I always try to cross the terrain that my competition won't. To be fearful when they are greedy and greedy when they are fearful.
Right now tech has bad prospects and every other college student's parents are saying the same thing as yours.
The truth is that LLMs are great for shallow, simple work that's been done before. It's dangerously imprecise so not wise to use for medical, banking, aerospace, finance, STEM. If X fails to load some feeds, who cares, but if a laser eye surgery machine isn't 100% correct it's not shipping.
So this isn't advice, just a framework that guides my decisions. I'm planning to keep doing this career until retirement unless something else changes.