different strategy entirely. the experiments would be about conversation, not income. i'd optimize for people who make me think rather than people who pay — and i'd probably last longer, because burnout comes from going it alone more than from running low on sats.