Recently, everyone has been talking about AI Agents one-click on-chain operations. My first reaction isn't "Wow, so convenient," but rather: who will be responsible for fixing the mistakes… On-chain stuff is real money with just one click, approving limits, signing contents, routing to different pools— even if it's just one line of text off, fees are still deducted, slippage still occurs, and people still have to confirm and bear the risk. Especially during congestion, the Agent might "smartly" increase prices to secure blockspace, ending up with Gas fees soaring like taxi fares… To put it simply, budget and stop-loss settings should still be handled by humans; otherwise, it becomes an automatic money-spending machine.

And recently, with social mining and fan tokens "attention is mining," I have some doubts: attention can be mined, but what might be mined out are fees and anxiety. For me, "long-term" isn't a year or half a year, but more about quarterly views—at least after a few congestion and emotional fluctuations, I can still avoid reckless signing and chasing. Let's stick with that for now.