I just reviewed a DAO voting proposal, which on the surface says "optimize budget / improve efficiency," but upon closer inspection of the attachment, I found that the incentives have already been fixed: who can receive subsidies, who has signing authority, who can join small discussion groups, basically laying out the power structure for the next year. On-chain data can sometimes be quite honest — a few days before voting, a bunch of small, scattered addresses appeared, and delegation relationships were quietly changing hands. Honestly, it's not that everyone suddenly cares more about governance; someone is just calculating probabilities.

Recently, AI Agents and automated trading have also been very popular. I'm actually more interested in the fact that they "automatically interact with proposals": storytelling can be exaggerated, but contract permissions and execution conditions shouldn't be sloppy, especially those that say "vote and it automatically executes / automatically disburses funds." If the security details aren't clearly written, I wouldn't dare to trust it. Anyway, my first instinct when looking at a proposal isn't to consider the vision; I first look for where the money and keys are.