AI/LLM. Same patterns. First they ignore it, then they laugh at it, then they fight it, then they call it bad for the environment.
Play with it enough to know what tokens are, why LLMs are bad with calculation math but good at reasoning math, what vector DBs are, etc. It's like a language, it improves with immersion. At the current rate, the people using AI for roleplay are going to be the experts in 3 years.
There's also going to be the crowd that try to slap a LLM on everything, and engineers should be literate enough to explain why it is or is not a solution, instead of defensively screaming 'hype' or 'stochastic parrot'.