We're discussing things with different mindsets on this forum. I'm sure your mindset isn't the same as mine.
I can do 2 years with nothing but learning to make money in 3rd year. Now I'm gonna be a lot richer in year 5 than someone that used a simple method and wasn't planning long term.
Btw algebra or any other thing is whatever you make out of it. Algebra can be a money making tool. And it'd be not a bad one probably if you could motivate people and get them to use it in practical areas, like AI for example.
These are good points you made. I recommend anyone wanting to take AI seriously learn multivariable calculus and linear algebra. There's tonnes of good resources out there, search MIT open courseware for a good multivariable calculus and linear algebra course, devote 3 months to each full time it should be enough (6 months total, full time). If you need to refresh your math first up to a high school level, use khan academy before venturing into those courses (you should be confident with single variable calculus first, understand intuitively how derivatives and integrals work, and why they're the inverse of one another. Understand the fundamental theorem of calculus. Don't just memorize formulae - terrible stuff that, you will learn nothing). After understanding these two fields (linear algebra & multivariable calculus), learn basic machine learning and how neural networks work. Learn how they work on an intuitive level such that you can create your own NNs without any libraries (no tensorflow, keras, pytorch etc, you may use numpy only). Look into Andrew Ng's original full python courses (not the recent dumbed down ones, but the 5 course series) - he goes over creating your own NNs from scratch using calculus & linear algebra, which will help you gain invaluable intuitions. Don't skip the maths and go straight into Andrew Ngs courses. You will learn nothing. Learn the math first so you can understand the "why" behind the code. Remember you're doing these courses not to write a production level AI, but to gain intuition as to how AI works.
Now you should be able to understand the basics of how modern NNs work, and from there, build up your understanding to more modern architectures like LLMs. You will gain great intuitions, especially when it comes to hyperparameter tuning, and why you might be changing different variables. Otherwise, you'll waste years just randomly changing hyperparameters and asking people what to do, without having any intuition. You should also be able to somewhat read research papers to gain ideas of new techniques, even if not understanding most of a paper, you should understand some of it.
In terms of the monetary aspects of AI. 1445990 - you mention you fine tuned AIs and still didn't find them useful for content generation. Really? I find that very hard to believe, unless incompetence was involved. Learn about overfitting and underfitting, understand how and why they occur from a mathematical perspective. I have a website ranking well in google with 100% AI generated fine tuned content, that is cited by other top sites. I perform no SEO whatsoever as I trust Google is now sufficiently advanced at ranking good websites correctly without the need for manipulation. I even label it as AI content. Google doesn't care, because it's designed for humans, and it's good quality content. Stop fooling search engines to make money, and start generating content that humans want. This is what Google cares about, they couldn't care less if an AI made the content, they only care the content isn't crap and is good quality. So stop passing it through "AI detectors". No one gives a shit about them. Google doesn't despite what people say. Blackhat SEO is becoming less relevant. I used to do all sorts of trickery to get high page rank backlinks. These days I don't bother with any of that. It's no longer 2010. Train your AI to write for humans, not bots. If you're doing keyword stuffing and other outdated techniques, update your knowledge.
Google is getting closer to fulfilling its original mandate: ranking sites that are good for humans. You can keep pushing against the grain with new tricks that will eventually become redundant, or you can just embrace their goal and write for humans. Let me tell you, the day Google penalizes good quality AI content is the day they're no longer a relevant search engine. People will always use a search engine that delivers them useful results, not a search engine that's biased to an AI, human, a dog or whatever other species created the content. So long as the content is good, that is all that matters. It doesn't matter how it was created.
