Recently, I’ve been seeing a bunch of AI Agents claiming to fully automate on-chain operations. My first reaction isn’t “save effort,” but rather “who’s going to take the blame”… When it comes to lending/leverage, there are still several key steps that humans need to oversee: first, cut off authorization/limits—don’t just give unlimited approval right away; then, things like liquidation thresholds—no matter how diligent the model is, it might overlook extreme volatility, so it’s best to leave a manual switch and hard stop-loss lines; also, for cross-chain and pool swapping, routing changes can cause slippage to spike, and agents can easily “choose the optimal path” to trap you. Recently, rumors about stablecoin regulation, reserve audits, and de-pegging have been circulating in the group, and during such emotional times, someone needs to hold the line: don’t let the agent automatically add positions or borrow in panic—better to go slower, first confirm the pool’s health and whether your drawdown can handle it. Anyway, I now prefer to let it be an assistant, not a driver.