It might be better in its own "Ask HN" post.

If you want to get into AI topics: start learning about LLMs, embeddings, vector stores, datasets, fine-tuning, evals, and how to use CLI dev tools like Claude Code.

Andre Ng helped build a great free resource with courses: https://www.deeplearning.ai/ (some are sponsored but still free)

I think a quicker pivot (job wise) would likely be Python web development. They already know Python. Then layer on LLMs when doing that job (with more "runway")