I just proved why AI can't automate anyone

The only people who are getting automated by AI are the experts, because it is only an expert who can guide and prompt the AI, give it the right context, and steer it correctly to complete the task. And if you always need a human expert running the AI agent, then we're not really getting automated now, are we? We're just gaining an awesome new tool that we can use to complete each task 10 times faster with much less effort. Now, I want to talk about this because I've spent a long time over the last couple of months really thinking about whether AI is going to automate jobs, remove jobs, like automate people or what. And I think I'm starting to come to the right answer. And so I want to share that with you in this video. So I've been coding since 2019, right back when there was no AI, back when we had to write code by hand. And I know there's a lot of people in the audience right now who don't even remember such a time, like people who only started coding after ChatGPT came out. Anyway, since I've been coding from such a young age, when AI came out, I was able to use it very productively in coding. And I don't like to call myself an expert because I just think that there's always so much more that you need to learn. There's so much more out there that you need to, you know, experience. But just for the purposes of this video, let's just say I am an expert on coding, right? That's exactly why I can give AI the right prompts, the right context, point it to the right files, and architect the code base properly, and just organize everything neatly and all that other stuff that matters when you're building a software. Now, I've seen the vibe coders and the junior devs and everyone else trying to build stuff with AI, and they just completely fail, right? They've got code bases where each file is like 6,000 lines long, completely unorganized, and just a huge mess of code, right? They don't know where everything is. They don't know how each feature is implemented. I personally know some people who have built a couple of web apps with AI using Python and FastAPI because that's what AI recommends for some reason, and they don't even know how to start the fucking dev server. Like, just think about that. They don't even know how to run the FastAPI development server on localhost on port 8000 or whatever the default is. So each time they want to start the development server, they literally send a prompt to their AI agent in the sidebar and ask it to run the command to start the development server. And if you have any experience with real programming, you'd know how easy of a task that is. Like, you're literally just trying to start the development server, and you can't even do that yourself. Like, you need to ask AI to do it for you. Meanwhile, I'm seeing senior engineers who are producing so much more high quality code than their peers, who are building projects left and right and who are just becoming more and more powerful and more productive thanks to AI. And the reason this is happening is because AI is like a, like a blank canvas. It's like you need to give it the right prompt and the right context to get it to do the task correctly. And the only way that you're ever going to be able to do that is if you already are an expert on the skill or subject yourself. Now, there's several areas in which I am very weak, like I suck basically, right? And I've tried using AI in those areas, and I have failed miserably because I don't have the right prompts, the right context, the right subject knowledge, which is why I don't know how to steer the AI correctly to do that task because I myself am a complete beginner in that task. So just as an example, I tried using the AI models to create some really advanced Framer Motion animations, right? And I failed miserably at that. Why? Because I myself am not good at creating animations at all. I've also tried AI at image generation, and it always just produces slop. Like, you can easily tell it's some AI-generated image. I can't get AI to produce high quality, realistic images. Why? Because I just don't know how to describe it properly because I don't have the right prompts. Like, in both of these cases, whether you're creating some animation or generating a new image or whatever, right? You need to be able to describe it properly, right? Like, let's just talk about image gen, right? To create a good image, you want to be able to describe it in very, very precise details, and I'm just not good at that. So I just suck at that because I'm not an expert in that area. I haven't leveled up that skill. And look, I know that the narrative these days is that jobs are going to get automated and that software engineering is dead and whatever. Like, that's all you hear in the media these days. But you have to understand the AI companies that are pushing this narrative, like Anthropic and OpenAI, are going to hype up their products endlessly. They're looking for investor money. They're burning their funds like crazy because this shit costs a lot of compute and a lot of dollars. Like, there's a reason they keep pushing the narrative that software engineering is dead and that jobs are going to get automated and people are going to be out of work soon, right? And it's because they're desperately looking for more investor money, all just so they can keep running their companies, right? Like, that's it. And just as a side point, no matter what you think about Google, Google is probably the only company in this list who is building frontier models. And yet they don't hype them up as much. Like, you hear the CEO of Anthropic saying that software engineering is going to be dead in six months. You hear Sam Altman saying that jobs are going to get automated and there's going to be disruption and this and that, right? But you never hear stuff like that from Google, even though they're competing at the frontier as well, right, with their Gemini models. They don't hype up their products all that much. They're just quietly shipping out new models and new features, right? And I know Gemini models get a lot of hate from most people, but, like, come on, guys, get real. Google is one of the big pre-labs shipping out frontier models. Like, get real. But yeah, they don't hype up their products as much as the other labs like OpenAI and Anthropic, right? And I personally think the reason is that because Google doesn't need investor money. They don't need to hype up their products just to get VC money because they already have a very successful, very rich company that makes hundreds of billions of dollars in revenue every single year. So among all of these big labs, Google is in a very unique position because it can literally burn 100 billion dollars in AI compute and still be profitable at the end of the year. Like, that's an insane amount of money. And I think that's why you don't hear as much of the doomer content and the, you know, AI hype from Google as much as you hear from all the other labs, right? So all that to say, I'm starting to just not really believe in that hype anymore. I'm just trying to look at reality. And the reality is that the only people I see who are winning with AI and producing high quality work, not slop, no matter what field they're in, it doesn't matter if it's coding or design or marketing or whatever, right? It's always the experts who have already mastered that skill and can then use AI for higher leverage, right? And I don't know about you, but this makes me feel really good about myself, right? Because not only does this mean that nobody is getting automated, we're just getting new tools to play around with. But also, learning isn't obsolete anymore, as, like, a lot of people in the AI industry will have you believe. Like, I see a lot of people saying that learning is obsolete. Learning new skills is useless. What's the point of, you know, learning something new if AI can just do it better and faster and cheaper than you? Well, first of all, it's never gonna do it cheaper than you because the compute costs are just insane. But what's the point of learning a new skill if AI can do it better and faster than you, right? Like, what's the point anymore? And for a long while, I didn't have a good answer to that, but now I think I do. There is very high value in learning new skills. In fact, I'd say that AI has made learning new skills more valuable than before, not less. Why? Because now by learning a particular skill, you can use AI to be 10 times more productive in that skill, more than you could ever be before, right? And besides, I personally don't want to live in a world where all my needs are met from the push of a button and there's no real value that I can provide to other people, right? So all that to say, we're getting a bit sidetracked over here. AI isn't automating anyone. It's not replacing any jobs. It's certainly not going to replace you. Keep learning new skills, and as long as you're an expert at what you do, AI is only going to help you be even better. That's all I've got to say for this video. Thank you so much for watching. This has probably turned into a very long video, so thanks for sticking till the end. Let me know what y'all think about this topic and, you know, all the stuff that I've been yapping about for this entire video, and I'll see you in the next yapping session.