Recently, everyone has been talking about AI Agents going on-chain to work automatically, honestly it can save a little bit of effort, but if it’s fully entrusted I’m still hesitant. Checking contract permissions, seeing if it can transfer your money arbitrarily, even the smartest agent needs someone to double-check these steps; and then there’s the signing part, once signed there’s no regret medicine, you have to keep an eye on it yourself. As for the speculation about testnet points, whether the mainnet will issue tokens… the agent might be good at completing tasks, but if the rules change at the end, all the effort is wasted, and it’s more frustrating. Recently, I’ve been practicing trading psychology: don’t treat “automation” as certainty, what you can do is think through the worst-case scenario of each step, and reduce the head-over-heels enthusiasm of blindly powering through.