Just finished watching the funding rate—it’s as drowsy as the night-shift security guard getting off work. Now there are more and more on-chain AI Agents: automatic market making, automatic arbitrage, even automatic governance vote casting. It looks pretty convenient. But honestly, if you really hand over full control to them, I’m not comfortable. For example, on-chain trading—slippage, front-running/race conditions, MEV “sandwich” attacks—set even one parameter wrong and the bot can end up causing you all kinds of trouble. I’ve seen an automated strategy where when the market moved, it bought at the highest point, then a stop-loss triggered, and the loss was worse than if I’d done it manually. So I think AI Agents can at most be used as assistance—when it really matters, a human still needs to make the call, like judging the trend, assessing position risk, and deciding whether to manually intervene.

Recently I’ve seen miners and validators getting squeezed hard by the MEV players, and retail traders are already complaining about unfair ordering. This thing, governed purely by machines, definitely won’t prioritize the interests of ordinary users. Anyway, with so much information, my own filtering is simple: I only look at changes in the funding rate and large on-chain transfers. As for the other flashy narratives, I’ll put them aside for now.