[zrm]

What if the "useful work" is to work against a collective action problem or similar, where the existing problem is that everybody wants something to happen but coordination is too expensive or there are "tragedy of the commons" type incentives in play?

[zrm]

The way the article attempts to address this is unsatisfactory:

> Secondly, this only works because the benefits are externalities, they are public and not owned by the miners. So there is no incentive for Miners to switch to such a system or even adopt such a system.

By this logic there is no incentive for Folding@home et al to exist either, but they do.

The incentive to switch to it is that it solves the collective action problem. That itself is another collective action problem, but it's an easier one. The difference to the miner between e.g. folding proteins and hashing is much smaller than the original difference between folding proteins and doing nothing, while still providing the full public benefit of protein folding.