Ask HN: How to get back into AI in 2025?

I feel completely left behind. Last time I was up to date, it was FastText and fine-tuning the last layer of a neural network. Now it's BERT, LLMs, DeepSeek, Gemini Flash, RAG, LoRA, vLLM, LangChain, agents, and whatever else I'm missing.

Where do I start to get back in? What’s the best way to update my knowledge without drowning in hype?