Actually, everyone understands that the most tempting part of staking isn't the small returns, but the illusion of "I also bought security along the way"… I was watching the flow of entrusted security projects on a few shared security platforms late into the night, and the more I looked, the more it felt like stacking playing cards: one layer goes wrong, and everything above it trembles. What's even more awkward is that the labels on many on-chain data tools are still criticized for being outdated or even misleading; you think you've figured out "who's running, who's building positions," but you might just be seeing the story others want you to see. Anyway, my current approach is very simple: the returns can stack, but slow down on adding positions, and first see if I can sleep through the worst-case scenario.