I was always excited about the possibilities recent developments in LLM-powered workflows provide. At the same time, I was also wary about the larger implications of these things, from the whole theft problem to wondering “I wonder when this thing is going to become self-aware”.