For a long time, I ignored AI and focused only on software, product engineering, and finance/business. Seeing the recent breakthroughs in AI, because of Large Language Models (LLMs), has made me a believer. The AI revolution has begun and will probably happen in my lifetime. I would hate to be stuck sucking my thumb.
That's why I decided to dedicate 60 days of my life to learning about AI as an experiment and see what's happening and whether I like working on it. Since I had time-boxed it to 60 days, I tried to cover a lot of ground without going into too much depth and coming up to speed with what's currently happening with LLMs.
I started with the fast.ai to briefly explain what happened. I tried some simple Deep Learning and Machine Learning techniques. Learned to implement the basics in Numpy and Pytorch. Then, I quickly moved to LLMs, RAG and fine-tuning.
I have always believed in finding a problem to solve to learn something, and it's always better to solve your problems first. Since, I am also into investing; I did some related things. First, I collected all annual reports and quarterly reports and tried to see if I could use LLMs to determine what the management was doing and whether they were doing what they said. Surprisingly, LLMs can do this exceptionally well. I spent some time collecting data, but it was worth it. Second, I want to talk with Warren Buffett and Charlie Munger, but I don't think I will get that chance. So I collected all of their publicly available talks, writings, etc and I tried out RAG and fine-tuning. This became a fantastic product which I launched.
I read the Hands-on Large Language models by Jay Alammar and Maarten Grootendorst which is a great book. I finally started reading and understanding research papers. It feels good to have made this progress.
While these 60 days of AI were a great way to start, I have miles to go. I am going into much more depth of understanding and implementing fundamentals while trying to build something useful using something I learn. Wish me luck!