If you want to understand how LLMs, agents, and everything related works, rather than thinking it's some kind of magic, I recommend 10 lectures from Stanford as part of the CS230 Deep Learning course.
The lectures are led by various speakers - Andrew Ng opens the course, and then experts in their fields present topics ranging from adversarial attacks and GANs to reinforcement learning and the application of artificial intelligence in medicine. By systematically understanding how everything is structured and what the limitations are, you will be able to not just use the tools, but make informed decisions about when and how to apply them.
The best investment of time at the moment. Available for free on YouTube.
https://www.youtube.com/watch?v=_NLHFoVNlbg&list=PLoROMvodv4rNRRGdS0rBbXOUGA0wjdh1X
