This is Part 2. I assume you've read Part 1: https://www.creativeworksofkno...
We're living in the age of AI, where you can utilize AI tools without understanding the mechanics behind them. However, this approach only maintains the status quo, and in the worst-case scenarios, you may remain average or even below average.
AI has the potential to elevate humanity as a whole, not just individuals. It provides an opportunity for those who are average or below average to rise above the norm. To seize this opportunity, you need to adapt quickly and efficiently to the changing environment.
Here are a few empirical tips from someone who has been in the field for a while. You can verify my credibility in Part 1.
Let's delve into what you need to understand before diving into the world of AI:
Acquire Programming Skills
This is a given. If you're questioning why, you might want to reconsider your journey into AI. Python should be your primary tool, unless you have a preference for another coding language for some reason.
Relying solely on GPTs to solve all your coding problems might seem tempting, but reconsider. Such an approach could lead to stagnation, keeping you at an average level or even below. For instance, you should be able to understand why GPTs produce certain lines of code and how to best apply them to your projects.
As a professional English-Korean translator and interpreter, I engage with GPTs on a level of camaraderie, even when producing text. GPTs can sometimes generate nuanced text, and without the ability to discern the subtleties, you run the risk of merely copying GPT text without truly understanding it. This risk extends to other use cases such as coding, image and video generation, and more. For instance, without a basic understanding of human anatomy, you might not be able to identify inaccuracies in AI-generated images of human figures beyond incorrect number of fingers.
Always ensure you have a solid grasp of the fundamentals.
Understand Basic Math
Math may not be everyone's favorite subject, unless you're a coding genius with a natural affinity for it, like Linus. However, you don't need to master all aspects of math, just the basics. For any programming, a fundamental understanding of arithmetic is essential.
In the field of AI, modern linear algebra is of paramount importance. Have you ever wondered why GPUs are as prized as gold in the AI era? It's due to vectorized computation, a concept deeply rooted in modern linear algebra. I use the term 'modern' to distinguish it from 'traditional' linear algebra, which leans more towards theory and is less applicable in the context of AI. While I'm not against a deeper understanding of theoretical foundations, practically speaking, the extent of modern linear algebra you encounter in deep learning is usually sufficient. It's a good enough checkpoint from which you can always dive further.
Another often-dreaded math concept is calculus. AI uses it to optimize training and prediction accuracy through algorithms like gradient descent, most of which are based on differentials. You should familiarize yourself with these basic concepts, not necessarily the more complex ones. Interestingly, one of the most efficient methods, the Newton method, was developed by Isaac Newton, one of the creators of calculus.
By learning these concepts, you'll make a quantum leap into a realm above the average.
Learn Statistics
As I mentioned in Part 1, statistically speaking, approximately 68% of coders fall within one standard deviation from the mean, categorizing them as average.
Do you fully comprehend this statement? Understanding statistical concepts is not only crucial in AI but also in everyday life. Essentially, everything I'm sharing in this essay could be seen as essential life hacks for a more successful and fulfilling life.
Most of today's AI models can be fundamentally characterized as highly optimized and sophisticated systems rooted in statistical methods. This description remains accurate until further advancements in AI algorithms emerge. Interestingly, if we consider the concept of a statistically optimized engine, our universe could arguably be seen as the most advanced example of this. However, that's a complex topic best saved for another discussion.
Essentially, aim to become a data scientist. The field of AI inherently assumes this role.
No matter what path you choose in life, trust me on this - invest time in learning statistics as much as possible. You'll be grateful for it later.
Broaden Your Horizons
My background is in linguistics, with a major in English and English-Korean translation/interpretation at graduate school. Additionally, I'm a self-taught programmer with nearly 40 years of experience. My proficiency varies across languages, but I have a solid understanding of Korean (native), English (expert), Japanese, Chinese, and French. Given the right mindset, I believe I can learn to comprehend any language, be it human or machine. This skillset significantly aids in handling Large Language Models (LLMs).
Drawing is a hobby of mine, and the skills I've acquired in this area greatly assist in handling image generative models. A tech-savvy artist can better manage image generative AIs like Stable Diffusion, Leonardo, and Midjourney. It's intuitive.
Until recently, I ran my own YouTube channel for over a decade, using professional-grade hardware and software. I have professional experience in the broadcasting industry and also founded my own internet company, which I managed for about a decade. With more than two decades of experience, I consider myself an expert in audio and video. This expertise is beneficial when approaching multi-modal models.
The key takeaway is to broaden your horizons. Regardless of your innate abilities, you can surpass your limitations through diverse experiences and expanded knowledge.
Never rest on your laurels. Never consider yourself finished with learning. Perfection is unattainable, but continuous improvement is always within reach. Embrace your imperfections and strive for growth.
Remember: Perfection is the enemy of progress.
Keep an Open Mind
Don't hinder your growth with a narrow-minded worldview. In the age of AI, it's easy to fall into the trap of subscribing to doctrines propagated by so-called gurus and experts. However, history shows us that even these figures can make critical errors.
Equip yourself with knowledge and keep your mind open to new ideas and perspectives. No one is perfect, not even deities.
Please note that my statements may risk oversimplification and overgeneralization of the issues at hand. Everyone's personal experience will differ; I'm merely sharing the path I've taken so far. Ultimately, the path you choose to follow is entirely up to you.
How open-minded are you? You might surprise yourself.
Regularly review your life and strive for improvement.
And remember, always keep pushing forward. Just keep doing it, but do even more.
You'll know you're heading in the right direction when all your time and efforts start contributing to your happiness.
In conclusion, embarking on a journey into the world of AI is not a task to be taken lightly. It requires a solid foundation in programming, a grasp of essential mathematical concepts, an understanding of statistics, and a willingness to continually broaden your horizons and keep an open mind. As we navigate through the AI era, these skills are not just optional extras, but essential tools for success. Remember, the goal is not to achieve perfection, but to strive for continuous growth and improvement. By embracing this mindset, you'll be well-equipped to rise above the average and make significant strides in the field of AI. So, keep learning, keep questioning, and most importantly, keep moving forward. Your journey in AI promises to be a rewarding one.
Good luck!