My current attitude towards testnet points is: it's okay to farm them, but treat it as practice exercises, not as an advance on salary. If you really want to set a stop-loss, it's simple—limit your time and effort first: for example, no more than half an hour a day, a few interactions at most, and stop if you exceed that; then set a cost cap: I would advise against buying accounts, recruiting others, or using scripts, because that feels less like "practice" and more like gambling on someone else coming to take over your expectations later.

Recently, I've been discussing social mining and fan token schemes—"attention is mining," to be honest, attention is indeed valuable, but the value belongs to the platform and project teams, not necessarily to me. Anyway, I just calculate based on the real annualized return approach: the time invested counts as a cost, and if the returns are uncertain, don’t get too caught up; if I get a bonus, great, if not, it doesn’t affect my sleep. That’s all for now.