Self driving interview

In honor of yesterday’s nonspecific point in the gradual arrival of self-driving cars, an interview with myself. Interviewer: It sounds like you’re pretty excited about self-driving cars. Weren’t you just saying that unemployment from AI is on some kind of very overlapping continuum with extinction from AI? Isn’t rooting for self-driving cars rooting for AI unemployment here, and thus extinction? Katja: Hmm. Well first I should say, I’m actually fairly neutral on unemployment in general from technology. If technology makes it overall easier to produce what we want, but empowers some people over others, that change in power might be a downside (or not), but if so, it’s one I’m inclined to solve with direct redistribution rather than having the people who would be disempowered do unnecessary busywork to ‘earn’ their living. Interviewer: Ok, so you think AI unemployment is different? Katja: Yes, because it involves the disempowerment of humans in general in favor of non-people entities whose empowerment has a decent chance of spelling our ruin. Doing things the hard way to avoid that happening isn’t busywork, it’s very valuable. I don’t usually want to take sides between different humans systematically—society seems probably best served by letting the most effective production methods win out in most cases. But sometimes there are entities who produce things efficiently, and you still shouldn’t trade with them because it empowers them. It’s a lot like not trading with Nazis (broadly—I’m not saying AI entities are evil in the same way, just that their empowerment has a good chance of leading to genocide or omnicide). Interviewer: Ok, but aren’t self-driving cars AI? Katja: Yes, but the class of entities I don’t want to empower isn’t ‘AI’ really—it’s more like ‘AI agents’. Though also, the processes that are creating them, such as LLM companies, which complicates things. Self-driving cars are narrow and not much like entities that can be empowered. And my understanding i