claw code operating system. Create a custom command that I'll go / new employee and it's going to walk me through asking questions about what I want the open claw agent to do, how I want it to behave. It's going to apply all the best uh techniques and stuff and then it's going to connect to the hosting API and use their pre-made setup that makes it very very easy to host and also manage these long term. And so beyond there is just my claude code AI operating system. I've got a couple different AI employees and they're getting better over time and then they're all hosted on Hostinger and they're kind of talking back down into my AI operating system. But the one thing you need to be very very cautious of is when you you have lots of these open claw agents distributed is that they just end up like sending you so much trash. Like if you don't have them really dialed in about and they're specific about how they're feeding information back to you. I think a lot of people are claiming that they've got like oh I've got 200 agents working for me. It's like shut up. You don't you really really don't. So let's stop let's stop lying about it. All right. Uh if you had 200 agents you just would not be able to even if a quarter of them were ones that you actually had to read the output from. even if we're just 10% of the of the agents in that 200 agents is actually stuff that you need to check and and you are essentially the bottleneck and this is what we're running into with agents is that humans become bottleneck but so when I have a new use case for say an AI employee within my workspace within my company I'll run that / new employee command actually I built that live in front of my accelerated members and that's something that they've got in the program there but it allows you to very easily and using claude code to create openclaw agents and deploy them to places like hostinger um so if you guys want to get any links or or tutorials on how to do that I've some down in the description, but it's just been a very good workflow for me to be able to do that kind of setup. Basically, Open Claw is still a valid option for a bunch of use cases. But I definitely want you guys to go and check out uh these cloud managed agents and just have a think have a think about the problems you're solving in your own business. Have a think about the problems you're solving for your clients and have just a think about the the broader problems that exist in society or around you that could be solved through this kind of incredible new platform that we have. uh because being able to resume sessions and give it a bunch of tools and give it access to a bunch of databases is incredibly powerful and you basically have your own claude code workspace that can be configured and set up in the cloud and then accessed whenever you want. Um and when you layer on something like a cognitive architecture, I think we have some very very exciting times ahead of us. So yeah, if you guys haven't seen my video recently on talking about why you need to unplug a bit from all of the AI news, check that out here. I certainly have been and it's just really really essential right now when things are so crazy. You have to find a bit of space. You have to pick something out and say, "This is the thing I'm betting on. This is the big thing that I want to take on." And that's what I've covered in this video up here. So, stay blessed. Thank you for watching, guys. and I'll see you in the next one.