This Claude Skill Will COMPLETELY CHANGE How You Work (AIOS)

I haven't opened my email, my calendar, or any of my dashboards first thing in the morning for weeks. And that's because instead I say four words to Claude, give me my coffee debrief. It knows more about my day than I do. And I'm not alone in this.

CEOs and managers at Fortune 500 companies are restructuring their entire organizations along these same principles. And that principle is an AI operating system. But don't worry because you don't need to know how to code or be super technical to know how to do this. And you also don't have to be running a Fortune 500 company to make it worthwhile for you.

If you can copy and paste, you can get this exact same process set up on your computer by the end of this video. I'm going to be breaking down three different things. what an AIOS actually is, why entire multi-billion dollar organizations are restructuring and firing over half their workforce to implement it. Starting with the simplest, most powerful entry point I've found, the coffee debrief. So, let's get into it.

So, what this actually looks like, I had to blur out a bunch of stuff um for sake of just like privacy reasons in my own company as well as my clients, but every morning um I get this message in Telegram that gives me a brief overview of everything going on. And then, uh linked here is a full Google doc that looks like this where it gives me, you know, my full schedule. Um checks out my YouTube. You know, I don't mind showing you this.

You can see this public information anyways. I can check out how my videos are doing. shows me um what's going on with all my voice agents. Shows me what's going on um in my uh Rise account, which is a software I use to track all my time. Um it shows me how my episodes are in that pipeline.

Um my inputs to outputs of like what I'm spending time on and what um results or analytics actually came from that. You you can see I had finals at school last week and so it's telling me like you took a lot of time at school but that was likely due um to finals and uh you can see all this time invested whether it meets my percentage requirements um and you know also the things it skipped or the noise it filtered that I don't have to muddy up my brain with um looking through because there is just so much noise in the current day today.

So, this video was um something I've been planning for a while, but was really inspired when I saw this tweet this morning by Brian Armstrong that now has over 8 million impressions mentioning that he has fired about 14% of his workforce. And so, um this is a huge staff cut and uh the reason that he cited for this was both AI capability uh and really not market conditions. uh and since they're a crypto firm, market conditions would be, you know, a big reason for staff cuts. And historically, that's why, you know, companies get smaller as the company uh itself isn't doing well.

But he said in this tweet that his um engineers are using AI to ship in days what used to take a team weeks, as well as people that aren't technical, they didn't hire to develop code, are shipping production code that's being used in Coinbase. and that he's looking to rebuild Coinbase as an intelligence with humans around the edges aligning it. And so some of the big structural changes he announced was he didn't want any more than five layers below the um seauite. And so um if you sort of think about an organization here, you have the CEO at the top and then maybe you have the seauite or some senior managers here and then that sort of just keeps breaking down and when you get these massive companies um you know you think about Nvidia with you know thousands tens of thousands of of employees or retail chains like Walmart you know there could be you know 10 15 layers um where you know a customer could have a problem here where they're actually interacting with this uh employee and it just like the information is so muddled.

It's like a game of telephone getting up to the top and um this AI AIOS um as a lot of people in the space are calling it uh is going to solve that and you know eliminate a lot of these layers uh as well as you know these leaders these people uh in charge up here would be able to take um more direct reports. So instead of this person reporting here, uh they'd report to this person. You know, the chain will look uh a lot more like this, a lot flatter. Uh as well as um there will be no pure managers.

So pretty much everybody on uh in the company will be um taking some sort of action. Their job won't be simply to manage and direct people. Uh as well as there'll be AI native pods and some oneperson teams. So that really shows you how powerful um you know these companies.

And I'm going to show you another one in a second. Uh think this technology is to you know make massive changes like this. Uh and they also say that you know these leaders must uh ship code themselves and really be um ingrained into you know the development process. And um I'll go over who this is important for in a second.

But I want to mention another example of this which is Jack Dorsey at Block. If you're not familiar Jack Dorsey is the founder of Twitter um and now owns Block which is a huge financial company again publicly traded. and he recently uh a month or two ago went pretty much viral for firing about half the company and their stock jumped a bunch and he had uh an entire he's made a really good podcast on it with Sequoia but um basically he said um single intelligences companies as many AGIs with humans on the edges aligning the system.

So it's the exact same principle um Brian Armstrong was harping on this morning. uh and they both sort of took the same action of um you know cutting a significant amount of their workforce. The whole idea and I'll go deeper into this uh in a second is that you have AI at the core of the company and then humans on the edges to actually make sure um you know their product gets out and is touching customers. So it's not AI servicing customers but AI at the core informing the humans um aiding the humans to uh serve customers out here.

And the emerging pattern we're pattern we're really seeing and I'm seeing um in my own company as I'm starting to use this is that the more context you feed to AI uh the more reach it has into your real systems real data feedback um the more autonomy uh it has to act. The result is smaller and cheaper teams because you're paying less people. There's fewer communication layers which is really what bloat and makes bigger companies less effective. uh and decisions get made much faster, which is incredibly important um in any business.

And so, moving on into why this matters, no matter who you are. So, let's say you're a CEO, and you don't have to be a billion-dollar CEO, but let's say um you know, you have uh you know, 15 to 100 employees, for example. Um what's happening is your competitors are quietly cutting their headcount and replacing it with AI capability. Uh and if they aren't, then you should be that competitor.

And the reality is the organizational charts. So basically an organizational chart is what we were looking at uh over here of who's reporting to who you know the structure of the organization uh in terms of human capital of 2027 will not look like the ones of 2024 or the years before. And really in Jack Dorsey's blog uh he mentioned how it's going to be significantly different um from you know really the way we've been doing it originated from Roman times uh you know 2,000 years ago. And so the window to act deliberately and really get ahead is now.

And um it's really important. This is something I'm really bullish on and I know a bunch of other people in the AI space are as well. Uh it also matters if you're a soloreneur. Um so you now have access to the same AI infrastructure as a 10,000 person company.

Uh and you can build your organization bottom up using this uh AIOS. And that way you don't have to fire half your staff once you've built a way bigger company. Or you could remove the ability to actually have to hire anyone at all. So if you're a soloreneur, I just think like the leverage gap has never been smaller in competing with bigger companies and maintaining and running an organization just by yourself.

It gives you so much more agency to um run your business in this way. Again, I know it's a bit it's not super concrete now. Uh I'll explain more about like what this AIOS is in a second, but just know like the solo operator of 2026 is more powerful than a 10person team uh in 2020. And then lastly, if you actually sell solutions like I do, if you run an AI agency, um literally every business owner is going to need help building this, it's really confusing. uh especially, you know, if they are a CEO of um a decently sized company or even a soloreneur, they really won't have the time to go out, learn, and implement all this themselves.

They'll need a lot of help, and you can be the person that's actually selling it to them. Um and the market for this just opened. Uh you can rely on these Whoops. you can rely on these uh case studies you're looking at here of some of the smartest people in the space uh taking the same action and uh using that in your sales pitch and you know just know that AI in a chat window is just a parlor trick everybody's seen chatbt uh business owners don't need help with that and really what wins is AI with context reach and autonomy um gives them the real leverage and they'll see you know real change of their business with that and you can be the one who builds it for them I know people who are selling this for tens of thousands of dollars honestly just to set up um you know it's like $10,000 and they're charging per employee after uh there's a huge bull market for this and um the reality is it's just really valuable for businesses and uh lastly I want to mention the winners are the ones giving um AI access to real data and real systems the people who aren't super scared about you know the privacy concerns or anything like that now obviously you don't want to violate any laws but the more context you've given AI I we've seen it time and time again, the more effective it is by like an exponential amount nearly.

Um it's so much it's so much better. It it's you really can't even compare. And um the reason that is is something I like to call the craft framework. And now I wanted to organize thinking about AIOS in this way because there's really a bunch of different layers on top of it.

And just know like a AIOS isn't a business model. Um it's not like an AI agency where you now um run an AIOS. It's a way of wrapping um around your business model in these layers that just makes it like 10 20 times more effective. Um and uh you'll see an example of that in a second with the coffee debrief.

Now uh the way to think about this is at the core you have your business model. So, if you're a fencing company, uh this would be, you know, your fencing company business model, your services business. Um, you know, if you're a crypto exchange like Coinbase, that would be at the core. Um, if you're a fintech company, uh, like Block, then that would be at the core.

If you're an AI agency or a content creator, that would be at the core here. And then you build your AIOS around your business model. So, first we have context. So, this is really just a bunch of information about you and your company or if you're building the AIOS for an employee um the employee and their company and where the company actually wants to go.

And so, um in here you would have some important things like I like exporting all my previous like chatbt history, my claud history because that gives um the AI a bunch of context on like what I'm actually using it for. um you know what I do in my day-to-day business as well as you want to give it like your broader company goals um who you are you know what sort of tone you want it to speak in and so that's what context is and something I want to stress is you don't need to build this perfectly right away this is something that will grow and get more layers uh as time goes on I like to think of it uh sort of like those rubber band balls if you know what I'm talking about so like this is what I'm talking about these rubber band balls and the more context you give it over time just the the bigger the ball gets that's and you know the stronger and more um rigid it gets.

And so coming back here that's the the first C of the craft framework and really the first layer is the context. So the AI actually knows who it's speaking to and what its objective is. Next you have reach which is where I like to connect um a bunch of different platforms that we use in our company. And this is where uh the coffee debrief really comes in handy.

So, this could be stuff like your Stripe dashboard, your CRM, um your calendar, uh your drive, um call transcripts. I use like Fireflies for all my calls. I have the AI note taker. Uh as well as like your email.

And that way uh when it doesn't have the contacts hardcoded in this section, it can go out and reach uh literally reach into um this next section, the reach section uh to get whatever in uh insights it needs, the stuff that'll change dayto-day. So like if you run an e-commerce store, you can reach into your Shopify and get that data so that it's always knows what's actually going on. And we won't cover um the FRT flow tuning uh in this video. I am creating a full course on the AIOS right now, but with Coinbase um you know announcing their big layoffs, I wanted to get this video out right now because I think it's so timely.

Um and so after this uh we have a which are actions or custom skills. uh which means it can actually you know with this information take action and return to you some sort of information or do some sort of task for you and so in the actions is where it actually starts to become really useful and a lot bigger than just chatbt where it's now literally saving you time and making your organization much more effective. And so we're going to be covering one of the actions in this video as well as how to set up um the context and the reach uh to make this action effective.

And the action we'll be covering is something I call the coffee debrief um which I think is the easiest entry point into the AIOS and um something really powerful and easy to set up and it's really just like the layup automation uh that you can add to your AIOS. And so this is basically a clawed skill that pulls from every system you use and gives you a complete morning briefing on demand. Uh if you are an employee, if you are the CEO, um if you are just like a grunt worker, you still have to come in uh every day and know what's going on in the business generally and what you actually have to achieve.

And that's what the coffee debrief is all about. Usually people will spend half an hour to an hour going through their email, making sure um they're checking up with things, maybe they fall down a rabbit hole and they don't get a full view of what's actually going on um inside of the company. So the way this works is that you run the coffee debrief skill and it goes it grabs from um whatever you have connected in this reach section. So, for example, for me in my own agency, I have a time tracker that literally tracks everything I do on my computer throughout the day, how much time I spend on it, what website I'm on.

I also make YouTube content like this. It'll grab my content analytics, tell me how my YouTube videos are performing. It'll check my calendar, tell me what meetings I have and how I need to prepare for them. Uh, it'll check my email, let me know if I have any urgent emails to catch up on or people I should follow up with.

Um, I have voice uh and sales tools. So, I run uh a bunch of voice agents specifically from uh what my agency used to be niched down to. And so, checking up with those clients, making sure that's all running smoothly. Um, meeting intelligence.

I have uh AI recorders on every meeting. And so, if I have a a second follow-up meeting uh that detected on my calendar, it also knows what happened in the last meeting automatically because it ran this debrief and it has full context of what I might need to prepare for it. It also has access to my CRM and my video scheduling letting me know like, hey, you should record this video uh today. And so it told me like today that I hadn't uh in 8 days record a video.

And so that's why I'm making this video today. Also shows me all my financials. So I know the financial standing of my company. And the trigger to do this is inside of Cloud Code.

I just say give me my coffee debrief. Uh it takes, you know, fiveish minutes. Um, it then delivers a telegram message and a Google doc and uh it mostly just returns information. Uh, and so I'm not saying it's sending out and automating everything I need to do for the entire workday.

Um, but this is the easiest way to feel the power of AI with full context. Uh, even though this isn't some crazy crazy automation um that saves you an entire day's work. Now, if you don't want to bother with all the setup and headache of actually implementing an AIOS and following the entire craft framework, feel free to book a call with me um at arose.ai/book. You can see that here.

And we can actually build this for you um if you want me and my team to work with you personally. But if you are willing to um go ahead and build it yourself, which I fully encourage, uh you do need a Claude code pro uh pro plan. Uh you can't use it on the free plan, unfortunately. And so um you can come here to claw.upgrade.

Uh you can get a plan here. And then uh what I recommend you do is you also uh click the link down below to join my free community and I will have uh the actual skill here of the coffee debrief. So you can just copy paste everything. You don't have to worry building it all yourself.

It will be on the top link here. If you're watching this video just as it releases, if you're watching this video later, just feel free to search up the title of this video and you will see the YouTube resource pop up. I link all the resources in there uh as you can see here uh in one of my previous videos. Now, I also highly recommend uh using Cloud Code via an IDE um which is an integrated development environment which is basically a way uh to actually visualize all the folders and all the code while you're using cloud code.

I personally use Google Anravity. You can download it here atravity.google.d download or/d download. When you download it, you should see a screen like this. Now, once you see this screen, what I want you to do is open a new folder. you can create a new folder.

So, I'm just going to name this one AI AIOS uh demo and then uh I will select it and you'll see it will open a new environment for me. Now, I do have a full tutorial on Google Anti-gravity and Claude Code. I know this can look a bit intimidating right now if you don't know what all this means. Uh so, you can go find those videos um on my channel.

I hopefully should also link them down below. But getting into this, you will um next step want to uh come here to extensions and download the cloud code for VS Code extension. Uh also a note, you can use any IDE if you prefer VS Code or cursor. Um I just prefer Google anti-gravity.

And you would install cloud code here. And then what that gives you access to is this little button here for Claude Code. Um which will actually open up Claude Code. And uh if you haven't if this is your first time setting it up, you will need to sign into Claude Code.

Uh so you would do that now. But once signed in, uh this is what your screen should look like. So next steps, you want to go grab you want to come to the YouTube resource where I will have linked um this coffee debrief uh GitHub repo. And please start it uh if you're here.

It helps me out and helps other people watching the video find it. And you can just copy this link up here. I've included all the information that you need. And once you have copied this up here, you could come back into Google Anti-gravity.

I'm going to um go back to the folders. I'm going to paste this here. And I'm going to tell it to set up the coffee debrief. Now, prior to this, I do recommend um exporting, you know, any chats you have with ChatVT or Claude um and throwing them in this folder as well. as well as uh you can come here create a new um context.md file and uh load in you know any files about your company or you know describe whatever information you need to about your company.

This will look different depending on what your company is, who you are, um what tools you store and so it's hard for me to actually give like a comprehensive overview of how to load context. just know uh you should air on the side of loading more context rather than less uh and only don't load things if like you're really really concerned about the legal or privacy reasons um regarding it. But um once you have handled that first layer uh you can just paste in this GitHub repo and then say can you set up this GitHub repo for me um make sure that uh you know it's set up according to the readme and then what I'm going to do is I'm going to go um onto plan mode.

Uh I always start every cloud code um instance or when I'm uh building something with plan mode. And let's let this run. Okay. So after thinking for a bit, uh it made a coffee debrief setup plan um where it will generate uh a daily morning uh briefing from uh a bunch of the different tools I actually use.

Now, if you see it mentioning tools in here that you don't use or don't want connected to here, um you want to mention that. now. And so I recommend reading through this, but um you uh it'll also mention, you know, what folder it wants to put that Google doc in. Um you know, if you have a content platform, um you know, some voice workspace names. So this will be really custom to your own business.

And uh one thing I also recommend adding that I'm doing right now is I'm having AI uh actually generate me a little contact doc. So, I'll come back when that's done and we can actually set up the skill together. So, um this new cloud chat just got done basically just scraping my website. So, I would include a lot more personal information uh if uh this wasn't a YouTube video I was making.

You can see this doc here um has a bunch of stuff about sort of my background uh you know my business, what it is, what we do, our pricing model. So, I'm going to download this here. Uh, and then what I'm going to do is I'm going to come uh throw this into uh anti-gravity uh right over here. And so now it has full context on this contact stock.

And so what I'm going to do is um in this debrief plan, I really don't want to set up all these tools right now. I just want to test it out. So I'm going to tell Claude what to do instead uh of this plan. Uh first I'm just going to mention um here I'm going to copy this path.

Am I not able to? There we go. Copy path. And I'm also going to say I don't want to set up all the tools you mentioned in the setup plan.

I just want to set up two. I want to connect um my Google Drive as well as um my notion content hub workspace. Are you able to do that for me? And so those are two tools that I think would be a really good example here um because they're heavily information based.

They don't share any sensitive info um like you know my private calls or my revenue numbers and they can really show off how it can grab a bunch of this information and actually run. Um once again uh I want to mention you know what the actual coffee debrief looks like and how hefty this is. Uh this is pretty much a full overview of my entire business day-to-day. Uh as well as you know something you can set up for your employees to get and it's cool because it's modular.

Um and the entire AI OS is because you know up at the the sea suite level you know they can have access to let's say you know everything the company uh has but you say you go down you know a level or two you have an employee down here maybe they only need like access to one thing you know maybe they're a developer so like having access to GitHub is really important for them but they don't need to be worrying about you know all the revenue numbers or they only need to know the revenue numbers for a specific project you can fully control, you know, the access people have in uh the AIOS and uh the coffee debrief skill.

So, coming back over here, uh we're going to wait until this is done and I'll show you uh when that's finished up. So, it just created uh a brand new plan. Let's read through it. Um you know, Google Drive uh in Notion and you know, creating the MV folder if you're not familiar.

Uh AEMV folder, we'll see it set up in a second. It's where you host all your private API keys and everything uh and so that you know they're completely safe. Um adding some MCP servers uh as well. And I actually will tell it I don't want to add the MCP servers but um and that has you know the manual steps we'll actually have to go through to grab all this information.

And so I'm going to tell it I don't want to set up MCP. Uh I just want the API connections. And then you can go ahead and run this. Now, the reason I don't want any MCP is very simple.

It really bloats up the context window of Cloud Code. If you're not familiar with what I'm talking about here, is that uh you know, as you jam more information in here, it gets a lot more difficult to actually um get better responses. And due to that, uh API calls just tend to be a lot more lightweight on the system overall than uh MCP. And so, uh, now, uh, the plan looks good.

There's no MTP in there. And so, I'm going to say, uh, yes and auto accept. And then I'm actually going to put this on bypass permissions, um, because I've honestly never had an issue with bypass permissions. And it runs a lot faster and doesn't keep bugging me and making me click uh, allow every time.

So, now you will see what this skill will actually do. Um after it sort of set up a bunch of this uh dotcloud folder with um some of the settings, it sets up this uh these environment variables where it says, you know, your name, what the output drive folder is, um content platform, uh notion database ID, uh some of the stuff, you know, it says content platform because that's just specific to me. So you will see the AI now prompts us to fill in this folder and it will fill in some of this stuff for us but we will have to go out and grab some specific uh API keys.

Now uh API keys if you're not familiar are basic basically um web authentication tokens. It's sort of like you know how you log in to Google or any other platform. You need a username and password. uh that's the same exact thought process, but this way you know something like Cloud Code is able to use that software now. And so the great thing about the AIOS and Cloud Code is it can connect to literally any platform that has an API which is pretty much every tool you use um with very few exceptions.

So now I'm going to go through and uh I'll bring you along to grab some of these APIs. So first uh it gives us instructions to go to console.cloud.google.com. at google.com. So we can just click this here and then what you want to do is just follow along. So it wants us to create a project um and enable the Google Drive API.

So um coming in here, let's see how we uh create a project. Um oh I already have a project up here, but I can create a new one. Um I'll just call this one uh demo. Uh I will then have uh you have to connect a billing account if you don't have one.

I already have one. So, I'll just create this here. And then boom, I have a project. Now, um I want to connect it to uh or enable the Google Drive API.

So, if I go here, I should be able to just search up Google Drive API. And I may already have it enabled. Um and I do, but if I didn't, I would just have to put API uh or hit enable here. um there would be a button there and then let's move on to the next step which is to create some OOTH credentials which are basically the signin authentication credentials um of the desktop app type and then we have to copy the client ID uh and put it into thev file. So, let me go here and um go back and search up OOTH and then hit uh credentials is what we would be looking for.

And um honestly, Google's is a bit more in-depth than a lot of other ones. And so, we want to create uh an ooth client ID. Our application type is desktop. It told us in there.

And then this name will just be uh demo AIOS. I'll hit create. And then I will have to um grab does it want both of those? Uh the client ID and secret.

So I will copy this client ID uh put it in here. And then I will also grab this client secret um and paste it in here uh as well. And then it wants me to run all these to get uh a Google refresh token. So I'm just going to ask if it can do this uh for me.

Can you do this for me? Uh I don't want to run these myself. A lot of the times the AI can actually go out and do things uh for you like this. And so it's always worth trying.

If it does mess up then you can just do it manually. And if it's if you're confused, um you can, you know, query it again and then, uh it will sort of give you more in-depth instructions. Uh overall, I go back and forth with the AI all the time. And so, uh it says it, uh can run the pip install, but the oath flow has two parts that require um my Google credential credentials.

Um so, let's see how this runs. So, it's telling me uh to do something I've already done, and that's because I just didn't save the MMV. So, I'll tell it check again. Uh I forgot to save the MV, but the credentials are inside of there.

So, you can see um it opened this automatically. I'm going to sign in with my email here. Make sure to choose um whatever your uh email is. Going to hit continue even though this isn't a verified app.

Going to hit continue. And uh it says this flow is completed. I can close this window. If I come back in here, uh it says it got the refresh token and it should throw it in the MV.

Uh and that it did. So now our Google Drive is fully connected to the AIOS um and ready for the coffee debrief skill. Now I just need to connect my notion. Uh this is a bit more simple.

Um it I just have to go to this link then new integration and then copy the po uh the token uh right here. So let me open this. Uh I already have a connection but um I'll create a new one and I'll say demo AIOS um installable in my notion. Hit create and then configure connection settings and then um I will see does it need uh I'll edit its access.

I can give it access to my content hub is what I want to give it access to here in notion. Um and then configuring it. Uh I will show this installation access token. Copy it.

Come back over here and then uh paste it in here. And so um that should be the token they need. I will hit CtrlS to save that. So if you ever see the the dot up there, that means it's not saved.

I'll delete that. And then uh you know, should be saved now. I'll say all right. I put my notion key in there. it'll be all good.

So, the cool thing it just did there is I only gave it the API key and not the database ID, but it was able to go in and find that itself. So, um, Cloud Code is very intuitive. And so, now it should be able to run my coffee debrief, run my full coffee debrief, um, create the Google doc, and output the link to the Google doc here in this chat. And now I'll wait for this to get done, and I'll come back to you with the results.

So, the AI just finished up. Um you can see the coffee debrief here. Uh once you click this, it will take you actually over to this Google Drive um where it uploaded it. And at first I actually created this HTML file.

So I had to open it with Google Docs. And this was the output. Now uh again it wasn't blacked out. I blacked out um a bunch of the client information here uh so that you know that wouldn't be shared.

But um you can see I had three critical tasks for the day. One was uh to actually film this video that you're watching. um as well as review and approve um some pending time uh rise time entries. So rise is like the um you know tracker I have on my computer for my productivity. It'll show me my schedule, all my emails and it actually gave me this really interesting format.

Uh as you saw before like my actual coffee debriefs don't look like this. They are much more plain and less fun. So I actually sort of prefer this although for some it might be distracting. It also gave me priority emails. um some uh you know not less prioritized emails, you know, tier 2 emails and uh the list goes on of all the things.

Um you know, sort of my business overview, a bunch of different information uh there. Looks like it actually went and grabbed some different API keys that I didn't explicitly give it access to because it has all like my rise. It actually has that all that data for my email. So, um that's pretty impressive uh that it it scraped through my email that far to find all the time entries.

But you can see here you can run this every morning. It'll create a new debrief. So you know tomorrow it'll be 56 and the day after that 57. Um it'll be you know different every time.

Uh specifically obviously uh you know the bigger the company the more you might have to debrief on. And those will all be stored um in the Google uh folder that you know the AI creates. And so um moving on from that I want to uh really zoom out again and look at what we just did. Um because the coffee debrief is the same architecture we talked about at the beginning of this video that massive companies like Coinbase and Block are scaling just shrunk down to one person.

Uh but that could obviously be scaled you know across an organization and every single employee could um be given or instructed to build their own coffee debrief that they can then customize themselves. Um and the three layers you saw in action of the AIOS. Um so coming back to here are is the context uh the reach and the actions. And uh diving deeper on that the context layer meaning it knows who you are, what you're building uh your goals and how you work.

Again I know I didn't show off putting a lot of context in here. Once I build the full course I'll go a lot more in depth into like what I actually recommend putting in there. And then obviously when I'm able to work with a client one-on-one and everything's private, we have an NDA signed. Um, you know, I can throw in all their information in there.

It's just difficult to show all of that on YouTube. Then you have the reach layer. Um, so we plugged it into the email. Uh, you know, it also has access to my Google calendar, but um, you know, my real one also has access to my time tracker, content analytics, CRM, uh, the voice tools I use to build my agents transcripts, um, all the project tracking and video tracking as well as my financials.

And then in terms of the actions layer, uh this was pretty light. Uh it really just generates a Google doc and mine personally sends me a telegram message as sort of like a notification summary. And so uh it's really a loop and everything just compounds. Uh and then you can obviously build on top of this actions layer uh by telling it to like reply to that specific email or help you draft email etc.

And so, um, in terms of comparing this directly to what Coinbase is doing, the coffee debrief is really just one person. Uh, it's about 10 integrations. Um, you can add more, it just will take longer for it to run. Uh, as well as it is only one daily output.

Uh, and it gives you sort of full context on your world, your work world. But um you know these AI native pods in Coinbase you know are thousands of people, hundreds of systems, continuous output and you know they're running a an AI with full business context. And so where does this go next? So the coffee debrief which is basically an AI with context on your specific world um and the craft layers with C R and a little bit of A.

Uh you just did this if you followed along which I highly recommend. Um you already did this. you you went into the community, you grabbed the skill, you signed up for Cloud Code or logged into Claude Code, you set this up. Uh, and the system already knows who you are and what you're building. And so that's really the foundation of all of this.

Uh, and I like to add the skill on top of it because it makes it a lot more concrete. Now, you can add some flow, um, some AI that runs on its own, which is a craft layer F. We can start to schedule cloud routines, um, give it phone first access via Telegram, uh, because the system does work without you asking at that point. You can also run crown jobs.

Again, in my full course, I will cover that. Um, as well as uh I'm really exploring this a lot more, finding good ways to to do the F stage because a lot of the ways now are just a bit clunky. Um, or I don't quite trust it or they're really expensive for people just beginning. And then lastly, tuning um which is AI that gets smarter the longer it runs where, you know, you can run weekly, you know, audits or level ups on the actual AI.

That's something Nate Herk um in his video uh explained that I thought that was a great idea as well as uh you can have an actual change log of it so that the system will like be introspective and uh you know better itself every time based on you know your feedback. So the end state is a real operating system for your business or your life. And so this is what Coinbase is building at company scale as well as block. And so you can build the same thing at whatever your scale is today.

Whether you're a CEO of a company with a thousand people or you're a soloreneur or you're someone, you know, this is the first video you've watched on AI and you're just looking to become more productive. And so, thank you so much for watching. Um, once again, make sure to um comment down below if you enjoyed and join my uh community for more information. I'll see you in the next video.