Agentic Operating Systems: Automate & Build ANYTHING!

Julian Goldie (0:00)

I want to show you the Agenetic Operating System that I've built using something I call the Time Machine Method. And this is the way that it's a paradigm that's really changed the way that I think about AI, how to use it, how to get the most out of it, and how you can really automate pretty much everything in your business step by step by using this simple three step process. And the way that I think about this is that every hour you spend on a repeat job, you'll spend again next week. That's what most people do, right? And then typically what they do is they hire someone one job at a time, track where the time goes. And here's the most important part that you need to figure out is how to build the automation around it, iterate until it works, and then move on to the next job. It's how, for example, I've built out the agentic operating system. So previously, for example, I was spending a lot of time doing SEO, right? And so what I actually built is this search OS system that automated that step by step. And I'll come on to that in a second. And then the next thing was to start automating video. So what we actually built over here is a powerful system where we can automate videos. Everything is saved. We can create avatar stuff if we want to over here as well. And we've even got the workspace where we can preview and see everything that we previously created over here, right? And we can just switch between them. We can see what we've built. Everything is ready to go, right? Now, inside this system, we're using the time machine method. Now, how does that work? Essentially, it works in three simple steps. Step number one is you want to track where you spend your time. Track where you're spending your time. So for me, for example, I'll do a time audit and I'll look at, okay, what am I spending my time on? And then what do I need to automate from there? The next step is to start automating. I'll show you how to do that in a second and how we've done that step by step. And then from there, you want to just iterate. So for example, I was spending time on SEO, then it did the video system. Then, for example, we created the research system that you can see here with Hermes Oracle. And everything was just automated step by step, week by week. Now, the thing that I would say here, this is a mindset shift that if you're watching this, you really want to think about, which is you want to stop doing the same week over and over again. And I see people doing this all the time, like even high level entrepreneurs. So the old way is like every week just looks the same, which means they're doing the same jobs by hand every single week. They're hiring someone or getting a tool for each one. They're automating at random, then abandoning stuff when it breaks, which is a big problem I've seen pretty much everywhere. And then finally, the AI forgets between chats. With this system, the new method, which is the time machine method with the agent OS, each week gets hours back. So I'll show you an example of something that we've just done, like very, very recently. What you can see, for example, here is like Opus 5.5 drops, and we create a system inside the agent OS that edits these amazing videos, as you can see right here. I mean, look how good this is. This is better than a video editor can do. And I know that because we had video editors until Opus 5.5 came out. And then as soon as it dropped, guess what? They were replaced by Opus 5.5. And so now every week doesn't look the same. Each week, we get hours back with systems like this, and it looks absolutely amazing. So that's what this game is all about, right? It's all about figuring out, okay, where are you spending your time, and how do you avoid doing that in the future? So three steps, and this pretty much repeats forever. And each step just takes one more job off your plate. And then the time you get back helps you work on the next loop. So for example, what we want to do first is track it. So the thing to note here is that you can't get back time that you don't see. So you want to log what you and your AI actually do every single day, and then look for the jobs that repeat and eat the most hours. And then ask your AI to read what you're spending time on, read that log, and then tell you what to automate next. So what I can do over here, for example, is I can say, okay, based on my Obsidian Vault, give me some ideas on what I should automate today. My Obsidian Vault is simply my memory system that we've got inside here. And so, for example, here, it's like, okay, you should do this, then this, then this. And it gives me very clear recommendations in terms of what I need to automate next. Now, after that, you want to think of a system that you can build around it. And this is not a one-off prompt. You don't want to think in terms of prompts. You want to focus on systems. So, for example, if we go back to this example, which you can see right here, and we look at this, when we're looking at this, for example, what you can see here is that we have built a system, a skill inside Claude and an automation tool that can basically automate what you're seeing right now. And the powerful thing about that is that I don't need to think in terms of prompts, I don't need to copy and paste anything. I can just create something like this at a click of the button inside the agentic OS. How do I do that? I just go down to the video agent. And so the way you want to think about this with an agentic OS is that you build a system around the job, not a prompt. So you give the job its own tap. So it's just a one-click system. You connect it to your real data and your memory system. And if it's anything that goes public, just have a human gate on it that proves it for you. So for example, inside our agentic OS, we have one job and one tab for pretty much everything. So if we go down here, for example, we have the SEO tab, right? And that's where all the SEO stuff lives, as you can imagine. Then if we go over here, for example, we've got the video tab, right? And we can switch between them. Then we also, for example, have the memory tab. And the memory tab lives down here. And we can just switch that anytime. And also, this plugs in real data. So for example, we have the search console. The search console is pulling in our latest data from SEO, from our websites. And then also, we can select the best model for the job as well. So for example, if we look over here, we've got Claude, we've got Hermes plugged in, we've got Antigravity, we've got Codex, even Kimi and GLM, whatever you want to use. And the point here is like, not every model is good at everything. And so you can switch between the models depending on the job and the automation building. Now, from here, we're moving into stage number three, and this is iterate. So you want to keep testing until this stuff works with value and then pick the next job. So the way that I built out the SEO system, it didn't look like this when I was first building it out, right? There's a lot of changes, a lot of tweaks that we made along the way. So for example, like the Search OS is actually something that I just built this week. Why did I build it? Because it was required inside the tab. It gives me, we've got so much data of SEO now, we wanted a better system for organizing our opportunities, our rankings, our AI visibility, and everything else. And so you use it and you fix what breaks. You turn every fix into a rule, so it never breaks the same way twice. And then when it runs without you, you then move to the next biggest job. Now, I like to look at this week by week because what I tend to find with AI particularly is that it gets overwhelming, it can become distracting.