Okay, this might be one of the coolest things that you can do with AI right now and almost nobody's actually doing this. You can run a real genuinely smart AI completely free on the computer that you already own. You don't need a subscription. You don't have any limits. You don't need loginins and you have no credits just running out. You can have local AI writing for you, answering your questions, and getting real work done. And the best part is it is completely private and it runs 100% offline. You can legitimately rip your Wi-Fi out of the wall and it will just keep going and nothing that you type will ever leave your computer. And I don't care if you've never downloaded an AI model before a day in your life or you're not even sure that your laptop can handle this. By the end of this video, you are going to know exactly which models are worth running and how to have your own private AI up and running today for completely free. By the way, I am not an engineer. I don't write code. I just run AI businesses. So, I'm coming at this probably way closer to you than the people who actually typically make these videos. And I went down this whole rabbit hole so you do not have to. So, in this video, I'm going to be covering what Local AI actually is in plain English, why it suddenly got good, who it's actually for, and who should just stick with Chat GPT and the exact models worth running for you and your business, and what hardware you need, and how to be installing all of this. And at the end of this, I'll just show some demos of running all of this completely offline. It is genuinely insane. Okay, so the normal way that you have been using AI, most likely through Chai GBT, Claude, Gemini, with all of those, the actual AI isn't on your phone and it's not on your laptop. Instead, it's just sitting in a giant data center somewhere. And when you type a message, your message is then traveling to that data center. So, the AI is going to be thinking and the answer it travels back to you. So, you're basically renting time on a brain that lives in somebody else's building. You do not own it. However, local AI just flips that. So instead of renting the brain, you download it onto your computer, onto your machine, and then it just runs right there on your machine for you. So the thinking, it happens on all of your proprietary hardware. So your message, it's not going anywhere. There's no company that's in the middle reading all of your information, all of your messages, and there's no meter running. And the way that this is even possible is something called open weights. So when a company is training an AI model, the end result, it's basically just one big file. And that file, this is the model. So some companies they keep that file locked up and they only really let you reach it through their application but a lot of them and some of the best ones now just release the file practically for anybody to downloading for free usually on GitHub. Those are open weights. That is what I mean. So open weight that just means the actual brain is going to be yours to keep. So you download it once and from then on it's just going to be a file on your computer just like a movie that you downloaded instead of a movie that you're streaming. Right? And that one difference. So, renting the brain versus actually owning it, that is what the whole rest of this video is going to be built on. Well, there's three real reasons, and I'll be pretty honest about each one. The first one is that the free models finally got good and are getting better. So, for years, these open models that you could download, they just weren't good enough to really bother with. So, most people, they didn't bother with them. But that has been changing very fast. So, these strongest open models right now, one of them, they're called GLM 5.2. I just recently made a video on it, and it's surprisingly close to the top paid models. on a lot of different tasks. Really quick before I continue, I do want to preface because you're going to be hearing it constantly in this video and you know outside of it as well. And it's actually very simple, but people they describe a model by its number of parameters like 7 billion or maybe even 70 billion. And parameters, they are just the model's brain cells. So they're like the little settings inside it that are just holding everything that it had learned. So more parameters that generally means a smarter model, but also it's going to be a heavier one that needs a bigger computer to run on. That's the whole idea with that. So when I say GLM 5.2 is 740 billion parameters, I'm just saying that it is an enormous brain, which is exactly why it is so capable and also why, spoiler, you and I cannot run that particular one at home right now. But we'll get to the ones that you can. Now, the second reason is the one that I actually think matters most, and it's just about control. So when you use a closed model like Chat GPT or Claude Fable or Opus 

The number one most common question I've been getting lately is about how I have my AI operating system organized. Questions about how I should be routing files and how I should have wikis organized and folders organized and where do I put client projects and so many questions around the organization of it. Because the problem is if you don't have it organized, then your agent is going to start to hallucinate things not only to you, but potentially in your skills and when you're having it build automations and stuff like that that could get pretty bad. So today I want to talk to you guys about these five different tricks that I've been using that have helped me keep my AI operating system super accurate, super upto-date, and it allows me to add more and more data every week without sacrificing quality or memory. So let's not waste any time and just get straight into today's video. All right, so here is an older version of my Herku project that I pulled in just to show you guys a quick skill that I built that I'll be giving you all for free. So look at this. I'm going to go in here and I'm going to run this OS audit skill. Now what's going to happen here is it's going to look through my entire project. It's going to read everything. It's going to look at all of the routing rules, and it's going to tell me all of the areas where there are things that are weak, where I need to make some improvements, where I need to update some data, that kind of stuff. Essentially, the deliverable of this audit is an audit. It says, "Hey, here are 10 things I noticed. Here are the 10 fixes. Do you want me to do them? Yes or no." It won't actually do anything yet. This is just kind of an exploratory phase. It will also create this folder at the root of your project if you don't have one. If you do, it's called audits. and then it will just chuck in a markdown file that tells you what it found and what the fixes are. And by the way, if you guys want to get this skill for free, just go to my free school community. Link for that is down in the description. Come into here, click on classroom, click on all YouTube resources, and you'll find it in there. So, while this is running, what I want to do real quick is talk about why this is so important, and then we'll come back and we'll look at the actual audit. So, what I want to talk about today before we get into these five tips are the different methods of context failures. So, four different failure modes of context and then two different context types. So understanding these six different things is going to make the audit make more sense and it's going to just help you in your day-to-day when you're using your AIOS way more. So the first thing is the four failure modes, poisoning, bloat, confusion, and clash. Basically meaning when your agent tells you something that you know is incorrect or maybe you don't even know, maybe you find out later that it's incorrect. When it makes a mistake because of the context, there are these four reasons. So let's start with poisoning. Poisoning means you have a false fact somewhere in the context. So imagine this is the context. Imagine this is a false fact that gets dropped in amongst these green facts which are the right ones. The agent looks in there and it will display that back to you or put it in the email to the customer or whatever it is because it was in the context. So the agent inherently is not going to intentionally lie. It's probably going to grab something and then use that as data. But the problem is the data set was poisoned with an incorrect fact. Now luckily poisoning is the easiest one to fix because basically it's just a matter of having some verification. So making sure that it fact checks everything with a web search or maybe two web searches or fact checks and cross- checks across your live database or maybe if it's not 100% confident it has to just human in the loop. You know what I mean? So poisoning is the easiest to fix. Now let's take a look at bloat. Bloat is when there's so much stuff. There's just way too much data. And this is what I think a lot of you guys start to feel as you scale up your AIOS's and you start to use them more and more. Now, this one is tough because as you can see, we all know about context rot as far as the window, but we also know about the idea of needle in the haystack, right? The agent is going to look at the data that it's currently loaded in in order to make some sort of decision. And if it has way too much to look at, then it's going to be really hard for it to actually pull out what's relevant and what's not. And there's just going to be some stuff that bleeds in, and you probably don't want it to. The tough thing about bloat is that it's a little bit harder to fix. But when I talk about expertise versus situational, that's something that's really going to help us out here. So just hold on to that thought for a sec. So anyways, that's bloat. Then we have confusion. This is where there

이 병화리는 미쳤습니다. AGHD 때문에 매일 할 일을 까먹는 사용자가 바이브 코딩으로 미친 프로그램을 만들었는데요. 일정을 등록해 두면 모니터에 포스티이 생기고 시간이 다가올수록 폭탄이 생깁니다. 시간 안에 일정을 끝내지 못하면 병아리가 나와서 사이렌을 울리는데 게다가 노션과 연동도 되니 투드리스트 관리하는 학생들에겐 도움이 될 겁니다. 본인도 모르게 쌓이는 찌꺼기 시간을 모두 없애 보고 싶다면 댓글에 맥스 남겨 주세요. 현재 3,000원에 평생 사용 가능한 링크 보내 드릴게요.