1/ in the big picture, just like prediction markets, decision markets are only useful when there are a lot of participants (liquidity) who care deeply about a subject, do deep research, and place bets. This might work for big public topics like elections, but it’s very hard to achieve for smaller, domain-specific markets where there’s information asymmetry, making the system very inefficient.

2/ This suboptimality is similar to the 'one token, one vote' governance model where people can 'buy votes'. In low-liquidity decision markets, people can 'buy votes' and game the system. For the market maxis who believe arbs will solve everything: that’s only true with sufficient participants/liquidity

3/ What’s interesting is that AI could help solve this inefficiency by participating in smaller markets. But then, this isn’t for every use case.

So is decision markets better than existing governance systems? Slightly. Is it better than quasi-centralized systems? Nop

4/ What's missing from DAOs are roles and responsibilities—ironically they should be 'more centralized'. This doesn’t sacrifice the decentralization ethos we all resonate with, because everything will be transparent and the public will be involved and have veto power