Before the advent of LLMs... the term machine learning (ML) was more widely used than the term AI. My main job is as a firmware engineer, and my main task is bring-up on Bare Metal SoC, so I don't actually actively use LLMs in my work that much now. Nevertheless, it's been almost 10 years since I first started studying AI, back in 2017.
At that time, LLMs hadn't come out yet (or maybe they had... I don't know when the research papers that form the basis of LLMs came out..ㅎㅎ) When people said AI, they were mainly talking about the implementation of artificial neural networks. It was a time when deep learning was starting to show real commercial results.
I didn't even know the mathematical theories (or rather, I had forgotten them all...) so I bought and read a linear algebra book. I studied up to the part about implementing the multi-layer perceptron structure of deep learning, but gradually lost interest and got busy with my main job, so I stopped studying after a few months.
Looking at the related books that were stuck on my bookshelf and half-read, I wonder what situation I would be in now if I had studied more seriously back then and tried to make a career change. Of course, it's been pretty good to continue with my main job so far, but with all the news about LLMs like Claude, GPT, Gemini pouring out every day, I feel a little uneasy.
