What's the hardest part of learning AI right now?

AI is moving incredibly fast.

Whether you're exploring LLMs, RAG systems, agents, fine-tuning, or traditional machine learning, what's the biggest challenge you're facing?

For me, one challenge is deciding what to focus on when new tools and frameworks appear every week.

Would love to hear how others are approaching it.