I'm working my way through the lessons and examples from the courses, learning a lot! And I was just thinking that at the time I'm writing this, and in my personal opinion, Google Gemini seems to be ahead of OpenAI in terms of cost effective and performance. Their latest models deliver incredible well at a fraction of the cost. So after cloning the cookbook repository to follow along with Dave's lessons, I started to change the code to use Gemini API using their Python SDK. It has been really useful for me for grasping even deeper all the knowledge because I had to rewrite the example code following a different API documentation to essentially achieve the same thing, and both are similar at they core but with small differences that made me have to think more carefully about each step. You could use OpenAI compatible bridge with Gemini, but I'd figured that doing the switch would be educational to say the least.