My newest thoughts about monetary economics, plus exhortation to find a new initial coin distribution paradigm.
I believe that:
- coin generation schedule should take into account the size of the economy;
- it is not necessary to waste 100% of the value of the new coins as energy - that was necessary in gold era because trustless generation could not be otherwise maintained;
- it might be possible to find a way to distribute this value trustlessly in a way that enhances the economy and adoption;
- none of the current POS are in a right track
(but I am glad to be proven wrong! 
).
I think I have an idea for a better way to do it. Let's say an altcoin fixes its sights on some long-term rate of currency supply inflation, say 10%.
They start bootstrapping by awarding 1 coin to the miner per block, for Proof-of-work, and never stop or reduce this amount.
Additionally, whenever someone actually makes a transaction, the age of the inputs used is applied and 10% compounded annually is added to the block; 1% annual for the coin age, or 1 month's interest, whichever's more, as a tx fee for the miner, and the rest back to the payer in the transaction. This includes coinbase transactions: The miner can turn over whatever inputs he likes, and get 10% annual-rate interest for them paid along with his block.
The "score" or basis of competition of blocks for proof-of-work purposes is the actual proof-of-work, plus some factor multiplied by the amount of interest being awarded in that block; So the miner who includes all transactions, and is also turning over a lot of his own old unspent txouts, needs to meet a lower threshold for proof-of-work to get a block, than a miner who leaves out transactions or who is using no stake to mine.
Over time, the interest awards and tx fees dominate the 1 coin per block award, until eventually the 1 coin per block is negligible relative to the currency supply and the currency inflation converges on 10%. The amount of resources that it's worth spending per block for mining very slowly converges on being proportional to 1 months' interest (at 10% annual) on the txouts spent per block.
Because everybody is motivated to turn over their currency frequently, and the blocks with larger interest payments are strongly preferred, a version of TAPS or Transactions-As-Proof-Of-Stake, becomes viable. Preferring chains on the same basis as preferring blocks (proof-of-work plus the same factor times interest awarded) should rapidly converge on blocks representing the greatest fraction of the money supply, with older txouts counting more. An absolute "lock point" beyond which no reorg could ever go, could occur as little as 24 hours in the past in a healthy economy where the money is moving fairly fast.
Overall, the reward for mining would grow very slowly, with no "extra" to reward early miners. And the incentive to mine would likely never justify the kind of giant hashing farms we see with Bitcoin. Conversely, someone who does devote massive hashing power to it, especially in early days when the interest awards aren't very important yet in picking which chain to go with, could make attacks fairly easily, because the justification for investment for honest miners wouldn't be very high.
