Someone asked me how much we can trust things like address labels and clustering. To put it plainly, you can only treat them like “weather forecasts”—not use them as if they were an ID card… When a large holder splits into dozens of addresses, or an exchange moves hot and cold wallets back and forth, you might interpret it as funds flowing back, but it could just be moving houses. In any case, what I look at labels for now is mostly to find “anomalies”: a sudden concentration going into a specific contract, or the same batch of addresses doing the same kind of action at the same time—this is more reliable than “this is smart money.”

Recently, AI Agents and automated trading have also been pretty popular. After there are more on-chain interactions, things get even more chaotic: a bunch of robots running at full speed, everyone can hype the narrative, but no one wants to look at security details. My approach is pretty old-school: I’d rather chase the hype less, and first check permissions, whether the contract is newly issued, and whether the funds are coming in a wave and going out a wave. When the market is cold, slowing down actually feels more reassuring.