> Which with cheap models like Jev as spam classifiers, this can now become a reality.

With LLMs on the other side trying to make spam less detectable by cheap models like Jev?

It's funny when people assume benefits from some tech go one way and baddies will surely never use it...