[kajumix]

How do you imagine this could be done? Is there an inexhaustible repository of "scientifically important problems" that can be solved every 10 minutes by some computing entity, and whose difficulty is adjustable -- a function of available compute power in the network, and are verifiable by others without human judgement involved?

> So why the hell are we not putting those to good use?

Some would argue that securing transactions that transfer value without trust is a good use

[arcticbull]

> Some would argue that securing transactions that transfer value without trust is a good use

Those folks haven't identified a good reason why being able to skirt sanctions, allow rogue states to accumulate value and to pay terrorists is worth "transferring value without trust." It's a pretty high bar, IMO.

[afriendofjungs]

> Is there an inexhaustible repository of "scientifically important problems" that can be solved every 10 minutes by some computing entity

Well, is the pool of mineable BitCoin hashes not limited? I'd say there are enough computational problems to be solved, certainly far more than there are BitCoins to be mined, for one.

As to trust and transactions, as far as I understand, every update of the BitCoin blockchain has to be agreed upon, meaning that it is verified by the network. To that end, you might as well verify useful computation, with the same mechanism.