You’re all talking again about how great parallel processing and sharding are—I’ll pour some cold water first: as lively as it may be, in the end it’s still the same two problems—where to put the money, and how to make sure you can’t lose it. As soon as the narrative gets a new skin, everyone charges in, but the old pits are still there: cross-chain bridges, message verification, and account permissions. When something really goes wrong, no one should pretend, “I didn’t think this would happen.”

Recently, AI Agents and automated trading have also been pretty hot. Put simply, you’re outsourcing your reflex speed to scripts, but safety is still on you to take the blame for. When I see all those demos of “automated on-chain interactions,” I instinctively start looking: is there a limit? Can you revoke approvals with one click? Did they spell out the exit path clearly? Otherwise, even if it’s parallel, it’s still just a parallel blow-up… And for voting, I can’t be bothered to listen to visions either—if the incentives and execution can’t back it up, then that’s that.