Got blocked on X for this AI take

You guys probably don't know this, but I'm a complete nobody online. Shocking, I know. I don't get a ton of engagement, let alone a ton of haters. But recently, I've been lucky enough to receive this lovely comment from this lovely middle-aged man. And the reason he's so triggered is because of some of my opinions on AI. And we're going to go through all of them in this video. And I'm going to make my stance clear. And in the process, maybe make a few hundred more people hate my guts for existing. And then you can go down in the comments right below and let me know your thoughts, whether you agree with me, whether you disagree with me, whether you want to know my location, come up and just punch me in the face. All of that's allowed. Okay, so Huston's Rules of Living Part 1. If I paid for something, I get to use it however I want. Obviously within the constraints of the law and everything, but I get to do what I want with it. Now, here's what this guy said. I'm going to read out the entire thing to you. Then I'm going to read my reply and everything. And then we'll, you know, see what I think about it. under-discussed aspect of AI. As soon as smart people realize their IP could be hoovered up by an LLM and redistributed to the world for free without them being reimbursed for it and potentially without them even being credited for it, they gave it all their best advice. Emphasis on being reimbursed for it because this part is going to be important in just a bit. Paid sub stacks, school groups, masterminds, and because of that, today's LLMs are completely incapable of delivering the type of cutting-edge strategic or tactical info that actually matters. If you know how to use these tools, you would not be saying something like that. There's ways to get over this entire hurdle, but I digress. Instead, they're regurgitating boilerplate advice that's been around for decades, combined with newer information from people not important enough to actually gate their content. Okay, so the whole issue over here is basically that people who have paid sub stacks and paid blogs and all that other stuff, course sellers community sellers and stuff have just gated all of their advice behind the paywall because LLMs can train on them without necessarily paying them for it And now here what I said Buy access to the paid hidden content, save it on your local machine, let Claude or whatever agent you're using train itself on all of it so that you can use that LLM to ask questions about the content, to do all the other things. I've done this with a few subchecks myself, not feeling guilty about it. yeah not feeling guilty about it you're a piece of shit and i hope you stay poor yeah that's what i got from this guy so yeah my reply probably hurt him a lot and to be fair i'm not saying go steal someone's content and that's not what i've done over here i'm saying go and buy the thing buy the course pay full price for it get it legally make sure the other person gets all the money they want for it, you know, that they're fairly compensated for their work. And then when you own that content, go ahead and do whatever you want with it, because you've already paid for it, right? Now you own it. And by the way, he blocked me. So yeah, must have touched a pretty big nerve over there. Look, if you get so triggered by a bunch of random words popping up on your screen that you have to go and press a big black block button, you've got bigger issues to deal with, my friend. And you should certainly not be on Twitter of all places, because half the people on this platform are just professional rage baiters and I'm not even counting myself in them. Anyway, let's go back. Now, I don't believe in stealing someone's content, okay? But I do believe that if I paid for it, I get to do whatever I want with it. That doesn't mean I'm going to go and, you know, resell it or just claim it as my own or something or some other bullshit like that, like some other people do. No. But in the age of AI, when you can have models that can act as learning partners and learning tutors, why wouldn't you let Claude help you to learn a new topic? And if information about that topic is only available in a paid Substack or some private course, but you paid for the entire thing, you paid for access to that knowledge why can you use Cloud or use DeepSeq or some other LLM to help you understand that topic better In fact I seen many Substack authors paid authors, create their own, you know, personalized AI chatbots trained on their own content and then selling those chatbots for a premium price. So clearly it's fine to train a chatbot on their content. I mean, they're doing it themselves. It's just that they're asking a premium price for all the work that they did, for the setup that they're providing, for the technical details that they handled, as in creating the entire chatbot for you, which is, you know, fair enough. You handled a lot of the work for me. But if you're an engineer and you can build a rag chatbot or some other agent trained on some, you know, external knowledge base, if you're an engineer and you can build that thing yourself, you don't have to pay that premium price. same reason why you don't pay for as many SaaS apps anymore because you can just wipe code your own solution using AI that works great for you and you don't have to pay for those apps anymore right you've already paid for all the content you've paid for access to that private content so now you can go and learn it however you want or you can build applications on top of it just obviously don't go and you know resell it or try to claim it as your own or break some copyright law or something, obviously not, don't do that. But using AI to help you learn a topic or ask better questions about the topic, I don't see anything wrong with that. And by the way, if you like this take, go ahead and press that like button, go ahead and press that subscribe button, and you're going to hear way more than this over the next coming days and weeks. Now, in the very near future, AI is going to be very, very heavily used in learning and education. Many course authors are already including AI chatbots and AI prompts in their courses, in their lectures, in their articles, whatever. So in essence, what you're paying for is one, the course content and also a custom chatbot trained on that course content So clearly this is the future of education and with how capable and how intelligent these models are now you leaving a lot of potential on the table if you not using them to help you learn new things Like you should absolutely be allowed to copy a long article paste it in the chat GPT, and ask it to explain some things to you, ask it to make some explainer video for you to follow along with, or ask it to explain specific concepts to you in like a fourth graders comprehension level or something or whatever, right? I see no reason why we should not be allowed to do that. By the way, there's this other guy on Twitter who actually liked my reply and wanted to learn how to implement something like this. I couldn't actually reply to him because I'm blocked by the original poster. So I had to quote him. And if you're wondering how to achieve the same result, the answer is pretty simple. You just have a bunch of rag embeddings on that content. Or you can have a tool like grep or some other tool like find. And you can have a cloud MD in that folder where you save all that content. And these are just two of the ways that I came up with on the spot. But obviously, if you want me to make a full video on this, how to, you know, train cloud or some other LLM like chat jibbiti or something on a private course content or some local knowledge base that you've got and how to ask questions and all that other stuff, how to basically train the LLM to answer questions about that particular knowledge base. If you want me to make a dedicated video on that, leave a comment down below, I'll do it. But these are just two ways that I came up with on the spot just to respond to this guy, right? You can have a rag embeddings chatbot, you can have a basic, you know, cloud.md describing the entire knowledge base, along with grep and other bash tools so that cloud can go and search across the entire knowledge base itself for ways to, you know, for things that it's looking for. But yeah, the main thing is just to the content stored somewhere cloud can easily read and find it. Doesn't have to be a local folder on your machine but that's usually the easiest way to pull this off. And a simple cloud.md file explaining how the directory works, where everything is, how it's structured, works great. Anyway, let me know who's right, who's wrong in this twitter drama down in the comments below and thanks for watching.